Relapse and acquired drug resistance in T-cell acute lymphoblastic leukemia (T-ALL) remains a significant clinical problem. This study was designed to establish a preclinical model of resistance to induction therapy in childhood T-ALL to examine the emergence of drug resistance and identify novel therapies. Patient-derived T-ALL xenografts in immune-deficient (non-obese diabetic/severe combined immunodeficient) mice were exposed to a four-drug combination of vincristine, dexamethasone (DEX), L-asparaginase and daunorubicin (VXLD). 'Relapse' xenografts were characterized by responses to drugs, changes in gene expression profiles and Connectivity Map (CMap) prediction of strategies to reverse drug resistance. Two of four xenografts developed ex vivo and in vivo drug resistance. Both resistant lines showed altered lipid and cholesterol metabolism, yet they had a distinct drug resistance pattern. CMap analyses reinforced these features, identifying the cholesterol pathway inhibitor simvastatin (SVT) as a potential therapy to overcome resistance. Combined ex vivo with DEX, SVT was significantly synergistic, yet when administered in vivo with VXLD it did not delay leukemia progression. Synergy of SVT with established chemotherapy may depend on higher drug doses than are tolerable in this model. Taken together, we have developed a clinically relevant in vivo model of T-ALL suitable to examine the emergence of drug resistance and to identify novel therapies.