A comparative study on tissue processing procedures for the immunohistochemical investigation of oral mucosal Langerhans cells.

An immunoperoxidase technique was used to compare wax-embedded tissue with frozen tissue for quantitative immunohistochemistry of oral mucosal Langerhans cells. Initial experiments using anti-CD1a, -HLADR and -S100 antisera showed that phenotype, fixative, antibody dilution and trypsinisation of the tissue section significantly affected Langerhans cell counts. Only the anti-HLADR antibody detected Langerhans cells in both frozen and wax-embedded sections. Some 38% of S100-positive dendritic cells were situated in the stratum basale, and 41-84% of these contained melanin as determined by double-labelling. Sections from 39 volunteers were then reacted with the anti-CD1a and -HLADR antibodies. The morphology of Langerhans cells was more dendritic in frozen sections, and the mean HLADR-positive Langerhans cells count in frozen sections was significantly higher than that in wax-embedded sections from the same individual. The intra-individual ratio of counts between frozen and wax-embedded sections was variable; hence, the apparent loss of HLADR antigenicity as a result of tissue processing was unpredictable. Counts of CD1a-positive Langerhans cells were consistently higher. We conclude that the use of anti-CD1a antibody on frozen tissue is the optimum method for quantitative studies of oral mucosal Langerhans cells, and that such studies performed on wax-embedded tissue may be unreliable.