Ultrastructural quantitative stereology on 'mixed' cell populations: problems and possibilities.

Contrary to homogeneous tissues, "mixed" tissues or cell suspensions are composed of different cell individuals. They have been characterized as heterogeneous cell populations (composed of cells of different cytogenetical source) and heteromorphous or inhomogeneous cell populations (of cells of the same type but of different individual functional state). Certain problems may arise during ultrastructural morphometrical investigations of heteromorphous populations. Because of the different size of planes of sectioning of individual cells, morphometrical results should be treated by means of modifications of the t-test [4]. Furthermore, the unambiguous classification of individual cell profiles necessitates a conscious selection of cell profiles containing a nuclear profile. This non-random sampling approach leads to systematic errors of computed parameters for the whole cell as the nuclear volume fraction and the specific cell surface area which should be corrected. Besides correction procedures derived from regular geometrical models we have presented correction methods for cells with non-spherical nucleus and a cell surface with marked surface projections [13, 21]. On the other hand, on heteromorphous cell populations, more detailed information about functional events can be gained, in comparison the customary stereological mean values, by means of frequency distributions of morphometrical data of individual cell profiles as well as by correlation of data of different cell organelles.