Neoadjuvant use of oncolytic herpes virus G47Δ prevents local recurrence after insufficient resection in tongue cancer models

Abstract

A complete resection of tongue cancer is often difficult. We investigate the usefulness of administering G47Δ (teserpaturev), a triple-mutated oncolytic herpes simplex virus type 1, prior to resection. G47Δ exhibits good cytopathic effects and replication capabilities in all head and neck cancer cell lines tested. In an orthotopic SCCVII tongue cancer model of C3H/He mice, an intratumoral inoculation with G47Δ significantly prolongs the survival. Further, mice with orthotopic tongue cancer received neoadjuvant G47Δ (or mock) therapy with or without "hemilateral" resection, the maximum extent avoiding surgical deaths. Neoadjuvant G47Δ and resection led to 10/10 survival (120 days), whereas the survivals for G47Δ alone and resection alone were 6/10 and 5/10, respectively: all control animals died by day 11. Furthermore, 100% survival was achieved with neoadjuvant G47Δ therapy even when the resection area was narrowed to "partial," providing insufficient resection margins, whereas hemilateral resection alone caused death by local recurrence in half of the animals. G47Δ therapy caused increased number of tumor-infiltrating CD8⁺ and CD4⁺ cells, increased F4/80⁺ cells within the residual tongues, and increased expression of immune-related genes in and around the tumor. These results imply that neoadjuvant use of G47Δ is useful for preventing local recurrence after tongue cancer surgery.

Keywords: G47Δ; head and neck cancer; immunotherapy; local recurrence; neoadjuvant therapy; oncolytic virus therapy; squamous cell carcinoma; teserpaturev; tongue cancer; tumor microenvironment.

Graphical abstract

Figure 1

Oncolytic activities of G47Δ in vitro and in vivo (A) In vitro cytotoxicity assay. Murine and human head and neck cancer cell lines were seeded in six-well plates at 2 × 10⁵ cells/well except for SCCVII cells (2 × 10⁴ cells/well). After incubation overnight, the cells were inoculated with G47Δ at various MOIs (0.01, 0.1, or 1). The number of surviving cells was counted daily and is expressed as the raw number (top) or as the percentage of the number of mock-infected control cells on each day (bottom). Data are presented as the mean of triplicates ± SD. (B) In vitro viral replication assay. Cells were seeded in six-well plates at 2 × 10⁵ cells/well. Triplicate wells were infected with G47Δ at an MOI of 0.01. At 24 and 48 h after inoculation, the cells were collected and progeny virus titered by plaque-formation assay using Vero cells. Data are presented as the mean of triplicates ± SEM. A horizontal line indicates the initial number of virus added to each well (2 × 10³ pfu/well). (C) Treatment of orthotopic tongue cancer with G47Δ alone in immunocompetent mice. C3H/He mice with established SCCVII tongue tumors were inoculated with G47Δ (1 × 10⁶ pfu) or mock (PBS) when tumor volumes reached 35–40 mm³ (day 0). (D) Efficacy of G47Δ assessed by survival curve. (E) Efficacy of G47Δ assessed by tumor growth. (F) Body weight of animals treated with mock or G47Δ. Data are presented as the mean ± SEM; n = 10 per group. Two-tailed Student’s t test at day 4 (A) and at day 6 (E and F) and log-rank test (D). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant.

Figure 2

Planning of the extent of tongue resection for neoadjuvant G47Δ therapy (A) Schematic drawings showing the areas of resection of the tongue according to the five resection types. Black lines indicate the resection lines. (B) Survival curve of C3H/He mice according to the resection type (n = 10 per group). (C) Body-weight changes in C3H/He mice according to the resection type (n = 10 per group). Results represent mean ± SEM. (D) Resection plans of tongue tumors for assessing the efficacy of neoadjuvant G47Δ therapy. (E) Histological evaluation of resected tumor and residual tongue based on the proposed resection margins. C3H/He mice with established SCCVII tongue tumors were subjected to resection and the resulting tissues were examined for residual microtumors. A resection with a margin of 1–1.5 mm results in almost complete removal of microtumors in the remaining tongue. In contrast, residual microtumors are observed in the remaining tongue after resection with a margin of 0–0.5 mm. Scale bars: ×2.5, 1 mm; ×20, 100 μm.

Figure 3

Efficacy of tumor resection with neoadjuvant G47Δ therapy (A) C3H/He mice with established orthotopic SCCVII tongue tumors were inoculated with G47Δ (1 × 10⁶ pfu) or mock (PBS) on day 0 when tumor volumes reached 35–40 mm³, and the tongues harboring tumors were partially resected on day 3 by hemilateral resection. (B and C) Representative histopathology of resected tongue tissues with tumors. (B) Resected tongue tissues with tumors without neoadjuvant G47Δ therapy. (C) Resected tongue tissues with tumors with neoadjuvant G47Δ therapy. Representative histopathology from each mouse is presented. Asterisks indicate tumor resections with sufficient margins defined as >1 mm between the resection line and the tumor boundary. Hematoxylin and eosin staining. Blue lines depict actual resection lines. Scale bars: 500 μm. (D) Efficacy of tumor resection with neoadjuvant G47Δ therapy assessed by survival curve (n = 10 per group). (E) Efficacy of tumor resection with neoadjuvant G47Δ therapy assessed by tumor growth (n = 10 per group). Results are the mean ± SEM. (F) Body-weight changes in mice with or without tumor resection and with or without neoadjuvant G47Δ therapy (n = 10 per group). Results are the mean ± SEM. Log-rank test (D) and two-way ANOVA followed by Tukey’s multiple comparison test (E and F). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant.

Figure 4

Efficacy of neoadjuvant G47Δ therapy prior to resections with insufficient margins (A) Modified resection plan of tongue tumors. Tumors treated with G47Δ were resected with purposely insufficient margins of 0–0.5 mm. (B and C) Representative histopathology of resected tongue tissues with tumors. (B) Resected tongue tissues with tumors without neoadjuvant G47Δ therapy. (C) Resected tongue tissues with tumors with neoadjuvant G47Δ therapy. Hematoxylin and eosin staining. Blue lines depict actual resection lines. Scale bars: 500 μm. (D) Survival curve. C3H/He mice with established orthotopic SCCVII tongue tumors were inoculated with G47Δ (1 × 10⁶ pfu) or mock (PBS) on day 0 when tumor volumes reached 35–40 mm³, and the tongues harboring tumors were partially resected according to the modified resection plans on day 3 (n = 10 per group). (E) Efficacy of neoadjuvant G47Δ therapy assessed by tumor growth (n = 10 per group). Results are the mean ± SEM. (F) Body-weight changes in mice with or without neoadjuvant G47Δ therapy (n = 10 per group). Results are the mean ± SEM. Log-rank test (D) and two-way ANOVA followed by Tukey’s multiple comparison test (E and F). ∗p < 0.05; ∗∗∗∗p < 0.0001; ns, not significant.

Figure 5

Intratumoral administration of G47Δ stimulates the host immune response (A) C3H/He mice with established orthotopic SCCVII tongue tumors were inoculated intratumorally with G47Δ (1 × 10⁶ pfu) or mock (PBS) on day 0, and tumors were collected on day 7. Immune cells infiltrating the tumors were evaluated by flow cytometry. Responders were defined as mice with tumors that were reduced in size on day 7 compared with day 0, and non-responders were defined as mice with tumors that did not change or increased in size. (B) Tumor-infiltrating lymphocyte populations. Responders and non-responders both showed significantly greater numbers of tumor-infiltrating CD8⁺CD4⁻ T cells compared with mock-treated mice (n = 4 per group). Tumor-infiltrating CD8⁺PD-1⁺ and CD8⁺TIM-3⁺ cell populations were comparable among the three groups. (C) CD39 expression in tumor cells. The median expression rate of CD39 in tumor cells tended to be low in G47Δ responders, although there were no statistical differences among three groups. (D) C3H/He mice with established orthotopic SCCVII tongue tumors were inoculated intratumorally with G47Δ (1 × 10⁶ pfu) or mock (PBS) on day 0, and tumors were resected on day 3 and residual tongues extracted on day 15. Immune cells in the residual tongues were evaluated by immunohistochemistry. (E) Sections of residual tongue were immunostained with indicated antibodies. Immunohistochemistry (IHC)-positive rates (the number of IHC-positive cells/total cells) for each immune cell type inside the residual tumor area (solid line) or outside of the area (dashed line) in mock- or G47Δ-treated mice were calculated automatically using PatholoCount (version 1.0; Mitani, Tokyo, Japan). Results of representative areas are shown. HE, hematoxylin and eosin. Scale bars: ×2.5, 1 mm; ×20, 100 μm. (F) Immunopositive immune cell counts inside or outside of the residual tumor area in mock- or G47Δ-treated mice. Results are mean ± SEM. One-way ANOVA followed by Tukey’s multiple comparison test (B and C) and two-tailed Student’s t test (F). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ns, not significant.

Figure 6

Inoculation with G47Δ significantly altered gene expression in the tumor microenvironment (A) C3H/He mice with established orthotopic SCCVII tongue tumors were inoculated with G47Δ (1 × 10⁶ pfu) or mock (PBS) on day 0, and tumors and adjacent tongue tissues were collected on day 5. Gene expression in the collected tissues was determined by RNA-seq. (B) Principal-component analysis. M1–M5, samples from mock-treated mice; R1–R5, samples from G47Δ responder mice; NR1–NR5, samples from G47Δ non-responder mice; G1–G5, samples from healthy tongue inoculated with G47Δ; T1–T4, samples from healthy tongue without G47Δ inoculation. (C) Microarray (MA) plots of differentially expressed genes (DEGs). Relative log expression was normalized based on lead count value using the DESeq2 software, and DEGs were extracted. One-way ANOVA was used to filter variable genes and the Wald test was used to extract DEGs. Genes with |log2 fold change| > 1 and p < 0.05 were selected. (D) Heatmap of DEGs in mock-treated mice (M1–M5), non-responders (NR1–NR5), and responders (R1–R5). Genes are categorized into four clusters. The gene labels of each cluster are described in Table S2. (E) Top 20 gene ontology enrichment terms of DEGs. Bar length indicates the number of DEGs in that gene class, and the color reflects the p value of the enriched term. The p value is higher from red to blue.