Comparison of in situ hybridization and immunocytochemistry for the detection of residual beta cells in the pancreas of streptozotocin-treated diabetic rats.

The relative efficacy of immunocytochemistry versus in situ hybridization in identifying residual beta cells was studied in rats with streptozotocin-induced diabetes. Consecutive sections of pancreas of streptozotocin-treated diabetic rats and control animals were alternately subjected to in situ hybridization (synthetic oligonucleotides complementary to rat preproinsulin mRNA) and immunocytochemistry (monoclonal antibodies to insulin). The results obtained with both methods were quantitated with the use of computer-assisted image analysis, and the ratio of cells positive by immunocytochemistry to those positive by in situ hybridization was determined. Under normoglycaemic conditions the values obtained by immunocytochemistry correlated well with those obtained by in situ hybridization (immuno/in situ > 95%). In the streptozotocin diabetic animals, however, immunocytochemistry resulted in a distinct underestimation of the number of residual beta cells (immuno/in situ < 80%). This difference was even more striking in small islet cell clusters (< 100 microns) (immuno/in situ 20%). These results suggest that in situ hybridization for prohormone mRNA is the method of choice for the identification of residual or regenerating beta cells with very low insulin content. Caution should be used when interpreting quantitative data in diabetic conditions that are based exclusively on immunocytochemical detection methods.