In a phase I study, 21 patients with metastatic adenocarcinoma of the gastrointestinal tract received the murine monoclonal antibody D612. This antibody is directed at a M(r) 48,000 antigen restrictively expressed on tumors of the gastrointestinal tract and to a limited degree on normal gastrointestinal mucosa. Patients received total doses of 10-180 mg/m2 administered as single or multiple doses of 1-100 mg/m2 over an 8-day period. Dose-limiting toxicity was secretory diarrhea. A single dose of 100 mg/m2 exceeded guidelines for maximal tolerated dose. Higher total doses were achieved in subsequent patients by using repeated administration of lower doses. Three of five patients receiving 60 mg/m2 for 3 doses (180 mg/m2 total dose) experienced grade 3 diarrhea and could not complete the prescribed course. The dose of 40 mg/m2 administered on days 1, 4, and 8 (total dose, 120 mg/m2) has been selected as the dose for phase II studies. The pharmacokinetics of D612 is best described by a one-compartment model with a mean t1/2 of 48 +/- 3 h (SEM). Eighteen of 21 patients developed human anti-mouse antibody (HAMA). Patients who developed high levels of HAMA demonstrated a more rapid clearance of the day 8 dose than those who developed low levels of HAMA. In all patients studied, a component of HAMA was directed at the D612 variable region. With one exception, serum from all patients with detectable antibody to the D612 variable region demonstrated anti-paratope reactivity. Thirty-four % of known metastatic sites demonstrated uptake of radiolabeled D612. There were no objective antitumor responses in this phase I trial. The antitumor effect of D612 in vitro has been shown to be potentiated by interleukin 2 and recombinant human macrophage colony-stimulating factor. A phase II study of D612 administered in combination with cytokines that enhance human effector function is presently ongoing.