Evaluation of murine liver transmission electron micrographs by an innovative object-based quantitative image analysis system (Cellenger)

Abstract

Background: Transmission electron micrographs are widely used to demonstrate tissue damage. However, the results are qualitative and dependent on the experience of the investigator. Recently, a new multiscale object-based quantitative image analyzing systems (Cellenger) has been introduced to study highly textured black-and-white images. It is unknown, whether this system permits the quantitative image analysis of electron micrographs of parenchymal tissue. Therefore, we analyzed whether the Cellenge system permits the quantitative evaluation of electron micrographs of murine liver under normal conditions and after ischemia-reperfusion injury. The results were compared with those obtained by conventional qualitative classification. -

Method: Transmission electron micrographs from murine liver that had been exposed to isolated reversible ischemia at hypothermic conditions of 4 degrees C, 15 degrees C, 26 degrees C and 37 degrees C, and of sham-operated animals, which served as controls (2 images per animal, n = 3 in each group), were analyzed qualitatively by an investigator with experience in electron microscopy. For quantitative analysis, the Cellenger was used and the following damage parameters were studied: ratio of area of endothelial cell nucleus to area of endothelial cell (N/C ratio), ratio of area of hepatocellular vacuoles to area of total hepatocyte cytoplasm (V/C ratio) and ratio of area of microvilli in the space of Disse to area of the sinusoids (M/S ratio). All values were sampled within one group (n=6) and the data given in [%] (MW +/- SEM). P-values were accepted as significant below 0.05.

Results: After normothermic ischemia, all quantitative damage parameter were significantly altered as compared to sham-operated animals (N/C 15 +/- 9% vs. 37 +/- 7%, V/C 18 +/- 4% vs. 0, and M/S 0 vs. 10 +/- 1%) and all hypothermia groups. The qualitative electron micrograph section analysis corresponded very well with these results.

Conclusion: We demonstrate that an multiscale object-based quantitative analysis of transmission electron micrographs from mouse liver under control conditions and after I/R provide accurate classification of relevant tissue damage parameter. The system is now ready to use for further applications within the field of highly textured electron micrographs.