Limitations of conventional approaches to identify myocyte nuclei in histologic sections of the heart.

Accurate nuclear identification is crucial for distinguishing the role of cardiac myocytes in intrinsic and experimentally induced regenerative growth of the myocardium. Conventional histologic analysis of myocyte nuclei relies on the optical sectioning capabilities of confocal microscopy in conjunction with immunofluorescent labeling of cytoplasmic proteins such as troponin T, and dyes that bind to double-strand DNA to identify nuclei. Using heart sections from transgenic mice in which the cardiomyocyte-restricted alpha-cardiac myosin heavy chain promoter targeted the expression of nuclear localized beta-galactosidase reporter in >99% of myocytes, we systematically compared the fidelity of conventional myocyte nuclear identification using confocal microscopy, with and without the aid of a membrane marker. The values obtained with these assays were then compared with those obtained with anti-beta-galactosidase immune reactivity in the same samples. In addition, we also studied the accuracy of anti-GATA4 immunoreactivity for myocyte nuclear identification. Our results demonstrate that, although these strategies are capable of identifying myocyte nuclei, the level of interobserver agreement and margin of error can compromise accurate identification of rare events, such as cardiomyocyte apoptosis and proliferation. Thus these data indicate that morphometric approaches based on segmentation are justified only if the margin of error for measuring the event in question has been predetermined and deemed to be small and uniform. We also illustrate the value of a transgene-based approach to overcome these intrinsic limitations of identifying myocyte nuclei. This latter approach should prove quite useful when measuring rare events.