Introduction The adjuvant treatment of colorectal cancer (CRC) has seen little improvement in terms of mortality of the disease in the last 40 years. There has been a resurgence in research into the use of monoclonal antibodies in the treatment of CRC. Carcinoembryonic antigen (CEA) is a useful target in cancer immunotherapy. The distribution of CEA in CRC differs from that in normal colorectal tissue. In normal colorectal tissue CEA is found only on the luminal surface of the cell which is inaccessible to intravenous antibody, whereas in CRC, CEA is found on all borders of the cell membrane and so becomes accessible to intravenous antibody. However, anti-CEA antibodies are prone to sequestration by circulating CEA. The anti-CEA antibody, PR1A3, binds only membrane-bound CEA and thus is able to overcome this problem. The aim of my research was to assess whether PR1A3 is suitable to be considered as a therapeutic agent in the treatment of CRC and what its mechanism of action might be. Methods The level of expression of CEA on a panel of cell lines was determined under different conditions using a solid-phase ELISA and FACS analysis. Humanized PR1A3 (hPR1A3) was assessed in a variety of in vitro cytotoxicity assays with colorectal cell lines expressing varying levels of CEA, using peripheral blood mononuclear cells and purified natural killer cells as sources of effector cells. The mechanism of action of PR1A3 was investigated by modifying the Fc fragment of the antibody and using antibodies to block the FcIIIA receptor on the effector cells. PR1A3 was also investigated in combination with a humanised A33 antibody. Results A panel of colorectal cell lines was found to have a range of CEA expression which could be upregulated in certain cell lines by growing the cell line beyond confluence and by treatment with the chemotherapeutic agent, 5-fluorouracil. The in vitro assays demonstrated hPR1A3 antibody-dependent and CEA-specific killing of tumour cell lines by human PBMC. The effect increased with increasing concentration of antibody and was lost by using the parent murine IgG1 PR1A3. Using 50μg/ml hPR1A3, tumour cell lysis was increased by more than 3-fold above spontaneous killing (p < 0.001) in a high CEA-expressing cell line. Both antibody-dependent and antibody-independent (spontaneous) killing was blocked by using whole antibody to the Fc-γIIIA receptor, although the spontaneous killing was restored when a F(ab')2 was used instead of whole antibody. hPR1A3 and the A33 antibody showed potential synergy when used in combination against a high-CEA and a moderate-A33 expressing cell line. Conclusion The monoclonal antibody hPR1A3 causes CEA-specific lysis of human colorectal cancer-derived cell lines in the presence of human PBMCs. This lysis is dependent on the dose of the antibody, requires a compatible Fc-receptor and is inhibited by blockade of the FcγIIIA receptor. These findings show that hPR1A3 can kill tumour cells by antibody-mediated cellular cytotoxicity (ADCC) and implicate NK cells as a major contributor to this effect. The results support the development of hPR1A3 for therapy of colorectal cancer.