Elimination of autofluorescence in immunofluorescence microscopy with digital image processing.

Abstract

Autofluorescence can be a very disturbing factor in immunofluorescence microscopy. We present here a method to eliminate autofluorescence. The method is based on the fact that most autofluorescent compounds have broad-banded excitation and emission spectra, whereas specific fluorescent probes have narrow spectra. Two images are recorded and digitized, one at a wavelength exciting both the fluorescent probe and the autofluorescent molecules, and one at a wavelength exciting only the latter. Subtraction of the autofluorescence signal from the total fluorescence signal, using a self-developed computer program, results in an autofluorescence-free image. The procedure is demonstrated for elimination of elastin-derived autofluorescence in human lung alveoli and for elimination of lipofuscin-derived autofluorescence in human heart muscle. The autofluorescence signal is positively correlated with tissue section thickness (r = 0.93; p < 0.0001), and can be used to correct the specific fluorescence signals for section thickness.