Image analysis of lymphoid cell differentiation in rat thymus throughout development.

In order to identify subtle changes in cell morphology and nuclear pattern modification during thymus ontogenesis, cell image analysis using System for Analytical Microscopy in Biological Applications (SAMBA 200) was applied in 9 stages of rat thymus development. The morphometric and chromatin parameters made it possible not only to identify automatically between the two main cell populations in the thymus gland (lymphoid and epithelial cells), but also to classify automatically 5 lymphoid sub-populations (lymphoid stem cells, lymphoblasts, large lymphocytes, medium lymphocytes and small lymphocytes). The evaluation of the 18 parameters during the lymphoid cell differentiation was studied in detail. The nuclear texture parameters made it possible to discriminate, in each cell subpopulation, 4 phases of cell cycle (G0, G1, S-phase, and cells in G2). Evaluation of the nuclear parameters of the cell cycle in each lymphoid sub-population was studied in this investigation. The results illustrate the high majority of the lymphoid stem cells at the 14-day-old embryo stage while in the 20-day-old embryo the small lymphocytes become the main part of the whole lymphoid population. From the continuously renewed modification of lymphoid nuclear image analysis we discuss the origin of thymus lymphocytes. Lymphoid cells can be distinguished into different functional states and the striking morphological changes appearing during cell differentiation are related with drastic structural changes occurring in chromatin pattern from undifferentiated lymphoid stem cells to small lymphocytes. Terminal cell differentiation is associated with inhibited cell proliferation. The relative increase of chromatin condensation and nuclear pattern heterogeneity which reaches an extreme in small lymphocytes is accompanied by a progressive diminution of the nuclear area during the successive differentiation of the lymphoid population. Using one parameter of the nuclear texture features from the co-occurrence matrix (as LM or CON) and one parameter of the nuclear textures from the run-length section matrix (as GLD or RPC) the image analysis can discriminate between the different states of lymphoid cell differentiation.