Phase II, investigator-initiated, single-arm study demonstrates the efficacy and safety of G47∆, a triple-mutated, third-generation oncolytic herpes simplex virus type 1 (HSV-1), in 19 adult patients with residual or recurrent, supratentorial glioblastoma after radiation therapy and temozolomide. The 1-year survival rate of 84.2%, the median overall survival (OS) of 20.2 months and progression-free survival (PFS) of 4.7 months after G47Δ initiation compare favourably with other treatments. The study findings are published by Prof. Tomoki Todo of the The Institute of Medical Science, The University of Tokyo in Tokyo, Japan and colleagues on 21 July 2022 in the Nature Medicine. G47∆ is the first oncolytic virus product that has been approved in Japan.

With a median OS of 20.9 months, glioblastoma has a poor prognosis despite the current Stupp regimen of radiotherapy plus temozolomide together with tumour-treating fields (TTF). This regimen with TTF is the current standard-of-care for newly diagnosed glioblastoma in the US, but its recommendation level varies among other countries. None of the current therapies can yet prevent recurrence of glioblastoma, not to cure it.  

G47Δ was confirmed safe in the first-in-human study when administered intratumourally, two doses within 2 weeks, to patients with recurrent glioblastoma. Preclinical studies show that G47∆ exhibits efficacy via two mechanisms: an immediate effect via virus replication and direct oncolytic activity, and a delayed effect via induction of specific antitumour immunity.

The authors explained that this phase II study was designed as a single-arm because in clinical studies for lethal diseases in Japan, setting a non-curable standard-care control arm would not be accepted, and also would be unethical, especially if the study was an academia-initiated, research grant-supported drug development such as this one. Further, because G47∆ needs to be administered by surgery, blinding treatment arms and allowing sham surgery on randomly selected patients would not be considered acceptable in Japan.

G47Δ was administered intratumorally by magnetic resonance imaging (MRI)-guided stereotactic surgery at intervals of 5–14 days for the first and second doses, and up to 6 doses at intervals of 4 ± 2 weeks for the third and subsequent doses. A total of 1 × 109 p.f.u. per dose in 1 ml of solution was divided equally and injected into 1–3 coordinates within a tumour using the Biopsy/Injection Needle device which was specifically designed for oncolytic virus injection. The device consists of an outer guide (cylinder), a biopsy needle and an injection needle, and either of the latter two needles can be inserted into the outer guide without moving the position of the guide. Therefore, the coordinates of biopsy and G47∆ injection were exactly the same. Biopsy always preceded G47∆ injection. G47∆ was injected manually and after injection, the injection needle was kept in place for 5 min before retraction to avoid reflux. From the second dose onwards, G47∆ was injected into viable tumour sites, depicted as contrast-enhanced portions on MRI, remaining from previous injections.

The study primary endpoint was the 1-year survival rate after G47∆ treatment initiation. Secondary endpoints included OS and PFS after initial G47Δ administration, tumour response for efficacy and adverse event frequency.

The primary endpoint of 1-year survival rate after G47∆ initiation was 84.2% (95% confidence interval, 60.4–96.6; 16 of 19). The prespecified endpoint was met and the study was terminated early. Regarding secondary endpoints, the median OS was 20.2 months after G47∆ initiation and 28.8 months from the initial surgery; PFS was 4.7 months after G47Δ initiation.

MGMT methylation status was available for 5 patients (26.3%) from referring hospitals; 3 were unmethylated and 2 were methylated. The paraffin-embedded slide sections of initial surgery provided from referring hospitals and the biopsy specimens from this study were not enough to extract sufficient amounts of DNA to perform methylation-specific PCR. Alternatively, the study team performed immunohistochemistry (IHC) for the expression of MGMT using the provided paraffin-embedded slide sections post hoc. MGMT IHC was negative in 11 of 19 patients (57.9%), positive in 8 of 19 (42.1%) and strongly positive in none. Median OS was not affected by MGMT expression, both after G47Δ initiation and from the initial surgery.

Since the status of IDH1 mutation was not included in the eligibility criteria, it was examined as a post hoc study, and 6 of 19 patients were found to have IDH1-mutated tumours. However, the difference in IDH1 mutation was shown to have no impact on the OS both after G47∆ initiation and from the initial surgery in this study.

On MRI, enlargement of and contrast-enhancement clearing within the target lesion repeatedly occurred after each G47∆ administration, which was characteristic to this therapy. Thus, the best overall response in 2 years was partial response in 1 patient and stable disease in 18 patients. Biopsies revealed increasing numbers of tumour-infiltrating CD4-positive/CD8-positive lymphocytes and persistent low numbers of Foxp3-positive cells.

The most common G47∆-related adverse event was fever followed by vomiting, nausea, lymphocytopenia and leukopenia.

Of 15 patients who experienced disease progression after G47Δ, 9 patients received bevacizumab every 4 weeks, 3 patients received bevacizumab after reoperation and 4 patients received extended field stereotactic radiotherapy. One patient with a temozolomide allergy received nimustine as second-line therapy. After symptomatic progression, 16 patients received steroids. At the time of writing the manuscript, 16 of 19 patients had died.

During this phase II study, G47∆ was designated as a SAKIGAKE (breakthrough therapy) Product, and further as an Orphan Regenerative Medicine Product for malignant glioma by the Japanese Ministry of Health, Labor and Welfare (MHLW), allowing fast-track review and approval. Hence, this phase II study served as a pivotal study, and led to the conditional and time-limited approval of G47∆ for malignant glioma by the MHLW on 11 June 2021 as a Gene Therapy Product, the first oncolytic virus drug in Japan.

The study population was rather small, with 97 surgeries in 19 patients, as this was an academia-initiated drug development for rare cancer in the Japanese medical system under Japanese regulations for gene therapy. It is planned that all patients using commercially distributed G47∆ be registered and followed, and clinical data evaluated against a control population of patients under supervision of the Japanese Pharmaceuticals and Medical Devices Agency in the next 7 years.

The authors commented that G47Δ is perhaps the first new drug since temozolomide and the first new treatment since TTF that shows a survival benefit for glioblastoma, and provides a potential cure in a proportion of patients. To date, G47Δ has been shown to be efficacious via the same mechanism of action in various solid tumours in vivo, including prostate cancer, gastric cancer, hepatocellular carcinoma, tongue cancer, oesophageal cancer, breast cancer, neuroblastoma and malignant peripheral nerve sheath tumour. Oncolytic HSV-1, including G47∆, does not infect normal bone marrow-derived cells, but recent studies show that even some haematological malignancies are susceptible to G47∆. Further, G47Δ has been shown to have augmented efficacy when used in combination with immune checkpoint inhibitors.

This research was supported in part by grants from the Japanese Agency for Medical Research and Development. 


Todo T, Ito H, Ino Y, et al. Intratumoral oncolytic herpes virus G47∆ for residual or recurrent glioblastoma: a phase 2 trial. Nature Medicine; Published online 21 July 2022. DOI: