Microarray as a model for quantitative visualization chemistry.

ABSTRACT

For visualization of proteins or nucleic acids, direct and indirect in situ fluorescence and absorption methods (immunohistochemistry and cytochemistry) have existed for many years. The authors describe a new experimental approach using microarray as a model to quantitatively compare both visualization methods. The spots obtained with the microarray robot had a progressive twofold decrease in concentrations and are used as objects with known amounts of DNA. Subsequent hybridization resulted in a direct fluorescence (DF) label or in hapten for indirect fluorescence (IF) and absorption. The results show that the image of the object in the IF method is larger than that in the DF method because of an edge effect, with stronger staining at the circumference. This leads to a higher plateau level and an 8- to 10-fold reduction in the detection threshold for IF compared with DF. These features are especially useful for one-color DNA-related microarray analysis, such as single nucleotide polymorphism, loss of heterozygosity, and mutation analysis, provided that the spots are not designed directly adjacent to each other, so that the edge effect is taken into account. The slope of the linear range for the IF method is much steeper than for the DF method, pointing to a narrow dynamic range in immunohistochemistry. It is noteworthy that the detection limit for absorption images after indirect immunoenzyme visualization is lower than for the DF images. The indirect immunohistochemistry semiquantitative absorption signal was at least similar compared with the DF fluorescence. In conclusion, an explanation for the difficulties experienced in quantitative immunohistochemistry is provided, and the data emphasize that in general, for daily pathology, semiquantitative patterns should suffice. Indirect labeling of DNA has useful characteristics for application in microarray analyses because of the large signal enhancement.