We evaluated the effect of six different histologic stains on the productivity of PCR amplification of DNA isolated from paraffin-embedded tissue samples. The tissue was collected from glass slides by microdissection techniques, whereby tiny portions of tissue are visually identified through a microscope and selectively resected for subsequent DNA extraction and PCR amplification. We found that the success of PCR amplification depended on the type of histologic stain that was used to facilitate microscopic visualization of the undeparaffinized tissue section. The best results were obtained with methyl green and nuclear fast red, while Wright's stain yielded less PCR product. Two other stains, Evans blue and light-green SF yellowish (also known as the counterstain for geomori methenamine silver stain) yielded sufficient PCR products; however, their staining characteristics did not afford satisfactory visualization of nuclear chromatin to discriminate between benign and malignant cells. Our most significant finding was that a commonly used histologic stain, hematoxylin, failed to produce DNA templates that could be consistently amplified by PCR. In conclusion, it is prudent to avoid hematoxylin stains when preparing tissues as starting material for PCR. Among the remaining five stains that were evaluated, the best choice depends on the differential staining characteristics of the cells to be dissected.

No abstract