MORPHOLOGICAL AND FUNCTIONAL CHARACTERIZATION OF THE RAINBOW TROUT GUT (ONCORHYNCHUS MYKISS) TO DEVELOP A PREDICTIVE IN VITRO INTESTINAL MODEL

To obtain an adequate comprehension of the intestinal morphology and function, I performed a detailed morphological evaluation of the intestinal epithelial cells lining the rainbow trout gut. This was obtained through histological, histochemical and immunohistochemical observation, accurate stereological analysis, morphometric measurements, and fine qualitative and quantitative characterization of goblet cells secreting activity at typical time points of in vivo feeding trials (50, 150 and 500 g). The gross anatomy of the rainbow trout gut corresponds to the general description of the organ in teleost fish. It consists of a proximal intestine with annexed pyloric caeca, easily distinguishable from the distal intestine, which presents a larger diameter, dark pigmentation, and circularly arranged blood vessels. Microscopically, the proximal intestine was characterized by simple folds whereas the distal intestine displayed a more complicated mucosa arrangement. In fact, this latter, was defined by the presence of peculiar complex folds from which other simpler originated. While only minor changes took place along the classical time points of in vivo feeding trials, two morphological and functional compartments not linearly distributed were observed along the trout gut. The first included the proximal intestine and the apical part of the complex folds of the distal portion. This was characterized by abundant and actively secreting goblet cells, no pinocytotic vacuoles, high expression of three defined functional enterocytes markers (PepT1, Sglt1 and Fabp2), low proliferation rate, few round apoptotic cells and an extended area of fully differentiated cells. The other, comprised the basal part of the complex folds and the pyloric caeca and was defined by few goblet cells with deflated mucus vacuoles, pinocytotic vacuolization, weak expression of the classical functional enterocytes markers, high proliferation and apoptotic rate and by a smaller extension of fully differentiated epithelial cells. The presence of these Chapter 2 23 distinct morphological and functional compartments suggest that digestive and absorptive function are mingled along the trout gut and that possibly in this species, the proximal intestine extend itself in the distal portion of the intestine to maximize its absorptive capacity [65]. To achieve a better knowledge of the cellular and molecular mechanisms implied in the intestinal homeostasis preservation and to further investigate the two different renewal rate previously observed, I performed a detailed qualitative description of the organization of the intestinal epithelial stem cell (IESC) niche and their regulatory molecules. To this end, the typical mouse ISCs markers (LGR5, HOPX, SOX9, NOTCH1, DLL1, and WNT3A) have been selected as target genes, and subsequently their expression have been localized along the different portions of the trout gut trough in situ hybridization. All the target genes were expressed also in rainbow trout intestine. However, considering their typical position in mouse, we highlighted substantial differences. Lgr5+ cells were rare but surprisingly in RT were located along the folds’ stroma rather than restricted in the crypt epithelium. The folds’ connective axis hosted also scattered notch1+ as well as wnt3a+ cells where they colocalized with lgr5+ cells. Interestingly, also in mouse these markers colocalize, even if this occurs in the epithelium lining the crypt base. Moreover, close to lgr5+ stromal cell population we observed elongated, epithelial cells expressing dll1. This marker in mouse is selectively expressed by Paneth cells at the crypt bottom. Despite the different spatial position of dll1+ cells in mouse and rainbow trout, it is interesting to note that in both species lgr5 and dll1 are expressed by cells topographically close each other, suggesting a functional interaction. Moreover, the epithelium lining the intestinal folds base in rainbow trout showed exclusively positivity to sox9. Particularly, here, we observed slender cells expressing sox9 at high level, strongly reminiscent of the crypt base columnar cells (CBCs) cells described in mouse intestinal crypts. Sox9 expression tended to disappeared outside del fold base to give way to hopx which was abundantly detected along the folds rather than restricted to the +4 position as described in mouse. Based on these observation, it is reasonable to assume that in this species Chapter 3 66 the functional equivalent of lgr5 is sox9, since sox9+ cells showed the typical location and shape of CBCs in mouse intestine. Whereas, lgr5 expressed by a stromal population along the folds could represent a specific mesenchymal subsets involved in the maintenance of the folds tip epithelium. To expand and integrate our previous findings describing the organization and the architecture of the RT stem cell niche, here I performed a detailed qualitative analysis of the distribution of the stromal component of the niche and in particular of telocytes (TC) as active player of the intestinal homeostasis maintenance. To this purpose, specific histological staining for the connective tissue and ultrastructural analysis have been performed. Furthermore, since the mouse is the species in which the mesenchymal component of the niche has been studied most, I selected the typical mouse TC markers (PDGFRα and FOLX1) as target genes, and I verified their topographical localization through fluorescence in situ hybridization. In addition, I studied their spatial distribution in relation with the epithelial components of the stem cell niche. Our results indicated the presence of slender elongated cells, distributed juxtaposed the enterocytes’ basement membrane creating an intricate mesh extending from the folds base to their apex. TEM analysis confirmed the identity of this cell population as TC; indeed, they possessed the distinctive features of this cell type, including a voluminous nucleus and limited cytoplasm from which long thin discontinuous branches develop. Moreover, in the close proximity of telocytes, I observed extracellular vesicles suggesting their functional implication in long distance cell-to-cell communication. In situ localization of pdgfrα transcripts highlighted two distinct pdgfrα+ cell populations, one displaying the typical morphology and position of TC, and another expressing pdgfrα at lower level and not possessing the distinctive TC morphological features and therefore considered as common fibroblasts. In situ localization of foxl1 transcripts revealed that foxl1+ cells were rare and distributed in the peri-epithelial space both at the folds base and along the folds length. In both locations, foxl1 was always co-expressed with pdgfrα, indicating the existence of a small functional telocytes subset. Basal TC were in the proximity of actively cycling intestinal stem cells, suggesting a possible interaction through short-range signaling. Their morphology and distribution at the basal and apical portion of the intestinal folds were identical to those observed in the mouse. This allows us to infer that also in RT, as in mouse, telocytes stimulate either cell proliferation or cell differentiation depending on their topographical location. Altogether, these results substantially improve our understanding of the molecular mechanism and of the cell types involved in the maintenance of intestinal homeostasis. As mentioned above, despite the complexity organization of the RT mucosa, the development of an in vitro tool able to retain similarities with the in vivo counterpart, would be a helpful tool for the rapid screen of several alternatives feeds implied in aquaculture. After several attempts, two novel cells line belonging from proximal (RTpi-MI) and distal (RTdi-MI) intestine of RT were successfully derived and characterized. Therefore, based on the extensive morphological and functional characterization of the RT mucosa in vivo, in this thesis several efforts have been also dedicated to the derivation of new stable cell lines. Then, they have been compared with the only continuous cell line that so far has been established from rainbow trout intestine (RTgutGC). Gene expression analysis demonstrated that all the cell line presented both epithelial and mesenchymal components. However, despite all of them displayed an epithelial component, this was significantly higher in the RTpiMI. Moreover, they also expressed the typical enterocytes’ functional markers (PepT1 and Sglt1) suggesting their suitability as a model of in vitro absorption. Furthermore, all the cell lines preserved both stem cells and differentiating cell types. High seeding density induced their differentiation into more mature phenotypes, as indicated by the simultaneously downregulation of intestinal stem cell-related genes (i.e., sox9, hopx and lgr5) and upregulation of alkaline phosphatase activity. Nevertheless, among the three cell lines most parameters that we analyzed were conserved, some significant differences were observed. Therefore, this suggests that some typical features that characterize the two main intestinal portions are also preserved in vitro. Moreover, since only a mild stimulus was able to induce their differentiation it would be interesting to further explore these mechanisms using more controlled and sophisticated culture systems.