Detection and estimation of mRNA levels using a nonlinear model in neurons labeled by in situ hybridization histochemistry.
In situ hybridization histochemistry (ISHH) is an anatomical technique used to monitor gene expression at the cellular level via detection of steady-state levels of mRNA. Previously, densitometric analysis of ISHH-generated autoradiographic material has provided a relatively quantitative measure of the level of a specific mRNA distributed in any given anatomical region. The present study details the development of a parametric modeling technique used to automate the quantitative aspects of ISHH. The ISHH experiments described here utilized a specific DNA probe complementary to mRNA molecules encoding the neuropeptide substance P and related tachykinin peptides. A nonlinear model was used to describe the dark-field intensity pattern of labeled neurons. The model's parameters were then employed in detecting individually hybridized neurons and in estimating levels of preprotachykinin mRNA and associated cellular areas. Total mRNA content was quantified by relating the intensities described by the models to those obtained from 14C autoradiographic standards. Finally, the algorithm's performance was evaluated by comparing these estimates to those obtained from manual grain counts of labeled neurons. Overall, the parametric model presented here facilitates the process of performing quantitative analysis of hybridized neurons based on predetermined and unbiased morphological criteria.