Due to its high nutritional value, global fish consumption has constantly grown in the past few decades, reaching all-time records. Whereas the production of capture fisheries has reached its plateau, aquaculture output has increased fivefold in the last 30 years, becoming the biggest source of the fish we eat [1
]. Nevertheless, intensive fish farming must constantly evolve in order to increase its sustainability, particularly through the search for feed with low environmental impact. However, many of the alternative and sustainable ingredients have a negative impact on the integrity and function of fish intestine [2
]. Therefore, it is important to increase our knowledge on its structure and on the mechanisms that regulate the mucosa renewal and repair.
Fish represent the largest group of vertebrates, including almost 21,000 species, more than all other vertebrates combined; they are characterized by a wide variety of anatomical and physiological differences [4
]. We recently examined in detail the intestinal structure of rainbow trout (RT; Oncorhynchus mykiss
], a member of the Salmonidae family, one of the most successful groups in aquaculture by virtue of its adaptability to a wide range of farming conditions [6
]. Following the nomenclature proposed by Bjørgen et al. [7
] for the Atlantic salmon (Salmo salar
), along the craniocaudal axis of the intestine, we first meet the pyloric caeca, over 50 hollow tubes that branch off the first segment of the mid-intestine, to increase its absorptive surface. These are followed by the rest of the thin and nearly transparent first segment of the mid-intestine that continues in the second segment of the mid-intestine, easily recognizable for its larger diameter and darker pigmentation due to the presence of circularly arranged blood vessels. Fish intestinal mucosa lacks two typical structures found in mammals: the crypts of Lieberkühn, and the villi [8
]. Throughout the length of the whole intestine, the mucosa raises in large folds, and, in the second segment, some of them are up to three times higher with smaller secondary folds branching out along their length. For their peculiar structure, these are known as complex folds [5
We previously described a peculiar distribution pattern of the different cell populations along rainbow trout intestinal mucosa [5
]. The epithelium covering the apical part of the complex folds of the second segment presented abundant and actively secreting goblet cells and no pinocytotic vacuoles in the first segment of the mid-intestine, whereas the pyloric caeca and the basal part of the complex folds in the rest of second segment of the mid-intestine were lined with an epithelium characterized by few deflated goblet cells and high pinocytotic vacuolization. Moreover, these two intersected districts showed distinct turnover rates: the first segment of the mid-intestine and the apical part of the complex folds were characterized by low proliferation and extensive differentiation, while the contrary occurred in the pyloric caeca, the basal part of the complex folds, and the rest of the second segment of the mid-intestine [5
The self-renewal capability of the intestinal epithelium is ensured by a dedicated population of multipotent intestinal stem cells (ISCs), which in mammals are in the well-demarcated space of the Lieberkühn crypts that form a highly specialized supporting niche [10
]. To date, by far the most detailed knowledge on ICSs and their interaction with the surrounding niche has been achieved in mouse [12
]. Two different stem-cell populations coexist in this species: the leucine-rich-repeat-containing G-protein-coupled receptor 5-expressing (LGR5+
) stem cells and the +4-stem cells so called because are located in the fourth position from the bottom of the crypt that selectively express homeobox only protein (HOPX) [16
cells, also called crypt base columnar cells (CBCs), are continuously cycling cells located at the crypt base intermingled among Paneth cells [13
cells are located in the fourth position from the bottom of the crypt, characterized by label retaining and slow cycling [7
]. They enter in action in response to a damage of the epithelium, and most recent data suggest that they can integrate the CBC population. ISC cell division originates a pool of highly proliferating cells known as transient amplifying (TA) cells that expand rapidly to produce terminally differentiated cells that migrate up toward the villus tip with the only exception of Paneth cells that migrate toward the crypt. On the LGR5+
cell surface, the NOTCH 1 receptor is also present, which interacts with its transmembrane ligand Delta-like protein 1 (DLL1) located on Paneth cells. The two together play a crucial role in regulating stem-cell proliferation and differentiation [19
]. ISC maintenance is also regulated by other pathways, such as WNT3A, which is a ligand of LGR5 and acts together with the Notch pathway to modulate stem-cell equilibrium between proliferation and differentiation [15
Whereas detailed knowledge of ISCs and their crypts is growing exponentially in mouse and in humans, very little is known in other species [11
]. Among fish, the organization and the architecture of the intestinal stem-cell niche were recently investigated only in medaka [8
] and in zebrafish [21
]. However, these are two small laboratory species and, as mentioned above, the taxonomic divergence among teleost fish does not allow direct extrapolations.
Given the complex nonlinear three-dimensional (3D) pattern of the RT intestine that is not observed in mammals, this manuscript had two aims. The first was to investigate whether the unusual distribution of cell populations reflects similar functional activities. To this end, we determined the protein expression along the different portions of the intestine of three well-defined functional markers of differentiated enterocytes: peptide transporter 1 (PepT1), sodium–glucose/galactose transporter 1 (SGLT-1), and fatty-acid-binding protein 2 (Fabp2). The second was to characterize the morphological and molecular structure of the RT intestinal stem-cell niche and to determine whether the different proliferative rates previously observed along the different mid-intestine portions correlates with a different organization and/or extension of the stem-cell population.
The first aim of this study was to investigate whether the complex nonlinear 3D distribution pattern that we previously described along the RT intestine [5
] reflects a similar distribution of functional activities.
Our preliminary morphological observations were fully coherent with those that we described previously despite, in the first case, animals being raised in standardized indoor conditions, whereas, in this experiment, animals were farmed in open natural-like conditions. This suggests that these morphological features are typical of this species and are not the result of intensive farming or of a specific diet. We did not observe any fold branching in the first segment of the mid-intestine, as described in previous works [5
], supporting the hypothesis that this phenomenon is not directly related to growth per se, as described by other authors [9
], but rather a consequence of stressful conditions [27
]. Indeed, in our previous study, fold branching occurred in correspondence with the increase of lipid concentration in the diet and with qualitative changes in mucin composition [5
], further supporting this thesis. All this, together with the fact that we did not observe inflammation features [26
] such as supranuclear vacuolization of the enterocytes of the first segment of the mid-intestine, nuclear position disparity, or the presence of intraepithelial lymphocytes, confirmed that the all animals were in perfect health and support the physiological significance of our findings.
Peptide transporter 1 (PepT1
), is a high-capacity, low-affinity transporter and is the main carrier responsible for the uptake of dietary peptides in mammals and fish [28
protein was localized along the brush border of the absorptive epithelial cells lining the mucosa folds, as previously described in rainbow trout alevins [29
]. The signal was highest in the pyloric caeca, while it was moderate in the first segment of the mid-intestine and in the apical part of the complex folds of the second segment of the mid-intestine; however, it was completely absent at their basal part and in the rest of the second segment of the mid-intestine. This is consistent and explains the steady decrease in PepT1
messenger RNA (mRNA) previously described in salmon [30
], rainbow trout [31
], and sea bass [32
]. Our results are also in agreement with previous observations showing a significant decrease in SGLT-1 in the second segment of the mid-intestine compared to the first portion and to the pyloric caeca, supporting the thesis that the rainbow trout displays three kinetic glucose absorption systems corresponding to the three different intestinal tracts [33
]. Moreover, the heterogeneous distribution of Fabp2 along the whole intestine that we observed is fully in agreement with a previous study conducted in salmon in which the authors investigated Fabp2 expression and localization along the different tracts [2
Overall, the correspondence between the complex nonlinear 3D distribution pattern that we described along the RT intestine [5
] is reflected only partially by a similar distribution of functional activities. This holds true only in the first segment of the mid-intestine and in the apical part of the complex folds on one side, which are always different from the rest of the second segment of the mid-intestine on the other. This further confirms the heterogeneous nature of the second segment of the mid-intestine mucosa, where the apical part of the complex folds has the same morphology and, presumably, extends the functions of the first segment. Such structure corresponds well to the nutritional needs of the RT short intestine, apt for processing a highly digestible, nutrient-dense diet, high in protein and low in carbohydrate [35
], extending the active nutrient transport area typical of the first segment of the mid-intestine [36
] in the more distal region.
The second aim of our work was to determine the mechanisms sustaining the higher proliferation rate that we previously observed in the pyloric caeca and in the second segment of the mid-intestine compared to the other sections. To this purpose, we characterized RT intestinal stem cells and their niche.
As in mammals, the RT intestinal epithelium undergoes a constant renewal even though it occurs at a slower pace [37
]. The renewal mechanisms are driven by an intestinal stem-cell population housed in a defined niche [13
]; however, in fish, both the cells and the niche are poorly characterized. In mammals, intestinal stem cells are located in the crypts [11
], while, in fish, they are found at the fold base [8
]. To identify RT intestinal stem cells, we used LGR5, the specific marker for rapidly dividing intestinal stem cells identified in mouse, where it is expressed in specific cells at the bottom of the crypts [10
]. Medaka is the only other fish species where lgr5
expression has been investigated, and its mRNA is confined to the base of the intestinal folds [8
]. On the contrary, we found lgr5+
cells exclusively within the lamina propria at the middle of the fold height. This suggests that lgr5
is unlikely to be a specific marker of RT intestinal stem cells. However, lgr5
expression in a stromal cell subpopulation housed along the villi axis was also recently described in mouse. These were identified as telocytes, and the ablation of these peculiar cells caused a perturbation of the gene expression pattern of enterocytes, suggesting that their function is to ensure the maintenance of an efficient epithelium at the villus tip [40
]. Telocytes have been identified as a subepithelial source of proliferative signals to the stem/progenitor cell compartment of the intestinal stem-cell niche in both mouse [41
] and human [43
]. Therefore, it is reasonable to assume that a similar function may also be played by these cells in RT.
While we did not observe lgr5+
cells at the base of RT intestinal folds, in this location, epithelial cells were expressing different levels of sox9
along the whole intestine length. This is consistent with sox9
expression within the mammalian crypts and in the basal portion of medaka [8
] and zebrafish [23
] intestinal folds. Mouse ISCs not only express lgr5
but also show a peculiar morphology for which they are named crypt base columnar cells (CBCs) [12
]. Interestingly, we observed elongated cells intensely expressing sox9
in the middle of the intestinal folds base, displaying the same typical position and shape of mouse CBC cells. In mouse intestine, CBC cells are interposed between other sox9+
cells; similarly, in RT, columnar strongly sox9+
cells were located among other cells expressing a lower sox9
signal. Moreover, sox9+
cells were much more abundant in the second segment of the mid-intestine and the basal part of the complex folds and in the pyloric caeca compared to the first segment of the mid-intestine and the apical part of the complex folds, consistent with the pattern of proliferating cell nuclear antigen (Pcna) expression previously described in this species [5
]. Notably, sox9+
cells did not coexpress other stem-cell markers and they sharply disappeared outside the intestinal fold base, where sox9
expression was substituted with that of hopx
. These observations support the hypothesis that this distinctive cell population represents the RT ISCs.
We also observed scattered, elongated sox9+
cells along the fold’s axis; therefore, according to our previous observations, they should not express Pcna and, thus, constitute a different subgroup. This is consistent with the identification of two SOX9+
cell populations in mouse, one located at the classical crypt base and the other located along the villus epithelium [45
]. The latter lacked the typical proliferating markers and simultaneously expressed chromogranin A, the specific marker of enteroendocrine cells. These findings suggested that. in mouse. SOX9 expression along the villus epithelium might detect terminally differentiated secretory cell type. Even if we were not able to verify the expression of chromogranin A in our samples, due to the lack of a reactive specific antibody, this hypothesis may be true in RT as well, because sox9+
cells displayed the same elongated morphology of enteroendocrine cells described in zebrafish [46
In mouse intestine, HOPX+ cells are rare and restricted to the +4 position of the intestinal crypt. Conversely, in RT, hopx+ cells were abundant, absent at the fold base, but detected in the epithelial cells along the fold. Few hopx+ cells were also found within the fold connective axis. Taken together, these discrepancies lead us to believe that hopx+ cells in RT cannot be considered the functional equivalent of quiescent ISC as in the mammal intestine. Since the hopx+ cell location pattern also matches Pcna expression, these cells are more likely to correspond to the transient amplifying population, an undifferentiated population in transition toward differentiation.
Consistently with the sox9
expression pattern, hopx+
cells were more abundant in the pyloric caeca and even higher in the second segment of the mid-intestine and in the basal part of the complex folds, while a smaller population was detected in the first segment of the mid-intestine and in the apical part of the complex folds of the second segment of the mid-intestine. Interestingly, in piglet intestine, hopx+
cells also did not follow a linear distribution along the whole intestine and were more numerous in the colon [11
NOTCH1 is another mouse ISC marker that is expressed in the crypt base, specifically by LGR5+ cells, inducing their differentiation toward the absorptive lineages. In RT, notch1+ cells were scattered along the fold connective axis, while it was also expressed by the same cells expressing lgr5. The combined expression of these two molecules must have a different function in this context, and further research is needed to clarify this aspect.
Close to lgr5+
stromal cells, elongated dll1+
epithelial cells were found in RT intestine. They were located along the fold length, and their proximity suggested an interaction between these two cell types. This is consistent with the expression of DeltaD, the dll1
homolog in the zebrafish intestinal epithelium [22
]. In mouse small intestine, Dll1 is expressed by Paneth cells at the crypt base [19
]. Despite their fundamental role, the presence of bona fide Paneth cells has not been detected in many mammals and fish species including RT [5
]. However, DLL1 is also expressed in mouse colon, where Paneth cells are absent [47
]. Taken together, these observations suggest the hypothesis that, in those species where typical Paneth cells are not present, another cell type may functionally substitute them, actively interacting with ISCs.
cells were also found within the stromal axis of the folds, but their functional role and their implications in the intestinal stem-cell niche are unclear. In zebrafish, in the absence of the Delta–Notch signaling pathway, secretory differentiation becomes the default, and the lateral inhibition of a functional Delta–Notch signaling restores a balanced mixture of absorptive and secretory cells [22
]. In the mouse intestine, Paneth cell maturation and differentiation are driven and promoted by Wnt signaling, the evolutionarily conserved pathway that contributes to stem-cell maintenance. Recent studies suggested a redundancy of Wnt sources; indeed, it is expressed by stromal and epithelial cells [48
]. In RT, similarly to mouse small intestine, wnt3a
is expressed by a subepithelial cell population. However, while, in mouse, WNT3A+
cells are in the pericryptal region, in RT, wnt3a+
cells were scattered along the fold connective axis.
Recent studies of mouse small intestine demonstrated that telocytes also express WNT3A [49
]. Interestingly, in RT, the stromal wnt3a+
cell population colocalizes with that expressing lgr5
. This expression pattern is consistent with that of mouse telocytes, supporting the hypothesis that telocytes are also active in the RT intestinal mucosa.
Therefore, our results indicated that mammal intestinal stem-cell markers are also expressed in the RT intestine; however, their localization, the niche architecture, and the interactions among these markers are not conserved. In fact, the trout functional equivalent of lgr5 seems to be sox9, since sox9+ cells showed the peculiar location and shape of CBCs in mouse intestine. On the other hand, while, in mammals, LGR5 is the specific crypt epithelial stem-cell marker, in trout, it is exclusively expressed by a stromal cell population along the fold. These cells could be telocytes, a specialized mesenchymal cell type that was recently indicated as a fundamental component of the intestinal stem-cell niche, able to support and regulate the epithelium interacting with nearby dll1+ cells.
The absence of typical Paneth cells and the presence of a specific epithelial cell population displaying an expression pattern similar to mouse Paneth cells suggests that this specialized cell type might be considered its functional equivalent.
Lastly, our data indicated that the high renewal rate displayed by pyloric caeca and by most of the second segment of the mid-intestine is fueled by a more extensive stem-cell population.