A Novel Proteomic Approach to Define Leukemia Cell Resistance to Farnesyltransferase Inhibitors.

FTase targets in three different myeloid leukemia cell lines (HL-60, K562 and NB-4) were labeled using the unnatural FPP analogue 8-anilinogeranyl diphosphate (AGPP) in a tagging-via-substrate approach. Antibodies specific for the anilinogeranyl moiety were used to detect AG-modified proteins. This highly effective labeling/detection method was coupled with two-dimensional electrophoresis (2-DE) and subsequent Western blotting. Identities of individual protein spots were determined by overlaying anti-AG immunoblots and immunoblots probed with antibodies specific for different known prenylation target proteins, such as RAS-family proteins. The clinically tested FTIs BMS-214,662 and L-778,123 as well as the nitrogen-containing bisphosphonate zoledronate were used to further validate the specificity of this labeling technique. As a proof of principle of the ability of this approach to identify target proteins important for FTI resistance, RNA interference (RNAi) was applied to investigate the role of one of the specific protein targets (K-RAS) which remained AG-labelled in the presence of FTI BMS-214,662. Following confirmation of K-RAS silencing by Western blotting, flow cytometric analysis of Annexin-V stained cells was used to quantify the effect of K-RAS silencing on BMS-214,662-induced apoptosis.

Metabolic labeling of prenylated proteins with anilinogeranyl (AG) occurred in a time-dependent manner. This method provided increased resolution of the prenylated proteome as we were able to detect as many as ten distinct protein spots corresponding to a single band at approximately 20 kDa observed by 1D SDS-PAGE. Some of the proteins we have identified using this overlay method include H-RAS, K-RAS, N-RAS, Rap1, RhoB, RhoC, pre-Lamin A, Lamin B, and LKB1. AG-specific signals decreased upon FTI treatment, further substantiating the specificity of this method. Additional evidence of the specificity of this approach was the observation that the bisphosphonate inhibitor zoledronate did not inhibit protein AG-labeling. Interestingly, this method allowed direct evaluation of specific FTI targets as demonstrated by the different efficacies we observed between the two FTIs studied here – BMS-214,662 and L-778,123. This method even allowed identification of subtle differences among the leukemia cell lines tested. Closer evaluation of the farnesylated proteome demonstrated clear differences between BMS-214,662 and L-778,123, consistent with earlier reports that FTIs elicit numerous cellular effects, including induction of apoptosis and cytostatic effects. While BMS-214,662 effectively inhibited farnesylation of the majority of larger molecular weight proteins, L-778,123 also blocked prenylation of smaller molecular weight proteins such as N-RAS and K-RAS. However, RhoB and RhoC remained farnesylated regardless of FTI treatment. BMS-214,662-induced apoptosis was substantially potentiated by RNAi knock-down of K-RAS expression.

This snapshot approach allowed simple, rapid and dynamic analysis of the complex farnesylated proteome in leukemia cells. Our results demonstrate that this method can be used to identify and validate specific inhibitor targets. Importantly, this approach also successfully identified proteins (e.g. K-RAS) which may be important for resistance to some FTIs, and thus may be useful in directing development of more efficient therapeutic regimens.

No relevant conflicts of interest to declare.