Background: Vaccinia virus is a tool for a novel cancer therapy that infects and lyses cancer cells. Clinical trials targeting solid tumors are currently ongoing, and promising results are reported. Although in vitro studies suggest that hematological malignancies have only modest susceptibility in general, we found that multiple myeloma shows an exceptionally high susceptibility to vaccinia virus, therefore, we focused on myeloma. Because tumor specific infection is the key for a successful oncolytic therapy, we modified thymidine kinase (TK) gene and B5R gene in the vaccinia genome to make it more tumor-specific. Efficacy and safety was determined by the in vitro and in vivo mouse infection models.Methods: Vaccinia TK is required for viral replication, but it is dispensable in tumor cells because high level of TK produced by the tumor can substitute it. In order to delete TK gene, we generated a plasmid vector which have a homologous sequence of TK (pTK), and the firefly luciferase (Fluc) gene with the PSFL110 promoter was inserted into the middle of TK gene. By the co-transfection of 143B cells with the pTK-Fluc and parental vaccinia virus LC16mO (derived from Lister clone 16), a homologous recombination occurred and the TK-deleted vaccinia virus was generated. Vaccinia B5R gene is essential for the viral transmission from infected cells to adjacent cells. To regulate B5R gene expression and inhibit viral transmission in normal cells, we inserted a let-7a target sequence into the downstream of B5R in the pB5R plasmid. Let-7a is a micro RNA of which expression is lost in tumor cells but maintained in most of the normal cells, therefore, we hypothesized that endogenous let-7a in normal cells will inhibit B5R, inhibit viral transmission, and improve safety. By the co-transfection of RK13 cells with the pB5Rgfp-let7a and TK-deleted vaccinia virus, a homologous recombination occurred, and TK-deleted let-7a-regulated vaccinia virus was generated.Results: To find out a good target for this therapy, we first infected different cell lines including Kasumi-1, NB4, NALM6, Jurkat, OM9;22, K562 (these are from leukemia), FL218, Daudi (lymphoma), MT2, TLO-m1 (ATL), and RPMI8226, U266, NCI-H929 (myeloma) with a vaccinia virus. Very interestingly, myeloma-derived cell lines showed an exceptionally high susceptibility (10-100 times higher than others), and this susceptibility was as high as solid tumors; therefore we chose multiple myeloma as the target. To confirm that regulation of B5R by the let-7a improves myeloma-specific infection, we infected a myeloma cell line RPMI8226 as well as normal skin fibroblasts with the recombinant B5Rgfp-let7a vaccinia virus. In vitro infection of a RPMI8226 cells showed a significant infectivity even with a very low titer (MOI=0.1), while control normal cells (human skin fibroblasts) were not infected at all. Without the let7a-regulation, the vaccinia virus infected both myeloma cells and normal cells at the same titer, suggesting that let-7a-mediated regulation clearly works. As an in vivo infection model, 1x107 RPM8226-Rluc cells were injected into immunodeficient mice (CB.17-SCID) subcutaneously. Four weeks later, 1x106 pfu of virus (Parent; TK-deleted; TK and let7a double regulated) was administered via an intravenous injection, and tumor volume and the amount of virus were determined weekly by the in vivo imaging system with renilla and firefly luciferases. Without any regulation, vaccinia virus infected not only myeloma cells but also normal tissues, and mice developed pock lesions in the ear, nose, mouth, foot, and tail, resulting in death within 21 to 24 days. TK-deletion significantly alleviated viral toxicity but still caused death 42 days after infection. In contrast, infection with TK and let-7a double regulated vaccinia virus was clearly limited in myeloma, and all mice remained alive and the size of myeloma shrunk continuously.Conclusions: These data suggest that the TK-let7a-double regulated vaccinia virus infects and kills myeloma cells specifically, and will be a good candidate for the future clinical application. DisclosuresNo relevant conflicts of interest to declare.