Neoadjuvant systemic oncolytic vesicular stomatitis virus is safe and may enhance long-term survivorship in dogs with naturally occurring osteosarcoma

Abstract

Osteosarcoma is a devastating bone cancer that disproportionally afflicts children, adolescents, and young adults. Standard therapy includes surgical tumor resection combined with multiagent chemotherapy, but many patients still suffer from metastatic disease progression. Neoadjuvant systemic oncolytic virus (OV) therapy has the potential to improve clinical outcomes by targeting primary and metastatic tumor sites and inducing durable antitumor immune responses. Here we describe the first evaluation of neoadjuvant systemic therapy with a clinical-stage recombinant oncolytic vesicular stomatitis virus (VSV), VSV-IFNβ-NIS, in naturally occurring cancer, specifically appendicular osteosarcoma in companion dogs. Canine osteosarcoma has a similar natural disease history as its human counterpart. VSV-IFNβ-NIS was administered prior to standard of care surgical resection, permitting microscopic and genomic analysis of tumors. Treatment was well-tolerated and a "tail" of long-term survivors (∼35%) was apparent in the VSV-treated group, a greater proportion than observed in two contemporary control cohorts. An increase in tumor inflammation was observed in VSV-treated tumors and RNA-seq analysis showed that all the long-term responders had increased expression of a T cell anchored immune gene cluster. We conclude that neoadjuvant VSV-IFNβ-NIS is safe and may increase long-term survivorship in dogs with naturally occurring osteosarcoma, particularly those that exhibit pre-existing antitumor immunity.

Keywords: antitumor immunity; immunotherapy; neoadjuvant; oncolytic virus; osteosarcoma; tumor microenvironment; vesicular stomatitis virus; virotherapy.

Graphical abstract

Figure 1

Vesicular stomatitis virus replicates in and kills three established canine osteosarcoma cell lines Virus titers and cell viability following infection of canine osteosarcoma cell lines: (A) OSCA-8, (B) OSCA-40, and (C) OSCA-78 with recombinant VSVs: VSV-GFP (green), VSV-hIFNβ-NIS (blue), or VSV-cIFNβ-NIS (pink) at an MOI of 0.03.

Figure 2

Design schematic of the VIGOR study Clinical procedures following enrollment of dogs with appendicular osteosarcoma are shown. This includes diagnostic imaging to evaluate for pulmonary metastases with either computed tomography (CT) or thoracic radiographs (CXR); pre-treatment biopsy; VSV (or placebo) treatment (study day 1); and standard of care amputation and chemotherapy (consisting of intravenous carboplatin every 3 weeks for six doses) that was started on study day 21.

Figure 3

Intravenous VSV therapy is safe, and a subset of VSV-treated dogs with osteosarcoma are long-term responders (A) Kaplan-Meier survival curve with event-free survival (EFS) of dogs treated with systemic VSV in the VIGOR study, compared with two contemporary control cohorts: NCI-COTC and UMN VMC cohorts. (B) Long-term response was defined as overall survival greater then 75^th percentile of survival from the COTC cohort (479 days). Similar proportion (26%) of dogs from the UMN VMC cohort were long-term responders. Thirty-five percent of dogs from the VSV-treated dogs from the VIGOR cohort were long-term responders (overall survival >479 days).

Figure 4

Increased tumor inflammation following VSV treatment in resected osteosarcoma tissues Inflammatory infiltrates were scored by a pathologist in baseline pre-treatment tumor biopsies and post-treatment amputation-resected tumor samples (collected on study day 10). (A) Comparison of Tumor Inflammation Score (TIS) showed higher TIS in post-treatment samples from VSV-treated dogs. (B) Where matched paired pre- and post-treatment tumor samples were available for assessment, we observed a significant increase in TIS in tissues from VSV-treated dogs between baseline and 10 days post-treatment (p = 0.0027, paired t test).

Figure 5

Acute viremia following systemic VSV infusion Detection of VSV-N RNA copies in whole blood, PBMCs, and plasma samples collected at indicated time points following VSV infusion shows detection of viral RNA in blood and viral localization primarily to PBMCs.

Figure 6

RNA-seq analysis of canine tissues does not show significant changes immune gene signatures (A) VIGOR RNA-seq analysis of osteosarcoma tumor samples (pre- and post-treatment), lung metastases (where available from necropsy samples), skin biopsy, and osteosarcoma cell lines (derived from osteosarcoma tumors) clustered by tissue type. Data are log base 2 transformed mean centered and filtered for genes with high standard deviation. Unsupervised hierarchical clustering resulted in 14 gene clusters including cell cycle, immune clusters, and cluster composed primarily of genes expressed in skin samples. (B) UMAP projection of samples present in VIGOR dataset. Clear separation of OS tumors, skin and cell lines is apparently consistent with heatmap representation of the data. Samples are labeled with colors present in the legend in (A). (C–F) Zoomed in regions showing the (C) skin-specific cluster, (D) immune 1 cluster composed of genes enriched in macrophage lineage immune genes, (E) immune 2 cluster composed of genes enriched in T cell lineage immune genes, and (G) cell cycle enriched genes. (H–J) GCESS values generated by summing the genes present in each cluster for each sample are plotted in box plots representing samples grouped by sample type, treatment, and time point for (H) skin-specific GCESS, (I) immune 1 GCESS, (J) immune 2 GCESS, and (F) cell cycle GCESS. Unsupervised hierarchical clustering resulted in 14 gene clusters including cell cycle and immune clusters. GCESS scores were not significantly different in pre- vs. post-treatment VSV- and placebo-treated tumor samples.

Figure 7

Survival outcomes correlate with pre-treatment T-cell GCESS (A) Comparison of survival outcomes between placebo-treated and VSV-treated dogs. Relationship between survival outcomes and (B) immune 1 GCESS (macrophage/monocyte), (C) cell cycle GCESS, and (D) immune 2 GCESS (T cell). These plots reveal a correlation between survival and T cell GCESS in pre-treatment tumor biopsy samples (E). Statistical significance was evaluated using Analysis of Variance (ANOVA). Notably, this analysis excludes the two dogs with non-osteosarcoma bone tumors (hemangiosarcoma and rhabdomyosarcoma) and the dog that succumbed to post-operative complications following amputation surgery.