Abstract. 2' and 7-polyethylene glycol esters of taxol were prepared and found to be essentially water soluble. The rates of hydrolysis of these compounds were measured under neutral, acidic and physiological (basic) conditions. The half-lives of 0 and N substituted 2'-esters are short enough to permit their use as prodrugs. Water soluble prodrugs of taxol (1) have been the object of several recent serious investigations. ~-s Thus far, these approaches have all been based on ester hydrolysis of poorly soluble (< 10 mg/ml) prodrugs. The purpose of those studies was to circumvent the use of potentially antigenic solubilizing agents such as Cremophor EL which is currently being employed for taxol infusion. 6 The most recent work in this direction is the study by Nicholaou and coworkers 4 who provide precisely designed taxol esters which possess strong electron withdrawing substituents (such as alkoxy) in the a-position of the ester in order to accelerate hydrolytic cleavage. Additional modification also provided anchiomeric rate enhancements. The solubility of these prodrugs varied from <0.1 to 1.2 mg/mL and were reported to have half-lives >8.3 h at pH 7.5 and 37~ A shorter in vitro half-life (1.5-2 h) was demonstrated for one compound when human plasma was employed. 7Polyethylene Glycol (PEG) a, an amphiphilic macromolecule, 9 imparts greater aqueous solubility to conjugates of hydrophobic organic compounds when the molecular weight of PEG is 2 kD or greater. ~~ By incorporating PEG as the o~-alkoxy group of an acid, taxol esters with highly enhanced water solubility should be produced. Thus, PEG 5000 carboxylic acid n (2) was coupled to taxol ~2 in 90-95% yield using diisopropyl carbodiimide (DIPC) and dimethylaminopyridine in methylene chloride to give exclusively the 2'PEG ester of taxol (3a).Tile solubility of 3a was estimated to be -> 666 mg/mL at ambient temperature, and as shown in Figure i, stability was observed for prolonged periods in Phosphate-saline buffer (PBS) at pH 5.8. Dissolution of 3a in PBS at pH 7.0 and 7.4 (physiological pH) resulted in slow release of taxol (t~=5.5 -t- 0.5h, pH 7.4) as was expected.13Preparation of the 7-PEG 5000 ester 3c was accomplished employing the 2'-methoxyacetate (MAc)ester (4)~~ as a blocking group (equation 2).Treatment of 3c with excess ethanolamine produced the desired ester 3d with no indication of ammonolysis occurring at the 7-position. Ester 3d was also found to be virtually water soluble. Kinetic studies of the hydrolysis of 3d were performed in similar fashion to those of the 2'-ester 3a (see Figure 2), and clearly demonstrate the stability of esters in the hindered 7-position.14 Therefore, the greater reactivity reported for taxol 7-'y-carboxy-o~-alkoxy-esters 4 must be due exclusively to neighboring group parllcipation. Additional evidence for this effect was provided by the 7-earboxamido-2'-taxol ester 3e prepared from acid 615 in similar fashion to 3d.The half-life of 3e was determined to be 3.0 -4- 0.5 h at pH 7.4 and constitutes the shortest half-life found for any of the esters examined in this study. (Figure 2).For purposes of extending the half-life of taxol esters, the use of less election withdrawing substituents in place of oxygen was explored. Reaction of PEG 5000 chloroformate I~ (7) with glycine ethyl ester followed by saponification resulted in a 90% yield of the PEG substituted acid 8.Esterification of taxol (1) with 8 using DIPC resulted in a 92% yield of 3f. The kinetics of hydrolysis of 3f are presented in Figure 2. As anticipated, a longer half-life was observed (t~ = 8.5 -I- 1.0 h) at pH 7.4.Both the 2'-esters (3a, 3e, and 30 and the T-ester (3d) were tested for in vitro activity according to published protocols 17, t8 using the murine leukemia cell lines P388 and L1210 (Table I). The 2'- ester prodrugs had ICso values essentially the same as unmodified taxol (5-18 nM) for both taxol-sensitive strainsindicating that taxol had been released from the prodrug. The 7'-ester exhibited reduced cytotoxic activity, but still retained an ICso in the 270 nM range.The kinetics of taxol release in rat plasma was also determined. Prodrug 3e was dissolved in water, added to the plasma (EDTA treated), and incubated at 37~ for various times. The plasma was extracted with ethyl acetate TM, and the concentration of taxol was determined by reverse-phase HPLC using a phenyl column. The t~ in plasma was 1.1 hours (Figure 3). The shorter half-life for the PEG-ester bond compared to pH 7.4 buffer is expected, since rodent plasma contains high non-specific esterase activities.Thus, the use of PEG for not only solubilizing taxol, but to afford controlled release of the drug over various periods of time, has been accomplished and may provide the ftrst viable method of aqueous delivery of this important drug