Enzyme kinetics of GTI-2040 (5'-GGC TAA ATC GCT CCA CCA AG-3'), a phosphorothioate ribonucleotide reductase antisense, were investigated for the first time in 3' exonuclease solution and human liver microsomes (HLMs), using the ion-pair high-performance liquid chromatogram method for quantification of the parent drug and two major 3'N-1 and 3'N-2 metabolites. Enzyme kinetics of GTI-2040 in 3'-exonuclease solution were found to be well characterized by the Michaelis-Menten model, using the sum of formation rates of 3'N-1 and 3'N-2 (approximately total metabolism) because of sequential metabolism. In HLMs, a biphasic binding was observed for GTI-2040 with high- and low-affinity constants (K(d)s) of 0.03 and 3.8 microM, respectively. Enzyme kinetics of GTI-2040 in HLMs were found to deviate from Michaelis-Menten kinetics when the total GTI-2040 substrate was used. However, after correction for the unbound fractions, the formation rate of total metabolites could be described by Michaelis-Menten kinetics. Using the free substrate fraction, the K(m) and V(max) of GTI-2040 were determined to be 6.33 +/- 3.2 microM and 16.5 +/- 8.4 nmol/mg/h, respectively. Using these values, in vitro hepatic intrinsic clearance (CL(int)) in HLM was estimated to be 2.61 +/- 0.56 ml/h. The CL(int) was then used to predict GTI-2040's in vivo intrinsic clearance in humans by a microsomal protein scaling factor, which gave a mean value of 182.7 l/h, representing 24.1% of the observed in vivo mean scaled hepatic intrinsic clearance of 758.7 l/h in patients with acute myeloid leukemia. We concluded that the saturable nonspecific binding of GTI-2040 in HLMs complicated the interpretation of its enzyme kinetics, and scaled intrinsic clearance from HLMs only partially predicted the in vivo intrinsic clearance.