Microfluorimetry of brain glycolipid. Theoretical and applied analysis of PAS fluorescence and dye concentration

A vertical epi-illuminated microspectrofluorimeter was employed to measure the relative concentration of brain glycolipids in tissue sections after the periodic acid pararosaniline-Schiff (PAS) staining reaction. This histofluorimetric analysis includes systematically varying section thickness, defining an extinction coefficient, and determining the range of fluorescent intensity from measurements made on PAS stained microtome sections of agarose. Coupling fluorimetry of PAS stained brain sections with fluorimetry of serial brain sections stained after extraction with chloroform-methanol (C-M) provided a relative estimate of PAS-reactive glycolipid present in a given region of white or gray matter. Biochemical analysis of C-M extracts from different sections of formalin fixed brain, processed similarly to the tissue sections, showed that the extracted PAS-positive material consisted primarily of glycolipids. The limitations imposed on the biochemical interpretation of the results are discussed. The method presented provides a sensitive technique for comparing the relative amounts of C-M extractable, PAS-positive material in brain tissue sections when appropriate statistics are employed. Linear regression analysis of theoretical data based on the nonlinear formula for absorption of Bouger-Beer produced a slope near unity and an intercept near zero when absorption was less than 20%. As the absorption increased, theoretical data generated lines with increasingly smaller slopes and higher intercepts. Correlation coefficients of these lines generated from theoretical data from small segments of the range and domain exceeded 0.99 (P < 0.05). Data derived from the absorption formula, ranging from 1 to 95% absorption, generated a line with a linear regression correlation of 0.94 (P < 0.05). Thus linear analysis of fluorescence vs concentration seemed valid to interpret changes in dye concentration over narrow ranges. But a linear model analysis of fluorescence intensity made erroneous estimates of dye concentration. A nonlinear least squares analysis used to determine the parameters of the fluorescent-absorption formula produced the best estimate of relative dye concentration.