Retargeted oncolytic viruses provoke tumor-directed T-cell responses

Abstract

Intratumoral application of oncolytic viruses effectively induce tumor-directed immune responses. However, their systemic application is typically insufficient to stimulate the required extent of tumor tissue inflammation to elicit antitumor immunity. We recently discovered evidence that this barrier can be overcome by effective molecular retargeting of viral infection.

Tumor-selectively replicating ‘oncolytic’ viruses are a promising and innovative option for the treatment of solid tumors. Different viral species such as vaccinia virus, herpes virus, adenovirus, vesicular stomatitis virus, and reovirus have been developed as oncolytic agents, or have been genetically modified to adapt viral replication to tumor cell-specific molecular alterations. Initially developed for direct tumor cell killing, it is now clear that oncolytic virotherapy exerts multiple antitumoral effects, e.g. by modulating the tumor microenvironment and tumor vasculature.^1,2 Similar to vaccination in-situ, oncolytic viruses induce tumor-directed adaptive immunity via the induction of extensive tumor cell lysis and cross-presentation of tumor-associated antigens occurring under strong inflammatory conditions.³ To further amplify this promising function, oncolytic viruses have been provided with additional immunostimulatory transgenes. In a recent Phase III trial (OPTiM study) in human patients with advanced melanoma, the GM-CSF expressing oncolytic herpes virus T-Vec has elicited a significant increase in durable response rates, thus demonstrating the therapeutic potential of virotherapy in human patients.⁴ In contrast to encouraging results after direct intralesional infiltration of the therapeutic virus, systemic application has proven remarkably ineffective in the clinic, currently limiting virotherapy to tumors that are easily accessible by percutaneous injections. Apparently, systemic application of unmodified oncolytic viruses is not sufficient to achieve critical thresholds of tumor transduction, cell lysis and inflammation needed for induction of tumor-directed immune responses. In order to accomplish effective tumor cell transduction following intravenous application of oncolytic adenoviruses, viral surface structures typically need to be profoundly modified to achieve improved biodistribution and tumor cell tropism. Bispecific adapter molecules have proven convenient and versatile tools to decorate the adenovirus capsid with an address label for tumor cells, mostly ligands or antibody single chain variable fragment (scFv) domains recognizing targets overexpressed on the surface of malignant cells. Corresponding adapters have been shown to substantially improve in vivo tumor cell infection after systemic administration.⁵ However, immunodeficient murine models are not conducive to a prediction of whether molecular retargeting strategies are sufficient to induce tumor inflammation and the induction of antitumoral immune responses.

To address this issue, we investigated molecular retargeting of oncolytic adenoviruses in immunocompetent mice with disseminated lung cancer using polysialic acid (polySia) as a novel molecular target.⁶ PolySia is a cell surface glycopolymer that is absent on normal epithelial cells in adults but is reexpressed during carcinogenesis on malignant cells of several clinically relevant tumor types including small cell lung cancer. For our study, we generated a molecular adapter comprising the ectodomain of the coxsackie adenovirus receptor (for binding to adenoviral fibers) and a polySia-recognizing scFv-domain, a construct amenable to use in both rodents and men due to the molecular identity of polySia allowing a direct transfer into clinical settings in the future. We showed that adapter-pretreatment of adenoviral vectors enabled effective and polySia-specific infection of several cell lines resistant to infection with unmodified virus. Upon systemic, intravenous delivery into tumor-bearing nude mice, the adapter improved virus uptake in subcutaneous polySia-expressing human tumors. Furthermore, consistent with other reports,⁵ coverage of the viral fibers with adapter protein reduced hepatic viral load and prevented hepatotoxicity. In order to investigate polySia retargeting-induced immune responses, we also established an immunocompetent and orthotopic murine model of disseminated lung cancer using polySia-positive, transgenic CMT64 lung cancer cells transplanted into syngeneic C57BL/6 mice. In this model, polySia-retargeting greatly enhanced the delivery of oncolytic adenovirus to lung nodules after systemic injection resulting in a substantial reduction of the pulmonary tumor load and improved survival. Interestingly, this finding could not be recapitulated in immunodeficient nude mice suggesting that tumor eradiction was primarily due to adaptive immune effects. Accordingly, histologic analyses revealed that improved transduction by polySia-retargeted oncolytic adenoviruses led to extensive infiltration of immune cells associated with large lytic tumor areas. To clarify the role of CD8⁺ T cells in mediating antitumoral activities, we focused on neoepitope-directed responses. Neoantigen-specific T cells have been shown to dominate autologous immune responses against human tumors.⁷ Furthermore, the correlation of the mutational load and responses to immune checkpoint inhibitors indicate that neoantigen-specific T cells play an important role in the impressive therapeutic success of these novel immunotherapies.⁸ In our study, we took advantage of whole transcriptome sequencing of the tumor cell line utilized and subsequent prediction of neoepitopes that can be specifically recognized by CD8⁺ T cells. We found that only tumor infection by polySia-retargeted oncolytic virus facilitated the induction of a significant CD8⁺ T-cell response against a predicted, tumor-specific neoepitope derived from the Y9H mutation in the Gsta2 gene. Finally, we showed by CD8⁺ T-cell depletion studies that antitumoral CD8⁺ T-cell responses play a decisive role in the observed, immunotherapeutic success of oncolytic virus retargeting. In summary, we demonstrated that molecular retargeting of oncolytic viruses can facilitate effective tumor transduction required for tumor inflammation and tumor-directed immune responses (Fig. 1). Apart from the role of CD8⁺ T-cell responses, further experimentation is necessary to understand how retargeted oncolytic virus infection supports the priming of antitumoral T-cell responses. It is also of particular interest to investigate the involvement of neoepitope-specific CD4⁺ T-cells which have been shown very recently to play an essential role in driving therapeutic immune responses against cancer.⁹ Since polySia can be expressed by immune cells under specific conditions,¹⁰ we also need to know whether polySia-retargeting may directly affect tumor-resident immune cells involved in homeostasis in the tumor microenvironment.

Nevertheless, our findings provide a strong justification for the intense efforts underway to molecularly retarget oncolytic viruses to achieve effective tumor cell transduction by systemic delivery.

Figure 1.

Molecular retargeting of oncolytic virus to provoke antitumor T-cell responses. Tumor therapy by systemic delivery of oncolytic virus is usually ineffective. In contrast, systemic application of adenovirus, retargeted to tumor cells using a polysialic acid-specific adapter, facilitates effective tumor inflammation subsequently leading to potent tumor-directed CD8⁺ T-cell responses.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.