Myxoma virus tropism in human tumor cells

Received: 5 April 2004; Accepted: 15 April 2004; electronically published: April 2004

Summary Myxoma virus is a species-specific poxvirus that causes myxomatosis in European rabbits but is nonpathogenic in other vertebrate species, including man. We show here that myxoma virus productively infects the majority (15/21) of human tumor cell lines tested from the NCI-60 reference collection. To assess for the potential involvement of virus host range genes, we screened several candidate gene knockout mutants of myxoma virus for permissiveness in these tumor cells. We observed that one particular myxoma virus variant, deleted in the ankyrin- repeat host range gene M-T5, was uniquely defective for replication in most of the human tumor cells that were permissive for the wild-type virus. Myxoma virus, therefore, exhibits specific tropism in a broad spectrum of human tumor cells and thus has the potential to be exploited as a novel oncolytic virus candidate.

DNA genome that ranges in size from 130 to 300 kilobase

pairs, and encode a wide spectrum of immunomodulatoryOncolytic viruses are traditionally defined by their proteins to help evade the host immune response (Seet etcapacity to selectively infect and kill cancer cells while al, 2003). Poxviruses replicate exclusively in thesparing non-transformed somatic cells (Bell et al, 2002; cytoplasm of the host cell, and therefore must encode theirChiocca, 2002; Hawkins et al, 2002; Mullen et al, 2002; own transcription and replication machinery (Moss, 2001).

Vile et al, 2002). Although the idea of using viruses to A number of poxviruses have been exploited as cancertreat cancer is not a new one, alternative viral vectors to vaccine vectors, including vaccinia virus, fowlpox andspecifically target transformed cells continue to be canarypox (Zeh et al, 2002; Menon et al, 2003). Thedeveloped as more is learned about the specifics of poxvirus large genome size allows for as much as 25individual virus tropisms (Stojdl et al, 2003; Balachandran kilobases of contiguous DNA to be inserted, thereforeet al, 2004; Tseng et al, 2004). For example, many enabling the expression of large eukaryotic genes and/orcandidate oncolytic viruses specifically utilize host cell gene clusters (Kaufman, 2003; Vanderplasschen et al,signaling pathways, such as p53 or Ras, that are activated 2003). This feature of poxviruses has been extensivelyor altered in neoplastic cells (Kirn et al, 2001; Nemunaitis exploited as numerous tumor antigens have beenet al, 2002).

expressed from poxvirus vectors, including CEAPoxviruses are among a group of viruses that have (Marshall et al, 2000) and prostate-specific antigen (Horigbeen used to kill tumor cells. For example, certain vaccinia et al, 2002). Poxviruses have also been used to delivervirus variants have been shown to replicate to a greater cytokines (Kaufman et al, 2002) or co-stimulatoryextent in transformed cells, however no poxvirus has been molecules (Horig et al, 2000) to activate dendritic cellsdescribed which exhibits a restrictive replication pattern and thereby increase the effectiveness of the vaccineonly in human tumor cells (Mastrangelo et al, 2002; Zeh et (Tsang et al, 2001). Human clinical trials are currently inal, 2002). Poxviruses are a family of large eukaryotic progress, utilizing novel and unique antigens expressed byDNA viruses that infect a wide range of vertebrates and poxviruses to combat a wide range of cancers (Hermistonarthropods (Moss, 2001). They contain a double-stranded

et al, 2002). However, the ability of poxviruses to infect and kill tumor cells specifically has been relatively unexplored.

Myxoma virus is a rabbit specific virus, which causes a lethal disease termed myxomatosis in the European rabbit, Oryctolagus cuniculus. In the early 1950s myxoma virus became the first example of a biological pest control strategy, when it was used in an attempt to control the disastrous feral rabbit population situation in Australia (Fenner et al, 1994). To date, myxoma virus infection of the European rabbit is one of the best models available for the study of pathogen-host interactions, and has allowed for the detailed investigation of viral anti-immune mechanisms as well as host immune responses (Kerr et al, 2002). One of the notable features of myxoma virus is its species specific ability to cause disease only in rabbits while being nonpathogenic for all other vertebrate species tested, including humans (Fenner et al, 1994). Despite this extremely narrow species host range myxoma virus can productively infect certain non-rabbit cells in vitro, such as immortalized baby monkey kidney fibroblasts (BGMK). Recent experiments from our lab indicate that myxoma virus can also infect primary murine cells genetically deficient in interferon responses (unpublished data), which prompted us to investigate the ability of myxoma virus to replicate in different classes of human tumor cells. Here we report that myxoma virus can productively infect the majority of human tumor cells tested from a wide spectrum of tissue types. Although the basis for this human cancer cell tropism of myxoma virus remains to be determined, we show that one particular viral host range gene (M-T5, an ankyrin-repeat protein previously shown to be required for myxoma replication in rabbit lymphocytes) is critical for virus replication in the majority of human tumor cells tested.

Dulbecco’s modified Eagle medium, DMEM, (GibcoBRL) was used to grow all of the cell lines. Baby green monkey kidney cells (BGMK) were grown in DMEM supplemented with 10% newborn calf serum (NCS) (GibcoBRL). All of the human tumor cell lines (from the NCI-60 reference collection) were propagated using DMEM supplemented with 10% fetal bovine serum (FBS) (Sigma). All cell lines were grown in medium containing 100 units/mL penicillin, 100µg/mL streptomycin at 37°C in 5% CO2. The cells used in this study are indicated in Table 1.

Table 1. Screening of Human tumor cell lines for permissiveness to infection with myxoma virus

Monkey

Rabbit

Human

Human

Human

Human

Human

Human

Human

Colo205 MDAMB435	Colon cancer Breast cancer	Human Human	� �

Human

The parental myxoma virus used in this study was designated vMyxlac, a version of myxoma virus, strain Lausanne (ATCC), containing the E.coli lacZ gene (under the control of the vaccinia late p11 promoter) inserted at an innocuous site between open reading frame (ORF) M010L and ORF M011L in the myxoma virus genome (Opgenorth et al, 1992). Three recombinant derivatives of myxoma virus were also used. vMyxlacT5-, a M-T5 knockout myxoma virus, had both copies of M-T5 replaced by lacZ (Mossman et al, 1996). Myxoma knockout viruses vMyxlacT2-and vMyxlacM11L-were constructed in the same manner as the vMyxlacT5-, with a lacZ gene replacing both copies of M-T2 (Upton et al, 1991) or the single copy of M11L (Opgenorth et al, 1992). The T5^-, T2^-and M11L^-knockout viruses are all unable to replicate in rabbit T lymphocytes (Macen et al, 1996; Mossman et al, 1996).

C. Infection of cell lines with myxoma virus and X-gal staining

Cells were infected at 90-95% confluency at a multiplicity of infection (moi) of 10, 1, 0.1, 0.01 or as otherwise indicated. Appropriate amounts of virus corresponding to the indicated moi was added to the cells, adsorbed for 1 hour and then the infection allowed to proceed in DMEM with 10% FBS. Infected cells were incubated in a CO2 incubator at 37°C for 48 hours. Cells were stained with X-gal (100mg/mL X-gal, 500mM Kferricyanide, 500mM Kferrocyanide, 100mM MgCl2 and PBS) for 4-8 hours after being fixed in neutral buffered formalin (NBF, [10% formaldehyde, PBS]) for 5 minutes.

For single step growth analysis, the appropriate virus at a moi of 5 was added to a cell monolayer at 95% confluency. The inoculum was allowed to adsorb for 1 hour, virus was removed and each well was washed three times with 1xPBS. Supplemented DMEM was added to the cells, which were then incubated at 37°C. Cells were collected by scraping following infection at the indicated time points: 1, 4, 8, 12, 24 and 48 hours. Following a 5 minute spin at 1500rpm the cells were resuspended in 100µL of hypotonic swelling buffer. To release virus from infected cells, each Eppendorf tube containing infected cells was frozen at -80°C and subsequently thawed at 37°C, this freeze-thaw cycle was repeated twice more. The lysed cells were sonicated in a cup sonicator for 1 minute to disaggregate virus complexes and then spun at 1500rpm for 5 minutes.

For multi-step virus growth curves, cells were infected at a moi of 0.01 and collected at the following time points: 12, 24, 48, 72 and 96 hours after infection.

Infectious virus at each time point was titrated on BGMK cells. Virus was diluted 1:20 in DMEM supplemented with serum and further serial dilutions were performed for each time point of each growth curve. The appropriately diluted virus was added to BGMK cells and allowed to adsorb for 1 hour, the virus was removed and DMEM supplemented with serum was added to each well. The infections were allowed to proceed for 48 hours, at which point the cells were fixed using NBF and stained with X-gal. Blue foci, indicating virus replication and spread, were counted and viral production per 10⁵ cells was determined. Titration of each time point was done in triplicate and graphed as average with corresponding standard deviation bars.

E. Western blot analysis to detect early and late viral gene expression

Western blot analysis was used to assess the expression levels of M-T7 and Serp–1, as prototypical early and late myxoma genes, respectively, in BGMK, HOS and 786-0 cells following infection with vMyxlac or vMyxlacT5-. Each cell line was infected at a moi of 5 and cells and supernatants were collected at 2 hours and 16 hours post infection. Supernatants were collected and concentrated 10 fold using 10K Omega centifugal concentrators (PALL Life Sciences). Cells were harvested by spinning at 3000rpm followed by three rounds of freeze/thaw for cell lysis. Proteins were extracted by spinning the lysed cells at 10000 rpm for 10 minutes and resuspending in 50µL of 1xPBS. The protein concentrations were determined by the Bradford assay, using a spectrophotometer (Beckman DU640). Equal amounts of protein (20µg) from cell supernatants and cell lysates were loaded in each lane and run on a 12% SDSPAGE gel to resolve M-T7 and 10% SDS-PAGE gel for the Serp1 blot. The separated proteins were transferred onto a nitrocellulose membrane (Amersham Biosciences) by a semi-dry transfer apparatus (BioRad) and probed with corresponding antibodies according to indicated conditions. The presence of MT7 and Serp1 proteins was visualized with horseradish peroxidase-conjugated second antibody using the enhanced chemiluminescence detection system (Amersham Pharmacia Biotech).

Cells were seeded in 96-well plates and allowed to grow to 95% confluency, corresponding to approximately 3.3x10⁴cells in each well. Each cell line was infected at a moi of either 0.01 or 5. Virus inoculum was adsorbed for 1 hour and infected cells were transferred to a -80° freezer at 24, 48, 72 and 96 hours pi and stored until further analysis. After the collection of all required time points, plates were subjected to three rounds of freeze/thaw. 50µL of lysed cells containing virus from each sample (well) was transferred into a new dish. To each sample well, 110µL of buffer A-β-mercaptoethanol mixture pH7.5 (100mM NaH2PO4, 10mM KCL, 1mM MgSO4, 50mM β-Mercaptoethanol) was added giving a final volume of 160µL. The components were mixed by inversion and the plate was incubated at 37°C for 5 minutes. 50µL of ο-Nitrophenyl-β-D-Galactopyranoside (ONGP) substrate was added to each well, the plate was covered with a lid and incubated at 37°C. The β-galactosidase activity, as measured by cleavage of the ONGP substrate, was read by a plate reader at 415nm at 10 and 15 minutes after the addition of ONGP. The results were graphed as average with corresponding standard deviation bars.

A. Myxoma virus has the ability to productively infect a wide spectrum of human tumor cells

To investigate the tropism of myxoma virus, a wide spectrum of human tumor cell lines from the NCI-60 reference collection were screened for productive infection with vMyxlac, wild type myxoma virus that expresses βgalactosidase. The results are summarized in Table 1 and Figure 1. The majority of the cell lines screened (15 of 21) were permissive to infection with myxoma virus, the exception being six cell lines that did not exhibit

Figure 1. Myxoma virus infects human tumor cells. Selected human tumor cells were screened for permissiveness to infection with myxoma virus. Cells were infected with vMyxlac at indicated multiplicities of infection and infection was allowed to proceed for 48 hours. Cells were then fixed and stained with X-gal. Foci were visualized using light microscopy. See Table 1 for a list of all the cells

screened in this study.

significant focus formation 48 hours following infection. A cell line was defined permissive to myxoma virus infection when visible foci were formed, that were comparable to those seen in the permissive control reference cell line, BGMK cells (Figure 1 upper row). A focus can be defined as a localized aggregate of myxoma infected cells and indicates a productive infection. In other words, the myxoma virus has infected a cell, efficiently replicated and spread to neighboring cells, thereby forming a distinct focus. Cells that were designated non-permissive (e.g. Colo205, HT29) failed to form discernable viral foci (Figure 1). Potential cytopathology in the nonpermissive cells was not investigated in this study. No direct correlation could be made between permissivity to infection with myxoma virus and cancer tissue origin or cell type.

To determine whether any specific virally encoded factor might be critical for the myxoma tropism observed, three of the known myxoma host range knockout viruses were selected for screening. They were vMyxlacT5-, vMyxlacT2- and vMyxlacM11L-, and the results of this screen are summarized in Table 2. The myxoma knockout viruses tested represent previously defined myxoma encoded host range genes, M-T2, M-T5 and M11L that were shown to replicate in rabbit fibroblasts or BGMK

Table 2. Permissiveness of selected cell lines to infection with myxoma virus Cell line Cell origin Species VMyxLac MT5-MT2-M11L

BGMK Kidney Monkey + + + + controls RK-13 Kidney Rabbit + + + + RL5 T-Lymphocyte Rabbit + --

HOS Osteosarcoma Human + + + + supportive PC3 Prostate cancer Human + + + + Caki-1 Renal cancer Human + + + +

M14		Melanoma	Human	-	-	-	-
abortive	MCF7	Breast cancer	Human	-	-	-	-
	COLO 205	Colon cancer	Human	-	-	-	-

HCT116 Colon cancer Human + -+ + restrictive 786-0 Renal cancer Human + -+ + SK-OV-3 Ovarian cancer Human + -+ + ACHN Renal cancer Human + -+ +

+ represents positive staining for X-gal and indicates formation of foci

-represents lack of X-gal staining

cells but are non-permissive in rabbit T-cells (Macen et al, 1996; Mossman et al, 1996). Ten representative human tumor cell lines were screened, originating from a wide variety of tissue types, including prostate, kidney, colon, ovarian, breast, bone and skin (Table 2). Cell permissiveness was determined based on the presence of detectable X-gal stained foci 48 hours pi. Cell lines where only isolated individual blue cells were observed were scored as non-permissive, however, it should be noted that in these cases myxoma virus still entered the cell and expressed β-galactosidase but was unable to spread to neighboring cells. Upon closer analysis of the data in Table 2, it was noted that the human tumor cells could be operationally divided into three groups based on the permissiveness to infection with vMyxlac and vMyxlacT5-. The first group (designated as supportive) was cells permissive to infection with both vMyxlac and vMyxlacT5- and included HOS, PC3 and Caki-1 cells. The second group (abortive) appeared to be completely non-permissive to infection with wild type myxoma virus as well as any of the knockout viruses and included M14, MCF7 and COLO205 cells. The last group of human tumor cells (called restrictive) which included HCT116, 786-0, SK-OV-3 and ACHN cells were permissive to infection with wild type myxoma virus, however, they were non-permissive to infection with the M-T5 knockout virus. M-T5 was the only host range gene tested that was shown to be required for myxoma virus infection of any group of the human tumor cells. This observation was the first indication of the significance of M-T5 in the host range of myxoma virus outside the rabbit system. In contrast to M-T5, the absence of either M-T2 or M11L had no effect on the ability of myxoma virus to infect any of the human tumor cells tested.

Three cell lines were chosen to best represent the three phenotypes observed that were differentially affected by M-T5 (Table 2). BGMK cells were utilized as a positive control, HOS cells represented fully permissive cell lines regardless of the presence or absence of M-T5, and 786-0 cells represented cell lines that specifically did not support infection with vMyxlacT5-. The results obtained in Table 2 were expanded by visualization of foci under a light microscope (Figure 2A) and quantification of virus titers obtained on indicated cell lines (Figure 2B). The virus titers of wild type myxoma virus obtained from 786-0 and HOS cells infected with vMyxlac were significantly lower than BGMK cells, emphasizing the infection efficiency differences between individual cell lines. It was also evident from the light microscope images that the lack of M-T5 was responsible for a decrease in focus size even in BGMK and HOS cells, and resulted in the complete absence of foci in 786-0 cells (Figure 2A). In 786-0 cells, the vMyxlacT5- virus titer was undetectable, therefore confirming the inability of vMyxlacT5- to productively infect this cell line. Thus, MT5 affects the replication efficiency in all the cells tested, but in some cases the defect was severe enough to completely prevent cell-cell spread of the virus.

C. M-T5 does not affect viral gene expression following high multiplicity infection

A standard approach to assess the nature of a block to poxvirus replication is to quantify early or late viral gene expression following infection. The three cell lines were infected with vMyxlac and vMyxlacT5- at a moi of 5 and samples were collected at 2 and 16 hours pi. Supernatants and lysates of infected cells were prepared for western blot analysis of early or late gene expression. To examine early gene expression, western blots were analyzed for the expression of M-T7, a 35kDa myxoma virus protein expressed and secreted early during infection. In BGMK, HOS and 786-0 cells M-T7 expression began as early as 2 hours and was readily detected at 16 hours pi with vMyxlac and vMyxlacT5infection, indicating that early viral gene expression occurred in all three cell lines infected with both viruses (Figure 3C). The M-T7 protein slightly differed in size when detected in cell lysates and cell supernatants because it is a secreted protein and undergoes glycosylation

Figure 2. Myxoma virus replication in BGMK, HOS and 786-0 cells. All three cell lines were infected at serial dilutions ranging from 10^-2 to 10^-8 of vMyxlac and vMyxlacT5-. At 48 hours post infection, cells were fixed and stained with X-gal. Foci were visualized by light microscopy (A) and counted to determine the viral titers in each cell line (B).Virus titers were calculated as focus forming units (ffu)/mL.

Figure 3. Demonstration of early and late viral gene expression following infection of three selected cell lines. Myxoma virus replication over the period of one replication cycle was investigated using high moi infection single-step growth curves in BGMK, HOS and 786-0 cells (A). Infectious virus progeny produced during the 48 hours time course was determined by titration on BGMK cells. Early (C) and late (D) viral gene expression following infection was investigated by resolving infected cell lysates or supernatants on SDS-PAGE gel and staining with corresponding antibodies. Early viral gene expression was determined by the expression of M-T7 (C) at 2 and 16 hours pi in the supernatant (lanes A) and cell lysate (lanes B). Late viral gene expression was determined by the expression of Serp-1 (D) at 16 hours pi in cell supernatants. The expression of beta-galactosidase over a 4 day timecourse following a high moi infection with vMyxlac or vMyxlac-T5 was determined to assess the ability of the two viruses to sustain late viral gene expression replication in all three cell lines (B). .

modifications during egress through the ER and Golgi. Nevertheless, the expression of M-T7 in all infections indicated that both viruses underwent successful viral binding, entry, uncoating and early gene expression in all three cell lines. Therefore, the non-permissive phenotype of the M-T5 knockout virus was attributed to an event following entry and early viral gene expression.

To complete the analysis of viral gene expression, cells infected with vMyxlac and vMyxlacT5- were analyzed in the same manner using the expression level of Serp1, a secreted late myxoma virus protein, as an indicator for late viral gene expression. In all three cell lines, Serp1 expression could be detected at 16 hours pi in both the vMyxlac and vMyxlacT5- infected cells (Figure 3D), indicating that both viruses reached the late gene expression stage. The lower amounts of Serp-1 secretion from HOS cells is not indicative of grossly lower levels of overall gene expression but may indicate a lower efficiency of the secretory pathway in HOS cells. This observation led to the conclusion that the block leading to the non-permissive phenotype in vMyxlacT5- infected 786-0 cells was after late gene expression.

To quantify the amount of late viral gene expression following infection, a micro β-galactosidase plate assay was performed. In this particular assay the activity of βgalactosidase, an enzyme engineered to be expressed by the virus under a late promoter, was assessed over a four day time period. BGMK, HOS and 786-0 cells were infected with vMyxlac and vMyxlacT5- at a moi of 0.01 (Figure 4B) or 5 (Figure 3B). The infections were allowed to proceed for 24, 48, 72 and 96 hours and ONGP, a substrate cleaved by the β-galactosidase enzyme, was added. Cleavage of ONGP could be quantified by measuring absorbance at a wavelength of 415nm. In control BGMK cells, both vMyxlac and vMyxlacT5behaved similarly to one another. At a moi of 5, both viruses exhibited progressively increasing β-galactosidase activity up until at least 96 hours pi (Figure 3B).

Figure 4. The ability of myxoma virus to replicate and spread following a low moi infection of BGMK, HOS and 786-0 cells. A low moi infection multiple step growth curve was performed in BGMK, HOS and 786-0 cells using both vMyxlac and vMyxlacT5- to investigate the ability of both viruses to infect and spread through the cell monolayer (A). Expression of beta-galactosidase was determined in the three cell lines infected with both viruses to quantify late viral gene expression over a 4 day time course (B).

At the lower moi of 0.01, the β-galactosidase activity was significantly lower during the first 48 hours pi when compared to a moi of 5 (Figure 4B). However, at 72 and 96 hours pi cells infected at a moi of 0.01 with both vMyxlac and vMyxlacT5-exhibited levels of βgalactosidase comparable to the higher moi infections (Figure 3B). In HOS and 786-0 cells, the induced βgalactosidase levels following infection with the two viruses at both moi of 5 and 0.01 appeared to be much lower than that observed in BGMK cells (Figure 3B, Figure 4B). Notably, β-galactosidase levels obtained for vMyxlacT5- infected HOS or 786-0 cells at a moi of 0.01 were extremely low, with little or no increase over the 96hour time course. This indicated that at a low moi, the vMyxlacT5- virus appeared to be markedly defective in both the “permissive” HOS and “nonpermissive” 786-0 cells.

D. Comparison of vMyxlac and vMylacT5- replication and spread

In order to quantitatively assess the ability of both viruses to infect and spread in each cell line, single step and multi step growth curves were performed in BGMK, HOS and 786-0 cells. Single step growth curves are performed at a high moi and indicate infectious virus progeny produced during a single replication cycle of the virus. Cells were infected with vMyxlac and vMyxlacT5at a moi of 5 and samples were harvested for infectious virus particles at 1, 4, 8, 12, 24 and 48 hours pi. All time point samples were titrated on BGMK cells by serial dilutions and stained with X-gal 48 hours pi to visualize foci. Infection of all three cell lines with vMyxlac produced growth curves closely resembling a classical poxvirus replication curve, reaching a minimum at approximately 8 hours pi followed by a continuous increase until 48 hours pi, at which point the virus yield had reached maximal levels (Figure 3A). Comparison of vMyxlac and vMyxlacT5- growth curves in the three cell lines revealed an interesting trend. Although BGMK and HOS cells were scored as permissive to infection with the vMyxlacT5- knockout virus, it was clear that the T5minus myxoma virus was to some degree defective in growth in both cells. The titers obtained at all time points were lower than for the vMyxlac titers, indicating a role for M-T5 in optimal virus replication (Figure 3A). Unexpectedly, the samples harvested in the nonpermissive 786-0 cells at all time points after infection with vMyxlacT5- contained detectable infectious virus particles that could be titrated on BGMK cells. As well, the growth curve of vMyxlacT5- infection of 786-0 cells followed the same classical poxvirus replication curve. Although the titers at 48 hours pi were considerably lower in the vMyxlacT5- infections than wild type myxoma infections (Figure 3A), infectious virus could nevertheless still be detected, therefore indicating some minimal level of virus replication following high multiplicity infection of 786-0 cells with the T5- knockout virus.

Multiple step growth curves are performed at a lower moi and for longer periods of time to measure multiple rounds of viral replication and cell to cell spread. All three cell lines were infected with vMyxlac and vMyxlacT5- at moi of 0.01. The samples were collected and harvested for infectious virus at 12, 24, 48, 72 and 96 hours pi, then titrated on BGMK cells. Wild-type myxoma virus underwent several rounds of replication in all three cell lines, evident by the progressive cell-cell spread over time (Figure 4A). A comparison of viral titers indicated that although all three cell lines were permissive to infection, the amount and efficiency of virus replication was approximately 1.5 logs higher in BGMK cells than in both of the human tumor cell lines (Figure 4A). BGMK cells appeared to be fully permissive to vMyxlacT5- infection as the titers of the knockout virus were comparable to wild type virus at 96 hours pi. HOS cells behaved in a similar manner to BGMK cells, with vMyxlacT5- producing viral titers approximately 1 log lower throughout the growth curve duration. However, in the 786-0 human tumor cell line, a dramatic difference in viral titers (approximately 2 logs) between vMyxlacT5- and vMyxlac was observed even as late as 96 hours pi (Figure 4A), indicating a significant defect in virus growth and/or spread in the absence of M-T5 expression. We conclude that M-T5 is required for optimal virus replication and generation of infectious progeny in all the cell lines tested, but the requirement is particularly stringent in the human tumor cells.

Myxoma virus has always been regarded as a rabbit specific poxvirus because it causes clinical disease only in rabbits; however it is not known which virally encoded genes or host determinants are responsible for this species specificity (Kerr et al, 2002). All attempts to demonstrate cell or species-specific receptors for poxviruses have been negative, and it is now presumed that intracellular events following binding and entry determine the tropism characteristics of a given poxvirus (Johnston et al, 2003). In this study we investigated myxoma virus tropism in the context of human tumor cells. To date, there are no reports of myxoma virus infection in primary human cells. A general survey of a panel of human tumor cells from the NCI reference panel was performed to determine permissiveness to infection with wild type myxoma virus (Table 1) as well as selected gene knockout viruses (Table 2). The knockout viruses that were tested represent three mechanistically distinct host range factors encoded by myxoma virus, namely, M-T2 (a TNF receptor homolog), M11L (a mitochondrially-located apoptosis inhibitor) and M-T5 (an ankyrin repeat protein) (Macen et al, 1996; Mossman et al, 1996). Of these, M-T5 appeared to be uniquely critical for the ability of the virus to optimally infect the human tumor cells tested (Table 2).

Three of the permissive cell lines were chosen to facilitate an investigation of the role of M-T5: BGMK cells, as a positive control cell line, HOS cells, representing the fully permissive group of human tumor cells that supported both wild-type myxoma and T5knockout infection and 786-0 cells, representing cells for which the presence of M-T5 was essential for myxoma virus infection and spread. Closer investigation of the three cell lines demonstrated pronounced differences in the ability of myxoma virus to productively infect and generate progeny virus. Foci size and density differed in all three cases, with BGMK cells forming classic myxoma virus foci, whereas HOS and 786-0 cell foci appeared much smaller in size (Figure 2A). ACHN and 786-0 (renal), HCT116 (colon), SK-OV-3 (ovarian) cells infected with vMyxlacT5- produced no observable blue foci or blue cells, indicating a complete absence of virus replication at 48 hours pi. Analysis of viral replication and production of infectious virions in 786-0 was performed by conducting single and multi-step growth curves (Figures 3A, 4A). In BGMK cells there appeared to be no difference in virus replication levels regardless of the presence or absence of M-T5. In the two human tumor cell lines, however, differences between vMyxlac and vMyxlacT5- were much more evident. For these cells, the absence of M-T5 caused a delay in virus replication and less infectious progeny was produced, yielding much lower virus titers than in cells infected with vMyxlac (Figure 3A, 4A). It was evident that in both HOS and 7860 cells, M-T5 was required for optimal myxoma virus infection and the production of any substantial levels of infectious progeny. However, it was still unclear at which stage during the replication cycle the vMyxlacT5- virus was hindered. Early and late gene expression was monitored in BGMK, HOS and 786-0 cells using western blot analysis, probing for M-T7 and Serp1, early and late myxoma virus genes respectively (Figure 3C, 3D). Both early and late gene expression appeared unaffected by the absence of M-T5 following high multiplicity infection of all three cell lines, leading to the conclusion that the block in virus replication must occur at some point following late viral gene expression. To investigate this further, a colorimetric assay measuring the expression of the LacZ gene, under a late promoter, was performed. This assay allowed us to quantify the amount of late viral gene expression that accumulated following single and multiple cycle growth. Consistent with data obtained from the growth curves, in BGMK cells both wild type and T5minus viruses appeared to behave the same, implicating similar amounts of late viral gene expression. However, in HOS and 786-0 cells the absence of M-T5 caused viral replication to significantly decrease over multiple replication cycles. Therefore, although late viral gene expression appeared to be normal in the initially infected cells when investigated at 16 hours pi following high multiplicity infection (Figure 3D), the colorimetric assay indicated that the absence of M-T5 indeed progressively hindered myxoma virus late gene expression at later times irrespective of the multiplicity. It may be that the function of M-T5 becomes increasingly critical at a later point in the replication cycle, during viral assembly or virus spread, and therefore is less obvious at 2 and 16 hours pi following high multiplicity infection. Although we conclude that M-T5 plays a critical role during myxoma virus infection of human tumor cells, it is also required for optimal virus replication even in permissive BGMK cells. At present, the known structural features of M-T5 (Figure 5) do not provide clues as to the precise mechanism of action of this viral host range factor.

Although the molecular basis for myxoma tropism in human tumor cells remains to be determined, the available evidence suggests that differences in intracellular

Figure 5. Schematic representation of the location and features of M-T5.

The 483 amino acid protein is present in two copies and located within the inverted terminal repeat regions of the myxoma virus genome. M-T5 contains 4 predicted palmitoylation sites (✳) as well as well as 7 predicted ankyrin repeats (lined bars). The predicted ankyrin repeats are situated at amino acids 32-63, 67-104, 105-140, 141-175, 177-213, 250-282 and 283-315.

signaling within the infected cell are critical for distinguishing the permissive vs. restrictive phenotype (Johnston et al, 2003). Oncolytic virus candidates have been selected to optimally replicate within transformed cells, and this work indicates that myxoma virus exhibits pronounced tropism for productive replication in the majority of human tumor cells tested. Further analysis of the determinants at this tropism should yield insights into the unique phenotype of human tumor cells, as well as the ability of myxoma virus to induce cell death of transformed cells. More importantly, the ability of myxoma to infect human tumor cells, the lack of preexisting immunity in humans and a large genome allowing for insertion of therapeutic genes all suggests significant potential for the exploitation of myxoma virus as a novel oncolytic virus platform.

We would like to thank Xuijuan Gao for technical assistance, John Barrett for reviewing the manuscript and Doris Hall for help with corrections. GM holds a Canada Research Chair in Molecular Virology. The lab is supported by the CIHR and NCI of Canada.

Balachandran S, and Barer GN, (2004) Defective translational control facilitates vesicular stomatitis virus oncolysis. Cancer Cell 5, 51-65.

Bell JC, Garson KA, Lichty BD, and Stojdi DF, (2002) Oncolytic viruses: programmable tumour hunters. Current Gene Therapy 2, 1-12.

Chiocca EA, (2002) Oncolytic Viruses. Nature Reviews/Cancer 2, 938-961.

Fenner F, and Ross J, (1994) Myxomatosis. In: GV Thompson, and CM King (ed.), The European Rabbit, the History and Biology of a Successful Colonizer, pp. 205-239. Oxford, New York, Tokyo: Oxford University Press.

Hawkins LK, Lemoine NR, and Kirn D, (2002) Oncolytic biotherapy: a novel therapeutic platform. Lancet Oncol 3, 17-26.

Hermiston T, and Kuhn I, (2002) Armed therapeutic viruses: strategies and challenges to arming oncolytic viruses with therapeutic genes. Cancer Gene Therapy 9, 1022-1035.

Horig H, Lee CS, and Kaufman HL, (2002) Prostate-specific antigen vaccines for prostate cancer. Expert Opinion in Biological Therapy 2, 395-408.

Horig H, Lee DS, Conkright W, Divito J, Hasson H, and LaMare M, (2000) Phase I clinical trial of a recombinant canarypoxvirus (ALVAC) vaccine expressing human carcinoembryonic antigen and the B7.1 co-stimulatory molecule. Cancer Immunology, Immunotherapy 49, 504

514.

Johnston JB, Barrett JW, Chang W, Chung CS, Zeng W, Masters J, Mann M, Wang F, Cao J, and McFadden G, (2003) Role of the serine-threonine kinase PAK-1 in myxoma virus replication. J Virol 77, 5877-88.

Kaufman H, (2003) The role of poxviruses in tumor immunotherapy. Surgery 134, 731-737.

Kaufman HL, Flanagan K, Lee CS, Perretta DJ, and Horig H, (2002) Insertion of interleukin-2 (IL-2) and interleukin-12 (IL-12) genes into vaccinia virus results in effective antitumor responses without toxicity. Vaccine 20, 1862-1869.

Kerr P, and McFadden G, (2002) Immune responses to myxoma virus. Viral Immunol 15, 229-46.

Kirn D, Martuza RL, and Zwiebel J, (2001) Replication-selective virotherapy for cancer: Biological principles, risk management and future directions. Nature Medicine 7, 781

Macen JL, Graham KA, Lee SF, Schreiber M, Boshkov LK, and McFadden G, (1996) Expression of the myxoma virus tumor necrosis factor receptor homologue (T2) and M11L genes is required to prevent virus-induced apoptosis in infected rabbit T lymphocytes. Virology 218, 232-237.

Marshall JL, Hoyer RJ, Toomey MA, Faraguna K, Chang P, Richmond E, Pedicano JE, Gehan E, Peck RA, Arlen P, Tsang KY, and Schlom J, (2000) Phase I study in advanced cancer patients of a diversified prime-and-boost vaccination protocol using recombinant vaccinia virus and recombinant nonreplicating avipox virus to elicit anti-carcinoembryonic antigen immune responses. Journal of Clinical Oncology 18, 3964-3973.

Mastrangelo MJ, and Lattime EC, (2002) Virotherapy clinical trials for regional disease: in situ immune modulation using recombinant poxvirus vectors. Cancer Gene Ther 9, 1013

Menon AG, Kuppen PJ, Van der Burg SH, Offringa R, Bonnet MC, Harinck BI, Tollenaar RA, Redeker A, Putter H, Moingeon P, Morreau H, Melief CJ, and Van de Velde CJ, (2003) Safety of intravenous administration of canarypox virus encoding the human wild-type p53 gene in colorector cancer patients. Cancer Gene Therapy 10, 509-517.

Moss B, (2001) Poxviridae: The viruses and their replication. In: DM Knipe, and PM Howley (ed.), Fields Virology, Fourth edn, pp. 2849-2883. Philadelphia: Lippincott Williams & Wilkins.

Mossman K, Lee SF, Barry M, Boshkov L, and McFadden G, (1996) Disruption of M-T5, a novel myxoma virus gene member of the poxvirus host range superfamily, results in dramatic attenuation of myxomatosis in infected European rabbits. Journal of Virology 70, 4394-4410.

Mullen JT, and Tanabe KK, (2002) Viral oncolysis. Oncologist 7, 106-19.

Nemunaitis J, and Edelman J, (2002) Selectively replicating viral vectors. Cancer Gene Therapy 9, 978-1000.

Opgenorth A, Graham K, Nation N, Strayer D, and McFadden G, (1992) Deletion analysis of two tandemly arranged virulence genes in myxoma virus, M11L and myxoma growth factor. Journal of Virology 66, 4720-4731.

Seet BT, Johnston JB, Brunetti CR, Barrett JW, Everett H, Cameron C, Sypula J, Nazarian SH, Lucas A, and McFadden G, (2003) Poxviruses and immune evasion. Annu Rev Immunol 21, 377-423.

Stojdl DF, Lichty BD, tenOever BR, Paterson JM, Power AT, Knowles S, Marius R, Reynard J, Poliquin L, Atkins H, Brown EG, Durbin RK, Durbin JE, Hiscott J, and Bell JC, (2003) VSV strains with defects in their ability to shutdown innate immunity are potent systemic anti-cancer agents. Cancer Cell 4, 263-275.

Tsang KY, Zhu M, Even J, Gulley J, Aarlen P, and Schlom J, (2001) The infection of human dendritic cells with recombinant avipox vectors expressing a costimulatory molecule transgene (CD80) to enhance the activation of antigen-specific cytolytic T cells. Cancer Research 61, 7568-7576.

Tseng J-C, Levin B, Hurtado A, Yee H, Perez de Castro I, Jimenez M, Shamamian P, Ruzhong J, Novick RP, Pellicer A, and Meruelo D, (2004) Systemic tumor targeting and killing by Sindbis viral vectors. Nature Biotechnology 22, 70-77.

Upton C, Macen JL, Schreiber M, and McFadden G, (1991) Myxoma virus expresses a secreted protein with homology to the tumor necrosis factor receptor gene family that contributes to viral virulence. Virology 184, 370-382.

Vanderplasschen A, and Pastoret P-P, (2003) The uses of poxviruses as vectors. Current Gene Therapy 3, 583-595.

Vile R, Ando D, and Kirn D, (2002) The oncolytic virotherapy treatment platform for cancer: Unique biological and biosafety points to consider. Cancer Gene Therapy 9, 10621067.

Zeh HJ, and Bartlett DL, (2002) Development of a replication-selective, oncolytic poxvirus for the treatment of human cancers. Cancer Gene Ther 9, 1001-12.

From left to right, first row Dr. Grant McFadden, Dr. Joanna Sypula From left to right second row, Dr. Yiyue Ma, Dr. Fuan Wang