Gastrointestinal stromal tumor: an ultrastructural investigation on regional differences with considerations on their histogenesis.Ultrastruct Pathol. 2010 May; 34(3):174-88.UP
Gastrointestinal stromal tumor (GIST) is the most frequent spindle cell tumor in the gastrointestinal tract and may arise from esophagus to rectum. The stomach is the most frequent site, followed by small intestine, rectum, and esophagus. There have been some regional differences reported in their histopathologic and clinical presentations. The purpose of this study is to compare ultrastructural features of GIST, according to its anatomic site, in order to provide additional data to support the current concept of its histogenesis. Fifty-four GISTs (27 from stomach, 23 from small intestine, and 4 from rectum) were included in the study. Histopathologically, gastric GISTs tended to be more frequently epithelioid, particularly those in children, while small intestinal GISTs (SISTs) were mostly spindly in all but three cases. All four of the rectal GISTs were spindly. Ultrastructurally, there seem to be considerable regional differences. In the majority of gastric GISTs, in both epithelioid and spindle types, tumor cells exhibited focal features of myoid differentiation evidenced by the presence of incomplete external lamina (EL) and/or focal accumulations of thin fibers with interrupted electron densities consistent with actin filaments. However, features of myoid differentiation were exceptional for SISTs and rectal GISTs, being present in only one example in each. Some gastric GISTs, particularly those having an epithelioid appearance, showed cell borders luxuriously decorated by long filopods (anemone cell features). Anemone cell features were also present in spindle cell types of gastric GISTs as well as SISTs, albeit it was simpler and less luxuriant. Skeinoid fibers were present in the majority of SISTs and rectal GISTs, but absent in all gastric GISTs except one. These differences appeared to be too significant to propose a uniform histogenesis for all GISTs. Nevertheless, on closer analysis, certain features could be identified to explain a line of differentiation in all GISTs ranging from (1) polygonal uncommitted epithelioid mesenchymal cells with cell borders decorated by luxuriant fimbria, to (2) spindly tumor cells with less prominent fimbria, or (3) cells with or without features of minimum myoid differentiation characterized by the focal presence of cytoplasmic actin fibers or incomplete EL or skeinoid fibers, which might represent an altered product of EL protein. These findings led the author to speculate that the probable primordial cells of GIST may be the primitive mesenchymal cells, which have the potential to differentiate into myoid cells. In this regard, it is important to note that the putative primordial cell of GIST, interstitial cells of Cajal (ICC), and intestinal smooth muscle cells have been shown to develop from the common progenitor cells of the primitive gut, and c-Kit plays a crucial role in the determination of their fate to differentiate to muscle cells or ICC. The author concludes that all GISTs derive from stem cells in the gut retaining some of the differentiation potential seen in primitive gut cells. One of the likely candidates for such cells in the intestinal musculature is ICC-DMP (interstitial cells of Cajal associated with deep muscular plexus) identified as ICC having smooth muscle features identified exclusively by electron microscopy. These cells have been shown to have some of the features of muscle cells by the presence of external lamina and less well-organized cytoplasmic filaments; they also express CD117 in the cytoplasm. Furthermore, recent studies demonstrated the presence of so-called progenitor cells of ICC, similar to ICC-DMP in appearance, expressing insulin-like growth factor and CD34, indicating their stem cell nature. The author proposes that all GISTs develop from the common progenitor cells similar to primitive gut cells, which may differentiate into tumor cells with more myoid features in the stomach (similar to so-called ICC-DMP) as well as spindle cells with less myoid features (similar to ICC-MP [interstitial cells of Cajal associated with the myenteric plexus] in the small intestine and rectum). ICC-DMP have been recruited in the group of ICC by electron microscopic technique alone without methylene blue stain and it is questionable whether they are part of ICC depicted by the ICC network originally shown by Dr. Cajal more than century ago. Recent discovery of their expression of insulin-like growth factors may indicate that they represent persisting primitive gut cells (gut stem cells), which may serve as the progenitor cells to GIST. It is also pointed out that in this era of ICC and GIST pandemonium, a minority of intestinal stromal tumors with mature smooth muscle features have been totally ignored; these now appear to belong to GISTs, representing the best differentiated example among the tumors developing from the same progenitor cells.