Abstract 1661Most cancer drug developments are focusing either in vitro target-based screening or cell-based phenotypic screening to identify potential compounds. The target-based screenings are powerful screening methods if the cancer relay on a single driver mutation. On the other hand, cell-based phenotypic screenings measure cell phenotypes such as growth inhibitory effect on established cancer cell lines. However, these established cell lines do not recapitulate human cancer in some aspects. For instance, most cell lines have the microenvironment-independent growth ability, while primary cancer cells can not survive in ex vivo culture. In this study, we established a new drug screening system targeting microenvironment-dependent primary lymphoma cells.First of all, we established primary lymphoma xenograft models, including three diffuse large B cell lymphoma (DLBCL), one follicular lymphoma, one intravascular large cell lymphoma (IVL), and one mantle cell lymphoma. Interestingly, lymphoma cells from the patient with severe extranodal invasion invaded extranodal organ also in mice. And pattern of invasion in tissue such as intravascular tumor invasion of IVL was maintained in the xenograft models. In addition, mRNA expression profiles were similar between primary lymphoma cells and the lymphoma cells obtained from NOG mice. Furthermore, the heterogeneity of primary tumor such as heterogenous expression of CD20 was maintained in the xenograft models. These data suggested that the lymphoma cells propagated in NOG mice kept the original patient's phenotype.Next, we investigated the microenvironment of lymphoma in primary lymphoma xenograft models. Fibroblastic reticular cell (FRC) and follicular dendritic cell (FDC) are reported to be important to support the survival of lymphoma cells in their microenvironment. In the spleen of DLBCL model mice, lymphoma cells colocalized with FRC, but FDC was not detected in NOG mouse. FRC is known to produce reticular fiber and forms fibroblastic reticular network (FRN). Lymphoma induced reticular fiber production and tumor formation on FRN were also observed in our system. Furthermore, co-culture with FRC cell line, BLS4, significantly enhanced viability of 2 out of 3 DLBCL cells obtained from primary lymphoma xenograft mice and enabled more than three weeks long-term ex vivo culture of them. These results indicated that FRC played an important role for lymphoma cell survival.Finally, we established a new drug screening system using this co-culture system. Primary DLBCL cells transplanted to NOG mice were collected and seeded on pre-seeded BLS4 in 96-well plates (lymphoma cell co-culture). Monoculture of BLS4 in 96-well plates was also prepared. Both lymphoma co-cultured cell and BLS4 monoculture cell were treated by 2613 kinds of pharmacologically active compounds for 72 h. DAPI-stained dead cells of lymphoma cells were counted by image analyzer, and the proliferation of BLS4 was measured by MTT assay. We calculated Drug Effect Index (DEI) by multiplication of dead cell number of lymphoma cells and MTT value of BLS4 and compared. The compound with the highest DEI was pyrvinium pamoate, oxyuricide. Pyrvinium pamoate aborogated the survival of lymphoma cells co-cultured with BLS4 dose-dependently. And subcutaneous tumor of primary lymphoma cells and BLS4 were diminished by single intratumoral injection of 20 mg/kg pyrvinium pamoate. These results indicated that our screening system could be performed in large scale and select drugs with anti-tumor activity to primary lymphoma cells. Screening against primary lymphoma cells sheds new light on lymphoma drug development. Disclosures:Sugimoto:Otsuka Pharmaceutical co ltd: Employment.