New cancer therapies with novel mechanisms of action (MOA) are needed that are not cross-resistant with approved therapies and are effective even if apoptosis is blocked. Engineering agents for multiple complementary MOA can improve efficacy (as seen with multitargeted tyrosine kinase inhibitors1,2,3,4). Engineered viruses have been developed for cancer treatment using diverse methods, including gene therapy (transfer of a therapeutic gene using a replication-incompetent virus)5,6,7 and cancer vaccines (expression of tumour antigens, co-stimulatory molecules or cytokines)8,9,10. However, gene therapy has failed to date in patients owing to inefficient delivery of the gene to sufficient numbers of cancer cells locally and systemically, and vaccines have been limited by the immune evasion of tumours and the exclusive reliance on host factors for vaccine efficacy in patients with advanced bulky cancers (Table 1).

Table 1 The evolution of virotherapeutics

By contrast, the inherent capacity of oncolytic viruses to infect, multiply within and subsequently lyse cancer cells11,12,13,14 has been exploited to address some of the limitations of non-replicating viral therapeutics. First-generation oncolytic viruses, such as reovirus15,16 and vesicular stomatitisvirus17,18, are inherently selective for cancer, whereas second-generation agents, such as adenovirus19,20 and herpes simplex virus21,22 deletion mutants, have been engineered for cancer selectivity. Two agents have been extensively tested in clinical trials (>400 patients to date): a first-generation reovirus (Reolysin, Oncolytics)23, and a second-generation adenovirus (dl1520, also known as Onyx-015, an E1B-55kD deletion mutant)24,25,26,27,28. Clinical trial data from these agents indicated that they were safe and selective for cancer, but therapeutic potency was limited after both direct intratumoral and intravenous injection29, and no systemic spread to distant tumours was detected. Anticancer potency, delivery and systemic spread therefore had to be improved.

Given both the potential and the limitations of each of these approaches, several groups asked whether it was possible to combine and optimize the best attributes of each approach into a single therapeutic agent. Such third-generation oncolytic virus therapeutics could realize the potential of these disparate therapeutic platforms. In this Perspective, we describe how poxviruses, such as vaccinia, were selected as the optimal oncolytic product pharmacophore. These viruses were then armed with therapeutic transgenes, such as the gene encoding granulocyte–macrophage colony-stimulating factor (GM-CSF), to stimulate antitumoral immunity. The resulting agents have been shown to be highly selective for cancers and have a high degree of systemic efficacy by multiple MOA in preclinical testing30,31,32,33,34. Recently published clinical trial data have confirmed these features in patients with advanced metastatic cancers who had previously failed multiple prior therapies35,36.

Poxvirus — an ideal pharmacophore?

The family Poxviridae consists of enveloped double-stranded DNA viruses that infect a wide range of vertebrates and invertebrates. The orthopoxvirus37,38 subfamily includes vaccinia virus, which has had a crucial role in one of the greatest achievements in medicine: the eradication of smallpox39. This historical role has also led to a detailed understanding of vaccinia biology and pathogenesis. In addition, the fact that more people have been deliberately infected with vaccinia than with any other infectious agent means that we have unprecedented information on its behaviour in humans, including which populations are at risk for rare adverse events and how to respond to these40,41.

The highly immunogenic nature of vaccinia infection, which produces a strong cytotoxic T-lymphocyte (CTL) response42 and circulating neutralizing antibody that can be detected many decades later43, was crucial to the successful use of vaccinia in smallpox eradication and has led to its continued use in immunotherapy. Vaccinia strains and related poxviruses that express foreign antigens have been used in many vaccine trials for protection against a range of infectious diseases44,45,46,47,48 and for treatment of established cancers49,50,51. Increasingly sophisticated strategies (including sequential prime–boost treatments with different poxviruses49 or co-expression of combinations of co-stimulatory factors10) have led to enhancement of immune responses. However, to date the antitumour effects of these approaches in clinical trials have been limited. This limited success to date may be due to the local immunosuppressive nature of the tumour microenvironment, as well as immune editing, which can lead to the escape of tumour cell subpopulations that do not express the target antigens. Finally, early-stage clinical trials have generally enrolled patients with advanced, bulky tumours. Nevertheless, the ability of the immune system to raise a response against antigens encoded by poxviruses has been clearly and repeatedly demonstrated.

Numerous inherent biological properties of vaccinia and other poxviruses make them ideal for development as oncolytic agents37. First, they replicate and lyse cells rapidly compared with other virus species — key factors for determining efficacy52. The first viral particles produced are secreted from cells within 8 hours and infected cells are destroyed 48 to 72 hours after infection. Second, poxviruses have broad tumour tissue tropism32; they do not require defined cell surface receptors for entry into target cells but instead they infect through several membrane fusion pathways53,54. Third, they do not integrate DNA into the host chromosome (which is a safety concern with other virus families) because they replicate in mini-nuclear structures in the cytoplasm called viral factories. Fourth, the viruses that are produced are of several distinct antigenic forms, including an enveloped virus (EEV) form that shrouds itself in a host cell-derived envelope that contains several host complement control proteins and few exposed viral proteins43,55,56. Poxviruses are therefore highly efficient at spreading to distant tumours within a host and moving unharmed through the bloodstream. Therefore, the systemic delivery of poxviruses and their spread between tumours is highly efficient32,33,57. Fifth, a number of approved or experimental antiviral agents are available to treat poxvirus infections in case of an adverse response (such as vaccinia immune globulin58, cidofovir59, ST-246 (Ref. 60) or certain tyrosine kinase inhibitors61). Finally, vaccinia viruses are able to accommodate multiple large transgenes, which enables a wide variety of genes to be added to engineer complementary MOA62.

Tumour selectivity of poxviruses

Poxvirus-derived agents can selectively replicate within and lyse cancer cells, both based on their natural biology63 and through genetic engineering. For example, myxoma virus naturally targets tumours by relying on disrupted type I interferon (IFN) induction in mouse tumour cells63 or increased Akt levels in human tumour cells64 in order to overcome the species barrier and replicate selectively in cancer cells. Vaccinia vaccine strains have also recently been shown to inherently target tumours32,65. This characteristic may be attributed to the fact that many of the hallmarks of cancer66, such as blocks in apoptotic pathways, deregulation of cell cycle control and immune evasion, are also optimal cellular conditions for successful poxvirus replication. Cancer cells are therefore more susceptible to viral infection than normal cells. In particular, vaccinia replication and spread is associated with activation of the epidermal growth factor receptor (EGFR)–Ras signalling pathway67, and agents that block this signalling pathway can inhibit vaccinia replication68. The EGFR–Ras pathway is activated in most human cancers66, indicating that vaccinia could be broadly applied in cancer therapy. The multiple different vaccine strains of vaccinia used during the smallpox eradication programme probably display unique properties owing to variations in the expression or functionality of different virulence genes between strains, and several different backbones have been used in the design of oncolytic agents, including Western Reserve32, Wyeth30, Copenhagen69 and Lister70 (Table 2).

Table 2 Wild-type vaccinia strains used as backbones for therapeutic agents

Recent research has focused on further enhancing the inherent cancer selectivity of these viruses. When viruses infect a cell they express a number of genes, some of which can effectively transform a normal cell, such that it loses the ability to arrest the cell cycle, undergoes uncontrolled replication and is blocked from entering apoptotic pathways (Fig. 1). Because the genetic mutations that occur in cancer have already created the optimal cellular state for virus replication, these viral transformation genes are often expendable for viral replication. Importantly, deletion of these genes from the viral genome greatly reduces the ability of the virus to productively replicate in most normal cells. A range of gene deletions with such properties have previously been described for vaccinia. For example, vaccinia encodes a thymidine kinase gene (TK) that, when deleted, leads to dependence of the virus on cellular thymidine kinase expression71,72. Cellular thymidine kinase, which is regulated by the E2f transcription factors, is transiently expressed during the S phase of the cell cycle in proliferating normal cells, but is constitutively expressed at high levels in the majority of cancers regardless of proliferation status73. Vaccinia also expresses an EGF homologue (vaccinia growth factor; VGF) that binds EGFR74,75,76. Because VGF is secreted from infected cells, it induces proliferation in both infected and surrounding non-infected cells. Therefore, vaccinia strains with deletions in both thymidine kinase and VGF (known as vvDD) show selective replication in cancers with an activated EGFR pathway that results in a higher therapeutic index than viral mutants with a deletion in either thymidine kinase or VEGF alone32,33. Vaccinia also expresses several genes — including two serpins77,78 and an inhibitor of cytochrome c release79 — that block apoptosis in infected cells. Deletion of combinations of these genes results in vaccinia mutants that have enhanced tumour selectivity34,80.

Figure 1: Selectivity of oncolytic vaccinia strains for tumour cells.
figure 1

a | In normal cells, the wild- type virus produces a range of gene products that adapt the cell for viral replication, including VGF (vaccinia growth factor), which activates the epidermal growth factor receptor (EGFR) pathway, thymidine kinase (TK), which produces a nucleotide pool for replication of the viral genome, and a variety of immunosuppressive proteins, such as B18R, which binds and sequesters type I interferon (IFN), thus blocking the IFN antiviral response. b | In normal cells, oncolytic strains that carry deletions in any of these genes are unable to adapt the cell for viral replication and replication of the virus is blocked. c | In tumour cells, common oncogenic mutations compensate for viral gene deletions and viral replication occurs. Viral transgene expression (such as granulocyte–macrophage colony-stimulating factor (GM-CSF)) can further enhance the oncolytic effect. CEV, cell-associated enveloped virus; EEV, enveloped virus; IMV, intracellular mature virus.

Vaccinia expresses many genes that encode products with immunomodulatory functions and deletion of these genes can increase tumour selectivity. The products of these genes include a secreted type I IFN-binding protein, which acts as a decoy receptor to sequester extracellular IFN81,82. Deletion of this gene attenuates replication in the presence of IFN in normal cells, in which type I IFNs have profound antiviral effects. Because cancer cells have frequently lost the ability to produce or respond to type I IFNs, they remain permissive to replication of vaccinia anti-IFN gene deletion mutants31.

The expression of multiple virulence genes by poxviruses enables a multitude of tumour-targeting deletions to be created, thus potentially allowing oncolytic strains to be tailored to specific and varied tumour phenotypes.

Diverse MOA with oncolytic poxviruses

The highly heterogeneous nature of most tumours and the discovery that the cancer stem cell may be fundamentally different from the rest of the cancer cells83,84 indicate that therapies that are capable of recognizing and destroying tumours through multiple MOA are needed. Most cancer treatment regimens therefore rely on combinations of different agents. Poxvirus therapies also have the ability to target and destroy cancer cells by multiple complementary mechanisms (Fig. 2). Direct infection of cancer cells results in cell lysis and death. The mechanism that underlies this cell death is likely to depend on the viral strain and tumour cell targeted, but seems to have features of both necrosis and apoptosis. In addition, poxviruses trigger several other changes in the tumour that are of therapeutic value.

Figure 2: Mechanisms of action of oncolytic vaccinia virus.
figure 2

Vaccinia delivered to the tumour through the vascular system can produce an antitumour effect through multiple mechanisms, which include viral infection and tissue destruction. This leads to release of cytokines (blue symbols), danger signals (yellow symbols) and antigens (red symbols) that can stimulate the innate and adaptive immune responses. Viral infection of tumour cells leads to replication of the virus and viral spread through and between tumours. Viral infection in and around tumour endothelial cells leads to vascular collapse. Endothelial cells are destroyed either as a result of direct infection with virus, or subsequent to infection of surrounding tumour cells, which leads to infiltration of neutrophils into the tumour and thrombosis.

Immune-mediated cell death. The highly destructive nature of a poxvirus infection results in the release of many cellular danger signals (danger-associated molecular pattern molecules; DAMPs)85,86 and viral danger signals (pathogen-associated molecular pattern molecules; PAMPs)87, as well as the release of both virus- and tumour-associated antigens at the site of infection within the tumour. The potent inflammatory response initiated by these factors must ultimately overcome tumour-mediated immune suppression to clear the virus. In addition, the production and release of tumour antigens in this highly immunostimulatory context seems to be able to induce an adaptive immune response against the tumour itself. Protective antitumour immunity has been shown after vaccinia infection of murine tumours in vivo31. Oncolytic poxvirus therapy may therefore be considered as a method to achieve vaccination in situ, with the adaptive immune response being potentially able to clear minimal residual disease and provide long-term surveillance against relapse. This process can be augmented by the expression of transgenes that encode relevant cytokines30,31,32.

Tumour vasculature shut-down. Another mechanism that oncolytic vaccinia can use to destroy non-infected tumour cells is the induction of vascular collapse within tumours. This phenomenon has been demonstrated in both preclinical31,88 and clinical settings35 and seems to be mediated by production of chemokines and cytokines, which attract neutrophils to the tumour, resulting in intravascular thrombosis and avascular necrosis88. In addition, after intravenous administration, vaccinia strains are capable of infecting tumour-associated endothelial cells31. The destruction of these cells may also contribute to vascular collapse. It is not yet clear why tumour-associated endothelial cells are selectively targeted by vaccinia, although the EGFR–Ras pathway can be activated in these cells.

Combination therapies. Because the mechanisms of tumour cell destruction that are produced by oncolytic vaccinia viruses are typically distinct from the apoptotic mechanisms induced by chemotherapy or radiotherapy, oncolytic viruses can function effectively in combination with these traditional therapies25,89,90. Oncolytic viruses are also capable of destroying tumour cells that are resistant to traditional chemotherapeutics. The immune-stimulatory and anti-vascular mechanisms of tumour cell killing that are mediated by poxvirus therapies may also lead to synergy with immunotherapies or anti-angiogenic therapies. By contrast, targeted therapies that inhibit genetic pathways in cancer may inhibit poxvirus replication; these agents must thus be combined sequentially rather than simultaneously. Combination treatment regimens with oncolytic poxviruses will therefore require optimization.

Systemic delivery

A major advantage of poxviruses over other oncolytic viruses is their ability to travel systemically through the blood30,31,32,33,57,65. Systemic poxvirus infection may be achieved by intravenous or intratumoral administration; intratumoral injection results in replication within tumours, which is followed by shedding of the virus into the blood and infection of distant tumours. These mechanisms for virus spread have been demonstrated both in preclinical models57 and, more recently, in patients with widespread metastases36. However, because the virus is unable to naturally target tumours at the level of cell entry, only a fraction of the inoculum initially infects tumour masses and subsequent replication within the tumour is required for efficacy.

In order to improve on the natural delivery mechanisms of vaccinia through the bloodstream, investigators have developed approaches to enhance the ability of vaccinia to evade premature removal by the host immune system. In non-immunized patients, removal of circulating virus by complement and reticulo-endothelial cell-based mechanisms predominates, leading to phagocytosis of viral particles by macrophages or liver Kupffer cells. In patients who have previously been immunized, mechanisms of viral removal by neutralizing antibodies and T cells also need to be circumvented. Although immune suppression (especially the targeting of neutralizing antibodies against vaccinia91) can enhance systemic viral delivery, this approach raises safety concerns and reduces the beneficial effects that are mediated by the antitumour immune response. The alternative approach of concealing the virus from the immune system within the bloodstream has therefore received significant attention. This might be achieved by coating the virus in cationic liposomes or polymers, thus increasing the time that the virus is in circulation92. This approach might also theoretically allow a virus to be re-targeted by the incorporation of peptides or antibody domains93. Another strategy is to take advantage of the natural life cycle of the virus by introducing mutations that increase production of the EEV form of vaccinia94,95. Because the EEV form has evolved as a 'stealth' virus, which is capable of evading recognition by complement55 and neutralizing antibodies56, enhancing the production of EEV is a natural way to improve systemic delivery. Infection of tumours in vivo with mutant vaccinia that is enhanced with the EEV form leads to notable improvements in systemic tumour targeting and efficacy, even in the presence of neutralizing antibodies57.

Ex vivo infection followed by intravenous infusion of 'carrier' cells is another method of concealing oncolytic viruses during intravenous delivery to tumours96,97,98,99,100. Tumour cells that are used for this purpose allow improved delivery of viruses to the tumour99. Alternatively, the analogy of cancer as a 'wound that never heals'101 may be used to correctly predict that a variety of immune cell types naturally traffic to tumours. Various immune cell-based therapies (such as T cell102, natural killer cell103 or natural killer T cell104) rely on direct interactions between adoptively transferred immune cells and tumour cells to be effective. These tumour-targeting immune cells are therefore ideally suited to be carrier vehicles to deliver viruses to tumours. Indeed, pre-infection of certain immune cells has been shown not only to effectively traffic a viral therapy to the tumour, but also to conceal the virus from an antiviral immune response and to result in synergistic antitumour effects96,98. This synergy was dependent on interactions between the immune cells and the infected cancer cells within tumours. This finding is consistent with data that show that an immune response raised against a tumour that has been infected by vaccinia can enhance the efficacy of oncolytic vaccinia viruses31.

Armed poxviruses

Another advantage of poxviruses over other types of viruses is their ability to encode a relatively large number of transgenes62. Because oncolytic viruses such as vaccinia preferentially replicate in cancer cells, transgene expression is highly tumour selective. In addition, our understanding of the molecular biology of vaccinia has allowed synthetic vaccinia promoters to be designed. These promoters can direct high levels of transgene expression105,106, which is tied to late steps in viral replication106, thus ensuring that a transgene is only expressed from tissues that support the replication of engineered vaccinia viruses. Additionally, external regulation of transgene function was found to improve both viral replication and efficacy: repression of transgene function prevented premature clearance, and subsequent activation of the transgene after sufficient levels of vaccinia infection had been reached in the tumour led to improved efficacy of the viral therapy107.

Careful selection of transgenes is crucial. Because the virus will ultimately destroy any infected cells, a secreted transgene product that is capable of producing a complementary bystander effect in surrounding non-infected cells is attractive. In addition, it is necessary to ensure that the gene products chosen do not have direct antiviral effects and do not result in clearance of the viral vector before it is capable of destroying the tumour.

A variety of transgenes have been shown to be effective when expressed from oncolytic viruses, including vaccinia. The products of these transgenes include cytokines30,31,32 and other factors that are capable of modulating or enhancing the immune stimulatory effect of the virus108; anti-angiogenic agents109 to complement the anti-vascular effects of the virus; agents that disrupt the extracellular matrix to improve viral spread within the tumour110; and prodrug-converting enzymes to selectively convert non-toxic prodrugs to toxic products within the tumour111,112.

It is also necessary to consider the natural virulence gene products of vaccinia when choosing transgenes because the virus expresses many gene products that might interfere with the functions of certain transgenes113. For example, IFNβ expression has been used in conjunction with the deletion of a viral gene whose product binds and sequesters type I IFN — a deletion that also targets the virus to tumours. This combination was shown to simultaneously restrict viral replication to tumour tissues and enhance antitumour effects31.

In addition to enhancing the antitumour effects and selectivity of the virus, transgenes can be used for imaging purposes114. Because reporter gene expression can be linked to viral replication115, imaging can be incorporated into the preclinical and clinical development of poxvirus therapies as an early indicator of therapeutic action. For example, expression of genes such as luciferase or those encoding fluorescent proteins can be combined with optical imaging in preclinical testing116. Genes such as the sodium iodide symporter117 or the human somatostatin receptor118 can be expressed from the virus and used in conjunction with uptake of the appropriate positron emission tomography tracers to image the level and biodistribution of viral infection in the clinic. Alternatively, viral expression of the transferrin receptor can be determined by magnetic resonance imaging119. In fact, the ability of oncolytic vaccinia strains to be targeted to tumours and systemically delivered has even been proposed as a tool to visualize and quantify tumour burden in animal models and in patients65. Furthermore, combining different deletions that allow viruses to be targeted to tumours with reporter gene expression might be a non-invasive method to predict the sensitivity of tumours to other targeted therapies on the basis of their susceptibility to specific viral replication.

Clinical results and proof of concept

Vaccinia has been used globally in hundreds of millions of humans as the live vaccine for the eradication of smallpox39. Serious adverse events are rare (approximately 1 per 1,000–40,000 vaccinations) and include vaccinia necrosum (in patients who are severely T cell deficient), encephalitis, myopericarditis and eczema vaccinatum40,41,120. This extensive clinical experience allows investigators to exclude rare patient populations who may be at increased risk for toxicity. Cancer treatment with vaccinia agents has included cancer vaccines, non-engineered oncolytic vaccine strains and, more recently, targeted and armed vaccinia products.

Many vaccinia cancer vaccine trials have been reviewed elsewhere121,122,123. Non-engineered live vaccinia vaccine strains were used for oncolytic and immunostimulatory effects in several trials in the 1970s and 1990s (n = 48 patients), primarily by superficial tumour injection in melanoma patients124,125,126,127,128. Treatment was well tolerated with only mild, transient, flu-like symptoms; tumour responses were noted at the injection sites but distant responses were rare. A subsequent pilot trial evaluated vaccinia instillation in the bladders of four patients with bladder cancer129; treatment was well tolerated and resulted in intratumoral viral replication and tumour necrosis.

Clinical data has now been published for one targeted and armed oncolytic poxvirus36,130 (Supplementary information S1 (table)). JX-594 (Jennerex Biotherapeutics), a Wyeth strain vaccinia with inactivation of TK and transgenic expression of GM-CSF, under control of a synthetic early and late promoter30, is the first of this type of targeted oncolytic poxvirus to be used in the clinic. Efficacy of this virus results from direct oncolysis, antitumoral immunity that can be augmented by GM-CSF and tumour vascular shutdown35. In a Phase I pilot trial, low doses of JX-594 were injected into superficial melanomas in seven patients, resulting in tumour responses at the site of injection, including two complete responses, in five of the seven patients (71%)130. Distant skin metastases also responded to JX-594, whereas visceral metastases did not.

A formal Phase I–II dose-escalation trial of JX-594 was recently completed in 14 heavily pretreated patients with advanced hepatocellular, colorectal or lung cancer or melanoma36. The enrolled patients had refractory primary or metastatic tumours in the liver that were histologically confirmed and amenable to image-guided intratumoral injections. The patients were enrolled sequentially in four dose cohorts and treated once every 3 weeks. They received a median of four cycles and a maximum of eight cycles of treatment. The patients experienced flu-like symptoms and transient dose-related thrombocytopenia. Grade III hyperbilirubinaemia was dose-limiting in patients at the highest dose level and 109 plaque-forming units was the maximum tolerated dose. No liver or other organ toxicity was reported.

As expected, replication-dependent dissemination of JX-594 in blood was observed for over 2 weeks after injection, with resultant infection of non-injected tumour sites36. GM-CSF expression resulted in significant increases in blood neutrophil concentrations at the maximum tolerated dose. Three patients had objective injected-tumour responses, six had stable disease and only one patient had progressive disease, as determined using RECIST (response evaluation criteria in solid tumours). By Choi criteria (by which a response is defined as a 10% decrease in tumour size or a 15% decrease in tumour density on computed tomography scan131), 80% of patients responded. The responses in patients with advanced refractory hepatitis B virus-associated hepatocellular cancer were particularly notable. In all three of these patients, JX-594 treatment resulted in tumour vascular shutdown, tumour destruction and suppression of underlying hepatitis B virus replication35. A Phase II trial is now underway in this patient population.

A Phase I trial of intravenous JX-594 administration is also underway. In addition, a Phase I trial with JX-929 (Jennerex Biotherapeutics), a vaccinia with deletions in both TK and VGF genes (that is, vvDD)112 has recently commenced (NCT00574977). Finally, JX-963, which is a vvDD virus product armed with GM-CSF32, is scheduled to enter a clinical trial in 2009 (Supplementary information S1 (table)).

Future prospects

The first targeted and armed oncolytic poxvirus, JX-594, has confirmed the potential of this novel class of products in patients with end-stage cancers that are refractory to treatment. For the first time in the virotherapy field, intratumoral injection and replication led to reproducible delivery of the virus to systemic metastases through the bloodstream.

Nevertheless, as with any new therapeutic class, hurdles remain. First, balancing the benefits and drawbacks of the host immune response will be paramount. Although antibodies and CTLs can clear the virus, efficacy can also be increased by both neutrophil recruitment and induction of cancer-specific cellular immunity. Although efficacy with repeat intratumoral dosing was feasible despite the production of antibodies to JX-594, repeat intravenous dosing has not yet been assessed in patients. Second, tumour resistance mechanisms, such as deposits of extracellular matrix blocking viral spread, need to be elucidated, and non-vascularized micrometastases need to be reached. Biosafety issues must be continuously addressed to ensure the safety of patient contacts. These issues include shedding of the virus to the environment (no shedding has been detected from patients treated with JX-594 to date); in this respect, the lack of airborne spread of vaccinia is a crucial advantage. Finally, as trials progress, patient selection will be vital to ensure optimum benefit and regulatory approvals, and genetic profiling might identify tumours that will be sensitive to treatment. For example, EGFR pathway activation and cellular TK levels might predict JX-594 or vvDD efficacy.

Other targeted and GM-CSF-armed virus product classes are also in clinical trials. For example, adenoviruses132 and herpes viruses (Oncovex; Biovex)133 are in clinical trials for patients with bladder cancer or melanoma, respectively. To date, investigators have reported clear safety and anticancer activity with both agents in Phase I and II trials. The lack of intravenous spread to distant metastases with these viruses may limit applications to locally advanced tumours. Because cross-resistance between these virotherapy products is unlikely, combination virotherapy could be explored in the future.

Although hurdles remain, the class of targeted and armed oncolytic poxviruses holds great promise, and these products might soon be a vital part of the cancer treatment armamentarium.