Date Published: January 31, 2012
Publisher: Hindawi Publishing Corporation
Author(s): Mark S. Ferguson, Nicholas R. Lemoine, Yaohe Wang.
Despite recent advances in both surgery and chemoradiotherapy, mortality rates for advanced cancer remain high. There is a pressing need for novel therapeutic strategies; one option is systemic oncolytic viral therapy. Intravenous administration affords the opportunity to treat both the primary tumour and any metastatic deposits simultaneously. Data from clinical trials have shown that oncolytic viruses can be systemically delivered safely with limited toxicity but the results are equivocal in terms of efficacy, particularly when delivered with adjuvant chemotherapy. A key reason for this is the rapid clearance of the viruses from the circulation before they reach their targets. This phenomenon is mainly mediated through neutralising antibodies, complement activation, antiviral cytokines, and tissue-resident macrophages, as well as nonspecific uptake by other tissues such as the lung, liver and spleen, and suboptimal viral escape from the vascular compartment. A range of methods have been reported in the literature, which are designed to overcome these hurdles in preclinical models. In this paper, the potential advantages of, and obstacles to, successful systemic delivery of oncolytic viruses are discussed. The next stage of development will be the commencement of clinical trials combining these novel approaches for overcoming the barriers with systemically delivered oncolytic viruses.
Cancer remains a major health problem and is the 5th leading cause of death worldwide . There have been many advances in the last few decades both in surgical care and chemoradiotherapy regimes. Certainly this has contributed to improved survival rates for commonly occurring cancers. However, relapse and disease progression are still all too common occurrences in modern medical practice. A variety of novel adjuvant therapies have been developed over the last decade, and oncolytic viruses have been particularly promising members of this cohort.
There have been many clinical trials of a variety of OVs delivered systemically, as summarised in Table 1. Oncolytic adenovirus was one of the first oncolytic viruses to be developed and licensed for treatment of cancer [8, 24]. The first generation of oncolytic adenovirus, ONYX-015 (also known as dl1520, H101 in China), is a genetically modified adenovirus with deletion of the 55 kD gene in the E1B region. Nemunaitis et al.  in 2001 performed a dose escalation study using this agent in patients with advanced carcinoma with lung metastases. They demonstrated that ONYX-015 was safe to deliver systemically with no toxicity up to doses of 2 × 1013 particles, but the study was not designed for objective tumour responses. Also commencing in 2001, a succession of studies delivered ONYX-015 via hepatic artery infusion for the treatment of metastatic colorectal carcinoma with liver deposits [11–13]. In the first of these trials, a phase I dose escalation study, one patient (9%) responded after combination therapy with conventional chemotherapy and two patients (18%) had stable disease lasting several months . In a larger phase II follow-up trial, three patients (11%) had partial responses, nine (33%) had stable disease, and eleven (41%) patients had progressive disease . A final phase II trial by this group demonstrated similar results to the previous studies with overall median survival of 10.7 months with two patients (8%) having a partial response and a further eleven (46%) having stable disease . Of those with stable disease the median survival was prolonged to nineteen months. In a different study, Small et al.  treated patients with hormone-refractory metastatic prostate cancer using a single intravenous infusion. Unlike ONYX-015, the adenovirus (CG7870) in this trial was modified so that E1A was under the control of the rat probasin promoter and E1B was under the control of the PSA promoter-enhancer, thus making it prostate specific. Results from this trial were disappointing with no complete nor partial responses, although five patients (22%) did have a 25% to 49% reduction in their serum PSA values.
There are many obstacles to successful systemic delivery of viruses; host defences limit most oncolytic viruses’ ability to infect tumours after systemic administration. Blood cells, complement, antibodies, and antiviral cytokines , as well as nonspecific uptake by other tissues such as the lung, liver and spleen, tissue-resident macrophages, and additionally poor virus escape from the vascular compartment  are the main barriers to systemic delivery of oncolytic viruses (Figure 1). Clearly, in order for this method to be effective, the virus must persist in the circulation without depletion or degradation while selectively infecting tumour cells.
Preexisting immunity is a major problem for systemically delivered viruses whether this has developed due to the ubiquitous nature of the virus, previous immunization, or prior oncolytic viral therapy. Vaccinia virus was used in the worldwide immunisation program for the eradication of smallpox and so many people who are now developing cancer have a preexisting immunity to this OV. Reovirus is universally present within the environment and as a result many people have immunity to it [26, 27]. Furthermore, White et al.  have demonstrated that the antibody titre to Reovirus increases dramatically after systemic delivery and others have shown that the presence of these antibodies significantly impairs effective intravenous administration [29, 30]. One simple strategy for overcoming this problem has been to sequentially deliver related viruses with different serotypes or chimeric viruses .
Complement activation is an important antiviral mechanism. Vaccinia virus in its EEV form incorporates host proteins within its membrane that may well prevent complement activation . Furthermore, it has long been established that Vaccinia virus secretes a variety of immune-modulating molecules. One of the major secreted proteins is Vaccinia complement control protein (VCP), which binds and inactivates C4b and C3b [54–56] thus inhibiting the classic and alternative pathways of complement activation. Furthermore, there is compelling evidence from a variety of viral infection models that complement activation induces various elements within the adaptive immune system [57–62]. Recent work has suggested that VCP dampens viral antibody responses and reduces the accumulation of CD4+ and CD8+ cells at the site of infection in a complement-dependent manner . This has led to at least one group using VCP to perturb complement activation outside the context of a Vaccinia infection  and raises the possibility of using it in combination with other OVs to block complement activation.
Viral infections stimulate a variety of cytokines to be produced (for review see Randall 2008 ). These include type 1 interferons (IFN), type 2 IFN, and type 3 IFN [70, 71]. Although these molecules have pleiotropic functions, the main effects are to promote apoptosis in virus-infected cells and induce cellular resistance to viral infection in noninfected cells . Additionally, they recruit elements of the adaptive immune system, such as dendritic cells, leading to potentially lasting immunity . Most oncolytic viruses express proteins that block these IFNs [74–76], or their downstream targets, but the anti-viral response is often still sufficient to prevent intra-tumoral spread of the OV.
It is known that many viruses are either filtered or taken up by the lung, liver, or spleen thus reducing systemic availability. Our group has demonstrated that the spleen is pivotal in the early clearance of systemically delivered Vaccinia virus (unpublished data by James Tysome et al.). Furthermore, up to 90% of Adenovirus type 5 is sequestered from the blood by Kupffer cells  and as a result this acts as a major obstacle for the systemic delivery of Adenovirus.
Adenovirus is known to bind to human erythrocytes [81, 82], and this reduces its therapeutic availability when delivered systemically. Furthermore, it is well known that the neovasculature within solid tumours is very chaotic and abnormally leaky with often markedly raised interstitial pressures leading to reduced viral penetration of the tumour mass. Oncolytic viruses are known to stabilize tumour vasculature directly improving tumour penetrance . Interestingly, other work has shown that the addition of antiangiogenic agents with oncolytic viruses can further normalise the vasculature and improve viral delivery in preclinical models [84, 85]. There is also emerging evidence that blockade of the Hedgehog signaling pathway can affect tumour vasculature . Thus a Hedgehog antagonist may prove to be an effective treatment in combination with a systemically delivered oncolytic virus or indeed incorporated within one as a transgene. Another potential agent that could be incorporated into an OV as a transgene is histidine-rich glycoprotein (HRG) particularly in the context of repeated systemic administrations of OV. This protein has been shown to normalise tumour vasculature through its ability to polarize macrophages from M2-like TAM phenotype to M1-like tumour inhibitory phenotype .
Microbubbles have been developed as a potential method for enhancing the systemic delivery of a variety of agents including oncolytic viruses. They were first developed to help deliver small molecules to target tissues [88–91]. Microbubbles are ultrasound contrast agents that contain high-molecular weight gases which are less soluble and do not diffuse easily, and as a result the microbubbles persist in the circulation for a few minutes passing through the microcirculation several times . Ultrasound-targeted destruction of the microbubbles allows focused release of the oncolytic virus at the tumour site, and a secondary effect is transient and localised increased cellular permeability which potentially can improve viral infection of the cancer cells . This technique has been used in vivo with Adenovirus successfully delivering the virus to the tumour site in mice [93, 94]. The technique has not yet been used with other oncolytic viruses.
To date, most pre-clinical studies examining systemic delivery of Vaccinia virus have used nude mice bearing xenograft tumours. It is now clear that there is a need to assess systemic delivery in an immune-competent model as host immunity is a major barrier. Indeed, results from our group have demonstrated that while Vaccinia virus can effectively infect tumour cells in nude mice after systemic delivery, infection of tumour cells cannot be achieved at similar levels in the immunocompetent model. Concurrently, work in our group revealed that depletion of macrophages by clodronate liposomes dramatically enhanced Vaccinia virus infection of tumours in immunocompetent mice after systemic delivery (unpublished data by James Tysome et al.). This almost completely restored the antitumour potency to the level seen in nude mice. However, clodronate liposomes nonselectively deplete macrophages and therefore potentially diminish any beneficial activity in the tumour microenvironment unrelated to viral clearance. Consequently, this necessitates a search for a novel, more selective agent that could interfere transiently with macrophage function and thus enhance the systemic delivery of Vaccinia virus.
To date, the systemic delivery of oncolytic viruses has been shown to be safe but not efficacious mainly due to immunological factors that facilitate rapid clearance of these agents. There is a range of novel methods that are being developed at a pre-clinical level to overcome these hurdles which have been reported to be successful in vivo mainly in murine models.