Abstract

Model Tissues for Enteroendocrine Research

There are many models currently being used to investigate EECs. In this review, we will concentrate on ex vivo and in vitro models. The majority of research in the field has been performed using animal tissues (mostly rodent and in particular murine), as these are readily available, relatively inexpensive, and make use of the multiple transgenic strains available. Ideally, for research into the role of EECs in health, human tissue would be used, but this is restricted by tissue availability, practicalities, and additional technical difficulties. Some studies have used pigs, as it has been suggested they are a more relevant model for humans; unlike rodents, for example, they express the hormone motilin. However, there are key digestive differences between the two species, including increased GI fermentation in pigs, which is likely to influence EEC hormone secretion. Pigs also have the disadvantage over rodents that due to their size, studies require larger facilities, more labor time, and increased animal housing costs.

In this review, we will discuss ex vivo model systems involving isolated intestinal perfusion and Ussing chambers and in vitro models such as cell lines, short-term primary cultures, and intestinal organoids (Fig. 2). The characteristics of each model with regard to throughput capability, tissue suitability, longevity, etc., have been summarized in Fig. 2a.

Open in a separate window

Figure 2.

Ex vivo and in vitro models for studying EECs. (a) Table showing the different features and applications of ex vivo (isolated perfused intestine and Ussing chambers) and in vitro (primary cultures, isolated EECs, and organoids) models described in the text. (b) Schematic of small intestine and colonic epithelium with the rare EECs (blue) found in both crypts and mature epithelium (villi and surface epithelium). (c–g) Diagrams showing the different model systems. (c) In isolated perfused intestine, solutions can be perfused via the intestinal lumen and vascular supply. (d) Ussing chamber setup, in which tissue forms a barrier between two compartments (integrity measure by electrodes). Solutions can be added to, and sampled from, either compartment. (e) Single-cell digest of intestinal epithelium, enabling EECs to be isolated by fluorescence-activated cell sorting (FACS) or Percoll density gradient. (f) Primary intestinal monolayers generated by epithelial digestion and seeding onto a matrix scaffold (black) to form a nonpolarized monolayer in which EECs are intermixed with other epithelial cell types. (g) Three-dimensional intestinal organoid structures resembling in vivo architecture (crypt and villi domains) with polarized epithelium containing all cell types including EECs.

Ex Vivo Studies

Isolated intestinal perfusion and Ussing chamber systems are classic ex vivo models that have been widely used to investigate EEC function in different regions of the GI tract (2–10). Intestinal perfusion can be performed on animals as small as mice and generally involves isolating a target gut segment, whereas it is physically retained within the abdominal cavity, enabling vessel cannulation while the heart is still beating. The gut segment under investigation is defined by the area supplied by the cannulated artery, and the lumen can be additionally cannulated and perfused (Fig. 2c). In further modifications, the method has been adapted to enable gut hormone sampling from the intestinal lymph (11). Perfused intestinal models can be used to monitor hormone secretion from different regions of the small intestine or the colon and are a useful tool to study EECs in the context of their surrounding environment, as their local neuronal and vascular circuits largely remain intact, thus supporting near-normal regulation of the EECs. The intestinal epithelial barrier is also maintained, allowing luminal and systemic application of nutrients or drugs to identify their directionality of action. A blood substitute without red blood cells is commonly used for the vascular perfusion, allowing larger sample volumes and more frequent sampling than is possible in in vivo studies on mice, and thus better resolution of hormone secretory dynamics. Hormone secretion has also been measured from isolated intestinal loops mounted in liquid paraffin, which can be readily perfused via the gut lumen but not through the vasculature. This method relies on the detection of hormones after they have diffused through the muscular layers and serosa into the surrounding space, where they collect in small aqueous pools adjacent to the paraffin (12). Although this is a simpler experimental setup than the vascularly perfused intestine, hormone diffusion is too slow to allow resolution of the time course of secretion, and the tissue is more prone to hypoxia due to the lack of continuous vascular perfusion with oxygenated medium. For obvious ethical reasons, isolated human intestinal perfusion studies are not possible, and most research is performed in rodents or sometimes in pigs, as they are considered a more human-relevant model. Isolated intestinal perfusion studies involve highly specialized surgical techniques and are of low throughput compared with the in vitro models, which are quicker, easier, and can be scaled up to enable the parallel examination of multiple test conditions.

Ussing chambers have also been extensively used to study hormone secretion from intact intestinal tissue pieces (2, 6–10). This method involves mounting tissue segments in specialized chambers that were classically used to investigate mucosal physiology and pharmacology (Fig. 2d). As with the perfused intestinal models, the epithelial barrier is maintained, enabling questions to be asked about whether drugs and nutrients target EECs from the luminal or basolateral direction (2, 10). In this system, the integrity of the tissue (epithelial barrier) can be assessed using electrodes that enable live assessment of the transepithelial resistance and potential difference of the mounted specimen. Hormone secretion in Ussing chamber studies can be monitored either by directly assaying hormone concentrations in the basolateral chamber or by measuring the electrical properties of the epithelium, which are altered when secreted hormones such as peptide YY (PYY) and glucagon-like peptide-1 (GLP-1) target their receptors on enterocytes or enteric neurons. Electrophysiological changes in response to cAMP elevation through forskolin/3-isobutyl-1-methylxanthine can also be used to assess the responsiveness/viability of the investigated tissue at the end of the experiment. When hormone concentrations are measured in Ussing chambers by immunoassay, optimization of the chamber volumes and tissue size is required to ensure that hormone concentrations are not below the detection limit of the assay. It is also sometimes necessary to strip away the muscle layers to improve diffusion of hormones and stimuli between the EECs and the basolateral fluid compartment.

A considerable drawback of both these ex vivo systems is the limited experimental window during which the tissue remains healthy. Unlike in vitro models, which can be kept for a few days in primary culture or maintained indefinitely as cell lines, the ex vivo model systems described previously remain viable for only a few hours. Upper small intestinal tissue mounted in an Ussing chamber, for example, has been shown to become compromised after just 2 hours, although more distal regions were slightly more resilient (13). This has contributed to a paucity of Ussing chamber studies examining EECs in the upper GI tract. Ussing chambers have been used to study human as well as mouse EECs, as the method only requires small tissues pieces that can be obtained under ethical approval from routine surgical operations that resect normal tissue as part of a clinical procedure. Nevertheless, studies on human intestine are limited by tissue availability and by the integrity of the epithelium, as human samples generally incur substantial delays between tissue collection and experimentation (14).

In Vitro Studies

Historically, a basic cellular tool for many researchers has been cell lines, immortalized cells that can be routinely passaged and maintained and scaled up for high-throughput studies or experiments requiring large cell numbers. There are multiple cell lines used as models for EEC research, including GLUTag, NCI-H716, STC-1, HuTu-80 and BON cells. Several cell lines have been used as models for GLP-1–producing l-cells, including GLUTag, a murine endocrine tumor-adherent cell line, STC-1, a rat adherent cell line and NCI-H716, a human-derived suspension cell line (15–18). Multiple disadvantages are associated with the use of cell lines in research: many lines are derived from tumors and therefore have genetic and morphological differences from their in vivo counterparts, they are a simplified system and grown without other cell types present, and they lack many physiological traits of their native counterparts with regards to cell morphology and polarity. At the transcriptomic level, GLUTag and STC-1 cells exhibited many similarities to native l-cells, but also a number of differences, including altered expression of key G-protein–coupled receptors (GPCRs) involved in l-cell sensing (19, 20). A recent analysis of different EEC lines also found substantial differences in hormone processing compared with native cells (18). Cell lines do, however, play an important role in EEC research, particularly for studies that require high cell numbers or high throughput.

In the last decade, techniques and protocols have been developed for the use of primary EECs in research. A number of groups have measured hormone secretion from intestinal tissue biopsies, but responses to stimuli are relatively weak in this setting (14, 21). In 2008, Reimann et al. (22) published the first protocol for the use of short-lived (days) two-dimensional (2D) primary cultures of the murine intestinal epithelium for studying EECs, which has recently been published as a video protocol (23) (Fig. 2f). This method enabled, for the first time, assessment of primary l-cells by electrophysiology and calcium imaging techniques, combined with measures of hormone secretion. Since publication, this technique has been successfully applied to intestinal tissues from other species including human and has supported the identification of many nutrient- and drug-activated pathways involved in the release of gut hormones from varying regions along the GI tract (2, 6, 7, 20, 24–26). Typical preparations from mouse or human tissue samples generate approximately one 24-well plate per 5- to 10-cm length of mouse intestine, so this method allows several drug treatments to be tested in parallel, giving a relatively high experimental throughput compared with ex vivo studies. Although the protocol can potentially be scaled up to 96-well plates or beyond, the method is ultimately limited by the sensitivity of the immunoassays used for hormone detection. The major drawbacks of this system are the short-lived nature of the culture, the isolation of EECs from some of the surrounding cell types, loss of apico-basal polarity, and lack of the epithelial barrier. In cultures derived from the small intestine, EECs only survive for a few days, whereas EECs from the colon were still functional after 10 days (22). No new EECs appear to form under these culture conditions (22, 23), so every preparation requires a fresh tissue sample, incurring associated animal costs and local availability of suitable mouse strains. Because these intestinal primary cultures do not generate an epithelial barrier or exhibit apico-basal polarity, they are not suitable for analysis of whether stimuli act via the apical or basolateral surface.

Once EECs have been completely separated from their neighbors by enzymatic or EDTA dispersal, they seem not to survive for long in culture, although calcium imaging was successfully performed in acutely isolated cholecystokinin (CCK)–producing cells by flow cytometry based on their cell-specific expression of GFP (27). By contrast, Raghupathi et al. (28) separated enterochromaffin cells, an EEC type that primarily secretes serotonin (Fig. 2e), using a Percoll density gradient and generated a 95% pure population of enterochromaffin cells that could be kept in culture for up to 4 days with a 70% survival rate (28–30). The enterochromaffin cells isolated by this process have been assessed using amperometry, quantitative RT-PCR (RT-qPCR), and calcium imaging. The drawbacks of this technique are similar to those described for primary intestinal cultures with the added uncertainty of the potential functional consequences of separating EECs from their neighboring enterocytes.

The use of GI organoids rather than primary cultures eliminates the need for a constant supply of freshly harvested tissue (Fig. 2g). Intestinal organoids are three-dimensional (3D) structures grown from stem cells and consist of organ-specific cell types that self-organize and show spatially restricted lineage commitment (31). These organoids are initiated from either pluripotent or adult stem cells and grown in specialized media that mimic the stem cell niche. Protocols to generate intestinal organoids from embryonic or induced pluripotent stem cells are under development, but still require refinement to optimize the differentiation and maturation of EECs (32, 33). It is relatively straightforward, however, to produce organoids from adult intestinal stem cells, which are coerced to form organoids by creating conditions that mimic the stem cell niche environment deployed during tissue self-renewal or damage repair. Indeed, the intestine is a self-renewing tissue containing abundant adult crypt stem cells and was one of the first tissues to be used to generate long-term organoids from mouse and human (34, 35). The 3D organoids can be maintained in culture by regular splitting and replating and can be cryopreserved in much the same way as a standard cell line. Depending on the intestinal region and species of origin, however, it is often necessary to change the culture conditions to promote the formation of fully differentiated EECs, and, once they have been induced to form a terminally differentiated state containing functional EECs, organoid cultures generally have a limited survival of up to a week, thus resembling the behavior of primary cultures derived from freshly harvested tissue.

Intestinal organoids have recently been used to investigate EEC function and represent a promising model for EEC research (36, 37). When generated from transgenic mouse models expressing fluorescent sensors and reporters driven by different gut hormone promoters, for example, they provide a replenishable source of identifiable EECs for single-cell analysis. Organoid cultures have been shown to contain different types of EECs and to retain their regional identity with regard to the profile of gut hormones produced (38). Other studies have shown that EECs generated within intestinal organoids are responsive to a range of physiological stimuli and are secretion competent (36, 37, 39, 40). Cells within 3D organoids have apico-basal polarity with their apical membranes located toward the organoid core. Apical surfaces are therefore largely inaccessible when stimuli are applied via the media, which is a potential limitation of the model, as most nutritional stimuli first reach their target EECs in vivo via the gut lumen. However, experiments using dextran labeling have suggested that organoids are leaky and that solutions can readily reach the apical membrane (40). We recently developed a 2D plating method for intestinal organoids to facilitate patch clamp electrode access, cellular imaging within a single focal plane, and measurements of GLP-1 secretion (36). Data from these studies showed that l-cells derived from murine organoids exhibited similar responses to l-cells in ileal primary cultures, thus validating the system as a good model for EEC research. This method is particularly exciting because organoids, like cell lines, can be genetically modified to produce lines with knockout, knockin, and overexpression of genes of interest. Similar genetic manipulations of human derived organoids would give this technique unique advantages compared with other ex vivo and in vitro models.

Interestingly, recent progress has been made in converting 3D organoid cultures to an open-faced polarized epithelium that can be used for transport studies (41, 42). One such technique has been shown to recapitulate the normal intestinal hierarchy and morphology, with the formation crypt/villus structures resembling in vivo tissue architecture (41). This method combined with genetically modified organoids will potentially open up new avenues for EEC research that are not currently possible with existing techniques.

Experimental Tools for EEC Research

Transgenic mice

The generation of transgenic reporter mice has enabled a dramatic increase in our understanding of EEC function and has helped to identify and validate a wide range of signaling pathways controlling hormone secretion from the intestine. Most genetically modified mouse models used in EEC research harness cell-specific promoters, such as the promoters for the hormones themselves (e.g., GLP-1, GIP, CCK, somatostatin, and ghrelin), chromogranin A (ChgA), EEC-specific transcription factors (e.g., NeuroD1), or tryptophan hydroxylase (Tph1, required for serotonin biosynthesis) (Fig. 3) (22, 26, 43, 49–57). These EEC-specific promoters are used to drive cell-specific expression of fluorescent reporters/sensors or Cre-recombinase. Interestingly, the ChgA-hrGFP reporter line seems specifically to label enterochromaffin cells even though ChgA is considered to be a marker of most EECs, although this mirrors our previous observation that ChgA staining was low in murine l-cells compared with other EECs (37, 56).

Open in a separate window

Figure 3.

Transgenic EEC models. (a) Table listing the current transgenic models used to identify different EEC types. (b) Glu-Venus (green) mouse tissue showing colonic epithelium (image by L. Billing), (c) an isolated ileal villus, and (d) isolated ileal crypts. (e) Ileal organoid derived from Glu-Epac2-camps mice, with l-cells highlighted by Epac2-camps expression (green). Scale bars, 100 μm. 5-HT, 5-hydroxytryptamine; SST, somatostatin.

When the fluorescent reporter is directly driven by a hormonal promoter, the detection of fluorescence in a cell should reflect current or recent activity of the hormonal promoter, as is seen with mouse strains such as GLU-Venus, GIP-GFP, CCK-GFP, and ChgA-hrGFP (22, 52, 56, 58). Hormone promoters driving Cre-recombinase, by contrast, are actually lineage tracers in the sense that even a brief activation of the hormone promoter and consequent transient expression of Cre-recombinase can activate Cre-inducible elements in the DNA, causing permanent genetic changes that persist for the lifetime of the cell and are passed on to daughter cells. If Cre expression is used to label an EEC population, it does not therefore necessarily reflect ongoing hormone expression. This can be useful when the endogenous cell-specific promoter is too weak itself to drive detectable fluorescence, because it is used instead to activate a fluorescent reporter that is driven by a much stronger promoter. Although, as with all indicators, one might be concerned that chronic overexpression of fluorescent reporters might be toxic or alter the dynamics of the monitored secondary messengers, we have not detected any differences in the responsiveness of our reporter lines to different secretory stimuli and observed similar Ca²⁺ dynamics when monitoring with GCaMP3 or Fura-2.

Transgenic mouse technologies have allowed researchers to identify the rare EEC cells among the rest of the intestinal epithelial cells, enabling application of a wide variety of analytical methods, including population and single-cell transcriptomics, electrophysiology, and single-cell imaging of intracellular signaling pathways.

Assessing Physiological EEC Stimuli by Combined Experimental Techniques

The models and experimental tools described previously have been used to identify a wide range of stimulants and suppressors of gut hormone release, site of interaction of nutrients with EECs, and some of the molecular pathways involved in nutrient detection and hormone secretion. Many of the findings from these studies are summarized in Fig. 4. As an example of how the different techniques have been combined to explore a molecular pathway, we will consider the question of how bile acids trigger GLP-1 secretion.

Open in a separate window

Figure 4.

Mechanisms underlying glucose, short-chain fatty acid (SCFA), bile acid, and amino acid sensing in EECs. (a) Glucose sensing in EECs, mediated by two pathways leading to cell depolarization and opening of voltage-gated calcium channels (VGCC), triggering Ca²⁺ influx and hormone secretion. Glucose cotransport with Na⁺ through apical sodium glucose-linked transporter 1 can trigger depolarization and secretion, whereas glucose influx via glucose transporter 2 (GLUT2) increases ATP levels and closes K_ATP channels, potentially modifying the effectiveness of other depolarizing stimuli. (b) SCFA sensing in EECs mediated by FFA2 Gαq signaling, thereby increasing Ca²⁺ and triggering hormone release. An unidentified pathway might also trigger EEC depolarization and opening of VGCC, as suggested by experiments on perfused intestine. (c) Bile acid sensing is mediated by basolateral G-protein–coupled bile acid receptor (GPBAR) 1 signaling through Gαs, increasing cAMP and triggering hormone secretion. GPBAR1 activation also promotes depolarization and opening of VGCCs. (d) Amino acid sensing in EECs is mediated by multiple pathways. Highlighted in the figure are calcium-sensing receptor Gαq signaling and electrogenic amino acid uptake, which trigger subsequent depolarization and opening of VGCCs.

Bile acids have been demonstrated to increase GLP-1 secretion from GLUTag cells, primary colonic l-cells, ileal organoid-derived l-cells, ileal tissue mounted in Ussing chambers, and perfused intestine (2, 5, 10, 36, 72). Knockdown studies in GLUTag cells and the use of tissues from knockout mice confirmed that the major receptor involved in the bile acid–triggered GLP-1 secretion is the G-protein–coupled bile acid receptor (GPBAR) 1 (also known as TGR5), which is highly expressed in primary l-cells (2, 72, 73). Consistent with the known Gαs coupling of GPBAR1, bile acids and the GPBAR1 agonist GPBAR-A increased intracellular cAMP levels in GLUTag cells transiently expressing the cAMP fluorescence resonance energy transfer sensor Epac2-camps, as well as in primary l-cells in organoids derived from the GLU-Epac2-camps mouse model (2, 36, 72). The use of 2D organoid cultures allowed electrophysiology assessment of the bile acid receptor in ileal l-cells, revealing that GPBAR1 agonism increased action potential firing and enhanced the activity of voltage-gated Ca²⁺ channels, thereby potentiating Ca²⁺ signals triggered by FFA1 agonists, as measured in organoid l-cells from GLU-Cre/GCaMP3 mice (36). A key question for the potential translational exploitation of GPBAR1 as a drug target for increasing endogenous GLP-1 secretion is whether small-molecule receptor agonists could target the receptor from the apical membrane, thus enabling the development of nonabsorbable agonists with low side effect profiles. Ussing chamber and perfused intestine experiments, however, demonstrated that bile acids only enhance GLP-1 secretion following their absorption across the epithelium and that nonabsorbable GPBAR1 agonists are only effective when applied from the basolateral direction (2, 5, 10). Nevertheless, activation of GPBAR1, particularly in combination with FFA1, provides strong stimulus to l-cells that deserves further exploration as a potential therapeutic strategy.

Conclusion

Compared with the situation 20 years ago, when studies on gut hormone secretion were largely restricted to the use of whole animals, perfused intestinal preparations, Ussing chamber recordings, and cell lines, EEC research has exploded recently both in terms of the techniques and protocols available to researchers and our understanding of the molecular pathways used by EECs for stimulus secretion and detection. The use of high-throughput systems such as cell lines and primary cultures has enabled the identification of a wide range of nutrient-dependent and nonnutrient-dependent pathways involved in the stimulation and inhibition of hormone release. Increased use of intestinal organoids is hoped to facilitate more translational human-based research and reduce the need for ex vivo animal-based studies. The new technologies that use organoid cultures to generate an in vivo type intestinal architecture with intact barrier function provide promising avenues for the study of human and murine EEC function.

Acknowledgments

Abbreviations:

References