Affiliations: [a] Department of Medical Microbiology, University of Odense, Odense, Denmark | [b] Department of Pathology, University of Odense, Odense, Denmark | [c] Institute of Medical Microbiology, University of Aarhus, Aarhus, Denmark
Note: [] Address reprint requests to Dr. Henrik Ditzel, Department of Medical Microbiology, University of Odense, 1. B. Winsløws Vej 19, DK-5000 Odense C, Denmark.
Abstract: Preliminary investigations suggested the importance of an evaluation of different tissue preparation methods frequently used for immunohistochemical analysis of human or murine monoclonal antibodies on human tissue. Colon adenocarcinomas and adjacent morphologically normal colon epithelia were analyzed with an indirect immunoperoxidase technique. Duplicate tissue specimens were (1) snap frozen and fixed in acetone, (2) formalin fixed and paraffin embedded, with or (3) without ensuing treatment with pronase, or (4) alcohol fixed and paraffin embedded. Three different human monoclonal anti-colon cancer IgM antibodies, COU-1, D4213, and F10279, were used in the present study. Endogenous immunoglobulin and the secretory-component-mediated IgM binding were blocked on frozen sections with Fab' anti-IgM and anti-SC antibody. Bound monoclonal antibody was detected with horseradish peroxidase-anti-IgM. COU-1 was found to stain frozen sections of all 25 cancer and adjacent normal colon epithelia. In contrast, on formalin-fixed, paraffin-embedded tissue, only 80% (20/25) of the colon cancer and 44% (11/25) of the adjacent normal colon epithelia were positive. After treatment of the formalin-fixed sections with pronase, all cancers and normal adjacent epithelia were stained, but the cancer cells were more intensely stained than the normal colon epithelial cells. On alcohol-fixed tissues, intense staining was found in all the colon carcinomas analyzed, whereas no staining was found of the adjacent normal colon epithelia, except for a few cells in some of the sections investigated. Antibody F10279 stained none of the frozen, acetone-fixed tissues or the alcohol-fixed and paraffin-embedded tissues analyzed, whereas it stained 71% (5/7) of the formalin-fixed adenocarcinomas and 43% (3/7) of the adjacent normal epithelia. Pronase treatment enhanced the staining intensity. Antibody D4213 showed similar staining of all cancers and normal colon epithelia analyzed with all the four preparation techniques. Procedures for the fixation and processing of human tissue for the immunohistochemical screening of human monoclonal antibodies are critical. Our results demonstrate the necessity of examining several tissue preparation techniques for the analysis of the tissue reactivity of monoclonal antibodies, and indicate that alcohol fixation may advantageously be included. Destruction or masking of antigens, unspecific binding of antibodies, or other technical problems may lead to erroneous conclusions concerning the antigen distribution or the specificity or selectivity of the antibody.
Keywords: human monoclonal antibodies, immunohistochemistry, tissue fixation, COU-1, colon cancer
