Use of frozen sections to determine neuronal number in the murine hippocampus and neocortex using the optical disector and optical fractionator

Stereology is an important technique for the quantification of neurons in subregions of the central nervous system. A commonly used method of stereology relies upon embedment of tissue in glycol methacrylates to allow production of sections that are resistant to shrinkage in thickness. However, the use of glycol methacrylates for stereology has several disadvantages, including severe constraints on the size of tissue that can be processed and the long duration of time often required for infiltration. We describe a novel method of stereology utilizing tissue sections cut in the frozen state. This new methodology relies upon the staining of sections as free-floating sections and upon the mounting of these sections onto slides with a water-based mounting media. Sections cut in the frozen state and processed by these methods undergo little or no shrinkage in thickness and are ideal for stereological cell counts utilizing either the optical disector or optical fractionator methods of stereology. We demonstrate that frozen sections can be utilized to estimate neuronal number with high degrees of precision and with low coefficients of error. Because large tissue blocks can be cut as frozen sections, this method expands the range of tissues that can be processed efficiently for stereology and readily allows quantification of neurons from multiple brain regions from the same tissue sections. We applied this new methodology to estimate neuronal numbers in the neocortex and hippocampus of 10-day-old mice. The method was useful for estimation of both large, sparsely packed cell populations, such as the neocortex, and small, densely packed cells, such as the dentate gyrus granule cells. Thus, frozen section methodology offers many potential advantages over the use of glycol methacrylate embedment for stereology. These advantages include expansion of the size of tissue blocks that can be processed, reduction in expended time and costs, and ability to quantify mulitiple brain regions from a single set of sections. (C) 2004 Elsevier B.V. All rights reserved.