Preclinical studies suggest that the biochemical effects of N-(phosphonacetyl)-l-aspartate (PALA), an inhibitor of aspartate carbamoyltransferase (ACTase), may increase the metabolic activation of 5-fluorouracil (5-FU) and enhance its cytotoxicity through both RNA- and DNA-directed mechanisms. In this Phase I trial, 22 evaluable patients with adenocarcinoma of the gastrointestinal tract were entered at escalating doses of 5-FU starting at 1150 mg/m2/day given as a concurrent 72-h i.v. infusion with a fixed dose of leucovorin (LCV), 500 mg/m2/day. The dose of 5-FU was escalated within patients according to individual tolerance, and then PALA at 250 mg/m2 was added 24 h prior to the initiation of the 5-FU/LCV infusion of the subsequent cycle. Dose-limiting mucositis and myelosuppression occurred during the initial cycle in 3 of 5 patients treated with 2300 mg/m2/day 5-FU; therefore, the recommended dose of 5-FU with concurrent LCV is 2000 mg/m2/day. Twenty-seven additional patients were then treated with escalating doses of PALA ranging from 375 to 2848 mg/m2, i.v., followed 24 h later by 2000 mg/m2/day 5-FU with high-dose LCV. Dose-limiting mucositis and myelosuppression occurred during the initial cycle in 2 of 3 patients entered at 2848 mg/m2 PALA. Dose-limiting mucositis and skin rash ultimately required both PALA and 5-FU dose reductions in 4 of 6 patients treated with 1899 mg/m2 PALA. Toxicity was similar, however, in patients receiving PALA at doses ranging from 375 to 1266 mg/m2. The mean steady-state plasma concentration of 5-FU at 2000 mg/m2/day was 6.5 ± 0.9 µm; patients with 5-FU levels >9 µm had a significantly higher incidence of serious gastrointestinal and hematological toxicity. Compared to each patient's own baseline, a significant trend for decreasing ACTase activity with increasing PALA dose was evident using cytosol isolated from peripheral blood mononuclear cells 24 h after PALA treatment (P2 = 0.01). PALA ≤844 mg/m2 failed to appreciably inhibit ACTase activity at 24 h in most patients; furthermore, a decrease in ACTase activity by >50% from baseline was seen in only 29% of cycles. More consistent inhibition of ACTase activity was seen with PALA ≥1266 mg/m2. Even with the highest PALA doses, however, ACTase activity returned to baseline by 96 h in most patients. In contrast, a modest decrease in plasma uridine levels was noted at all PALA doses, but the decrease was ≥50% in only 21% of cycles at 24 h. PALA ≤ 1266 mg/m2 could be safely combined with a 72-h i.v. infusion of 5-FU 2000 mg/m2/day with LCV 500 mg/m2/day starting 24 h after PALA. Because the delivered 5-FU dose intensity for patients entered at or above 1750 mg/m2/day in this trial was similar at PALA doses ≤ 1266 mg/m2, we have selected 1266 mg/m2 for future studies.