| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
NEUROREGULATION AND MOTILITY
Gastrointestinal Pathobiology Research Group, Yale University School of Medicine, New Haven, Connecticut
Submitted 6 December 2005 ; accepted in final form 24 January 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
3.5 ± 0.62-fold vs. unstimulated cells (EC50 of 10–9 M). Octreotide, acetylcholine, and GABAA inhibited serotonin secretion with IC50 values of 3 x 10–11, 3 x 10–10, and 2.9 x 10–10 M, respectively. Gastrin had no effect on serotonin secretion. The naive EC cell transcriptome revealed highly expressed EC cell marker genes, the absence of marker genes for other small intestinal cell types, and a receptor profile that included cholinergic, adrenergic, dopaminergic, serotoninergic, GABAergic, and prostaglandin receptors. We were able to isolate homogeneous preparations (>99%) of live ileal EC cells and demonstrated regulation of serotonin secretion as well as established the normal EC cell transcriptome. Application of this methodology to normal and diseased human ileum will facilitate the elucidation of the pathophysiology of EC cells.
fluorescence-activated cell sorting; serotonin; secretion; small intestine; carcinoid
40%) (44). Other endocrine cells identified in this location include enteroglucagon/peptide YY (PYY) (EG/L cells), neurotensin (N cells), somatostatin (SST; D cells), gastrin (CCK cells), and pro-
-melanocyte-stimulating hormone (MSH) cells (48). Serotonin [5-hydroxytryptamine (5-HT)] is synthesized in the ileal EC cell by a mechanism involving the EC cell-specific enzyme tryptophan hydroxylase (TPH) (52) and stored in secretory vesicles via a process mediated by vesicular monoamine transporter 1 (VMAT1) and proton gradients. It is released by mechanisms that are not well defined in response to a number of stimuli including mechanical stimulation or acidification of the gut lumen (25, 52). Serotonin is released from the basolateral membrane of the ileal EC cell and exerts a paracrine effect on adjacent enterocytes and smooth muscle cells as well as passing into the circulation as a classic endocrine regulator (19). More recently, it has become apparent that in addition to serotonin, ileal EC cells synthesize and secrete substance P and guanylin (8, 44). Both of these agents have been identified as modulators of intestinal secretion and motility (8, 53), but little is known of the regulatory events associated with their release from the ileal EC cell. Nevertheless, the ileal EC cell, by virtue of producing at least three small intestinal regulatory agents, exhibits a pivotal role in the physiology of small intestinal function and most likely in its disease states such as irritable bowel syndrome and carcinoid tumors (11, 34, 53).
It is apparent that the ileal EC cell is one of the primary NE regulatory cells of the small bowel, and, given its wide distribution and large numbers, it may occupy as dominant a role in the small intestine as that of gastrin in the stomach (36). To date, however, the ileal EC cell has proved difficult to study given its scattered location amid a wide variety of different cell types. Serotonin release has been examined in a number of ex vivo preparations, and these studies have demonstrated that it is mediated via a number of agents including mucosal stimulation by fat and PYY (52) and cholinergic (acetylcholine) and adrenergic (e.g., norepinephrine) nerves (52). In addition, pituitary adenylate cyclase-activating polypeptide (PACAP) and vasoactive intestinal polypeptide (VIP) appear to have potent inhibitory effects on serotonin release (16). The interpretation of such observations is restricted by the fact that intact mucosal preparations preclude, to a large extent, the determination of direct effects of specific agents on the ileal EC cell itself. Furthermore, little is known about the roles and interactions of the other five ileal NE cell types that may participate in modulating either mucosal or EC cells (52). Given the heterogeneous nature of the mucosa within which the ileal EC cell is distributed, precise characterization of the cell would be facilitated by developing a homogenous ileal EC cell preparation that can be used to identify its transcriptome and define its response to stimuli.
We and others have reported methodologies of varying effectiveness for the isolation and short-term culture of a number of different gastrointestinal endocrine cells from animal models including Mastomys (2, 6, 24, 26, 42). With the use of elutriation and Nycodenz gradient centrifugation of everted normal and carcinoid-containing stomachs, 85–95% pure preparation of gastric EC-like (ECL) cells can be obtained (26, 31). Studies of purified gastric ECL cells have demonstrated the key regulatory role of the histamine enzyme histidine decarboxylase in gastric ECL cell histamine secretion, identified the presence of gastrin CCK2 and PACAP receptors and defined their key roles in proliferation and secretion (28, 31, 32, 50), and delineated the presence of SST receptor 2 (SSTR2) and the inhibitory role of octreotide in both secretion and proliferation (4, 45). With the use of a somewhat similar approach, preparations of neurotensin and enteroglucagon cells have been obtained from the canine ileum (2, 6). Assessment of these cells demonstrated a key role of cAMP activation by forskolin and 
-adrenergic receptor activation by isoproterenol in the secretion of neurotensin and enteroglucagon (2). More recently, Schafermeyer et al. (46), using counterflow elutriation and step density gradient centrifugation, have reported preparations of rat ileal EC cells with 15–20% contamination. Preparations of these cells were used for both in vitro serotonin release experiments and transcriptional profiling. The presence of cell contamination with other NE cells (47), as well as the high chemical doses (10–5–10–4 M) required to elicit serotonin release in these cells (46), suggests the need to develop homogenous ileal EC cell preparations.
More recently, a fluorescence-activated cell sorting (FACS) approach has been utilized to obtain homogeneous (99% pure) parietal and ECL cell populations from the rat gastric mucosa (29, 38). This methodology has not been applied to ileal cell preparations. In the present study, we outline the application of intestinal eversion, digestion, Nycodenz gradient centrifugation, and FACS using anti-TPH and acridine orange (AO) labeling to derive a pure population of Mastomys ileal EC cells and define their transcriptome and functional regulation in vitro.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Mastomys Tissue Specimens
We developed the methodology using ileal tissue from the multimammate mouse (Mastomys natalensis). Although our eventual goal is to define the human EC cell phenotype, we used this model because we have extensive experience in isolating and characterizing the gastric equivalent of the EC cell—the ECL cell—from this animal (26, 50), and the close phylogenetic relationship of Mastomys with the mouse allowed the use of murine Affymetrix GeneChip for transcriptome analysis of Mastomys EC cells (26). Mastomys with equal sex distribution were used. Mastomys (n = 4 animals/preparation) were euthanized by CO2 inhalation and cervical dislocation. We used the terminal 15 cm of the Mastomys small intestine (total length = 48 ± 4.3 cm, duodenum to cecum) as representative of the ileum (40). This was removed and divided into two to three pieces, which were everted and end ligated. All animal studies were approved by Yale University (IACUC no. 2004-10904).
Cell isolation. Everted, end-ligated ilea were filled with pronase E (0.7 mg/ml) and collagenase (0.25 mg/ml) and preincubated in DTT-HBBS for 30 min to remove mucus (6, 46). Cells were then separated from the muscularis by alternative switching between a calcium-containing respiration medium and a chelating digestive medium (with stirring) (the F0 fraction; Table 1). Ileal mucosal (IM) scrapings were retained to provide a baseline to assess enrichment.
|
Preparation for FACS.
TPH-POSITIVE CELLS.
Cells from the F0 and FN fractions were fixed in ice-cold methanol (–20°C, 100%, final concentration = 50% solution) for 5 min, washed three times with PBS, and blocked with 0.1% BSA before being immunostained with mouse monoclonal anti-TPH antibody (0.1 µg/ml, Calbiochem, overnight at 4°C). Cells were then washed and stained for 20 min (at room temperature) with FITC-labeled secondary antibody (1:100).
AO-POSITIVE CELLS.
Live cells from the FN fraction were incubated with AO (2–200 µM) for 30 min (at room temperature) before being sorted.
FACS.
All fractions were filtered (50-µm pore size) before being sorted. FACS analysis (high-speed FACS Aria, Yale Cancer Center) was used to identify and sort cells. Excitation was at 488 nm (activation of FITC- or AO-labeled cells), and sorting was achieved by gating on side scatter (dense, small cells of an estimated 8–12 µm size) and an emission of 532 ± 15 nm. We collected positive cells over 30 min (the FFACS fraction). To demonstrate that AO uptake was due to the generation of acidic spaces within granules, aliquots (n = 3) were preincubated with 20 mM NH4Cl for 30 min before the AO addition and FACS.
Immunostaining.
TPH-POSITIVE CELLS.
After FACS, an aliquot of FITC-TPH-positive cells (20–50,000 cells) were pipetted onto frosted microscope slides. Fluorescence microscopy was used to visualize and count the number of positive cells to give the percentage of TPH-positive cells after FACS. A second aliquot of these FACS-sorted cells were incubated with chromogranin A (CgA) antibodies (1:200; goat, Santa Cruz Biotechnology) overnight at 4°C. After being washed (0.1% Tween-PBS), cells were stained with secondary antibody (horseradish peroxidase-conjugated anti-goat, 1:100, DAKO) for 1 h at room temperature and visualized with a fluorescent chromogen (Cy5-tyramide, NEN Life Science Products) to give the percentage of CgA-positive cells (NE cells in each preparation).
AO-POSITIVE CELLS.
Live AO-FACS-sorted (50,000 cells/antibody) EC cells were fixed in methanol and pipetted onto frosted microscope slides. Preparations were stained with anti-CgA (mouse, 1:1,000, DAKO Cytomation), anti-TPH (mouse, 0.1 µg/ml, Calbiochem), or anti-serotonin (mouse, 1:20, DAKO Cytomation) antibodies overnight at 4°C. After being washed (0.1% Tween-PBS), cells were stained with secondary antibody (FITC-conjugated Cy5-anti mouse/rabbit, 1:100, Promega) for 1 h at room temperature. Cells were visualized and counted to assess the percentage of NE or EC cells in each preparation. As a control for nonspecific staining, primary antibodies were excluded. For confocal microscopy, micrographs were recorded using a confocal microscopy system equipped with three Ar488, Kr568, and HeNe633 lasers (TCS-SP, Leica; Mannheim, Germany) with a x40 objective.
EC CELLS IN SMALL INTESTINAL MUCOSA.
Small intestinal mucosa were obtained after the animals were euthanized, washed in PBS, and immersed in 4% formaldehyde overnight (28). Fixed tissue was mounted edge-on in paraffin blocks. Deparaffinized sections were studied using hematoxylin and eosin staining for the assessment of EC cell numbers by immunostaining with TPH (as above). The numbers of TPH-positive cells were assessed by examining sections (magnification: x450) using a square grid of 10 x 10 mm (American Optical; Buffalo, NY) and noted as the numbers of cells per 1-mm2 zone. Only the areas with full-thickness mucosa were measured. An average of 10 grids from each section were counted. The cell density of TPH-positive cells was calculated as follows (expressed as cell numbers/mm2):
 |
Electron microscopy. Cells were evaluated using the semithin technique (39). Isolated cells (500,000 cells) were pelleted and fixed with 3% gluteraldehyde in 0.1 M cacodylate buffer (pH 7.4). After osmification, cells were dehydrated with graded alcohol and embedded in Epon. Sections were evaluated with a Philips 300 transmission electron microscope. EC cells were distinguished from other NE cells by the presence of pleomorphic granules (oblong, ovoid, kidney shaped, triangular, or U shaped) that can be identified ultrastructurally (7).
PCR. Three markers of EC cells were assessed using real-time PCR (CgA, VMAT1, and TPH). For gastrin, PACAP, SST, and GABA receptor expression, CCK2 receptor, PACAP receptor 1 (PAC1) and VIP/PACAP receptor 1 (VPAC1), SSTR2, and GABAA and GABAB receptors (respectively) were assessed using a real-time approach (26). All primer probe sets were inventoried Assay-On-Demand probe sets obtained from Applied Biosystems. To detect contaminating NE cells, neurotensin, SST, glucagon, and PYY transcripts were amplified by RT-PCR.
RNA EXTRACTION AND CLEANUP.
RNA from the various fractions (IM, F0, FN, and FFACS) was extracted using TRIzol (Invitrogen), and then cleaned up using the Qiagen RNeasy Kit in conjunction with the DNeasy Tissue Kit (Qiagen) to minimize contaminating genomic DNA. The cleaned-up RNA (2 µg) was converted to cDNA using the High-Capacity cDNA Archive Kit (Applied Biosystems).
REAL-TIME PCR PROTOCOL.
Real-time RT-PCR analysis was performed using Assay-On-Demand products according to the manufacturer’s suggestions. Cycling and fluorescence detection were performed using the ABI 7900 Sequence Detection System. Only Assay-On-Demand primers designed to hybridize to exon:exon boundaries (TaqMan Gene Expression Assays Product Guide) were used, thereby minimizing the possibility of amplifying genomic DNA. Controls without reverse transcriptase were included in triplicate in each real-time RT-PCR experiment to ensure the absence of genomic DNA contamination. Cycling was performed under standard conditions as outlined in the TaqMan Universal PCR Master Mix Protocol, and the Standard Curve Method (ABI User Bulletin No. 2) was used to determine relative transcript levels. GAPDH was used to normalize data.
RT-PCR.
PCRs were performed in a 25-µl volume containing 0.2 mM of each dNTP, 2.0 mM of 1x reaction buffer (Roche PCR core kit), 0.2 M of forward and reverse primers, 0.5 units of Taq DNA polymerase (Roche), and 100 ng of template DNA. Thermal cycles consisted of 35 cycles of 1 min at 94°C, 1 min at annealing temperatures (50–55°C), and 1 min at 72°C, followed by an extension for 10 min at 72°C. Amplification of GAPDH transcript (5) in all samples was used as a positive control. Samples were loaded on 4.5% Supra Sieve agarose gels (American Bioanalytical) and visualized by ethidium bromide incorporation (0.5 µg/ml) under ultraviolet light.
Serotonin content. Serotonin content was measured and quantified as a percentage of the total protein content. Briefly, 50,000 cells were pelleted by centrifugation at 300 g for 5 min. Total protein was extracted by homogenization of the cells in lysis buffer [50 mM HEPES (pH 7.4), 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl2, 1 mM EDTA, 100 mM NaF, 10 mM Na4O7P2, 1 mM Na3VO4, 10 mg/ml leupeptide, and 1 mM PMSF]. The concentration of protein (in mg/ml) was assayed using the Bio-Rad Protein Assay. Serotonin content (in ng/ml) of each fraction was then measured using a commercially available serotonin ELISA (Diagnostika) according to the manufacturer’s instructions for serum samples. Serotonin content was defined as nanograms of serotonin per milligram of protein.
Culture, serotonin secretion, and cAMP experiments.
Live AO-FACS-sorted cells were divided into aliquots (3 x 104 cells/100 µl), placed in culture media [serum-free Ham’s F-12 medium (GIBCO-BRL) supplemented with antibiotics (100 units penicillin/ml + 100 µg streptomycin/ml, Sigma-Aldrich)], plated into 96-well collagen I-coated plates (Becton Dickinson), and maintained in a humidified atmosphere at 37°C in 5% CO2.
After 2 h, a time course for serotonin secretion in response to the -adrenergic receptor agonist isoproterenol (10–6 M) was evaluated. The latter agent was chosen based on its reported agonist effect on ileal neurotensin- and enteroglucagon-secreting cells (2, 6). Cells were stimulated with isoproterenol or media alone (control) for 0, 3, 5, 15, 30, 60, 90, and 120 min. To confirm that secretion represented a secretory process rather than cell damage or lysis, lactate dehydrogenase (LDH) release into media was measured using a commercially available LDH assay (CytoTox-ONE Homogenous Membrane Integrity Assay, Promega) according to the manufacturer’s instructions for cultured cells.
Having determined the optimal time course for secretion, the effects of isoproterenol on intracellular cAMP were measured using a commercially available ELISA kit (R&D Research). Briefly, 5 x 104 cells/100 µl in culture media were stimulated with isoproterenol (10–10–10–6 M) for 15 min, and cAMP was assayed and expressed as picomoles per milliliter.
Individual serotonin secretion dose-response curves for forskolin, isoproterenol, acetylcholine chloride, GABAA, PACAP-38, and gastrin (all 10–12–10–6 M) were then established. The cAMP activator forskolin was utilized because it has previously been used as a positive control for ileal neurotensin and enteroglucagon cell secretion (2, 6). To determine whether naive EC cell secretion was subject to influence by SSTR activation, the effects of octreotide (10–12–10–6 M) on isoproterenol (EC50 = 5 x 10–10 M) were also evaluated. Media were collected from each well after 15 min of stimulation, and serotonin secretion was measured using a commercially available serotonin ELISA (Diagnostika). Serum-free growth medium was used as a blank solution. The data for serotonin secretion were expressed as percentage changes from baseline secretion, and values are means ± SE for each concentration. LDH secretion was measured in parallel experiments to verify that EC cell serotonin secretion did not represent a damage-related release. The viability of cells was also assessed using Trypan blue exclusion. Equal volumes of Trypan blue and the cell suspension were mixed and, after 5 min, loaded on a Brightline hemocytometer. Cells (1 x 103) were visualized, counted using light microscopy (magnification: x40), and expressed as percentages of viable cells.
GeneChip examination. To obtain the naive EC cell transcriptome, we utilized Affymetrix profiling of AO-FACS-sorted preparations.
GENECHIP HYBRIDIZATION.
For each experiment (n = 3), 2–3.5 µg of high-quality (ratio >1.9) total RNA isolated from AO-FACS-sorted cells or RNA isolated from the IM (whole mucosal scraping from the terminal 15 cm of the ileum) were provided to the Keck Affymetrix facility to perform cRNA labeling, hybridization, and data analysis. A standard protocol (Keck Affymetrix Core Facility, Yale) was used before hybridization to the Mouse Expression 430 2.0 chip (39,000 transcripts).
GENECHIP DATA ANALYSIS.
The hybridized arrays were scanned using an Affymetrix GeneChip 3000 Scanner. Arrays were scaled to an average intensity of 500, and hybridization intensity data were converted into presence/absence calls for each gene using GCOS software. The signal intensity data were statistically analyzed using S-plus 2000 software (Mathsoft; Cambridge, MA) to obtain correlation coefficients and scatter plots for the evaluation of the reproducibility and quality of the array analysis. We used both the normalized array intensities and present calls to identify the full complement of transcripts in these preparations. This pattern of genes was used to define the EC cell transcriptome. In a separate analysis, a Wilcoxan rank test was used to compare expression levels in the EC cell preparations with those in the IM. Genes with relatively high expression (>2-fold) were identified and catalogued for the EC cell.
Statistical analyses. Results are expressed as means ± SE; n indicates the numbers of preparations. Data are represented as the fold differences vs. basal values. Statistical significance was calculated by the two-tailed Student’s test for paired and unpaired values as appropriate, with a probability of <0.05 representing significance.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Nycodenz gradient centrifugation. In the intact IM, the density of TPH-positive EC cells was 0.6 ± 0.23% by immunostaining. The serotonin content of the mucosa was 3.5 ± 0.6 ng serotonin/mg total protein. After eversion and pronase-collagenase digestion, the F0 fraction contained 2.7 ± 1.2% TPH-positive cells and 18 ± 0.5 ng serotonin/mg total protein (Table 1 and Fig. 1A). The process of digestion (development of a crude mucosal fraction) therefore provides an "apparent" enrichment for serotonin-containing cells. The most likely explanation is via loss of enterocytes with a subsequent "apparent increase" in EC cells. Nycodenz gradient centrifugation produced a 33-fold enrichment of serotonin-containing EC cells above that in the F0 fraction. The FN fraction had a serotonin content of 102 ± 12 ng/mg protein. For both the F0 and FN fractions, >99% and 97% of cells, respectively, were viable as assessed by Trypan blue exclusion.
|
To produce live EC cells, we evaluated two different methodologies, fluoro-PACAP receptor binding and AO uptake. The fluoro-PACAP-labeling strategy has been previously utilized to isolate gastric ECL cells (38). Using the same conditions [fluoro-PACAP (0.1 µM); 30 min of pre-FACS incubation], we were unable to identify PACAP-positive Mastomys ileal EC cells (Fig. 2C). However, by increasing the labeling concentration of fluoro-PACAP to 10 µM, a discrete EC cell population could be identified. We considered that the ability to collect cells using an increased fluoro-PACAP concentration might reflect different levels of expression of the PACAP receptor subtypes on the EC cell compared with ECL cells and subsequently confirmed this by RT-PCR (21). Nevertheless, we considered fluoro-PACAP labeling (although successful; produced 96% pure EC cells) a suboptimal sorting strategy because it induces ECL cell function (secretion and proliferation) (38). Thus a fluoro-PACAP-sorted cell population would represent a prestimulated EC cell cohort and limit the ability to establish "true" baseline (nonstimulated) parameters of EC cell function.
|
|
To further evaluate and confirm sequential increase of the EC component of fractions, we performed real time RT-PCR to detect and quantitate EC-specific (TPH and VMAT1) and NE-specific (CgA) transcripts (Fig. 1B). Both NE transcripts as well as the EC-specific transcript TPH were sequentially enriched, with the FFACS fraction (98–100% EC cells) showing the highest transcript expression for all markers (24- to 34-fold vs. the IM).
Immunohistochemical and Ultrastructural Characterization of Homogenous EC Cell Fractions
The ultrastructure of FFACS cells was evaluated using electron microscopy. Specific characteristics of EC cells (7, 49) could be confirmed, including the presence of 1) well-developed Golgi apparati, numerous mitochondria, and rough endoplasmic reticula; 2) vacuolated granules consistent with secretory function; and 3) large numbers of variously sized electron-dense vesicles consistent with NE secretory function (Fig. 3C). Confocal microscopy of FFACS cells dual stained with AO and TPH antibodies demonstrated colocalization of TPH and AO within the cytoplasm of these cells (Fig. 3D).
Identification of Contaminating Cells in the FFACS Fraction
To identify the cell type of the 1% contaminating cells in the live cell preparations, we initially used monoclonal CgA (DAKO, 1:1,000) staining and demonstrated that 100% of cells in this FACS-sorted fraction were CgA positive (Table 1). This indicated the 1% contamination to be of NE cell origin. The enrichment techniques employed principally used size/density indexes and would be predicted to preferentially enrich NE cells, which are similar in size and density (7, 49). Although contamination represented only a very small percentage (<1%), we sought to identify the nature of these NE cells. According to intact mucosal immunohistochemistry, we considered the possibilities to include EG/L, N, D, CCK, or pro--MSH cells (48). Accordingly, we utilized PCR to identify the specific contaminant NE cell types. In contrast to the F0 fraction and IM, AO-FACS-sorted fractions did not contain SST, glucagon, gastrin, or PYY transcripts (Table 2). This fraction did, however, contain the neurotensin transcript, suggesting that the contaminating cell type was a neurotensin-producing cell. We further confirmed the absence of D cells by immunostaining the FFACS fraction with an anti-SST antibody (1:2,000, ICN Biomedicals). In the FN fraction, 4.3% of cells were SST positive, whereas SST levels were undetectable in the FFACS fraction (Fig. 4, A and B). The absence of D cells in the 99% EC preparations is an important criterion because it ensures no extraneous inhibitory effect in the assessment of the proliferative/secretory profile of the EC cell (46).
|
|
To identify receptors that might be responsible for the modulation of normal EC cell function, we used real-time PCR to examine acetylcholine, GABA, gastrin, PACAP, and SST receptors. A comparison of the IM with the FFACS fraction demonstrated transcripts for acetylcholine M4, GABA-
2, GABA-1, VPAC1, and SSTR2 in both mucosal samples and FACS-sorted preparations (Fig. 5). PAC1 and CCK2 could not be detected in the latter preparations; results consistent with the EC cell transcriptome.
|
Viability (assessed by Trypan blue exclusion) after the addition of 2, 20, or 200 µM AO and FACS was >94%, 67%, and 57%, respectively. Similarly, LDH secretion using the CytoTox-ONE Homogenous Membrane Integrity Assay was significantly elevated in the 20 and 200 µM AO-FACS-sorted fractions [595-nm optical density (corrected for blank) = 0.494 ± 0.032 and 0.651 ± 0.102] compared with the 2 µM AO-FACS-sorted fraction [595-nm optical density (corrected for blank) = 0.278 ± 0.018, P < 0.01, >70% elevation in LDH secretion].
Given the low viability of the 200 µM fraction, this preparation was not examined further, and the lower AO concentration was utilized in subsequent studies. Basal (unstimulated) serotonin release from the FN fraction and 2 and 20 µM AO-FACS-sorted cells assessed in short-term culture (2 h) was 1.09, 1.04, and 0.26 ng–1·50,000 cells–1·h, respectively. The decrease in serotonin release in the 20 µM fraction presumably reflects vesicular damage caused by the high dose of AO and was confirmed by the associated high levels of LDH. On the basis of these measurements (unstimulated serotonin and LDH release), this fraction was therefore not further evaluated.
Serotonin responses were assessed in AO (2 µM)-treated EC cells to isoproterenol (10–6 M) to establish the time course of secretion. Significant serotonin secretion was noted as early as 3 min in the media, reached a peak (55 ± 5%) at 15 min, and was detectable up to 120 min (>20% vs. media-stimulated cells) (all evaluated time points, P < 0.05 vs. unstimulated cells; Fig. 6A). An examination of LDH levels over this time period (120 min) demonstrated no effect of isoproterenol (Fig. 6B).
|
-adrenergic receptor ligand. At 15 min, isoproterenol stimulated cAMP levels by 3.5 ± 0.62-fold vs. unstimulated cells, with an EC50 of 10–9 M (Fig. 6C). This effect was reversed by preincubating the cells for 15 min with 2',5'-dideoxyadenosine (1 µM), an adenylate cyclase inhibitor (51). This demonstrates that the cAMP pathway is activated by this adrenergic receptor ligand in the EC cell. To develop a profile for the regulation of EC cell serotonin secretion, we initially assessed a number of agents previously reported to be effective in intact IM or heterogeneous cell preparations (gastric and ileal) containing NE cells (45, 55). Dose responses (at 15 min) were developed for secretagogues: the neural agents, isoproterenol, PACAP, acetylcholine chloride, and GABAA, and the hormonal agents, gastrin and the SST agonist octreotide. Forskolin was also examined to assess the contribution of direct adenylyl cyclase activation in the stimulation of EC cell serotonin secretion.
Forskolin stimulated EC cell serotonin secretion with an EC50 of 5 x 10–10 M, confirming cAMP signaling pathway involvement in serotonin secretion (Fig. 7A). Isoproterenol stimulated EC cell serotonin secretion with an EC50 of 4.5 x 10–10 M (Fig. 7B). The SSTR2 ligand octreotide inhibited basal serotonin secretion by 30% (IC50 = 3 x 10–11 M; Fig. 7C). Isoproterenol (EC50)-stimulated serotonin secretion was inhibited by octreotide with an IC50 of 3 x 10–9 M (Fig. 7D). PACAP-38 stimulated serotonin secretion (EC50 = 1.15 x 10–9 M; Fig. 7E), but gastrin had no effect on EC cell serotonin secretion (Fig. 7F). Inhibitors of basal serotonin secretion included acetylcholine chloride (IC50 = 3 x 10–10 M; Fig. 7G) and GABAA (IC50 = 2.9 x 10–10 M; Fig. 7H). Isoproterenol-stimulated secretion (EC50) was inhibited by acetylcholine chloride with an IC50 of 6 x 10–10M (Fig. 7I) and by GABAA with an IC50 of 4 x 10–10 M (Fig. 7J). LDH levels were not altered by any of the evaluated ligands.
|
90% inhibition above basal). The Naive EC Cell Transcriptome
We (26) previously determined the transcriptome of Mastomys gastric ECL cells using an Affymetrix approach (Mouse U74A chips). Using the same technique but with newer expression GeneChips (Mouse Expression Set 430 chips), we now determined the gene expression profile of AO (2 µM)-FACS-sorted Mastomys EC cells and demonstrated the feasibility of quantifying gene expression from FACS-sorted ileal NE cells. As a comparator, we measured gene expression in the IM. In our preliminary studies, we noted both a significantly lower yield of total RNA from ethanol-fixed TPH-FACS-sorted cells [<2 ng/µl (106 cells)] compared with AO-FACS-sorted cells [200 ng/µl (106 cells)] as well as a higher 3'-to-5' GAPDH ratio (>7 vs. <5 for AO-FACS-sorted cells) on test chips, making this preparation unsuitable for RNA studies. We therefore focused on measuring the transcriptome in AO-FACS-sorted cells using two techniques. Initially, we identified the genes categorized as present in the FACS preparations to define the transcriptome. As a secondary analysis, we also defined genes that were overexpressed in these preparations compared with the IM. Sample comparison was undertaken using DNA Chip Analyzer (dCHIP) version 1.3 to identify differentially altered genes using the lower 90% confidence bound of fold change (>1.2 fold), and data were analyzed using the unpaired t-test (P < 0.05). Genes that were present and overexpressed in the EC cell transcriptome are the principal focus of this study.
A total of 8,125 probe sets were present in the AO-FACS-sorted fractions vs. 6,759 in the IM scraping. The latter included marker genes expected for the spectrum of different intestinal cell types including villin (enterocytes), mucins (goblet cells), and defensins (paneth cells). Marker genes for NE cells (chromogranins) were also expressed. When the two databases were compared, 722 genes were identified as upregulated (levels >2-fold, P < 0.05) in the AO-FACS fraction and 1,210 genes were identified as significantly downregulated.
Highly expressed genes (>4-fold) in the AO-FACS-sorted fractions included CgA and CgB (>6-fold, P < 0.005), TPH (>5-fold, P < 0.001), VMAT1 (>4.5-fold, P < 0.01), and the serotonin transporter (>6.4-fold, P < 0.01) (all markers of the EC cell; Table 3). Both substance P (>4.6-fold, P < 0.03) and guanylin (>4.9-fold, P < 0.02) were also expressed. Marker genes for other NE cells (SST, glucagon, and CCK) were absent. Neurotensin transcripts were variably expressed but present in two-thirds of the EC cell preparations. Receptors including SSTR2, SSTR5, and VPAC1, but not PAC1, CCK2, or neurotensin receptors, were identified. Levels of marker genes for other cell types in the ileum including villin (epithelial); mucin-2, -3, -4, and -5 (goblet); the defensins (1, 4, 5, 6, and 8) (paneth); CCR9 and CD4 (T cells); and CD45 and genes encoding lymphocyte antigen complexes (lymphocytes) were all absent. These results indicate that the gene expression profile of the FACS-sorted preparations is consistent with that expected for a homogenous preparation of intestinal EC cells.
|
1), dopaminergic (D2), serotoninergic (5-HT2C), GABA (A and B), and prostaglandin (E1/4 and F) receptors with an absence of histaminergic receptor subtypes. Growth factor receptors included IGF-I, IGF-II, transforming growth factor (TGF)-1, and FGF2. Growth factors included nerve growth factor-, stem cell growth factor, and members of the PDGF, TGF-, and EGF families. A few olfactory or amino acid transporter transcripts were noted, including OR72, SLC21a11, and SLC39. A number of cell cycle genes including A2, D2/3, G, H, I, and L as well as PCNA, cdk2, and the cyclin inhibitors P16INK4A and P27KIP1 were identified. Markers of carcinoid malignancy including melanoma antigen family D2 (MAGE-D2) and metastasis-associated 1 (MTA1) were not identified in these cells (27), but the mitosis-associated transcript NAP1L1 was expressed. We also noted that heat shock proteins (heat shock proteins 10, 27, 70, and 84) were expressed in EC cell preparations. These genes are presumably associated with stress responses and potentially represent activation by the cell preparative technique (Nycodenz gradient centrifugation and FACS). This indicates that this subset of genes may therefore not be present in situ in the EC transcriptome. |
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
-adrenergic, muscarinic, and GABA compounds. This information allows for the preliminary construct of a functional model of naive ileal EC cell serotonin secretion (Fig. 8).
|
80%) has been reported (46). This methodology entailed isolation of rat ilea, enzymatic digestion, and EDTA chelation. The resulting crude mucosal sample was separated by counterflow elutriation and enriched by performing step density gradient centrifugation to yield an 80% pure preparation. In our present study, we adopted a FACS technique to collect cells that specifically expressed the rate-limiting enzyme of serotonin synthesis (TPH), expressed a PACAP receptor, or were AO positive and produced a 99% pure ileal EC cell preparation. All preparations were preenriched for ileal EC cells by Nycodenz gradient centrifugation (EC cell enrichment 80%). The use of TPH immunostaining resulted in >98% positive preparations of ileal EC cells, fluoro-PACAP in >96% EC cells, and AO in >99% EC cells. In addition to producing a homogenous preparation, we sorted live ileal EC cells by using either fluoro-PACAP or AO to prelabel these cells. Initially, PACAP-positive Mastomys ileal EC cells could not be identified using the same concentrations and conditions that were used to isolate gastric ECL cells (38). Increasing the fluoro-PACAP concentration 100-fold resulted in identification of a discrete sortable cell population. These studies were discontinued because 1) cell preparations did not reach levels of homogeneity; 2) the predominant high-affinity PACAP receptor (PAC1) was not expressed on ileal EC cells; and 3) prelabeling with this agent, which is known to induce gastric NE cell function (38), could potentially have the same effect on ileal EC cells, which would limit the validity of physiological studies.
Increasing concentrations of AO resulted in an easily detectable population of positive cells that could be sorted by FACS. However, this increase in fluorescence was negated by both a decrease in cell viability (from 94% to 57%, 10-fold increase in LDH measurements) and unstimulated serotonin release (a 5-fold decrease at 20 µM concentrations). On the basis of this information, a concentration of 2 µM was determined to be effective (nondeleterious) in producing functional ileal EC cells that could be cultured and examined in vitro. With the use of this strategy, serotonin secretory responses to a number of candidate neural and hormonal agents could be quantified and compared with the Nycodenz gradient-enriched fraction (80% EC cells). This was undertaken to both quantitate serotonin secretion and establish the effect of contaminating NE cells on ileal EC cell function.
Serotonin secretion was specifically examined because prior studies have identified it as the dominant secretory product of the ileal EC cell and a key regulator of intestinal secretion and motility (10, 12, 20). Apart from the observation that serotonin is one of the main secretory products of ileal EC cells (8, 52), evidence has accumulated to suggest that it may be of clinical relevance in a number of gastrointestinal problems including diarrhea, motility abnormalities, and irritable bowel syndrome. (11). Of particular relevance is the observation that many intestinal carcinoid tumors are of EC origin and that serotonin is a key bioactive amine involved in the genesis of the carcinoid syndrome and thus implicated in the diarrhea and motility abnormalities that characterize this disease (34, 35). Plasma serotonin levels and urinary 5-hydroxyindoleacetic levels are regarded as the primary diagnostic criteria in the establishment of the diagnosis of carcinoid neoplasia (34). In addition, studies in humans have demonstrated that an infusion of 5-hydroxytryptophan (the upstream precursor of serotonin) induces diarrhea and abdominal cramping (22). Delineating the physiology of the ileal EC cell and its secretory regulators is thus of considerable scientific and clinical relevance.
In short-term culture, FACS-sorted ileal EC cells were stimulated to secrete serotonin by the adenylyl cyclase activator forskolin, demonstrating receptor-independent cAMP-mediated serotonin secretion (Table 4). The dual 1-adrenoreceptor agonist and cAMP activator isoproterenol induced serotonin secretion within 3 min, suggesting that a rapid neurally mediated event might be a component of the ileal EC cell response profile. This would certainly be consistent with the sudden onset of "carcinoid crisis" and the associated massive release of serotonin engendered by a provocative epinephrine-mediated event (contrast injection, general anesthesia). Both serotonin secretion and cAMP release exhibited dose-dependent responses with EC50s of 4.5 x 10–10 and 1.2 x 10–9 M, respectively. The time courses for secretion and cAMP activation are similar in this cell type (data not shown) and suggest that in ileal EC cells, -adrenergic receptor activation (via at least the cAMP pathway) plays a role in serotonin secretion. This effect on secretion is similar to that on neurotensin release from canine ileal NT cells (2).
|
Two of the best-characterized regulators of gastric ECL cell histamine secretion, gastrin and PACAP (42, 45, 55), were next examined in the FFACS fraction. Gastrin had no effect on ileal EC cell serotonin secretion, consistent with the demonstration of the absence of the CCK2 transcript. A direct effect of gastrin on the ileal EC cell is thus unlikely. It is therefore possible that pentagastrin when used as a provocative test of carcinoid tumor amine secretion (34) produces its effect via an intermediate mechanism, which may include pentagastrin-stimulated release of adrenergic compounds (1). This is consistent with the observed serotonin responses initiated by the -adrenergic mimetic isoproterenol, noted in the present study.
In contrast to gastrin, PACAP stimulated ileal EC cell serotonin secretion with an EC50 of 1.2 x 10–9 M. This effect is presumably mediated via the VPAC1 receptor because transcripts for this PACAP receptor subtype but not PAC1 were identifiable. The stimulatory effect of PACAP was in contrast to other studies using ex vivo preparations, which have demonstrated that PACAP (and VIP) inhibits serotonin release (16, 30). This reported difference in effects reflects the complexity inherent in the use of models that contain multiple cell systems and the likelihood of the activation of intermediate cellular regulatory circuitry. A similar experience was noted in dissecting out the effect of gastrin in isolated gastric ECL cells and isolated gastric parietal cells as opposed to the data that emerged from the use of Ussing chamber gastric mucosa or isolated gastric glands (23). Overall, these may be an upstream effect whereby PACAP/VIP stimulates the release of an unidentified inhibitory agent (e.g., SST) in a fashion analogous to that reported in an isolated gastric cell study (23).
Acetylcholine chloride, an important regulator of cellular muscarinic responses, inhibited basal serotonin secretion from FFACS cells with an IC50 of 3 x 10–10 M and inhibited isoproterenol (5 x 10–10 M)-stimulated serotonin secretion with an IC50 of 6 x 10–10 M (Table 4). Our observations of the inhibition of serotonin secretion are supported by the present GeneChip studies demonstrating muscarinic (M4) receptors on these cells as well as the identification of M2 and M4 receptors in rat EC cell preparations (47). In the isolated canine NT cell studies of Barber et al. (2, 3), the muscarinic M3 receptor agonist carbachol inhibited epinephrine-stimulated neurotensin secretion. This suggests that ileal NE secretion may share a common inhibitory pathway via muscarinic receptor activation.
GABA is a major inhibitory neurotransmitter in the mammalian nervous system (14) and functions through a paracrine or autocrine effect via GABAA receptor activation (13). More recently, it has become apparent that this agent may play a similar role as a signaling molecule in endocrine organs. Thus GABA and GABA receptors have been identified in peripheral endocrine organs, including growth hormone cells of the anterior pituitary lobe (17), in pancreatic islets (43), and in the testis of human and rodents (18). It is thought that GABA may regulate the synthesis and release of hormones in an autocrine/paracrine manner (13). In Mastomys ileal EC cells, GABAA dose dependently inhibited basal and adrenergic-stimulated serotonin secretion with IC50s of 3 x 10–10 and 4 x 10–10 M, respectively. These observations are supported by our present GeneChip studies demonstrating the presence of ileal EC cell GABAA receptors as well as another report (47) of the identification of GABA receptors in rat EC cell preparations. These results are in contrast to an earlier study (46) that demonstrated no effect of GABAA on EC function in an 80% pure ileal EC cell preparation. The absence of an effect in the latter preparation may reflect a degree of cell heterogeneity in that it was a mixed preparation contaminated with a spectrum of cells from the IM (47). In the present study, a comparison of secretory responses in FACS-sorted ileal EC cells and cell preparations enriched by gradient centrifugation only (80% pure) we noted that in the latter (impure) preparations both a lower serotonin secretory responses (50% of FACS-sorted cells) as well as different responses (stimulation by acetylcholine and GABAA in contrast to inhibition) to physiological agents. These differences emphasize the limitations of an impure ileal EC cell preparation and confirm the physiological relevance of a homogenous cell preparation in the assessment of function.
The response of EC cells to GABAA raises the consideration of the efficacy of pharmacological targeting the GABAA receptor with specific classes of benzodiazepines that can stimulate GABAA receptor function to inhibit EC cell function.
Neither the SST transcript nor SST immunoreactivity could be demonstrated in the FACS-sorted preparations. In contrast, the FN fraction expressed SST transcript and SST-containing cells were readily identifiable (5% of the preparation). A comparison of secretory responses in FFACS and FN preparations demonstrated a number of marked differences. The stimulatory effect of a maximal dose of isoproterenol was greater than two times higher in the pure cell preparation compared with the FN preparation, whereas no inhibitory effect of acetylcholine could be demonstrated in these cells. Inhibitory responses to maximal concentrations of GABAA and octreotide were noted in the FN preparation, but these were not as marked as in the pure EC cell preparation. We interpreted this as a reflection of actively secreting SST cells (or an unknown other inhibitory cell type) that hinder serotonin release under short-term culture conditions in the FN preparations. It would thus seem optimal to utilize pure preparations of EC cells to define the physiology of the EC cell.
Neurotensin, a peptide with a variety of putative biological functions of as-yet-unproven clinical significance, has also been noted to be involved in the regulation of intestinal motility (52). The identification of the neurotensin transcript in FACS-sorted preparations indicates that the N cell is a component of the <1% contaminating population. In the examination of the EC cell transcriptome, using the same approach that we used to define the transcriptome of Mastomys gastric ECL cells but with a more recent version of mouse chips (Mouse Expression Set 430 chips) (26), neurotensin receptors could not be identified. The absence of these receptors suggests that it is unlikely that the EC cell would respond to this peptide, and, hence, it seems doubtful that neurotensin might perturb EC functional assessment.
Further analysis of the FACS-sorted cell transcriptome revealed a receptor profile (SST, PACAP, acetylcholine, and GABA) consistent with the physiological results of our study. In addition, serotoninergic (5-HT2C) and tachykininergic (TAC2) but not histaminergic receptors were present on the EC cells, suggesting that these cells may be autoregulated by serotonin and substance P. The presence of olfactory receptors suggests that the EC cell may also have a defined sensing function particularly because EC cells have apical cytoplasmic processes that access the bowel lumen (49). Overall, the spectrum of growth factors and receptors identified is consistent with previous reports in small intestinal (EC-derived) carcinoids (37, 54), whereas the presence of cell cycle genes suggests that these cells may have the potential to proliferate and are not terminally differentiated as other specialized cells in the small intestine [e.g., epithelial cells (33)].
The definition of the EC cell transcriptome can also be utilized as a template to identify transcripts altered in malignant EC cells. Of note, therefore, is the fact that a number of marker genes that we have previously identified as being associated with malignant human EC cells (carcinoid tumors), particularly the adhesin MAGE-D2 and MTA1 (27), were absent in the naive Mastomys EC cell preparations. Elucidation of the naive human EC cell transcriptome will potentially 1) identify novel markers of EC cell neoplasia and 2) define pathways that are aberrantly expressed during neoplasia. We have previously used the delineation of the naive gastric ECL cell transcriptome as a template for the identification of neoplastic pathways in an animal model of gastric carcinoid disease (26).
In summary, we produced a homogenous preparation (>99%) of viable ileal EC cells from Mastomys that do not contain contaminating SST cells. The cells can be maintained in short-term culture and exhibit receptors for PACAP and SST but not gastrin or histamine. They respond to adrenergic agents and PACAP with serotonin secretion via a mechanism that is at least in part cAMP mediated. Serotonin secretion was inhibited by octreotide, acetylcholine, and GABAA. Gastrin has no effect on EC cell secretion. These studies, in 99% pure Mastomys EC cells, are consistent with studies of short-term cultured rat ileal enteroglucagon and neurotensin cells that secrete enteroglucagon-like and neurotensin-like immunoreactivity in response to the -adrenergic agonist isoproterenol and the cAMP activator forskolin. In our studies, serotonin secretion was reversed by the muscarinic agent acetylcholine chloride and SST, a phenomenon shared with EG/L and NT cells (2, 6). These data suggest that NE cells of the ileum may represent a group of sensory and regulatory cells that are subject to common adrenergic, muscarinic, and SST modulating influences (Fig. 8).
In conclusion, these studies provide the methodology by which the rodent ileal EC cell can be defined and its responses to stimuli quantified. It is likely that the application of this methodology will enable the evaluation of the human ileal EC cell and an assessment of its biological function in both health and disease. Overall, these studies will allow for a precise characterization of the ileal EC cell and provide an opportunity to identify the regulatory pathways of this cell system and its role in small bowel physiology. In addition, it provides an opportunity to assess the potential role of the EC cell in complex and ill-understood diseases such as irritable bowel syndrome and refine understanding of EC cell-derived carcinoid tumors that have hitherto eluded precise cellular and biological assessment.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
