Recent advances in molecular virology have enabled investigators to construct viruses that selectively destroy cancer cells (oncolytic virotherapy). Genetically engineered viruses belonging to different viral families have been evaluated for their potential as therapeutic agents in the treatment of malignant tumors [1-4]. Efficient replication, cell lysis and spread of HSV, and their natural broad host range make them attractive candidates as oncolytic viral agents [5-7]. Furthermore, the recent availability of cloned HSV genomes into bacterial artificial chromosome vectors greatly facilitates the rapid construction of new recombinant viruses carrying multiple transgenes of interest [8-10]. Tumor treatment with oncolytic HSV has been shown to induce anti-tumor immune responses [11-15]. Although the majority of people are seropositive for HSV-1, oncolytic virotherapy with HSV is not limited by pre-existing anti-HSV immunity [16,17], and in at least one example, preexisting immunity to HSV-1 enhanced anti-tumor immune responses [18].

Recently, the NV1020 oncolytic herpes simplex virus type-1(HSV-1) was shown to have significant promise for the treatment of many different types of tumors in preclinical studies in experimental animals as well as in human clinical trials [17,19-22]. The main advantage of this virus over other HSV oncolytic viruses is that it expresses one of the two original γ₁34.5 genes allowing the virus to replicate more efficiently, while safety is not compromised [23-26]. The γ₁34.5 gene is a major neurovirulence gene and an inhibitor of cellular apoptosis. Deletion of this gene drastically attenuates the virus and restricts viral growth to cancer cells because of their lack of intact apoptotic mechanisms [27,28]. Preclinical studies in mice as well as phase I/II human trials have revealed that oncolytic HSV-1 viruses having both γ₁34.5 genes deleted did not spread efficiently within tumors [29]. In contrast, deletion of one of the two γ₁34.5 genes drastically attenuated the virus, while allowing improved virus replication and spread in tumor cells [23-25]. The NV1020 was originally constructed for vaccine purposes and it contains HSV-2 viral sequences coding for glycoproteins gD, gG, gI and gE to facilitate production of anti-HSV-2 immune responses [24].

HSV can be transmitted from cell-to-cell by causing limited amounts of virus-induced cell fusion, thus avoiding the extracellular environment. Specific mutations within viral glycoproteins are known to greatly enhance virus-induced cell fusion. Specifically, syncytial mutations that cause extensive virus-induced cell fusion can arise in at least two of the glycoprotein genes: the UL27 gene, encoding glycoprotein B (gB) [30-32], and the UL53 gene, coding for glycoprotein K (gK) [33,34]. Work in our laboratory has shown that gK functions as a heterodimer with the UL20 viral protein and the UL20/gK heterodimer is necessary for virus-induced cell fusion [35,36].

The HSV-1 oncolytic virus Onc was constructed based on the NV1020 genomic arrangement with the exception that there are no genomic re-arrangements and no HSV-2 genes inserted within the viral genome. Recently, we reported that the OncSyn virus carrying a syncytial mutation in gB, enabling the virus to spread among cells by virus-induced cell fusion, replicated efficiently in breast cancer cells in vitro and drastically reduced tumor volumes in vivo [37]. In this study we constructed and tested the OncdSyn virus, which in addition to the gBsyn3 mutation also carried the gKsyn1 mutation known to enable the virus to fuse even difficult to fuse cells [38]. Intra-tumor injections of either virus effectively reduced tumor volumes as well as inhibited tumor metastases to internal organs.

Figure 1 Schematic representation of the genomic structures of the oncolytic recombinant viruses OncSyn and OncdSyn. (a) Representation of the prototypic arrangement of the HSV-1 genome with the unique long (UL) and unique short (US) regions flanked by the terminal (more ...)

Figure 2 Plaque morphology of the HSV-1 (F), OncSyn and OncdSyn viruses. Nearly confluent Vero (a-c) and 4T1 (d-f) cell monolayers were infected with wild-type HSV-1(F) (a, d), OncSyn (b, e) and OncdSyn (c, f) viruses. Individual viral plaques were visualized (more ...)

Figure 3 Comparative kinetics of viral replication of wild-type HSV-1(F) and mutant viruses OncSyn and OncdSyn grown on Vero and 4T1 cells. Near confluent monolayers of Vero (a) and 4T1 (b) cells were infected at an MOI of 2 with each virus, incubated at 37°C (more ...)

Figure 4 Intra-tumor treatment with OncSyn and OncdSyn viruses. (a) Balb/c mice were implanted subcutaneously in the interscapular area with 1 × 105 viable 4T1 cells. Tumors were measured using a digital caliper at defined time intervals prior and after (more ...)

The metastatic potential of the primary 4T1 tumor to internal organs after oncolytic or control therapy was assessed by gross and microscopic pathological examination of internal organs. In the first experimental protocol described above, mouse tumors were allowed to grow to approximately 80–90 mm³ and mice were sacrificed at 42 days post tumor cell implantation. In this experiment, mouse lungs from all three groups of mice (PBS, OncSyn, OncdSyn) had numerous metastatic foci, which were too numerous to be accurately counted (not shown). However, tumor foci in liver and spleen were substantially reduced in OncSyn and OncdSyn-treated mice in comparison to PBS-treated control mice (Table 1, Fig. 5). Specifically, all mice in the PBS group had metastatic nodes in liver, spleen, or kidneys. Some of the mice had tumors in all three organs (Fig. 5). Importantly, there were no metastatic tumors observed in the kidneys of virus-treated mice (Table 1).

Table 1 Metastatic nodes in internal organs

Figure 5 Gross pathological examination of metastatic tumor nodules on internal organs. Liver lobe, spleen, and kidney from a PBS-treated mouse carrying metastatic tumors (arrows) evaluated by gross pathological examination. Panel shows internal organs derived (more ...)

To better assess the potential of oncolytic virotherapy to reduce metastatic tumors in internal organs, a second experiment was performed in a similar fashion to the previous one with the exception that in the new experiment tumors were allowed to grow to approximately 35–40 mm³ in volume and mice were sacrificed at day 33 post tumor cell implantation after treatment with either OncdSyn or PBS. Lungs of OncdSyn-treated mice appeared to be practically devoid of metastatic tumors with only two mice having two nodes each. In contrast, all PBS-treated mice had multiple metastatic tumors in their lungs (Table 2, Fig. 6a and 6b). These results were confirmed by pathological examination of paraffin-embedded lung sections stained with Hematoxylin & Eosin (H&E) staining, which revealed the absence of tumors in OncdSyn samples, while PBS-treated control samples had numerous visible tumor foci (Fig. 6c–f).

Table 2 Metastatic nodes in lungs

Figure 6 Therapeutic effect of OncdSyn virus on lung metastases. (a, b) Gross appearance of excised lungs of representative mice from PBS control and OncdSyn treated groups. (c-f) Lung tissues were stained with H&E and examined. Representative stained (more ...)

The oncolytic HSV-1-based virus NV1020 has shown strong promise for treatment of different tumors in animal models and human clinical trials [17,19-22]. To facilitate the construction of recombinant viruses carrying multiple transgenes of interest, we cloned the NV1020-like HSV-1 recombinant virus OncSyn into a bacterial artificial chromosome (bac) vector. The OncSyn virus specifies a syncytial mutation in gB (Arg-to-His change at aa 858) that increases its ability to spread in tumor cells via virus-induced cell fusion [37]. In this study, we introduced the syncytial mutation syn1 (Ala-to-Val change at aa 40) within the gK gene to further enhance the fusogenicity of the new virus OncdSyn. The OncdSyn virus reduced primary tumor sizes and inhibited metastases to distal organs in the 4T1 syngeneic mouse model system.

The syn1 mutation within gK has been shown to produce extensive virus-induced cell fusion in all cells tested. In comparison, the gBsyn3 mutation produced virus-induced cell fusion in most cells, but it was unable to fuse certain hard to fuse cells, such as Hep-2 cells derived from human laryngeal carcinoma [38]. Therefore, to further increase the ability of the OncSyn virus to fuse all types of cells, we generated the OncdSyn virus carrying both the gBsyn3 and gKsyn1 mutations. As expected, the OncdSyn virus caused extensive virus-induced cell fusion and fused Hep-2 cells, while the OncSyn virus did not (not shown). Furthermore, the OncdSyn virus caused more extensive fusion than OncSyn in both Vero and 4T1 cells. The OncdSyn virus appeared to produce intact syncytia that remained attached to the cell culture flasks, while the OncSyn virus-induced syncytia contained infected single cells, which detached easier than the OncdSyn-infected syncytia. This phenomenon has been previously observed for the gB and gK syncytial mutations and it is probably due, in part, to the extensive virus-induced cell fusion caused by the gK syncytial mutation, which appears to also fuse internal membranes such as nuclear membranes in addition to plasma membranes of cells (Kousoulas, unpublished). Viral titers of the OncdSyn virus were lower in Vero cells than titers of the OncSyn virus and substantially lower than titers of the OncSyn virus in 4T1 cells. Typically, HSV-1 syncytial mutants produce lower viral titers than their parental wild-type viruses, most likely because of their direct effect on cellular membranes. In this regard, the increased ability of the OncdSyn virus to cause extensive virus-induced cell fusion is probably responsible for the observed decrease in viral titers in comparison to the OncSyn virus.

Defects of viral replication and spread in mouse cancer cells have been described in the literature for oncolytic herpesviruses [11,14]. HSV-1 does not replicate efficiently in mouse cell lines [41,42] most likely because it cannot as efficiently utilize the mouse nectin-1 receptor, which is approximately 5% different in its amino acid sequence to the human nectin-1 receptor [43]. Nectin-1 is also known to facilitate virus-induced cell fusion and virus-spread [44]. Consequently, both OncSyn and OncdSyn viruses replicated much less efficiently in 4T1 cells than in Vero cells. In this regard, the limited replication and spread of these viruses in 4T1 cells would be expected to adversely affect their oncolytic ability in 4T1-derived tumors in vivo. Previously, we reported that the OncSyn virus effectively reduced primary human breast cancer tumors in nude mice [37]. The disadvantage of the MDA-MB-435S human breast tumors is that these tumors would be rapidly eliminated if they were implanted in immunocompetent mice. Therefore, we chose the 4T1/Balb/c mouse model system for additional testing of both the previously constructed OncSyn virus as well as the newly constructed OncdSyn virus. Both OncSyn and OncdSyn viruses substantially reduced the growth of 4T1 tumors compared to the PBS controls, despite the fact that these viruses did not efficiently replicate in 4T1 cells in cell culture. Apparently, viral replication and infectious virus production in cell cultures did not correlate with the oncolytic efficacy of these viruses, because the OncdSyn virus reduced tumor volumes equally-well with the OncSyn virus, despite the fact that OncdSyn replicated approximately less than half a log than the OncSyn virus in 4T1 cells. Therefore, the relative increased ability of the OncdSyn virus to destroy tumors in vivo must be attributed to its enhanced fusogenicity.

Multiple murine tumor models have been used as preclinical settings for therapeutic purposes. The 4T1 mammary carcinoma model has several distinct advantages to be used as such model. It is regarded as a highly physiological, clinically-relevant mouse model that closely resembles stage IV human breast cancer in its properties [40]. 4T1 cells are considered to be very weakly immunogenic (relative antigenic strength is less than 0.01 with 9.9 being the most immunogenic) [45,46], and they spontaneously metastasize to distal parts of the body [40,47]. Metastatic tumor foci in liver and spleen were substantially reduced in OncSyn and OncdSyn-treated mice in comparison to PBS-treated control mice. Reduction of metastatic foci in internal organs (lung, spleen, kidney and liver) was dependent on the size of the original 4T1 tumor, as well as the time of necropsy post implantation of tumor cells. Specifically, there was drastic reduction in tumor foci detected in lungs when the initial tumor size treated with the virus was approximately 35–40 mm³ and necropsies were performed at 33 days after tumor implantation. Furthermore, lungs appeared to have the same number of metastatic foci with PBS-treated controls when the initial treated tumors where 80–90 mm³ and necropsies were performed at day 42 after tumor implantation. This metastatic pattern revealed that lungs were the primary metastatic site of the subcutaneous implanted 4T1 cells. Regardless of the size of the initial tumor treated and the time of necropsies post tumor implantation, it was evident that OncSyn and OncdSyn viruses appeared to efficiently reduce the growth of the primary tumor as well as substantially inhibit or eliminate formation of metastatic foci.

It is highly likely that reduction of the primary tumor after oncolytic virotherapy with the OncSyn and OncdSyn viruses is responsible for the observed reduction in the formation of secondary tumor foci, since treatment of the smaller (35–40 mm³) tumors appeared to drastically reduce lung metastases. Alternatively, it is possible that anti-tumor immune responses were elicited by exposure of tumor antigens after destruction of 4T1 cells within the primary tumor by the OncSyn and OncdSyn viruses. In this regard, a fusogenic oncolytic HSV-1 Synco-2D was reported to elicit anti-tumor immune responses when studied in a similar animal model of mammary carcinoma utilizing 4T1 cells [14]. A strong T-cell response was reported also by an HSV-2 derivative oncolytic virus FusOn-H2 effectively treating primary and metastatic mammary tumors in vivo [15].

Overall, our results showed that both OncSyn and OncdSyn viruses can efficiently reduce the primary and metastatic growth of 4T1 tumors established in immunocompetent mice. It is expected that these viruses would be even more efficacious against human breast cancer tumors by virtue of the fact that they can replicate substantially more efficiently (more than one log) in human than mouse cells. The availability of both OncSyn and OncdSyn viruses as bacterial artificial chromosomes will enable the generation of additional recombinant viruses that carry multiple anti-tumor and immunomodulatory transgenes, which could further enhance the anti-tumor efficacy of these viruses.

To determine the replication kinetics of the viruses, one-step growth kinetics were performed as described previously [48,49]. Briefly, nearly confluent monolayers of either Vero or 4T1 cells were infected with each virus at an MOI of 2 at 4°C for 1 h. Thereafter, virus was allowed to penetrate for 2 h at 37°C. Any remaining extracellular virus was inactivated by low-pH treatment with phosphate buffered saline at pH 3.0. Cells and supernatants were harvested immediately thereafter (0 h) or after 12 or 24 h of incubation at 37°C. Virus titers were determined by endpoint titration of virus stocks on Vero cells.

The authors declare that they have no competing interests.

AI performed most of the experiments and participated in drafting the manuscript, VNC participated in the construction and characterization of the viruses, AB was involved in the design and conduction of in vivo studies, ATD participated in pathological analysis and interpretation of data, MTK performed the statistical analyses, KGK was overall responsible for the project and for the preparation of the manuscript.

This work was supported by the grant "Novel Cancer Treatment Modalities" from the Louisiana Governor's Biotechnology Initiative (GBI), Louisiana Board of Regents and by the grant R01 AI43000 from NIH:NIAID to K.G.K. The project was also supported by the Louisiana Gene Therapy Program of the LSU Health Sciences Center, New Orleans. The authors gratefully acknowledge BioMMED's support by the LSU School of Veterinary Medicine and helpful discussions with Marlene Orandle and Karen Peterson (Department of Pathobiological Sciences, School of Veterinary Medicine, Louisiana State University).