Optimization of non-isotopic in situ hybridization on formalin-fixed, paraffin-embedded material using digoxigenin-labelled probes and transgenic tissues.

The sensitivity of non-isotopic in situ hybridization (NISH), particularly on formalin-fixed, paraffin-embedded (FFPE) clinical tissues, has been the subject of controversy. Generally, NISH has been regarded as being less sensitive than radiolabelled procedures, although some reports have contradicted this. Accordingly, tissues from mice which were transgenic for variable amounts of the human alpha-1-antitrypsin gene were used to optimize the NISH procedure and to estimate the sensitivity. This approach showed that prolonged incubation of slides in final substrate resulted in high sensitivity--about 13 kb of target DNA. However, this prolonged incubation crucially depended on achieving minimal non-specific background staining. Many factors affected the degree of background staining, but five were particularly important. First, the method of mounting cut sections onto slides. Second, the length of the probe (ideally less than 400 bp). Third, the procedure for proteolytic digestion. Fourth, the denaturation technique, and fifth, the quality of the dextran sulphate used in the hybridization mix. The optimized protocol showed variable patterns of mRNA distribution in the transgenic mouse livers, while DNA distribution appeared uniform.