Targeting an anti-cancer drug to tumors should increase the Area Under the drug concentration-time Curve (AUC) in tumors while decreasing the AUC in normal cells and should therefore increase the therapeutic index of that drug. Anti-tumor drugs typically have half-lives far shorter than the cell cycle transit times of most tumor cells. Tumor targeting, with concomitant long tumor exposure times, will increase the proportion of cells that move into cycle when the drug concentration is high, which should result in more tumor cell killing. In an effort to test that hypothesis, we conjugated a natural fatty acid, docosahexaenoic acid (DHA), through an ester bond to the paclitaxel 2'-oxygen. The resulting paclitaxel fatty acid conjugate (DHA-paclitaxel) does not assemble microtubules and is non-toxic. In the M109 mouse tumor model, DHA-paclitaxel is less toxic than paclitaxel and cures 10/10 tumored animals, whereas paclitaxel cures 0/10. One explanation for the conjugate's greater therapeutic index is that the fatty acid alters the pharmacokinetics of the drug to increase its AUC in tumors and decrease its AUC in normal cells. To test that possibility, we compared the pharmacokinetics of DHA-paclitaxel with paclitaxel in CD2F1 mice bearing approximately 125 mg sc M109 tumors. The mice were injected at zero time with a bolus of either DHA-paclitaxel or paclitaxel formulated in 10% cremophor/10% ethanol/80% saline. Animals were sacrificed as a function of time out to 14 days. Tumors and plasma were frozen and stored. The concentrations of paclitaxel and DHA-paclitaxel were analyzed by LC/MS/MS. The results show that DHA targets paclitaxel to tumors: tumor AUCs are 61-fold higher for DHA-paclitaxel than for paclitaxel at equitoxic doses and eight-fold higher at equimolar doses. Likewise, at equi-toxic doses, the tumor AUCs of paclitaxel derived from i.v. DHA-paclitaxel are 6.1-fold higher than for paclitaxel derived from i.v. paclitaxel. The tumor concentration of paclitaxel derived from i.v. paclitaxel drops rapidly, so that by 16 h it has fallen to the same concentration (2.8 microM) as after an equi-toxic concentration of DHA-paclitaxel. In plasma, paclitaxel AUC after an MTD dose of DHA-paclitaxel is approximately 0.5% of DHA-paclitaxel AUC. Thus, the increase in tumor AUC and the limited plasma AUC of paclitaxel following DHA-paclitaxel administration are consistent with the increase in therapeutic index of DHA-paclitaxel relative to paclitaxel in the M109 mouse tumor model. A phase I clinical study has been completed at The Johns Hopkins Hospital to evaluate the safety of DHA-paclitaxel in patients with a variety of solid tumors. Twenty-one patients have been treated to date. The recommended phase II dose is 1100 mg/m(2), which is equivalent to 4.6 times the maximum approved paclitaxel dose on a molar basis. No alopecia or significant peripheral neuropathy, nausea, or vomiting have been observed. Asymptomatic, transient neutropenia has been the primary side effect. Eleven of 22 evaluable phase I patients transitioned from progressive to stable disease, as assessed by follow-up CT. Significant quality of life improvements have been observed. Thus, DHA-paclitaxel is well tolerated in patients and cures tumors in mice by targeting drug to tumors.