Bcr-Abl, the causative molecular abnormality of CML3 is known to activate numerous intracellular signaling proteins and pathways. These include ras, raf, phosphatidylinositol 3-kinase, AKT, STAT (signal transducers and activators of transcription), and various antiapoptotic proteins (1). The activation of these pathways and proteins is dependent on the tyrosine kinase activity of the Bcr-Abl protein. Thus, targeting the Bcr-Abl tyrosine kinase with imatinib, a specific inhibitor, has been a highly successful clinical strategy (2). The impetus to add a second agent that targets one of these downstream signaling pathways has come from the observation that relapses in CML patients in the advanced phase of the disease are common (3, 4). Preclinical studies of farnesyl transferase inhibitors have shown activity against CML cell lines and in animal models (5, 6). In the initial Phase I clinical trial of the farnesyl transferase inhibitor, R115777, 2 of 3 CML blast crisis patients had partial responses (7). Although a subsequent study treating 6 blast crisis patients failed to observe a response, 6 of 10 chronic phase patients achieved transient complete hematological responses with single agent R115777 (8). In this issue, Nakajima et al.(9) demonstrate that the farnesyl transferase inhibitor, SCH66336, enhances the antiproliferative effects of imatinib against Bcr-Abl-expressing cells, including imatinib-resistant cells. These results are consistent with those reported by Hoover et al.(10). Both groups demonstrate that cells resistant to imatinib because of amplification of Bcr-Abl remain sensitive to SCH66336 and that cotreatment of these cells with imatinib plus SCH66336 leads to enhanced antiproliferative or proapoptotic effects (9, 10). Interestingly, Hoover et al.(10) also tested cells that are resistant to imatinib because of the expression of a Bcr-Abl kinase domain mutation, T315I, that is completely insensitive to imatinib. Although this mutant is sensitive to SCH66336, the addition of SCH66366 to imatinib yielded no increase in antiproliferative effects (10). These results raise critically important issues of how and when to optimally combine molecularly targeted agents. Although this discussion will focus on CML and Bcr-Abl signal transduction, this paradigm could be applied to any agent that targets signaling pathways. In deciding whether it is rational to add an agent to imatinib that targets a downstream signaling pathway, one would first want to know if the relapse mechanism was Bcr-Abl-dependent or-independent. In the case of Bcr-Abl-independent relapse, the Bcr-Abl kinase would remain inhibited by imatinib as would all of the pathways activated by the Bcr-Abl kinase. Thus, resistance would be driven by molecular abnormalities other than Bcr-Abl. Therefore, targeting pathways downstream of Bcr-Abl would be predicted not to be of utility, unless the additional molecular abnormalities happened to activate similar pathways to Bcr-Abl. In CML blast crisis patients who do not respond to imatinib therapy, many of them have resistance mechanisms that are Bcr-Abl-independent and would fall into this category (11). In the case of Bcr-Abl-dependent relapses, the Bcr-Abl kinase would have been reactivated as the cause of imatinib resistance. Thus, all of the Bcr-Abl pathways would also be reactivated. In this case, it would be anticipated that an agent targeting a downstream pathway would be useful. This is exactly the case as described by Nakajima et al.(9) and Hoover et al.(10). Bcr-Abl cells that are resistant to imatinib because of Bcr-Abl amplification or point mutation are sensitive to SCH66336. Whether or not adding this second agent to imatinib results in improved …