Optimization of endocrine pancreas fluorescence analysis using machine methods

Abstract

The study aims to establish the appropriate parameters of UV excitation intensity and permanent excitation time on the pancreatic islets photobleaching, the ratio of the intensity of the useful signal in the region of interest to the intensity of nonspecific background fluorescence. Materials and methods. The pancreas of three adult Wistar rats was fixed in Bouin solution (20 hours) and poured into paraplast after standard histological processing. The study was carried out on paraffin sections of the pancreas. The islets’ insulin and glucagon were determined by immunofluorescence method using monoclonal antibodies (Santa Cruz Biotechnology). The immunofluorescence reaction was studied using an AxioImager-M2 fluorescent microscope. AxioVision digital image analysis system was used for fluorescence imaging, and ImageJ 64-bit image analysis system was used for image quantification. 30 pancreatic islets with an area from 3000 µm2 to 5000 µm2 (8–13 % of the frame area) were analyzed. Results. Measurements carried out at constant values of hormone concentration in endocrinocytes showed a different estimate of the average fluorescence intensity for insulin and glucagon, which depended on the intensity of UV radiation. As the intensity of UV radiation increased, the average fluorescence intensity in the region of interest for insulin and glucagon increased, but when the camera exposure was corrected, it became almost the same. Regardless of this, the intensity of nonspecific background fluorescence increased monotonically. The use of the ratio of the logarithms of the background fluorescence of the drug and the fluorescence of endocrinocytes in the calculations gives a stable estimate of the relative concentration of hormones, which does not depend on the intensity of the selected UV radiation regime, as well as on the duration of UV irradiation of the drug. This makes it possible to neutralize the effect of photodynamic discoloration of the preparation caused by continuous irradiation. Methods for machine selection of the region of interest by various algorithms of the ImageJ program lead to different estimates of its area, integral, and average fluorescence values. At the same time, the result closest to the “ideal” interactive method of highlighting the area of interest for insulin and glucagon was shown by Otsu’s algorithm. Conclusions. In immunofluorescent examination of the pancreas, a moderate UV radiation mode should be selected, exposure correction of the CCD camera before taking each frame, and the total time for examining the visual field of the sample should be limited to 1–2 minutes. To highlight the area of interest for insulin and glucagon in automatic analysis, it is recommended to use the Otsu algorithm. To obtain a quantitative estimate of the average fluorescence intensity in the region of interest, it is recommended to use the ratio of the logarithms of the background fluorescence of the drug and endocrinocytes in the calculations.