Replicating reoviruses with a transgene replacing the codons for the head domain of the viral spike.

The capacity to modify the reovirus genome facilitates generation of new therapeutic reoviruses. We describe a method for generating replication-competent reoviruses carrying a heterologous transgene. The strategy is based on the expanded-tropism reovirus mutant jin-3, which can infect cells independent of the reovirus receptor junction-adhesion molecule A (JAM-A). Jin-3 harbors a mutation in the S1 segment, resulting in a G196R substitution in the tail of the spike protein σ1. The use of the jin-3 tail-encoding S1 segment allows replacing the codons for the JAM-A-binding head domain by up to 522 nucleotides of foreign sequences, without exceeding the size of the wild-type S1 segment. We inserted the codons for the porcine teschovirus-1 2A element fused with those encoding the fluorescent protein iLOV. Replicating rS1His-2A-iLOV reoviruses were generated by co-transfection of expression plasmids for all reovirus segments. These reoviruses contain the S1His-2A-iLOV segment in the absence of the wild-type S1 segment. Density-gradient centrifugation confirmed the association of the σ1-tail fragment with the capsid. Both JAM-A-positive and -negative cells exposed to the rS1His-2A-iLOV reoviruses exhibited iLOV fluorescence, confirming the jin-3-derived expanded-tropism phenotype. These data demonstrated the feasibility of generating decapitated replication-competent T3D reoviruses carrying a heterologous transgene.

Mammalian orthoreovirus T3D infects U-118 MG cell spheroids independent of junction adhesion molecule-A.

In the canonical pathway, infection of cells by the wild-type mammalian orthoreovirus Type 3 Dearing (T3D) is dependent on the interaction of the viral spike protein σ1 with the high-affinity cellular receptor junction adhesion molecule-A (JAM-A). We previously demonstrated that the human glioblastoma cell line U-118 MG does not express JAM-A and resists reovirus T3D infection in standard cell culture conditions (SCCC). Heterologous JAM-A expression sensitises U-118 MG cells to reovirus T3D. Here we studied reovirus infection in U-118 MG cells grown in spheroid cultures with the premise that cells in such cultures resemble cells in tumours more than those grown under standard adherent cell culture conditions on a plastic surface. Although the U-118 MG cells in spheroids do not express JAM-A, they are susceptible to reovirus T3D infection. We show that this can be attributed to factors secreted by cells in the spheroids. The concentration of active extracellular proteases cathepsin B and L in the medium of spheroid cultures was increased 19- and 24-fold, respectively, as compared with SCCC. These enzymes can convert the reovirus particles into a form that can infect the U-118 MG cells independent of JAM-A. Taken together, these data demonstrate that infection of tumour cells by wild-type reovirus T3D is not strictly dependent on the expression of JAM-A on the cell surface.

Heterogeneous reovirus susceptibility in human glioblastoma stem-like cell cultures.

Glioblastoma (GB) is a devastating disease for which new treatment modalities are needed. Efficacious therapy requires the removal of stem-cell like cells, these cells drive tumor progression because of their ability to self-renew and differentiate. In glioblastoma, the GB stem-like cells (GSC) form a small population of tumor cells and possess high resistance to chemo and radiation therapies. To assess the sensitivity of GSC to reovirus-mediated cytolysis, a panel of GSC cultures was exposed to wild-type reovirus Type 3 Dearing (T3D) and its junction adhesion molecule-A (JAM-A)-independent mutant, jin-1. Several parameters were evaluated, including the fraction of cells expressing the JAM-A reovirus receptor, the fraction of cells synthesizing reovirus proteins, the number of infectious reovirus particles required to reduce cell viability, the amount of infectious progeny reovirus produced and the capacity of the reoviruses to infect the GSC in 3-dimensional (3D) tumor cell spheroids. Our data demonstrate a marked heterogeneity in the susceptibility of the cultures to reovirus-induced cytolysis. While in monolayer cultures the jin-1 reovirus was generally more cytolytic than the wild-type reovirus T3D, in the 3D GSC spheroids, these viruses were equally effective. Despite the variation in reovirus sensitivity between the different GSC cultures, our data support the use of reovirus as an oncolytic agent. It remains to be established whether the variation in the reovirus sensitivity correlates with a patient's response to reovirus therapy. Moreover, our data show that the expression of the JAM-A receptor is not a major determinant of reovirus sensitivity in 3D GSC cultures.

A strategy for genetic modification of the spike-encoding segment of human reovirus T3D for reovirus targeting.

Human Orthoreovirus Type 3 Dearing is not pathogenic to humans and has been evaluated clinically as an oncolytic agent. Its transduction efficiency and the tumor cell selectivity may be enhanced by incorporating ligands for alternative receptors. However, the genetic modification of reoviruses has been difficult, and genetic targeting of reoviruses has not been reported so far. Here we describe a technique for generating genetically targeted reoviruses. The propagation of wild-type reoviruses on cells expressing a modified sigma 1-encoding segment embedded in a conventional RNA polymerase II transcript leads to substitution of the wild-type genome segment by the modified version. This technique was used for generating reoviruses that are genetically targeted to an artificial receptor expressed on U118MG cells. These cells lack the junction adhesion molecule-1 and therefore resist infection by wild-type reoviruses. The targeted reoviruses were engineered to carry the ligand for this receptor at the C terminus of the sigma 1 spike protein. This demonstrates that the C terminus of the sigma 1 protein is a suitable locale for the insertion of oligopeptide ligands and that targeting of reoviruses is feasible. The genetically targeted viruses can be propagated using the modified U118MG cells as helper cells. This technique may be applicable for the improvement of human reoviruses as oncolytic agents.

Transient infection of freshly isolated human colorectal tumor cells by reovirus T3D intermediate subviral particles.

Reovirus T3D preferentially kills tumor cells expressing Ras oncogenes and has shown great promise as an anticancer agent in various preclinical tumor models. Here, we investigated whether reovirus can infect and kill tumor cell cultures and tissue fragments isolated from resected human colorectal tumors, and whether this was affected by the presence of endogenous oncogenic KRAS. Tissue fragments and single-cell populations isolated from human colorectal tumor biopsies were infected with reovirus virions or with intermediate subviral particles (ISVPs). Reovirus virions were capable of infecting neither single-cell tumor cell populations nor small fragments of intact viable tumor tissue. However, infection of tumor cells with ISVPs resulted in transient viral protein synthesis, irrespective of the presence of oncogenic KRAS, but this did not lead to the production of infectious virus particles, and tumor cell viability was largely unaffected. ISVPs failed to infect intact tissue fragments. Thermolysin treatment of tumor tissue liberated single cells from the tissue and allowed infection with ISVPs, but this did not result in the production of infectious virus particles. Immunohistochemistry on tissue microarrays showed that junction adhesion molecule 1, the major cellular reovirus receptor, was improperly localized in the cytoplasm of colorectal tumor cells and was expressed at very low levels in liver metastases. This may contribute to the observed resistance of tumor cells to reovirus T3D virions. We conclude that infection of human colorectal tumor cells by reovirus T3D requires processing of virions to ISVPs, but that oncolysis is prevented by a tumor cell response that aborts viral protein synthesis and the generation of infectious viral particles, irrespective of KRAS mutation status.

Immunosuppression promotes reovirus therapy of colorectal liver metastases.

Mortality due to colorectal cancer (CRC) is high and is associated with the development of liver metastases. Approximately 40% of human CRCs harbor an activating mutation in the KRAS oncogene. Tumor cells with activated KRAS are particularly sensitive to Reovirus T3D, a non-pathogenic oncolytic virus. The efficacy of virus-based therapies may be positively or negatively modulated by the host immune system. This study was designed to assess the effect of immunosuppression on Reovirus T3D oncolysis of established colorectal micrometastases in the liver. Mouse C26 CRC cells harbor a mutant Kras gene and are susceptible to Kras-dependent oncolysis by Reovirus T3D in vitro. Isolated C26 liver tumors were established in syngenic immunocompetent BALB/c mice by intrahepatic injection. Reovirus T3D therapy was given as a single intratumoral injection in control mice and in cyclosporin A-treated immunosuppressed mice. Tumor growth was analyzed over time by non-invasive bioluminescence imaging. The outgrowth of established CRC liver metastases in immunocompetent mice was efficiently but temporarily inhibited with a single injection of Reovirus T3D. Immunosuppression with cyclosporin A markedly increased and prolonged the therapeutic effect and allowed complete Reovirus T3D-induced tumor eradication in a subpopulation of the mice. We conclude that Reovirus T3D is an effective therapeutic agent against established C26 colorectal liver metastases and that immunosuppression enhances treatment efficacy. Cancer Gene Therapy (2006) 13, 815-818. doi:10.1038/sj.cgt.7700949; published online 10 March 2006.

Creation of immune 'stealth' genes for gene therapy through fusion with the Gly-Ala repeat of EBNA-1.

A major obstacle in gene-therapy protocols is T-cell-mediated destruction of transgene-expressing cells. Therefore new approaches are needed to prevent rapid clearance of transduced cells. We exploited the Gly-Ala repeat (GAr) domain of the Epstein-Barr virus nuclear antigen-1, since the GAr prevents cytotoxic T-lymphocyte-epitope generation. Here we show that three different enzymes (viz. the E. coli LacZ gene encoded beta-galactosidase, firefly luciferase, and HSV1 thymidine kinase) fused with the GAr retained their function. Moreover, linking GAr with beta-galactosidase successfully prevented recognition of GAr-LacZ-expressing cells by beta-galactosidase-specific CTL. Nonetheless, vaccination with a GAr-LacZ adenovirus or with an allogeneic cell line expressing GAr-LacZ resulted in the induction of beta-gal-specific CTL. This demonstrates that the GAr domain does not inhibit cross presentation of antigens, but only affects breakdown of endogenously synthesized proteins. These data demonstrate how the GAr domain can be exploited to create immuno'stealth' genes by hiding transgene products from CTL-mediated immune attack.