Cell membrane reactivity of MIB-1 antibody to Ki67 in human tumors: fact or artifact?

Ki67 immunohistochemistry is a widely used marker of the tumor proliferative fraction. Apart from the nuclear staining of dividing cells, MIB-1 monoclonal antibody was also found to stain the cell membrane of some tumor types. Indeed, such membrane reactivity was proposed as a diagnostic feature of hyalinizing trabecular tumor (HTT) of the thyroid. To verify the diagnostic role of Ki67 membrane pattern, 6 HTTs, 8 pulmonary sclerosing hemangiomas (SH), and 6 other human tumors with MIB-1 cell membrane immunoreactivity were stained by immunoperoxidase with 5 different anti-Ki67 antibodies in different experimental conditions. We show here that the cell membrane reactivity reported in HTT is produced only by MIB-1 and not by other antibodies to Ki67 (including commercially available mouse and rabbit monoclonal antibodies). In addition, this peculiar pattern is obtained only if the reaction is performed at room temperature, because automated immunostainers which operate at 37 degrees C do not produce any MIB-1 membrane localization. The same findings were obtained in the other 6 tumors. Conversely, sclerosing hemangioma of the lung did not produce any MIB-1 cell membrane reactivity in our hands. A cross-reactivity of the MIB-1 monoclonal antibody with an epitope expressed at the cell membrane level (rather than an artifact) seems the most likely explanation for this finding, because the immunoreactivity is generally intense and uniform in the membrane positive tumors. We conclude that when Ki67 immunohistochemistry is used for diagnostic purposes in a suspected HTT, only MIB-1 clone at room temperature should be employed.