Detection of circulating tumor-associated antigen depends on the domains recognized by the monoclonal antibodies used: N-terminal trimmed EpCAM-levels are much higher than untrimmed forms.

The measurement of tumor-associated proteins is of high diagnostic value in the follow-up of cancer patients. Most tests ignore that various forms of the protein can exist; especially in epithelial cancers and the soluble receptors they produce. We choose EpCAM as model-antigen to analyze whether tests recognizing different domains of the protein give different results in patients' sera. EpCAM-reactive autoantibodies are present in the sera of patients with colorectal carcinoma, however little is known about the existence and possible relevance of circulating soluble EpCAM protein. Most monoclonal EpCAM-antibodies recognize the first EGF-like repeat and fail to detect N-terminal trimmed protein. We developed a novel ELISA to determine the concentration of serum EpCAM with mAbs recognizing the second EGF-like repeat. In 59 healthy controls, EpCAM concentrations ranged from 232 to 8893ng/ml (mean 1525ng/ml). Levels of EpCAM in 412 patients with adenocarcinoma were somewhat higher with concentrations ranging from 176 to 36,259ng/ml (mean 1971ng/ml). In direct comparison, the untrimmed protein specific ELISA detected lower levels and frequencies as compared to the EGFII-specific ELISA. Only sera with less than 1μg/ml circulating EGFII-EpCAM (66% of the sera) contained EpCAM-specific IgG antibodies. The absence of IgG antibodies in the sera with more than 1μg/ml circulating EpCAM was not due to immune complex formation. Anti-EpCAM IgA and IgM antibodies did not show such a correlation. It will be important to assess whether the presence of high levels of circulating EGFII-EpCAM is associated with side effects in patients given immunotherapy.