Mouse models have markedly improved our understanding of cancer development and tumor biology. However, these models have shown limited efficacy as tractable systems for unbiased genetic experimentation. Here, we report the adaptation of loss-of-function screening to mouse models of cancer. Specifically, we have been able to introduce a library of shRNAs into individual mice using transplantable Emu-myc lymphoma cells. This approach has allowed us to screen nearly 1,000 genetic alterations in the context of a single tumor-bearing mouse. These experiments have identified a central role for regulators of actin dynamics and cell motility in lymphoma cell homeostasis in vivo. Validation experiments confirmed that these proteins represent bona fide lymphoma drug targets. Additionally, suppression of two of these targets, Rac2 and twinfilin, potentiated the action of the front-line chemotherapeutic vincristine, suggesting a critical relationship between cell motility and tumor relapse in hematopoietic malignancies.