Viruses in the treatment of malignant glioma

Therapy for cancers of the CNS remains inadequate. Approximately 15,000 new cases of malignant glioma, the most common tumor of the CNS, are diagnosed annually in the USA [1]. The prognosis for these patients is extremely grim; with a median survival of 10–12 months, nearly all patients die from their disease within 2 years [2]. No curative treatment exists and disease prognosis has remained essentially unchanged over the past 50 years. Traditional therapies of maximal surgical resection, chemotherapy and radiation are only marginally effective in extending progression-free survival and improving quality of life. At present, malignant gliomas annually claim some 12,000 human lives. Thus, investigation of novel treatment modalities is urgently needed.

While gliomas usually remain confined to the CNS, a hallmark feature of the disease process is that, at the time of clinical presentation, it is probably already widely disseminated throughout the CNS. Whereas large focal lesions are usually the cause of symptoms, these micrometastases preclude the possibility of a curative surgical resection [3]. Even patients surviving historical attempts at treatment by complete right hemispherectomy ended up dying of disease recurrence [4].

Oncolytic viruses may represent an effective treatment of such invasive disease. Viral reproduction leading to tumor-cell lysis in one location leads to the release of viral progeny able to spread to and infect distant malignant cells. This process of self-amplification continues until either all cancer cells are destroyed or the virus is eliminated by the host immune system. Although effective oncolytic viruses must avoid inducing a potentially inactivating humoral immune response, the immune system may play a role in virally mediated destruction of tumor cells. Cancer cells infected with the virus can be recognized by elements of the cell-mediated immune system where cells presenting viral proteins through the major histocompatibility complex are marked for destruction by cytotoxic T lymphocytes [5].

In addition to being able to infect and destroy malignant cells efficiently, oncolytic viruses also must not pose a threat to normal tissue; as viral oncolytic therapy in the CNS bears the risk of neurocomplications, such as acute inflammation, meningitis/encephalitis, seizures and elevated intracranial pressures. Numerous strategies have been employed to accomplish tumor-cell specificity, including altering viral tropism for normal tissue, deleting viral virulence genes whose function is dispensable in cancer cells or placing viral genes under the control of tumor-specific promoters. Currently, clinical trials using oncolytic herpes simplex virus type 1 (HSV-1) and adenovirus (Ad) against malignant glioma are being conducted and active investigations of oncolytic poliovirus, HSV-1, Ad, reovirus, vesicular stomatitis virus and Newcastle disease virus continues [6].

Herpes simplex virus type 1

Genetic HSV-1 recombinants were the first to be considered for use in glioma oncolysis as this virus is well characterized, the wildtype form displays spontaneous tropism for glioma cells and antiherpetic therapies (acyclovir, gancyclovir) exist, which can be used in the event of viral toxicity. Genetic manipulation and/or deletion of neurovirulence genes, such as γ₁34.5, which reverses innate immune responses leading to host cell inhibition of protein synthesis, or UL39, which allows nucleotide synthesis to occur in nonmitotically active cells, creates viruses that infect and replicate preferentially in malignantly transformed cells [5]. Studies in cell lines and animal models of glioma demonstrated a survival advantage, prompting clinical trials. Phase I trials demonstrated the safety of these attenuated HSV-1s but tumor response rates were disappointingly low (0–22%) [7].

Later studies revealed that γ₁34.5 deletion not only conveyed attenuation of neurovirulence but it also required for efficient viral replication in many glioma cell lines. Viral protein synthesis in many glioma cells is inhibited in the absence of γ₁34.5, impairing the ability to infect and out-compete glioma cells [5]. Methods of increasing the efficacy of oncolytic HSV-1s conditionally expressing γ₁34.5 in glioma cells is an area of current investigation.

Adenovirus

The first oncolytic virus under clinical investigation was the ONYX-15 Ad. This virus differs from the wildtype Ad by harboring a deletion of the E1B-55K gene [8]. The product of this gene binds to and targets p53 for destruction, preventing p53-mediated apoptosis of infected cells. Theoretically, ONYX-15 was rendered able to only replicate in cancerous cells lacking functional p53. Clinical trials also demonstrated that the virus can be safely administered to glioma patients but the clinical response was again variable [9]. Some patients survived over 1 year after their disease recurred. In addition, the use of ONYX-15 has been considered for other tumor types and Phase II and III trials have either taken place, with variable results, or are still ongoing [10].

Recent studies offer explanations for the variable results seen with ONYX-15 viral oncolytic therapy. First, unlike HSV-1 or poliovirus, there is an apparent lack of tropism for most cancer cells, and second, Ad (which is not a neuropathogen) was unable to infiltrate CNS tissues distant from the inoculation site. While the receptor for the Ad5 serotype of ONYX-15, Cocksackievirus/Ad receptor (CAR), was highly expressed in selected cancer-cell lines, it was found to be minimal or absent on most cancer cells freshly obtained from tumor specimens [11]. Other studies have correlated a loss of CAR expression with increased malignant potential and it is also now apparent that ONYX-15 does not necessarily replicate preferentially in tumor cells [10]. Novel oncolytic Ads currently under research attempt to overcome these potential problems by expanding the role of secondary Ad receptors, integrins α_vβ₃ and α_vβ_5, introducing alternative genetic manipulations to achieve conditional replication in cancerous cells and genetically modifying viral-capsid proteins to shield the virus from neutralizing antibodies [12].

Poliovirus

Re-engineered polioviruses have also shown promise. Poliovirus exhibits natural tropism for glioma cells expressing its cellular receptor CD155, a cell adhesion molecule, ectopically expressed in many cancers [13]. Poliovirus neuropathogenicity has been attributed in part to cell type-specific function of an internal ribosomal entry site, which allows 5´ end-independent, cap-independent initiation of translation [14]. Replacement of this element with that from the related human rhinovirus type 2 yields an attenuated version named ‘PVS-RIPO’ that is unable to propagate in neuronal cells [15]. However, the recombinant virus retains the ability to grow in glial neoplasms [15]. Selectivity of engineered attenuated polioviruses for glioma cells is due to altered conditions for protein synthesis in cancerous cells. These promote efficient translation of the viral genome in glioma cells independent of mitotic activity or the cell cycle. Meanwhile, translation of the viral genome is thwarted in the normal CNS, eliminating neurovirulent properties inherent to poliovirus.

This virus has been administered safely to mice transgenic for CD155 and was devoid of neuropathogenic properties after intraspinal inoculation into Cynomolgus macaques [16]. It has demonstrated efficacy in animal studies with anaplastic astrocytoma and with glioblastoma multiforme neoplastic meningitis [17]. CD155 is universally expressed in established glioma-cell lines and was also abundant in primary explant cultures of glioma tissue from patients undergoing craniotomy [13]. The latter were also highly sensitive to PVS-RIPO oncolysis. Peripherally administered PVS-RIPO is highly efficient against subcutaneous xenografts: virus injected at one tumor site eliminated tumor in the contralateral flank of bilaterally tumor-injected athymic mice. However, peripheral virus administration was not effective against intracerebral tumors and intratumoral delivery may be required for such inaccessible targets [15]. PVS-RIPO is currently in investigational new drug-directed toxicology studies in preparation for Phase I clinical trials against recurrent glioblastoma multiforme.

Future directions

There are additional case reports and preclinical studies using other viruses for oncolytic therapy or using oncolytic viruses synergistically with other current therapies. Studies combining ionizing radiation and oncolytic viruses have yielded evidence that changes in the cellular environment induced by radiation may aid the ability of viruses to replicate and/or spread within the tumor [18]. Equally important is that no evidence exists that oncolytic viruses are antagonistic to standard chemotherapeutic drugs and that the effects of treatment with both may be additive. Modulation of the host immune response against the oncolytic virus may also increase efficacy. Cyclophosphamide pre-administration may act to inhibit clearance of viral particles by the host [19].

Other avenues of current research include using oncolytic viruses dually as vehicles for gene therapy or as immunomodulatory agents. Gene therapy aims to deliver foreign DNA specifically to tumor cells aimed at either halting disease progression by providing a deficient cell-cycle control gene or by increasing sensitivity to other treatments. Many groups have attempted to use the HSV 1-thymidine kinase (TK) system, in which the ganciclovir-sensitizing HSV-1-TK gene is delivered to target cells. Subsequent treatment with ganciclovir halts the division of successfully transduced cells [20]. Clinical trials using this system have taken place but results have been less robust than preclinical data would have suggested, failing to extend overall survival time in patients with malignant glioma [21].

Viral vector vaccines are another area of active research. Viruses harnessed as vectors deliver heterologous immunogenic material into host cells with the goal of inducing an immune response against cells expressing viral products [22]. Viruses engineered to express tumor-associated antigens can provoke a tumor-specific immune response by infecting professional antigen-presenting dendritic cells [23]. The viruses mimic a natural infection and provide the necessary costimulatory danger signals to induce an immune response against the tumor-specific antigen. The immune system is thus stimulated to attack the antigen-expressing tumor cells through direct lysis by activated T cells and by antibody-mediated indirect killing. Phase I and II trials of viral-vector vaccines against numerous tumor types have been conducted. While the vaccines have displayed an excellent safety profile, efficacy remains unproven.

While oncolytic viruses and other forms of viral therapy are not presently part of the standard treatment of malignant glioma, much promising research has been conducted. Currently, large clinical trials have not yet been conducted to determine their true efficacy. Similarly, research into using oncolytic viruses in combination with other established and/or experimental treatment modalities show great promise but are still under preliminary investigation.