Patient-derived xenografts (PDX) have become reliable tools for preclinical research for many types of cancer. We recently reported the establishment of 4 human diffuse-large B-cell lymphoma (DLBCL) PDX in mice arising from patient prostate carcinoma tissue (Wetterauer,C. et al. 2015). All PDX were EBV associated and of ABC subtype as determined by the new algorithm of Choi et al. 2009. In contrast, none of the donor patients showed clinical or histological evidence of lymphoma disease. To address the question why lymphoma engraftment takes place in immunodeficient mice whereas patients do not show clinical evidence of disease, we compared growth behavior of the DLBCL PDX in 4 different mouse strains harboring different immunological deficiencies. Furthermore using 3 different implantation routes the dissemination behavior of these models into different immunological niches was evaluated. Results were compared to a similar study carried out with an EBV negative DLBCL PDX of a secondary, cerebral lymphoma (ABC-subtype).None of 4 EBV+ DLBCL PDX showed tumor growth after implantation in NMRI nu/nu (B- cells+, T cells-, NK cells+, n=5 mice/PDX) or 1147-F mice (B cells-, T cells+, NK cells+ n=5 mice/PDX). Subcutaneous implantation into NOD/SCID mice (B cells-, T cells-, NK cells+ n=5 mice/PDX) revealed tumor growth for 2 of 4 lymphomas displaying take rates of 80% (4 out of 5 mice) and 100% (5 out of 5 mice) within one model. Infiltration of murine NK cells in tumor tissue was determined by IHC (analyses ongoing). All 4 EBV+ lymphoma models showed stable growth in NOG mice (B cells-, T cells-, NK cells- n= 10 - 12 mice/PDX) with a take rate of 100% and an average passaging time of 20 days, ranging from 14 to 26 depending on the specific model. Taken together these results indicate that tumor growth of EBV+ lymphomas in mice depends on the NK cell activity as well as the B- and T cells status of the recipient mouse. These cell populations control tumor engraftment in mice, which might explain the absence of lymphoma disease in donor patients. In contrast the PDX of a clinically apparent EBV- DLBCL patient was able to circumvent NK cell block in NMRI nu/nu mice and showed similar take rates (100%, 5 out of 5 engrafted mice) but prolonged passaging times (35 days) as compared to NOG (26 days).Comparison of tumor cell engraftment after subcutaneous (s.c.), intravenous (i.v.) and intratibial (i.t.) injection of tumor cells in NOG mice revealed disseminated tumor growth for the EBV- DLBCL PDX exclusively when injected i.v. or i.t.. After subcutaneous implantation of lymphoma material, no tumor cells could be detected by flow cytometry or IHC in lymph nodes, blood, spleen, brain, liver, lung or bone marrow of recipient mice. However, lymphoma cells could be detected in liver and bone marrow of all (n= 8) i.v. and i.t. injected mice. Currently we are analyzing the dissemination behavior of the EBV+ lymphomas after injection into the different immunological niches as described above.In summary these data highlight the importance of the tumor microenvironment in lymphoma engraftment and dissemination in vivo. Accordingly we could show for our EBV associated lymphoma models that NK-, T- and B cell population control engraftment in vivo and are possibly involved in controlling EBV+ transformed lymphatic cells in patients. Furthermore, we showed that injection into the biologically relevant niche enables lymphoma cells to form multiple tumor nodules in distinct organs in vivo. These models contribute to a better understanding of the interplay between lymphoma cells and their microenvironment. They will thereby facilitate the discovery of new targets for innovative anti-lymphoma treatment strategies. DisclosuresTschuch:Oncotest GmbH: Employment. Klingner:Oncotest GmbH: Employment. Lenhard:Oncotest GmbH: Employment. Haapaniemi:Biositehisto Oy: Employment. Schueler:Oncotest GmbH: Employment.