Carcinoembryonic antigen production, secretion, and kinetics in BALB/c mice and a nude mouse-human tumor model.

Carcinoembryonic antigen (CEA) is currently being used as a target antigen in the radioimmunodetection of cancer. Circulating CEA may adversely affect the outcome of such studies by formation of intravascular immune complexes. The following studies were undertaken to expand our knowledge of the production, secretion, and pharmacokinetics of CEA, since these factors should have a direct bearing on the serum levels of CEA encountered in radioimmunodetection. The production of CEA was assessed in nude mice given implants of the T-380 CEA secreting human colon tumor. Serum CEA rose linearly as the tumors enlarged; however, the concentration of CEA per g of extracted tumor remained constant throughout the weight range studied. The secretory rate of the T-380 tumor was determined by surgically removing all blood flow to the liver and gastrointestinal tract of the nude mouse model. This procedure removes the known sites of CEA degradation. Serum CEA levels rose progressively following surgery, the values being directly related to the tumor size. The secretory rate was also proportional to tumor size but was a constant 13.8 +/- 3.6 (S.D.) ng/g tumor/hr when expressed on a per g tumor basis. To determine if the serum levels of CEA observed in patients could be due to unique differences in the clearance rates of each patient's CEA, serum from three patients with CEA levels of 2150, 709, and 58 ng/ml was administered i.v. to groups of mice at the original and diluted concentrations. The kinetics of all samples followed a single exponential clearance pattern with a half-time of about 2.5 hr. This was dramatically different from the kinetics of tumor-extracted CEA which exhibited a multiexponential pattern, the first component having a half-time of 3 min. These data suggest that CEA secreted by a tumor is in some way different from that adhering to the tumor. If the secreted CEA truly has a monoexponential clearance with a fixed rate as the experiments suggest, the absolute values of serum CEA are either entirely a function of the tumor secretory rate, or else the product having the short half-time is not measured in serum samples obtained from patients.