By Rajiv S. Magge, MD
Assistant Professor of Neurology, Weill Cornell Medicine, Weill Cornell Brain Tumor Center
Dr. Magge reports no financial relationships relevant to this field of study.
In a Phase I/II trial, convection-enhanced delivery of an engineered poliovirus yielded similar median overall survival compared to historical controls. However, a small subgroup of treated patients had extended survival past two years.
Desjardins A, Gromeier M, Herndon JE 2nd, et al. Recurrent glioblastoma treated with recombinant poliovirus. N Engl J Med 2018;379:150-161.
Glioblastoma continues to be one of the deadliest forms of cancer and represents one of the most common primary brain tumors. Standard treatment is comprised of maximal safe resection followed by concurrent radiation therapy with temozolomide chemotherapy, as well as subsequent adjuvant chemotherapy. However, these tumors invariably progress, and effective treatments for recurrent disease remain limited. There have been significant advances in the use of immunotherapy to combat cancer with the enhancement of the body’s own host defenses. Certain types of systemic cancer, including melanoma, non-small cell lung cancer, and bladder cancer, have been shown to be highly responsive to checkpoint inhibitors, which enhance T-cell mediated tumor cell killing. Preliminary data indicate only limited efficacy of checkpoint inhibitors in glioma, likely because of its relatively bland mutational profile as well as its highly immunosuppressive microenvironment. Additional interventions probably are necessary to increase the immunogenicity of glioblastoma. One exciting avenue is the use of neurotropic viruses to boost the host immune response or even directly infect and lyse tumor cells.
Desjardins et al investigated the use of PVSRIPO, an engineered live attenuated poliovirus vaccine, for treatment of recurrent glioblastoma. Poliovirus is neurotropic and enters cells via CD155, which is upregulated in glioblastoma and expressed in antigen-presenting cells. The hope is that infection of neoplastic cells contributes to cytotoxicity and activates antiviral inflammatory responses in addition to contributing to sustained inflammation by infection of antigen-presenting cells. Importantly, the internal ribosome entry site of their engineered poliovirus vaccine was replaced with that of human rhinovirus type 2, preventing infection of healthy neuronal cells as well as improving specificity of the virus for glioma tumor cells.
The group at Duke University enrolled consecutive adult patients with recurrent supratentorial glioblastoma (WHO grade 4) who had measurable disease on imaging. A stereotactic biopsy was done first to confirm pathology, followed by catheter implantation into the tumor, and infusion of the PVSRIPO over 6.5 hours. Over a five-year period, 61 patients received a dose of PVSRIPO, initially in a dose-escalation phase and then in a dose-expansion phase. The single dose-limiting toxicity was a grade 4 intracranial hemorrhage after catheter removal (after treatment at fifth dose level). All tissue samples (available for 43 patients) stained positive for CD155 immunohistochemistry, which is the route of entry for the poliovirus. During the trial, two patients died, possibly related to treatment — one had a seizure in the setting of cerebral edema and tumor progression 4.8 months after viral infusion, and the second died 10.5 months after infusion from complications of intracranial hemorrhage while receiving both anticoagulation and bevacizumab. The most common treatment-related toxicities included grade 1 or 2 headache, hemiparesis, seizure, dysphasia, and cognitive disturbance. Patients with neurologic symptoms were limited to a maximum dose of dexamethasone 4 mg daily, but most patients were started on low-dose bevacizumab empirically for symptom control. In addition, most patients in the study were treated with additional therapies such as therapeutic-dose bevacizumab, temozolomide, and lomustine.
The median overall survival of the treated patients was 12.5 months (95% confidence interval [CI], 9.9-15.2), similar to the 11.3 months (95% CI, 9.8-12.5) in the historical control group of patients treated at Duke. However, the 21% overall survival rate at two years in the treated group remained steady at the three-, four-, and five-year mark, outperforming the control group (only 4% survival rate at three years).
COMMENTARY
This study is impressive in its size, scope, and use of a novel technique in targeting glioblastoma cells with a highly neurotropic virus. The engineered viral vaccine, PVSRIPO, enters tumor cells through CD155, but, importantly, does not infect healthy neuronal cells because of the replacement of the internal ribosome entry site. It is one of the largest trials to date (with extended clinical follow-up) investigating viral therapy in glioblastoma. Although the overall survival rate in treated patients was similar to their historical control group, a subset of patients (n = 8, 21%) had extended survival past two years. As with most studies, it is difficult to identify why these specific patients were more responsive to the trial treatment. Only one of the tumors in this group carried an IDH mutation, which defines a subgroup of glioma with significantly improved survival and response to treatment. Furthermore, six of the eight had MGMT-methylation at the time of initial diagnosis or treatment — this alteration in glioblastoma portends better survival and increased sensitivity to alkylating agents such as temozolomide. The answer may lie in the full genetic profile of the most responsive tumors.
A significant obstacle in interpreting the results is the variability in treatment after the initial single viral infusion. More than 47 of the enrolled patients received bevacizumab (including three of the extended survival group) post-infusion for symptom control. Several studies have shown that bevacizumab does not extend overall survival in glioblastoma; however it can extend progression-free survival and result in significant reduction in enhancing disease, which complicates response assessments based on radiographic imaging. In addition, most of the treatment group (again including three of the eight patients who lived past 24 months) received additional interventions such as therapeutic-dose bevacizumab, temozolomide, and lomustine.
Other similar trials investigating the use of neurotropic viruses in large clinical cohorts include DNX-2401 (DNAtrix) and Toca 511/Toca FC (Tocagen). Like the PVSRIPO, these viral therapies may infect and lyse cancer cells directly, as well as elicit strong inflammatory responses while countering tumor-induced immunosuppression. Even if these treatments are as effective as hoped at the tumor cell level, obstacles remain with convection-enhancing delivery, especially ensuring full distribution of the infusion across the entire tumor bed. In addition, many patients are not candidates for catheter-based treatment, often because of multifocal disease or subependymal involvement with close proximity to the ventricles (which suggests possible leptomeningeal tumor as well as prevents safe catheter implantation). However, convection-enhanced delivery of neurotropic viruses remains an exciting and novel technique that warrants further investigation.
