In vitro models of intestinal epithelial cell differentiation - Springer Link

14 downloads 0 Views 685KB Size Report
Cell Biol Toxicol 2007; 23: 241–256. DOI: 10.1007/s10565-006-0175-0. C Springer 2006. In vitro models of intestinal epithelial cell differentiation.
Cell Biol Toxicol 2007; 23: 241–256. DOI: 10.1007/s10565-006-0175-0

 C Springer 2006

In vitro models of intestinal epithelial cell differentiation P. Simon-Assmann1,2 , N. Turck3 , M. Sidhoum-Jenny1,2 , G. Gradwohl1,2 and M. Kedinger1,2 1 Inserm U682, 2 Universit´e Louis Pasteur, Facult´e de M´edecine, Strasbourg, France; 3 present address: Centre M´edical Universitaire, Univ. Gen`eve, Geneva, Switzerland Received 5 September 2006; accepted 23 October 2006; Published online: 12 December 2006

Keywords: cell culture, endoderm, established cell lines, human colonic cancer cells, immortalized cells, intestine Abstract The intestinal epithelium is a particularly interesting tissue as (1) it is in a constant cell renewal from a stem cell pool located in the crypts which form, with the underlying fibroblasts, a stem cell niche and (2) the pluripotent stem cells give rise to four main cell types: enterocytes, mucus, endocrine, and Paneth cells. The mechanisms leading to the determination of phenotype commitment and cell-specific expressions are still poorly understood. Although transgenic mouse models are powerful tools for elucidating the molecular cascades implicated in these processes, cell culture approaches bring easy and elegant ways to study cellular behavior, cell interactions, and cell signaling pathways for example. In the present review, we will describe the major tissue culture technologies that allow differentiation of epithelial cells from undifferentiated embryonic or crypt cells. We will point to the necessity of the re-creation of a complex microenvironment that allows full differentiation process to occur. We will also summarize the characteristics and interesting properties of the cell lines established from human colorectal tumors. Abbreviations: CFTR, cystic fibrosis transmembrane conductance regulator; FAE, follicle associated epithelium; GFP, green fluorescent protein; FHI cells, fetal human intestinal cells; IEC, intestinal epithelial cells; L-PK, L-type pyruvate kinase; MAP kinase, mitogen-activated protein kinase; MTX, methotrexate; ngn3, neurogenin 3; ZO-1, zonula occludens-1

The intestinal epithelium as a model in cell biology The intestinal epithelium represents a particularly appropriate system for the study of the fundamental processes involved in cell differentiation (for reviews Potten et al., 1997; Stappenbeck et al., 1998; Roberts, 2000). In early phases of development, the embryonic intestine is composed of a multistratified endodermal cell layer surrounded by the mesenchyme. Cell division during em-

bryonic development gives rise to several characteristic epithelial cell phenotypes that emerge according to a specific chronology. Three cell types of differentiated cells (enterocytes, goblet cells, and endocrine cells) appear before the initiation of crypt formation that starts shortly before birth. At this stage, Paneth cells differentiate in the crypt region. In the adult intestine, the monolayer of epithelial cells is organized into crypt-villus units in which proliferating cells are confined to the crypts, where the pool of stem cells is

242

Figure 1. Structure of the adult small intestine and aspect of the epithelial cells. View of a freshly isolated villus/crypt unit (left) and schematic representation of the crypt-villus axis (right). Putative stems cells (orange) reside in the crypt region where proliferation occurs (yellow). The epithelial sheet is in a continuous upward movement along the villus axis where differentiation takes place. The four main differentiated cell types (blue) found in the intestine are depicted at the bottom of the figure; the endocrine cell is delimited by a dotted line. Enterocytes, mucus cells (goblet cells), entero-endocrine cells migrate upward the villus, whereas Paneth cells migrate to the bottom of the crypts. Cell shedding occurs at the tip of the villi. Myofibroblasts shown to trigger cell differentiation are present underlying the crypt region.

located (Figure 1). The intestinal epithelium constitutes a system of constant and rapid renewal. The different immature cell types differentiate progressively as they migrate out of the crypts toward the tip of the villi, to be finally extruded into the lumen, except Paneth cells, which stay in the crypt region. The distal part of the intestine, the colon, is characterized by elongated glands lined by a simple epithelium in which mostly absorptive cells (colonocytes) and goblet cells are present. It was clearly shown that dynamic reciprocal interactions between the epithelial and mesenchymal tissues are necessary for morphogenetic and differentiation processes that occur

during organogenesis and crypt-villus migration (Simon-Assmann and Kedinger, 1993; Yasugi and Mizuno, 1996; Kedinger et al., 1998). The main cell type found in the differentiated small intestine is the enterocyte. As mentioned above, this cell type derives from the embryonic intestinal endoderm and later in the adult organ from the intestinal stem cells located in the crypt region. Enterocytes are highly polarized cells generated by their contact with the basement membrane and with the adjacent cells. At the apical part of the cell, the brush border membrane, composed of well-organized microvilli rich in transporters and hydrolases,

243 ensures the absorptive and digestive functions. Expression of specific microvillar hydrolases (such as lactase, sucrase-isomaltase, dipeptidylpeptidase IV) is the most reliable indicator of intestinal cell differentiation in vitro (Neutra and Louvard, 1989). Although numerous studies have already been performed, the exact impulses that are necessary to generate such a highly polarized cell from an undifferentiated epithelial cell remain to be fully elucidated. Yet, it is known that this polarized state is acquired gradually and involves numerous factors such as transcription factors, the contact with fibroblasts, secreted factors, and extracellular matrix molecules (Kedinger et al., 2000; Teller and Beaulieu, 2001; Sancho et al., 2004). The interaction of cells with the extracellular matrix is of fundamental importance in many biological processes such as development, cell growth, and cell differentiation. This extracellular matrix network includes the interstitial matrix (found in connective tissue) and the basement membrane. The basement membrane is a specialized, sheetlike, extracellular matrix structure that is found underneath the epithelia in particular (Miner and Yurchenco, 2004). The intestinal subepithelial basement membrane has been shown to contain the major ubiquitous components found in other organs such as laminins, type IV collagen, nidogen, and perlecan (Beaulieu, 1997, 1999; Simon-Assmann et al., 2003). The characteristics and properties of these molecules as well as their assembly have been reported (Li et al., 2003; Yurchenco et al., 2004). The effects of extracellular signals on the cell behavior are mediated via cell-surface receptors such as integrins (Brakebusch et al., 2002). The tissue specificity in the response of epithelial cells to the basement membrane can be explained, at least partly, by the existence of laminin and type IV collagen isoforms. Laminins are heterotrimeric glycoproteins composed of genetically distinct α, β, and γ chains and 15 resulting isoforms have already been identified (Yurchenco et al., 2004). In vivo, the distribution of the laminin isoforms

is tissue-specific and developmentally regulated, suggesting that they are functionally distinct. In the adult intestine, different laminin isoforms are expressed in the subepithelial basement membrane, namely laminin-1 (α1β1γ1), laminin-2 (α2β1γ1), laminin-5 (α3β3γ2), and laminin10 (α5β1γ1). Importantly, these isoforms are specifically distributed along the crypt-villus axis: laminin-1 and laminin-2 are found in the proliferative crypt region while laminin-5 and laminin10 are found underneath the differentiated cells (Simon-Assmann et al., 1998, 2003; Teller and Beaulieu, 2001). Knowledge of the cellular and molecular biology of the developing and differentiating enterocytes has evolved during recent decades, owing to experimental approaches that allow in vitro differentiation, mostly of the enterocytic cell lineage. In this review, we will not go exhaustively through the in vitro techniques and resulting data but rather give examples of enterocytic differentiation obtained from developing or adult intestine. For readers seeking more detailed information, excellent reviews have been published (Kedinger et al., 1987; Neutra and Louvard, 1989; Zweibaum et al., 1991; Pageot et al., 2000). We will also discuss the possibilities, using these cell models, of obtaining the other cell types found in the intestine, namely, the mucus, endocrine, and Paneth cells.

Intestinal embryonic epithelial cells Embryonic cell survival and differentiation are improved by the extracellular matrix The idea of deriving differentiated cells from the embryonic intestine is based on two arguments: first, during fetal stages the intestine is rapidly dividing; and second, the four epithelial cell types emerge from the endoderm. Attempts to induce and maintain enterocytic differentiation from endoderm (separated from the mesenchyme by collagenase treatment) have failed mostly owing to the poor survival of these cells

244

Figure 2. In vitro differentiation of intestinal endodermal cells. Embryonic intestinal endoderm is separated from the surrounding mesenchyme by collagenase treatment. (1) Fate of the endoderm cultured on plastic, showing a limited polarization of cells: (a) phase-contrast-microscopy; (b) electron-microscopy. (2) On a basement membrane substratum combined to collagen I, endodermal cells exhibit small apical brush borders (c) electron microscopy) and lactase immunoreactivity (d). (3) Better survival and differentiation are achieved in co-culture experiments composed of endodermal microexplants seeded on top of preformed monolayers of fibroblastic or mesenchymal cells (e). Basement membrane formation is visualized by a deposition of electron-dense material (arrow in f) and laminin staining (g) at the endodermal–mesenchymal interface. Differentiation of endodermal cells is defined by the presence of well-formed microvilli (h) and expression of lactase at the apical pole of the cells (i). (e, endodermal cells; m, mesenchymal or fibroblastic cells.)

245 in culture (Figure 2). Indeed, the endodermal microexplants, which spread on culture dishes, rapidly died. Yet the use as a substratum of reconstituted basement membrane (containing high amounts of laminin) to which type I collagen was added allowed cells to acquire a polarized state characterized by the presence of apical brushborder microvilli. This onset of differentiation, although limited, was confirmed by the expression of digestive enzymes. But again the survival time of these primary cultures is rather limited (Kedinger et al., 1989). Epithelial-mesenchymal cell co-cultures allow a morphological and functional differentiation The use of a cellular substrate that mimics the in vivo environment allowed more valuable data to be obtained as far as proliferation and differentiation of the endodermal cells are concerned (Figure 2). This was provided by a feeder layer of living fibroblastic cells from intestine or skin on which intestinal endodermal cells were seeded. These co-cultures remain viable for up to 2–3 weeks. Following the deposition of a basement membrane-like structure (visualized by the expression of laminin molecules) at the interface between fibroblastic and endodermal cells, expression of brush-border enzymes was achieved. Furthermore, sucrase, a developmentally lateexpressed enzyme, was induced in the endodermal cells when glucocorticoids were added to the co-cultures (Kedinger et al., 1989; Simo et al., 1992). The induction of epithelial differentiation depends on the contiguity of vital fibroblasts since fibroblast-derived matrices or irradiated fibroblasts were ineffective. The need of heterologous cell–cell contacts for intestinal epithelial cell differentiation can be explained by the need for de novo synthesis of basement membrane molecules required to trigger cell polarity and differentiation. The use of interspecific tissue recombinants combined with species-specific antibodies allowed determination of the precise chronological expression and deposition at the intestinal

basement membrane region of individual matrix components. This strategy allowed us to conclude that some molecules are strictly of epithelial or mesenchymal origin and others (mostly laminin isoforms) are of dual origin (for a review see Simon-Assmann and Kedinger, 2000). Induction of gene expression in cultured embryonic epithelial cells Manipulating gene expression is a powerful method for analyzing gene function in a variety of organ systems including the intestine. Production of mouse models is a lengthy process and requires suitable promoter elements to direct gene expression in vivo. Thus, the goal is to develop culture systems that should allow the maintenance of viable intestinal cells and gene introduction. This last requirement can be provided in two ways: electroporation and adenoviral infection. Lowvoltage electroporation was shown to be an efficient technique for manipulating gene expression in embryonic gut explants, although expression was found in only 5–20 % of the embryonic epithelial cells (Abud et al., 2004). Higher numbers of intestinal epithelial cells were shown to express transduced GFP or GFP-tagged fusion proteins using a closely similar approach (Tou et al., 2004). However, the efficiency is variable depending on the method used or the tissue analyzed. This is exemplified by the fact that in the avian embryonic stomach the efficiency is higher, reaching up to 50% of the cells (Fukuda et al., 2000). The second method is the adenovirus-mediated infection of cells. Validation of the use of this technique is provided by experiments carried out in our laboratory to test the ability of the transcription factor Ngn3 to trigger the endocrine cell differentiation in the intestine. Preliminary data show that intestinal E14.5 endodermal microexplants infected with an adenovirus-ngn3/GFP construction are highly sensitive to the infection (Sidhoum-Jenny et al., personal observation). Indeed, ectopic adenoviralmediated expression of the pro-endocrine gene ngn3 (Jenny et al., 2002) in endodermal progenitor

246 cells, triggers entero-endocrine differentiation at the expense of enterocyte development. Recently, it was shown that the combination of adenoviral infection and dispase digestion (which slightly separates the epithelial cells, allowing accessibility of the adenovirus receptors) significantly improves the introduction of exogenous genes into the intestinal endoderm (Quinlan et al., 2006).

Nontransformed intestinal epithelial cell lines Cells displaying crypt-cell characteristics Quaroni was the first author to report the establishment of long-term epithelial cell cultures from rat intestines, which were named IEC cells (Quaroni et al., 1979). These cells were derived from different levels of the intestinal tract and were obtained by keeping the explants floating in the medium in the presence of collagenase to reduce the growth of mesenchymal cells. Individual colonies were then transferred to culture dishes of increasing diameters. These cell lines were extensively characterized and have become a widely used model for crypt cells in culture (for references see Kedinger et al., 1987; Quaroni et al., 1999). Indeed these cells are characterized by the specific expression of antigenic determinants of intestinal crypt cells; they are nontumorigenic in nude mice and remain undifferentiated in culture. Yet they are insensitive to exogenous treatments, as glucocorticoids or insulin—known to stimulate intestinal enzyme activities in vivo—did not modify expression levels of the brush-border enzymes (Quaroni et al., 1999). More recently, intestinal epithelial cell cultures were obtained from 17- to 19-week fetal human intestine using thermolysin as a dissociating enzyme. These cell lines retain the ability to express specific cytokeratins and, like the IEC cells, present typical proliferative crypt cell characteristics. Like their rodent counterparts, they are unable to undertake a differentiation program in vitro (Perreault and Beaulieu, 1996).

Need of stromal and vascular factors to differentiate Cytodifferentiation of IEC cells could be achieved by placing these cells in a microenvironment that mimics the in vivo situation: i.e., in conditions providing interactions of cells with intestinal mesenchyme and systemic factors brought about by grafting the cells under the rodent kidney capsule (Figure 3). Indeed, when sheets of IEC cells were associated with fetal intestinal mesenchyme and subsequently grafted, they achieved cytodifferentiation and morphogenesis into a villus epithelium. Interestingly, the four main intestinal cell types—absorptive, goblet, endocrine, and Paneth cells—were identified by electron microscopy. Furthermore, this morphological differentiation is accompanied by the expression of typical brush-border hydrolases, such as sucrase and lactase (Kedinger et al., 1986). These data establish that epithelial–mesenchymal interactions combined with an appropriate humoral microenvironment are necessary to allow differentiation of crypt cells.

Immortalized intestinal epithelial cell lines Derived from transgenic mice There is good evidence that transgenic mice carrying an oncogene placed under the control of a tissue-specific gene provide a powerful tool for producing immortalized cell lines of different origins. As a example, a strain of transgenic mice carrying the SV40 Tag (simian virus 40 large tumor antigen) under the regulatory sequences of the L-type pyruvate kinase (L-PK) gene was used to derive immortalized mouse intestinal cell lines (Bens et al., 1996) as well as renal proximal and collecting duct cell lines (Vandewalle, 2002). The L-PK gene is physiologically expressed in the liver, kidney, small intestine, and endocrine pancreas; its transcription is controlled by a high-carbohydrate diet. A clone of intestinal

247 cells, m-ICcl2 was derived from microdissected explants of the lower part of the intestinal villus axis from 20-day-old fetuses of L-PK/Tag1 transgenic mice. The morphological and immunocytochemical results indicate that the m-ICcl2 cells are slightly polarized with rudimentary brush borders and tight junctions (ZO-1 staining). The intracellular accumulation of sucrase-isomaltase and the expression of the poly-immunoglobulin receptor and CFTR chloride channel activity in m-ICcl2 confirm their crypt phenotype (Bens et al., 1996). In agreement with these characteristics of immature cells, preliminary data from our laboratory suggest that they can be differentiated toward the entero-endocrine lineage upon forced expression of the pro-endocrine gene ngn3. These cells were also recently used for the study of transcription activity, allowing the identification of important activating elements such as Cdx in the chemokine CCL25 promoter (Ericsson et al., 2006). Derived from virus-infected primary cell cultures

Figure 3. Importance of stromal and vascular factors in inducing differentiation of IEC cells. At top, scheme of the association and grafting procedure: Crypt cell cultures (IEC-17 cells) were sandwiched between two fetal embryonic intestinal mesenchymes (∗ ). The associations were then grafted under the kidney capsule of adult animals and harvested 9 days later. In these conditions, morphogenesis and cytodifferentiation can be achieved, although in a limited number of grafts. The luminal side of the graft reveals the presence of villi lined by a monostratified epithelium (a) scanning electron microscopy; (b) light microscopy. Electron microscopy shows the presence of a well-defined basement membrane (arrow in c) and the presence of the four main epithelial cell types: enterocytes (characterized by the apical brush-borders in (d), mucus cells (e), endocrine cells (f), and Paneth cells (arrow in g). For further details see Kedinger et al. (1986).

Immortalization of primary cultures can also be achieved directly with viral oncogenes. This process has been further refined by the use of a temperature-sensitive mutant of the SV40 Tag, which allows growth of the immortalized cells at the permissive temperature and induction of their differentiation by a shift to a higher temperature. Using this strategy, Quaroni and Beaulieu (1997) have derived conditionally immortalized intestinal epithelial cell lines from human fetal small intestines (called tsFHI cells). The data showed that, at 39◦ C, differentiation-related alterations were indeed noted, such as p21 induction, irreversible loss of proliferative potential, and increase in frequency of differentiated cells.

Colorectal tumor cell lines Efforts to develop normal colon cell lines were disappointing owing to lack of viability and loss of the differentiation markers. Since 1964, when

248 Jorgen Fogh established the first human colon carcinoma cell line, HT-29, a large number of cell lines have been derived from human colon cancers (Fogh and Trempe, 1975; Fogh et al., 1977). In 1977, 127 cultured human tumor cell lines were derived from colon carcinomas that vary widely in their degree of differentiation. Like the tumor of origin, these cell lines exhibit high diversity in the type/state of differentiation, in their proliferation, and in metabolic properties. Under standard cell culture conditions most cell lines do not differentiate. They have been paid increasing attention in recent years, following the finding that depending on the culture conditions, some of these cell lines were able to express differentiation features that are characteristic of mature intestinal cells such as enterocytes or mucus cells. We will describe two well-known cell lines, namely, Caco-2 and HT-29 cells, that are able to exhibit interesting differentiation properties in culture, and the HRA-19 cell line that is able, under certain circumstances, to give rise to endocrine cells. The Caco-2 cell line A well-differentiated cell line. Caco-2 cells were derived from a relatively well-differentiated tumor. They spontaneously undergo a typical enterocytic differentiation, a process that is growthrelated (Figure 4). Indeed, at early stages of culture the cells are undifferentiated, while at confluence they form polarized monolayers of cells joined by tight junctions and present well-developed apical microvilli. Although derived from adult human colon (where microvillar hydrolases are not present), Caco-2 cells express enzymes— disaccharidases and peptidases—that are typical of normal small-intestinal villus cells; they also transport ions and water toward the basolateral membrane, forming domes in culture (Pinto et al., 1983). Twenty-six clones were isolated from early and late passages of the parental Caco-2 cell line. These clones differ dramatically in the levels of sucrase-isomaltase and rates of glucose consumption; other differentiation-associated proper-

ties were not modified. The clones, such as the TC7, that express high amounts of sucrase, show a homogeneous apical distribution of the enzyme (Chantret et al., 1994). Linked to all these properties, this cell line is of potential interest for studies related to molecular events associated with cell polarity, enzyme biogenesis, and transport properties, for example. Yet, one needs to keep in mind that these highly enterocyte-like differentiated cells are not normal cells but malignant cells carrying mutations in several genes such as p53, APC, β-catenin and Smad4 (Gayet et al., 2001). Differentiation into M cells. One first example illustrating interesting properties of the Caco-2 cell line is given by the experiments of Kerneis et al. (1997). This group aimed to study the transport of antigens and microorganisms through the Peyer’s patches in the intestine. These structures are organized lymphoid follicles—with a germinal center containing B lymphoid cells that is surrounded by a T cell area—covered by a specialized monostratified epithelium, the follicle-associated epithelium (FAE), in which specialized M cells are found. These cells have an important activity in the transport of antigens that are taken up in the lumen and presented to the underlying lymphocytes where the immune response takes place. M cells are morphologically characterized by a lack of apical brush border and the presence of an invaginated basal plasma membrane. It was postulated that M cells could either be derived from stem cells like the other intestinal epithelial cells or result from a phenotypic conversion of differentiated enterocytes as a consequence of interaction with the lymphoid microenvironment (for a review see Kerneis and Pringault, 1999). One original way to discriminate between these two hypotheses was to perform co-cultures mimicking the in vivo situation using transwell chambers. In practice, freshly isolated lymphocytes were added to the chamber facing the basal membrane of differentiated Caco-2 cells. Under these conditions, a conversion of the differentiated Caco-2 cells into cells sharing common properties of M cells was observed:

249

Figure 4. Spontaneous differentiation of the human Caco-2 colonic cancer cell line in culture. The Caco-2 cells progressively differentiate in culture as a function of the culture time. In early phases of the culture (3–4 days), Caco-2 cells display only few apical microvilli (a) and do not express sucrase immunoreactivity (b). At confluence, numerous and dense microvilli are present (c) and sucrase activity is present at the apical cell surface of numerous cells (d). (a) and (c), electron micrographs; (b) and (d), immunofluorescence staining on nonpermeabilized cells.

striking disorganization of the brush border, loss of enzyme activity, and concomitant acquirement of the capacity to transport bacteria such as Vibrio cholera. These results using the human colon cancer cell line Caco-2 allow to conclude that differentiated absorptive enterocytes are able to convert to M cells (Kerneis et al., 1997; Kerneis and Pringault, 1999). Regulation by extracellular matrix molecules. The regulation of intestinal epithelial cell differentiation by extracellular matrix proteins has been

approached using the Caco-2 cell line in particular. For example, it was shown that sucrase activity was found to be significantly higher in cells grown on the basement membrane constituent laminin than in cells cultured on interstitial type I collagen or plastic (Basson et al., 1996). As mentioned above, different laminin isoforms are expressed in the intestine and one question raised was whether there is a specificity in the cellular response in relation to a defined laminin substratum. It was found that the laminin isoforms are indeed functionally distinct at early stages of culture, with laminin-1

250 stimulating intestinal differentiation in contrast to laminin-2 or laminin-10. Indeed, sucrase was already detected at the apical surface of Caco-2 cells grown for 3 days on laminin-1, but was absent in cells cultured on the other two substrata. In parallel to sucrase stimulation, a strong nuclear immunoreactivity of Cdx2 was observed. Cdx2 is an intestine-specific transcription factor, known to activate transcription of genes encoding typical intestinal differentiation markers, in particular sucrase. These data allowed us to conclude that laminin-1 is the only laminin isoform studied able to accelerate differentiation of Caco-2 cells at early stages of culture, before the cells have deposited their own matrix (Turck et al., 2005). In order to clarify the mechanisms underlying specific cell responses to laminins, a proteomic approach was used to investigate differences in nuclear protein expression of Caco-2/TC7 cells cultured on laminin-1 or not so cultured. The strategy employed was based on a high-resolution 2-D polyacrylamide gel electrophoresis followed by the identification of the differentially expressed spots by mass spectrometry and validation of the data by immunodetection or RT-PCR analysis of the potentially interesting molecules. This technique was used in a first step on nuclear proteins from proliferative and differentiated Caco2/TC7 cells: 400 protein spots were detected and 85 spots corresponding to 60 different proteins were identified (Turck et al., 2004). Among the differentially expressed nuclear proteins, nucleolin was found to be of potential interest as it was specifically associated with the proliferative stage and was drastically downregulated as differentiation proceeded. Similarly, exogenous presence of laminin-1 decreased nuclear nucleolin level in Caco-2/TC7 cells. Knockdown of nucleolin expression by the small interfering RNA strategy mimicked the effect of laminin-1 as it resulted in the induction of cell polarization and differentiation. It was also shown that both effects of laminin1 on Caco-2/TC7 cells, induction of sucrase and loss of nuclear nucleolin, are mediated by a β1integrin dependent cascade that implicates activa-

tion of the p38 MAPK pathway (Turck et al., 2006 and Figure 5). The HT29 cell line The differentiation potential varies according to the culture conditions. The human HT-29 adenocarcinoma cell line is considered to be a pluripotent intestinal cell line. These cells have been shown to have a very high rate of glucose consumption. Under glucose supply and in the presence of serum (classical standard conditions), HT29 cells are undifferentiated. They grow as a multilayer of unpolarized, undifferentiated cells and do not express markers of functional epithelial cells. However, HT-29 cells are able to express various differentiation characteristics under the influence of culture medium changes or of differentiation inducers (Figure 6). Various experimental means were used to differentiate the HT-29 cells: glucose-deprivation, substitution of glucose by galactose, and use of inosin or uridine as carbon sources (for reviews see Neutra and Louvard, 1989; Zweibaum et al., 1991). In all these conditions, HT-29 cells will differentiate, expressing brush-border enzymes; although present, their activities are much lower than those observed in the normal small intestine or in the Caco-2 cells. In serum-free medium, 50% of the HT-29 cells differentiated as goblet-like cells with the presence of mucins. Derivation of various differentiated cell clones. The interest of the HT-29 cell line has been further extended owing to the isolation of different cell clones (Figure 6). The first example is provided by the HT-29/16E clone obtained after a preliminary long-term treatment of the parental HT-29 cell line with sodium butyrate. This clone has the ability to differentiate into mucus-secreting cells (Augeron and Laboisse, 1984). When HT-29 cells are treated with increasing concentrations of the anticancer drug methotrexate (MTX) they form, when adapted at 10−7 mol/L MTX, a mixed population of absorptive and goblet cells; at higher

251

Figure 5. Nucleolin, a nuclear RNA binding protein, is deregulated by basement membrane laminin-1 in Caco-2/TC7 cells. (A) Nuclear expression of nucleolin is dependent on the substratum: nucleolin staining is present as small nuclear dots in Caco-2/TC7 cells cultured on plastic dishes or on laminin-10 (a laminin isoform), while a decrease of immunoreactivity was obvious in cells cultured on laminin-1. (B) Nucleolin inhibition induces differentiation of Caco-2/TC7 cells: a small interfering RNA strategy, used to knock down nucleolin expression, leads to a clear induction of the morphological polarization of the cells (arrows point to the brush border membrane) and of sucrase as compared to control cells such as Caco-2 cells cultured on laminin-1 (not illustrated). The scheme summarizes data obtained by Turck et al. (2005, 2006) showing that the β1-integrin-mediated adhesion of cells to laminin-1 leads to signal transduction that implicates the p38 MAP kinase and to the modulation of differentiation gene expression via loss of nucleolin and induction of Cdx2.

concentrations (10−6 and 10−5 mol/L) goblet cells secreting gastric mucins are exclusively found. Unexpectedly, increasing the MTX concentration to 10−4 and 10−3 mol/L resulted in a shift of differentiation from the mucus to the absorptive phenotype. It is of note that this phenotypic change coincides with a major genetic event, namely, a strong amplification of the gene coding for the target enzyme of MTX, dihydrofolate reductase (Lesuffleur et al., 1991a). Using clones isolated from HT29-MTX 10−5 cells, it has been shown that the resistance of cells to high concentrations of MTX is restricted to cells committed to enterocytic differentiation (Lesuffleur et al., 1998). Likewise, stepwise adaptation of HT-29 cells to another anticancer drug, 5-fluorouracil, resulted in the emergence of adapted differentiated subpop-

ulations according to two phenotypes: polarized dome-forming cells and goblet cells that secrete colonic mucins. Resistance to 5-fluorouracil is acquired through gene amplification of thymidylate synthase (Lesuffleur et al., 1991b). All these cell lines are valuable models that should allow determination of the factors controlling the commitment of specific cell types, namely goblet or absorptive cells. The HRA-19 cell line has the ability to differentiate into endocrine cells in xenografts As shown before, Caco-2 and HT29 cell systems provide excellent models to study enterocyte and mucus-like functions. Differentiation into endocrine cells—another major cell type found in

252

Figure 6. Human colonic cancer HT-29 cells are able to differentiate into different cell types according the culture conditions. Schematic representation showing that from a heterogeneous undifferentiated multilayer of HT-29 cells, different populations of cells can be obtained displaying distinct properties. HT29-Glc− , cells grown in a glucose-deprived medium; HT29-FU and HT-29-MTX, cells grown in a medium in which 5-fluorouracil or methotrexate was added at the indicated concentrations. Scanning electron microscopy shows the presence of dense microvilli at the apical surface of HT29 cells grown in absence of glucose ( c: cells). For further details see Lesuffleur et al. (1991a, b) and Augeron and Laboisse (1984).

the intestine—was not observed in the previously mentioned cell lines. Kirkland (1986) established the HRA-19 cell line from a primary adenocarcinoma of the rectum. These cells have certain characteristics that have not been reported for the

other human colorectal adenocarcinoma cell lines, particularly the ability to differentiate into endocrine cells under in vivo conditions. In vitro, the cells form structurally polarized heterogeneous monolayers, with a slight proportion of cells

253 expressing apical alkaline phosphatase activity. When the HRA-19 cells were injected to nude mice (xenografts), the developed tumors had a typical adenocarcinoma histology. Of interest is the fact that endocrine cells were detected, generally located in the basal region of the multilayered neoplastic epithelium. It was proposed that this cell line is derived from malignant progenitor cells that retain the ability to differentiate (Kirkland, 1986). Human tumor colorectal cell lines are potent models in intestinal cell biology Linked to the variety of human colorectal tumor cell lines that differ in their characteristics and properties, these cell lines have been shown to be useful models for studying the different facets of cell biology. The impressive number of results obtained with cultured cells reflects the potential of these models. They were or are still used for cell polarity studies; investigation of the synthesis, targeting, and processing of apical membrane glycoproteins (such as hydrolases); understanding the formation of the polarized cytoskeleton (apical microvilli); elucidating the mechanisms linked to endocytosis and transport of macromolecules; and defining the adhesion/penetration of bacteria and viruses. Numerous receptors are carried by the human colorectal cell lines which transmit signals to the cells in order to control cell maturation and function: these include receptors for regulatory molecules originating from the neuroendocrine system (such as vasoactive intestinal peptide, neuropeptide Y), receptors for glucocorticoids, and receptors of cell adhesion molecules allowing studies on cell signaling. Often, the presence of a particular receptor in a cell line may depend on the degree of cell differentiation. Regulation of intestinal ion transport can also be approached by studying the effect of secretagogues. Of course, these established cell lines are also widely used as model systems for the study of biological pathways in the development and treatment of cancer. Cell lines present the same reper-

toire of genetic alterations as primary tumors (Gayet et al., 2001) and could be useful for relating biological and pharmacological changes with particular genetic alterations. Furthermore, they can contribute to the understanding of the dynamics of cell–extracellular matrix interactions involved in cell migration, proliferation, and tumorigenicity (Stutzmann et al., 2000). As an example, a link between laminin expression and tumor growth was reported using HT-29 cells. Indeed, a laminin-α1 expression vector was introduced into the HT-29 cells in which this chain was not expressed. In spite of similar growth properties with the control cells in vitro, the laminin-α1 transfectants showed significantly increased tumor growth when injected into nude mice, with increased recruitment of host stromal and vascular cells. These data emphasize the importance of laminin-1, in the context of epithelial/stromal cell interactions, as a chemoattractant of both stromal and vascular cells (De Arcangelis et al., 2001).

Conclusion Various in vitro models have been established in order to study, in the intestine, the process(es) involved in cell generation, migration, and differentiation from a stem cell population. As emphasized in the present review, culture cell models have already proved to be valuable tools for the study of interesting questions related to cell differentiation and function. But so far no models have been developed allowing the generation in culture of the four cell types characteristic of the intestine, which are enterocytes, mucus cells, endocrine cells, and Paneth cells. Surprisingly, primary cultures or cell lines originating from the embryonic endoderm or from the crypt region, which are populated by pluripotent stem cells, give rise spontaneously to only one epithelial cell population, namely, the enterocyte. This enterocytic differentiation is limited, owing mostly to the poor survival of cells in primary cultures but also to the obvious lack of undefined factors allowing

254 differentiation. One way to accelerate this process is to provide a fibroblastic support, which needs to be in direct contact with epithelial cells. Yet, placed in a more complex microenvironment context where 3-dimensional organization is achieved and in long-term grafting conditions, embryonic endoderm or crypt cells are able to derive the full repertoire of differentiated epithelial cells. This can be explained by the need for sequential implication of transcription factors, of humoral factors brought by vascularization, and of the extracellular matrix molecules, some of which are secreted by fibroblasts. Some tumoral colorectal cells have the ability to be permanently maintained in culture and exhibit enterocyte-like phenotypes and/or mucus-secreting functions according to the experimental conditions. Owing to their properties, these cells are particularly well adapted for the study of a large panel of mechanistic cell biology approaches and of cell–microorganism interactions, as well as for screening of the pharmacological effects of new drugs.

Acknowledgments Part of our work is at present supported by a grant of the Association pour la Recherche sur le Cancer (Grant number 3666 to P.S.A.), Inserm (Avenir Grant to G.G.), the Juvenile Diabetes Research Foundation (Center Grant to G.G.), and the EU (Integrated project 6th FP “Betacelltherapy” to G.G.).The authors thank Dr. T. Lesuffleur (INSERM U763, Paris) for sharing comments and figures on the human colorectal cell lines. We are very grateful to people in the laboratory for their work over the years. We warmly thank L. Eleuterio for preparation of the manuscript and for artwork.

References Abud HE, Lock P, Heath JK. Efficient gene transfer into the epithelial cell layer of embryonic mouse intestine using low-voltage electroporation. Gastroenterology. 2004;126:1779–87.

Augeron C, Laboisse CL. Emergence of permanently differentiated cell clones in a human colonic cancer cell line in culture after treament with sodium butyrate. Cancer Res. 1984;44:3961– 9. Basson MD, Turowski G, Emenaker NJ. Regulation of human (Caco2) intestinal epithelial cell differentiation by extracellular matrix proteins. Exp Cell Res. 1996;225:301–5. Beaulieu J-F. Recent work with migration/patterns of expression: cell-matrix interactions in human intestinal cell differentiation. In: Halter F, Winton D, Wright NA, eds. The gut as a model in cell and molecular biology. London: Kluwer Academic; 1997:165– 79. Beaulieu J-F. Integrins and human intestinal cell functions. Front Biosci. 1999;4:310–312. Bens M, Bogdanova A, Cluzeaud F, et al. Transimmortalized mouse intestinal cells (m-ICcl2) that maintain a crypt phenotype. Am J Physiol Cell Physiol. 1996;39:C1666–74. Brakebusch C, Bouvard D, Stanchi F, Sakai T, Fassler R. Integrins in invasive growth. J Clin Invest. 2002;109:999–1006. Chantret I, Rodolosse A, Barbat A, et al. Differential expression of sucrase-isomaltase in clones isolated from early and late passages of the cell line Caco-2—evidence for glucose-dependent negative regulation. J Cell Sci. 1994;107:213–25. De Arcangelis A, Lefebvre O, M´echine-Neuville A, et al. Overexpression of laminin α1 chain in colonic cancer cells induces an increase in tumor growth. Int J Cancer. 2001;94:44–53. Ericsson A, Kotarsky K, Svensson M, Sigvardsson M, Agace W. Functional characterization of the CCL25 promoter in small intestinal epithelial cells suggests a regulatory role for caudal-related homeobox (Cdx) transcription factors. J Immunol. 2006;176:3642–51. Fogh J, Trempe G. New human tumor cell lines. In: Fogh J, ed. Human Tumor Cells in vitro. New York: Plenum Press; 1975;115– 41. Fogh J, Fogh JM, Orfeo T. One hundred and twenty-seven cultured human tumor cell lines producing tumors in nude mice. J Natl Cancer Inst. 1977;59:221–5. Fukuda K, Sakamoto N, Narita T, et al. Application of efficient and specific gene transfer systems and organ culture techniques for the elucidation of mechanisms of epithelial–mesenchymal interaction in the developing gut. Dev Growth Differ. 2000;42:207– 11. Gayet J, Zhou XP, Duval A, et al. Extensive characterization of genetic alterations in a series of human colorectal cancer cell lines. Oncogene. 2001;20:5025–32. Jenny M, Uhl C, Roche C, et al. Neurogenin3 is differentially required for endocrine cell fate specification in the intestinal and gastric epithelium. EMBO J. 2002;21:6338–47. Kedinger M, Lacroix B, Marxer A, Hauri HP, Haffen K. Fetal gut mesenchyme induces differentiation of cultured intestinal endoderm and crypt cells. Dev Biol. 1986;113:474–83. Kedinger M, Haffen K, Simon-Assmann P. Intestinal tissue and cell cultures. Differentiation. 1987;36:71–85. Kedinger M, Bouziges F, Simon-Assmann P, Haffen K. Influence of cell interactions on intestinal brush border enzyme expression. In: Kotyk A, Skoda J, Paces V, Kostka V, eds. Highlights modern biochem. Zeist: VSP International Science Publishers; 1989;1103–12.

255 Kedinger M, Duluc I, Fritsch C, Lorentz O, Plateroti M, Freund JN. Intestinal epithelial–mesenchymal cell interactions. Ann NY Acad Sci. 1998;859:1–17. Kedinger M, Freund J-N, Launay JF, Simon-Assmann P. Cell interactions through the basement membrane in intestinal development and differentiation. In: Sanderson IR, Walker WA, eds. Development of the gastrointestinal tract. London: B.C. Decker; 2000;83–102. Kerneis S, Pringault E. Plasticity of the gastrointestinal epithelium: the M cell paradigm and opportunism of pathogenic microorganisms. Semin Immunol. 1999;11:205–15. Kerneis S, Bogdanova A, Kraehenbuhl JP, Pringault E. Conversion by Peyer’s patch lymphocytes of human enterocytes into M cells that transport bacteria. Science. 1997;277:949–52. Kirkland SC. Endocrine differentiation by a human rectal adenocarcinoma cell line (HRA-19). Differentiation. 1986;33:148–55. Lesuffleur T, Barbat A, Luccioni C, et al. Dihydrofolate reductase gene amplification-associated shift of differentiation in methotrexate-adapted HT-29 cells. J Cell Biol. 1991a; 115:1409– 18. Lesuffleur T, Kornowski A, Luccioni C, et al. Adaptation to 5fluorouracil of the heterogeneous human colon tumor cell line HT-29 results in the selection of cells committed to differentiation. Int J Cancer. 1991b; 49:721–30. Lesuffleur T, Violette S, Vasile-Pandrea I, Dussaulx E, Barbat A, Muleris M, Zweibaum A, et al. Resistance to high concentrations of methotrexate and 5-fluorouracil differentiated HT-29 coloncancer cells is restricted to cells of enterocytic phenotype. Int J Cancer. 1998;76:383–92. Li S, Edgar D, Fassler R, Wadsworth W, Yurchenco PD. The role of laminin in embryonic cell polarization and tissue organization. Dev Cell. 2003;4:613–24. Miner JH, Yurchenco PD. Laminin functions in tissue morphogenesis. Annu Rev Cell Dev Biol. 2004;20:255–84. Neutra M, Louvard D. Differentiation of intestinal cells in vitro. In: Functional epithelial cells in culture. New York: Alan R. Liss; 1989;363–98. Pageot LP, Perreault N, Basora N, Francoeur C, Magny P, Beaulieu J-F. Human cell models to study small intestinal functions: recapitulation of the crypt–villus axis. Microsc Res Technique. 2000;49:394–406. Perreault N, Beaulieu J-F. Use of the dissociating enzyme thermolysin to generate viable human normal intestinal epithelial cell cultures. Exp Cell Res. 1996;224:354–64. Pinto M, Robine-Leon S, Appay MD, et al. Enterocyte-like differentiation and polarization of the human colon carcinoma cell line Caco-2 in culture. Biol Cell. 1983;47:323–30. Potten CS, Booth C, Pritchard DM. The intestinal epithelial stem cells: the mucosal governor. Int J Exp Pathol. 1997;78:219–43. Quaroni A, Wands J, Trelstad RL, Isselbacher KJ. Epithelioid cell cultures from rat small intestine characterization by morphologic and immunologic criteria. J Cell Biol. 1979;80:248–65. Quaroni A, Beaulieu J-F. Cell dynamics and differentiation of conditionally immortalized human intestinal epithelial cells. Gastroenterology. 1997;113:1198–213. Quaroni A, Tian JQ, Goke M, Podolsky DK. Glucocorticoids have pleiotropic effects on small intestinal crypt cells. Am J Physiol. 1999;277: G1027–40.

Quinlan JM, Wu W-Y, Hornsey MA, Tosh D, Slack JM. In vitro culture of embryonic mouse intestinal epithelium: cell differentiation and introduction of reporter genes. BMC Dev Biol. 2006;6:24. Roberts DJ. Embryology of the gastrointestinal tract. In: Sanderson IR, Walker WA, eds. Development of the gastrointestinal tract. London: B.C. Decker; 2000;1–12. Sancho E, Batlle E, Clevers H. Signaling pathways in intestinal development and cancer. Annu Rev Cell Dev Biol. 2004;20:695– 723. Simo P, Simon-Assmann P, Arnold C, Kedinger M. Mesenchymemediated effect of dexamethasone on laminin in cocultures of embryonic gut epithelial cells and mesenchyme-derived cells. J Cell Sci. 1992;101:161–71. Simon-Assmann P, Kedinger M. Heterotypic cellular cooperation in gut morphogenesis and differentiation. Semin Cell Biol. 1993;4:221–30. Simon-Assmann P, Kedinger M. Tissues recombinants to study extracellular matrix targeting to basement membranes. In: Streuli C, Grant M, eds. Methods in molecular biology. Totowa, NJ: Humana Press; 2000:311–9. Simon-Assmann P, Lefebvre O, Bellissent-Waydelich A, Olsen J, Orian-Rousseau V, De Arcangelis A. The laminins: role in intestinal morphogenesis and differentiation. Ann NY Acad Sci. 1998;859:46–64. Simon-Assmann P, Bolcato-Bellemin A-L, Turck N, et al. Basement membrane laminins in normal and pathological intestine. In: Galle PR, Gerken G, Schmidt WE, Wiedenmann B, eds. Disease progression and carcinogenesis in the gastrointestinal tract. London: Kluwer Academic; 2003:223–39. Stappenbeck TS, Wong MH, Saam JR, Mysorekar IU, Gordon JI. Notes from some crypt watchers: regulation of renewal in the mouse intestinal epithelium. Curr Opin Cell Biol. 1998;10:702– 9. Stutzmann J, Bellissent-Waydelich A, Fontao L, Launay JF, Simon-Assmann P. Adhesion complexes implicated in intestinal epithelial cell–matrix interactions. Microsc Res Technique. 2000;51:179–90. Teller IC, Beaulieu J-F. Interactions between laminin and epithelial cells in intestinal health and disease. Expert Rev Mol Med. 2001;3:1–18. Tou L, Liu Q, Shivdasani RA. Regulation of mammalian epithelial differentiation and intestine development by class I histone deacetylases. Mol Cell Biol. 2004;24:3132–9. Turck N, Richert S, Gendry P, et al. Proteomic analysis of nuclear proteins from proliferative and differentiated human colonic intestinal epithelial cells. Proteomics. 2004;4:93–105. Turck N, Gross I, Gendry P, et al. Laminin isoforms: biological roles and effects on the intracellular distribution of nuclear proteins in intestinal epithelial cells. Exp Cell Res. 2005;303:494– 503. Turck N, Lefebvre O, Gross I, et al. Effect of laminin-1 on intestinal cell differentiation involves inhibition of nuclear nucleolin. J Cell Physiol. 2006;206:545–55. Vandewalle A. Immortalized renal proximal and collecting duct cell lines derived from transgenic mice harboring L-type pyruvate kinase promoters as tools for pharmacological and toxicological studies. Cell Biol Toxicol. 2002;18:321–8.

256 Yasugi S, Mizuno T. Mesenchymal influences as microenvironmental factors regulating morphogenesis and cytodifferentiation of gut epithelial cells. In: Kramer B, Rawdon B, eds. Embryos, endocrine cells and the neural crest. Witwatersrand: Witwatersrand University Press; 1996;99–113. Yurchenco PD, Amenta PS, Patton BL. Basement membrane assembly, stability and activities observed through a developmental lens. Matrix Biol. 2004;22:521–38. Zweibaum A, Laburthe M, Grasset E, Louvard D. The use of cultured cell lines in studies of intestinal cell differentiation and func-

tion. In: Field M, Frizzell RA, eds. Handbook of physiology. The gastrointestinal system. American Physiological Society; 1991:223–55.

Address for correspondence: P. Simon-Assmann, Inserm U682, Development and Physiopathology of the Intestine and Pancreas; 3, Avenue Moli`ere, 67200 Strasbourg, France. E-mail: [email protected]