Primary effusion lymphoma (PEL) is an aggressive B-cell lymphoma most commonly diagnosed in HIV-positive patients and universally associated with Kaposi's sarcoma-associated herpesvirus (KSHV). Chemotherapy treatment of PEL yields only short-term remissions in the vast majority of patients, but efforts to develop superior therapeutic approaches have been impeded by lack of animal models that accurately mimic human disease. To address this issue, we developed a direct xenograft model, UM-PEL-1, by transferring freshly isolated human PEL cells into the peritoneal cavities of NOD/SCID mice without in vitro cell growth to avoid the changes in KSHV gene expression evident in cultured cells. We used this model to show that bortezomib induces PEL remission and extends overall survival of mice bearing lymphomatous effusions. The proapoptotic effects of bortezomib are not mediated by inhibition of the prosurvival NF-κB pathway or by induction of a terminal unfolded protein response. Transcriptome analysis by genomic arrays revealed that bortezomib down-regulated cell-cycle progression, DNA replication, and Myc-target genes. Furthermore, we demonstrate that in vivo treatment with either bortezomib or doxorubicin induces KSHV lytic reactivation. These reactivations were temporally distinct, and this difference may help elucidate the therapeutic window for use of antivirals concurrently with chemotherapy. Our findings show that this direct xenograft model can be used for testing novel PEL therapeutic strategies and also can provide a rational basis for evaluation of bortezomib in clinical trials.