Double-label confocal laser-scanning microscopy, image restoration, and real-time three-dimensional reconstruction to study axons in the central nervous system and their contacts with target neurons.

The current double tracing-double confocal laser-scanning method was developed to reconstruct identified nerve fibers and their contacts with identified target neurons in the rat brain in three dimensions. It intends to fill the gap between conventional light microscopic and electron microscopic neuroanatomic tracing. The steps involved are as follows: (1) injection of two neuroanatomic tracers--Phaseolus vulgaris leucoagglutinin (PHA-L) to label fibers innervating a particular brain area and Neurobiotin to label prospective target neurons in that area; (2) immunofluorescence detection of the labeled fibers (fluorophore Cy5, infrared emission), together with fluorochromated avidin detection of the taken-up Neurobiotin (Cy2 or Alexa 488; green emission); (3) acquisition of Z-series of confocal images at high magnification with a laser-scanning microscope using the laser lines 488 nm and 647 nm; and (4) computer-processing and three-dimensional reconstruction of the labeled fibers and the presumed target dendrites. Rotation on the computer of the three-dimensional reconstructed fibers and dendrites along all three spatial axes enabled the authors to determine whether "true" or "false" contacts occur. In a true contact no space was present between the apposing structures, whereas a false contact consisted of two differently stained structures close to each other but separated by a narrow, optically empty space. One important phenomenon in the three-dimensional reconstruction of double-stained structures that needed correction was "twin image mismatch"--i.e., the observation that a three-dimensional reconstruction of a small test object (double-stained on purpose) produced two slightly shifted objects, each associated with its particular fluorochrome. To measure the actual twin image mismatch of the confocal instrument and to obtain accurate correction factors the authors took in each session in which they obtained image series of the real experiments, with both laser wavelengths Z-series of images of multifluorescent microspheres (500-nm diameter) and of thin, double-stained fibers. Given the small dimensions of the structures of interest, i.e., synaptic contacts, it is necessary in this type of research that the optical characteristics of the imaging system--e.g., the alignment errors and chromatic aberration that produce twin image mismatch--be precisely known.