INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

HCO₃ SECRETION by the duodenal mucosa alkalinizes the viscoelastic mucus gel adherent to the epithelial cell surface, providing an important first line of duodenal mucosal defense against hydrochloric acid in the intestinal lumen. In addition, centrally elicited vagal activation of the secretion presumably provides anticipatory protection against acid discharged from the stomach. Vagally mediated stimulation of duodenal alkaline secretion has been demonstrated in all species studied (8, 10). Several studies (15, 19, 22, 27, 29) of intact animals and isolated mucosa have also sought to characterize the mechanisms of both extrinsic vagal and enteric nervous control of secretion. In these studies, it has been shown that the nonselective muscarinic antagonist atropine inhibits the stimulation of mucosal HCO₃ secretion by the cholinomimetics carbachol and bethanechol in anesthetized rats (19, 27, 29) and guinea pigs (22) in vivo as well as in rabbit mucosa in vitro (15). There is, however, some controversy with respect to the type and subtypes of cholinoceptors involved in mediating this stimulation. One such study in the rat (27) provided evidence for involvement of nicotinic as well as muscarinic M₁ receptors. Further evidence for the involvement of M₁ receptors is the inhibition of basal secretion in human subjects (21) and of secretion stimulated by carbachol in the guinea pig (22) by the M₁-selective antagonist pirenzepine. Another study in the rat (29), however, suggested the involvement of muscarinic M₂ receptors. Interestingly, atropine also inhibits secretion stimulated by vasoactive intestinal polypeptide (VIP) in the guinea pig in vivo (22) and in rabbit mucosa in vitro (15). This suggests the possibility that there are interactions between muscarinic- and peptidergic-mediated stimulation of duodenal mucosal HCO₃ secretion.

Despite the physiological importance of duodenal HCO₃ secretion and its regulation, neither the duodenal enterocyte (duodenocyte) signaling mechanisms nor the specific receptors present on these cells have been characterized. Because muscarinic agonists have several possible sites of action in the intestine, including ion-transporting enterocytes, postsynaptic axons, and presynaptic terminals in the enteric nervous system, it is likely that one site of action will dominate in a particular experimental protocol. The use of isolated duodenocytes would thus seem a prerequisite for the demonstration of actions of muscarinic and other stimuli on duodenocyte receptors at the cellular level and for the evaluation of agonist-responsive intracellular signaling pathways.

The aim of the present investigation was thus to develop an agonist-responsive duodenocyte model and to begin to characterize the suitability of this model for defining cell-specific intracellular signaling pathways. We report the development of primary culture models of rat and human duodenocytes that retain agonist responsiveness for several days in culture. We further demonstrate, by measuring changes in intracellular calcium concentration ([Ca²⁺]_i) within single cells, the presence of receptors for both cholinergic agonists and CCK (CCK-8) on the same duodenocytes. Pharmacological analyses at the single-cell level further indicate that the cholinergic receptors that linked to the Ca²⁺ signaling pathway are of the M₃ subtype.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Isolation and primary culture of rat duodenocytes. Male rats (Lewis × Dark Agouti, 150-250 g; from Uppsala Biomedical Center breeding) were raised in a conditioning unit under standardized temperature and light conditions (21-22°C; 12:12-h light-dark cycle) and had access to pelleted food (Ewos, Södertälje, Sweden) and drinking water ad libitum. Animals were fasted in mesh-bottomed cages with free access to water for 20-24 h before cell isolation. All experiments were approved by the Ethics Committee for Experiments with Animals at Uppsala University.

2

Isolation and primary culture of duodenocytes from human biopsies. Duodenal biopsies (2-3 from each individual, ~1-2 mm in diameter) were obtained from eight patients (7 female and 1 male, median age 51 yr, range 39-71 yr) undergoing upper endoscopy at the Department of Surgery, Uppsala University Hospital and found to have endoscopically normal duodenal and gastric mucosae. No antiulcer drugs (H⁺-K⁺-ATPase inhibitors or histamine H₂-receptor antagonists), cholinergic antagonists, or sedatives had been administered for at least 5 days preceding endoscopy. The project was approved by the Ethics Committee of the Medical Faculty at Uppsala University, and all subjects provided written informed consent. Biopsies were immediately placed into 5 ml of sterile DMEM-F12 (plus 15 mM HEPES and L-glutamine, no HCO₃) containing 2.5% FCS and 0.01 mg/ml gentamicin, transported to the laboratory, rinsed, and minced in the same medium at room temperature within 20-40 min after removal. After washing, the minced tissue was suspended in enzyme-containing solution and cells isolated and placed in primary culture as described above for the rat duodenal mucosa.

Digitized video image analyses of duodenocytes in primary culture. Changes in [Ca²⁺]_i were measured in fura 2-loaded cells, using the dual-wavelength excitation ratio technique (31) as previously described (20). Cells that had been cultured for 1-5 days on Matrigel-coated glass-bottomed 35-mm dishes (20) were loaded with fura 2 (8 µM fura 2-AM for 30 min at 37°C), rinsed several times, and placed in a perfusion chamber on the warmed (37°C) stage of a Zeiss Axiovert microscope. Fluorescence in single cells was monitored at 340 and 380 nm excitation and 510 nm emission and quantitated with a GenIIsys light-amplified charge-coupled device (CCD) 72 video camera (DAGE-MTI, Michigan City, IN) using Inovision IC300 software (Research Triangle Park, NC) supported by a SUN 4/330 computer (Palo Alto, CA). Changes in [Ca²⁺]_i were expressed as the ratio of fluorescence intensities at 340 and 380 nM, as described previously (20). Images of cells and cell colonies were acquired with a DAGE 72 CCD camera and printed from the computer monitor with a UP 5000P printer (Sony).

Chemicals and drugs. CCK-8 (fragment 26-33) was from Peninsula Europe (Merseyside, UK). Atropine sulfate, EGTA, BSA (fraction V, pH 7, sterile 35% solution), sodium bicarbonate (sterile 7.5% solution, cell culture tested), carbachol (carbamylcholine chloride), collagenase type IA, DTT, DMEM-F12 (1:1 mixture with L-glutamine and 15 mM HEPES, without sodium bicarbonate), HEPES, gentamicin, and VIP (porcine) were obtained from Sigma Chemical (St. Louis, MO). Fura 2-AM was from Calbiochem (La Jolla, CA). Dispase II was purchased from Boehringer (Mannheim, Germany), Matrigel from Collaborative Biomedical Products (Bedford, MA), and FCS from Harlan Sera-Lab (Loungborough, UK). The muscarinic antagonists atropine, methoctramine tetrahydrochloride, 4-diphenylacetoxy-N-methylpiperidine methiodide (4-DAMP), and pirenzepine dihydrochloride were from Research Biochemicals (Wayland, MA). The M₃-selective antagonist 11-({2-[(diethylamino)methyl]1-piperidinyl}acetyl)5,11-dihydro-6H-pyrido(2,3-)(1,4)benzodiazepin-6-one (AF-DX 116BS) was a generous gift from Karl Thomae (Biberach, Germany).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Primary culture of rat and human duodenocytes. Initial cell fractions isolated from duodenal collagenase digests were enriched in columnar cells with villus enterocyte morphology (Fig. 1A). These cells, when attached in small groups, exhibited spontaneous active movement between 1 and 3 days of culture (not shown). In contrast, cellular fractions obtained after successive triturations were composed predominately of cells with rounded, vacuolated morphologies (Fig. 1A) characteristic of crypt cells (30). These vacuolated cells retained such morphologies for 2-3 days. Within 3-4 days of plating, small colonies of epithelial cells with distinct morphologies began to form and by days 7-10 "domes" characteristic of actively secreting cell types were evident (Fig. 1B). Importantly, duodenocytes did not attach to the matrix if collagenase was not used during the preparation procedure. Cells also did not attach as well if the mucosa was dissected away from the underlying substratum by blistering compared with our standard scraping protocol.

View larger version (77K): [in this window] [in a new window]   View larger version (102K): [in this window] [in a new window]   Fig. 1.   Duodenal enterocytes (duodenocytes) can be maintained in primary culture for several weeks, and cells with villus-like morphologies are present for at least 24 h in these cultures. A: duodenocyte fractions with villus and crypt-like morphologies are readily identified between 8 and 24 h of culture. Left: 2 villus cells after 8 h in culture. Middle: after 24 h, some villus cells become rounded, but others retain a columnar appearance. Right: after 24 h in culture, cells from crypt fractions are often found to contain secretory vacuoles. B: duodenocytes begin to divide and form discrete colonies after several days in culture. Shown are images from 2 different 4-day cultures of rat duodenocytes. Such colonies remained viable for at least 3 wk of culture. Bar, 20 µm. Left: arrow indicates dome formation in center of colony. Right: no domes are evident in this colony.

Effects of carbachol on [Ca²⁺]_i signaling in single duodenocytes. Carbachol consistently increased [Ca²⁺]_i in rat duodenocytes in primary culture. Moreover, the carbachol-induced increase in [Ca²⁺]_i in these cells was reproducibly observed even after several days of culture (Fig. 2). Two types of signaling patterns were observed in duodenocytes. In the first, [Ca²⁺]_i spiked rapidly and returned to basal or near basal values within 3-5 min after carbachol addition (Fig. 2). In the second, which was predominately observed in cells with enlarged vacuoles as described for secreting crypt cells (30), the initial Ca²⁺ transient was followed by slow, rhythmic oscillations in [Ca²⁺]_i of lower amplitude. The oscillatory patterns varied between cells and oscillations were usually not sustained. A typical pattern and frequency is shown in Fig. 3A in two cells from the same duodenal preparation. Selected images of these fura 2-loaded cells are shown in Fig. 3B.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Although most experiments were performed within 24-48 h after cells were placed in culture, carbachol continued to induce increases in intracellular calcium concentration ([Ca2+]i) in rat duodenocytes even after several days in primary culture. In contrast, cells with fibroblast-like morphologies did not respond to cholinergic stimulation. Shown are recordings of [Ca2+]i signals (10 µM carbachol present as indicated), as measured with fura 2, in 5 duodenocytes and 2 fibroblasts (not responding). Measurements were performed on the 5th day of culture.

View larger version (16K): [in this window] [in a new window]   View larger version (35K): [in this window] [in a new window]   Fig. 3.   Example of rhythmic oscillations in [Ca2+]i induced by carbachol in a subpopulation of rat duodenocytes with large secretory vesicles. Top: recordings of [Ca2+]i signals before and after addition of 10 µM carbachol. Bottom: selected images of cells from which recordings were obtained. Bar, 5 µm. A: transmitted light image before carbachol addition. B: fura 2 fluorescence ratio before carbachol addition. C: increase in fura 2 fluorescence ratio during peak carbachol response (darkened edges of image were off scale). D: transmitted light image at the end of the experiment showing expansion of cellular vacuoles.

6

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Similar to rat duodenocytes, human duodenocytes responded to increasing concentrations of carbachol in a dose-dependent fashion. For recordings depicted, duodenocytes were isolated from a biopsy from a 40-yr-old female and cultured for ~24 h before loading with fura 2. Data shown are from 5 duodenocytes and are representative of 30 cells (6 experiments with biopsies from 6 subjects aged 36-66 yr). Note the more sustained [Ca2+]i plateau in response to carbachol compared with the transient increase in rat duodenocytes (Figs. 2 and 3).

Effects of muscarinic receptor-selective antagonists. To define muscarinergic receptor subtypes present on duodenocytes, we tested several selective antagonists, including the M₃-selective antagonist 4-DAMP, the M₂-selective antagonists AF-DX 116BS and methoctramine, the M₁-selective antagonist pirenzepine, and the nonselective antagonist atropine. In an initial series of experiments, nanomolar concentrations of 4-DAMP were found to dose dependently inhibit the increase in [Ca²⁺]_i induced by 10 µM carbachol in human duodenocytes (Fig. 5). In other, less extensive experiments with rat duodenocytes, nanomolar concentrations of 4-DAMP were similarly effective in suppressing the increase in [Ca²⁺]_i in response to 10 µM carbachol. Interestingly, higher concentrations of 4-DAMP reduced the "steady-state" increase in [Ca²⁺]_i more potently than the initial, transient peak in [Ca²⁺]_i (Fig. 5D).

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Carbachol-induced increase in [Ca2+]i in human duodenocytes is potently suppressed by nanomolar concentrations of the muscarinic M3-selective receptor antagonist 4-diphenylacetoxy-N-methylpiperidine methiodide (4-DAMP). Shown are recordings from 5 duodenocytes that were obtained from a duodenal mucosal biopsy of a 57-yr-old female with normal duodenal mucosa. Cells were cultured for ~24 h before Ca2+ measurements were performed. Data are representative of results from 37 cells isolated from 3 female patients (36-57 yr of age).

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Differential effects of extracellular Ca2+ chelation on carbachol-induced elevation of [Ca2+]i in the presence and absence of 4-DAMP. Compared with the response to carbachol alone, a submaximal concentration of 4-DAMP (3 nM) caused a slight delay in the initial carbachol-induced increase in [Ca2+]i and strongly suppressed the plateau phase of the response (compare A and D). Chelation of extracellular Ca2+ with EGTA suppressed the plateau phase of the carbachol response but not the initial transient increase in [Ca2+]i (compare C and D). In contrast, 4-DAMP almost completely abolished the response to carbachol when EGTA was present (compare B and C). Taken together, these results and those in Fig. 5 suggest that both phases of the carbachol-induced increase in [Ca2+]i are suppressed by 4-DAMP. The EGTA data further suggest that the release of Ca2+ from intracellular store(s) is potently affected by the inhibitor. Recordings are from 6 duodenocytes that were originally isolated from the same human mucosal biopsy as used in Fig 5. Cells were in primary culture for ~24 h before use.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   In contrast to 4-DAMP, micromolar concentrations of the M1-selective antagonist pirenzepine (pir.) are required to suppress the carbachol-induced increase in [Ca2+]i. Duodenocytes from a 57-yr-old female patient were placed in primary culture for ~24 h before Ca2+ measurements. Data are representative of results from 22 cells in primary culture initially isolated from duodenal biopsies from 2 female patients (ages 36 and 57 yr).

View larger version (29K): [in this window] [in a new window]   Fig. 8.   Similar to the M1-receptor antagonist pirenzepine only very high (micromolar) concentrations of the M2-selective antagonist AF-DX 116BS (AF-DX) were capable of suppressing the carbachol-induced increase in [Ca2+]i in duodenocytes. Shown are recordings from 6 duodenocytes that were isolated from a duodenal biopsy of a 57-yr-old female and cultured for ~24 h. Data are representative of results with 24 cells in 4 different experiments (4 subjects, 36-57 yr).

View larger version (25K): [in this window] [in a new window]   Fig. 9.   Comparison of differential effects of the muscarinic receptor antagonists 4-DAMP, AF-DX 116BS, pirenzepine, methoctramine, and atropine. Shown is a continuous recording of [Ca2+]i in a single human duodenocyte. Data demonstrate the reproducibility of [Ca2+]i responses of human duodenocytes (compare with recordings in Figs. 5, 7, and 8) as well as the expected inhibition by atropine and lack of inhibition by methoctramine, another M2-selective antagonist.

Effects of CCK and VIP. CCK released postprandially from the duodenal mucosa is an interesting potential stimulus of intestinal secretion; however, CCK actions at the enterocyte level have not been characterized. The octapeptide of CCK, consisting of the carboxy-terminal eight amino acids (CCK-8), is the most biologically potent small peptide of CCK that has been isolated. Similar to carbachol, CCK-8 increased [Ca²⁺]_i in rat (Fig. 10) as well as in human (Fig. 11) duodenocytes. Comparing the [Ca²⁺]_i response of CCK-8 (10 nM) with that of 10 µM carbachol in rat duodenocytes (Fig. 10), it appears that the responses are of similar magnitude. Moreover, although the CCK response has not yet been studied in depth, the data suggest the [Ca²⁺]_i response to CCK in human duodenocytes is rapidly downregulated as was previously observed with respect to the CCK/gastrin response in the gastric parietal cell (5). Thus, in Fig. 11, the response to 500 nM CCK-8 was smaller than that to the preceding exposure to 50 nM CCK-8. Furthermore, 1 nM CCK failed to elicit an increase in [Ca²⁺]_i when administered after a higher dose of CCK had been removed. When 1 nM CCK was added initially, a response could be elicited (not shown).

View larger version (17K): [in this window] [in a new window]   Fig. 10.   CCK-8 induces an increase in [Ca2+]i in rat duodenocytes. Shown are representative recordings from 2 cells cultured for ~24 h. Cells were first stimulated with carbachol to show specific response patterns for the sampled cells. After removal of carbachol, cells were then stimulated with CCK-8. Although oscillations in [Ca2+]i were not observed in all cells, similar initial spike responses to CCK-8 were detected in 19 rat duodenocytes isolated from 6 different animals.

View larger version (15K): [in this window] [in a new window]   Fig. 11.   Human duodenocytes also respond not only to carbachol but also to the peptide hormone CCK-8 with an increase in [Ca2+]i. In contrast to carbachol, the response to CCK-8 is rapidly downregulated (compare with Fig. 4). Shown is a representative recording from a single duodenocyte from an ~24-h primary culture of duodenocytes isolated from a normal mucosal biopsy of a 66-yr-old female. Similar responses to CCK were obtained in 12 human duodenocytes from 3 separate biopsy experiments.

9

6

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The results presented in this study demonstrate, for the first time, that viable, agonist-responsive enterocytes from both crypt and villus regions of human and rat duodenal mucosae can be isolated and maintained in primary culture for several days. In addition, dividing colonies of agonist-responsive cells with epithelial-like morphologies, originating from the initial cell isolates, could be sustained in culture for at least 3-4 wk. Because of the possibility of actions at several sites of muscarinic and other secretory agonists in the intact animal in vivo, the use of isolated duodenocytes seems a prerequisite for the demonstration of the presence of enterocyte-specific receptor subtypes as well as for defining the intracellular signaling pathways activated by these agonists. In developing the models used here, cells were isolated only from young adult rats and from adult human biopsies. Thus potential problems with undeveloped expression of receptors that might occur with use of cells from immature (fetal) animals or from transformed cells lines were avoided.

The current model was developed using modifications of a technique previously used to establish and characterize a primary culture model of gastric parietal cells (3, 20). An important factor in the initial cell isolation steps was the omission of EDTA and the use of a relatively gentle trituration technique, rather than vibration, to isolate villus and crypt-enriched fractions (see Ref. 6 for a review of previously used intestinal cell isolation techniques). The use of a mild enzymatic digestion protocol and the inclusion of DTT to break mucus disulfide bonds were further found to be important prerequisites for cell attachment. Previous studies with gastric parietal cells have shown Matrigel to be an ideal attachment matrix for this cell type (3, 20). In this study, although cell attachment factors were not systematically examined, a thin coating of dilute Matrigel was also found to be a reliable attachment matrix for both human and rat duodenocytes. It should be noted that in an earlier study, we found collagen I to be a superior attachment factor for acutely isolated, single duodenocytes that had been exposed to Ca²⁺ chelators (16). It remains to be determined whether or not these differences are the result of differences in cell isolation techniques or are related to the plating conditions with cell clumps vs. single cell isolates.

Our results with duodenocytes in primary culture show that cells with crypt and villus-like morphologies from both humans and rats respond to carbachol with a dose-dependent increase in [Ca²⁺]_i. These data contrast with previous results obtained with human jejunal biopsies and in rat enterocytes isolated using a hyaluronidase digestion technique (13) in which there was no detectable response to cholinergic agonists in fura 2-loaded cells. The divergence in stimulatory pathways and transport characteristics between duodenal mucosa, capable of transporting HCO₃ at high rates (8), and jejunal mucosa may contribute to this difference. It would seem likely, however, that the improved responsiveness is due to the fact that duodenocytes are 1) isolated by a more gentle enzymatic technique, 2) not exposed to Ca²⁺ chelators, and 3) placed in primary culture to allow for recovery from the initial insults of tissue dissociation. Interestingly, we (unpublished observation) and others (13, 28) have found that the cAMP signaling pathway is much less sensitive to initial Ca²⁺ chelation and harsher enzymatic digestion protocols.

On the basis of our pharmacological analyses in which a pattern of inhibition characteristic of the M₃ receptor subtype was observed (i.e., 4-DAMP  pirenzepine > AF-DX 116BS) (2), we conclude that M₃ muscarinic receptor subtypes are present on both crypt and villus-like cells and are linked to the Ca²⁺ signaling pathway in these cells. To date, five different muscarinic receptor subtypes have been cloned (2). M₁ and M₃ receptors have been shown to be linked to phosphoinositol turnover and the Ca²⁺ signaling pathway and are generally stimulatory. M₂ (and M₄ receptors) are linked to adenylyl cyclase and are generally involved in mediating inhibitory effects. Within the intestine, muscarinic agonists have several well-documented sites of action, including the ion-transporting enterocytes, postsynaptic axons, and presynaptic terminals in the enteric nervous system. Previous studies of the neurohumoral control of duodenal mucosal HCO₃ secretion (9, 21, 27, 29) have suggested that cholinergic stimulation of the duodenal secretion involves muscarinic M₁ and M₂ receptors as well as nicotinic receptors. The central nervous system and the peripheral sympathetic ganglia provide additional possible sites of action in the intact animal, and one site of action may dominate in a particular experimental protocol. For example, the muscarinic M₁-receptor antagonists pirenzepine and telenzepine did not inhibit but caused a dose-dependent increase in duodenal mucosal secretion HCO₃ secretion in a study in anesthetized rats (27). Pretreatment with the adrenoceptor antagonist phentolamine prevented the increase in secretion, suggesting that pirenzepine and telenzepine acted in a stimulatory manner by antagonizing M₁ transmission in sympathetic ganglia thereby decreasing postsynaptic -adrenoceptor-mediated inhibition of the secretion.

The possibility of direct regulation of duodenocytes by muscarinic M₃ receptors, as demonstrated in the present study, has not been considered previously. However, cholinergic actions mediated by this receptor subtype have been demonstrated in several tissues (2), including stimulation of H⁺ secretion by gastric parietal cells (24, 32) and ¹²⁵I efflux in the T84 colonic cell line (7). In addition, pharmacological characterization of the carbachol response in isolated rat distal colonic mucosa has suggested the involvement of muscarinic M₃ receptors (23).

In the present study, the pattern of carbachol-induced increase in [Ca²⁺]_i in rat duodenocytes was typically transient, returning to basal or near basal values within 3-5 min after the initial peak response. In contrast, in human duodenocytes the initial transient peak was very reproducibly followed by a sustained plateau in [Ca²⁺]_i. This biphasic Ca²⁺ response to carbachol is characteristic of many other nonexcitable cell types and a substantial amount of evidence indicates that the initial spike in [Ca²⁺]_i is the result of release of Ca²⁺ from an intracellular storage site(s), whereas the later sustained phase is due to the influx of Ca²⁺ across the cell membrane. This also appears to be the case for duodenocytes, because as shown here chelation of extracellular Ca²⁺ leads to the abolishment of the plateau phase but not the initial increase in [Ca²⁺]_i elicited by carbachol. Thus our results with submaximal inhibitory concentrations of 4-DAMP and the extracellular Ca²⁺ chelator EGTA suggest that this M₃-receptor antagonist is a more potent inhibitor of carbachol-stimulated release of intracellular Ca²⁺ than of Ca²⁺ entry. These data are in line with a previous report by Wilkes et al. (32) who found that 4-DAMP is a more potent inhibitor of intracellular Ca²⁺ release (EC₅₀ ~1.7 nM) than influx (EC₅₀ >30 nM) in the gastric parietal cell.

Although not observed in human duodenocytes, we found that a subpopulation of rat duodenocytes with enlarged vacuoles, as described for crypt-like secreting enterocytes (30), responded to carbachol and CCK-8 with an initial transient increase in [Ca²⁺]_i followed by rhythmic oscillations in [Ca²⁺]_i of lower amplitude (~ min). The reason for these differences between rat and human Ca²⁺ signaling patterns is presently unclear but may be related to either differences in maturity or to secretory status of the cells that were studied. Agonists that provoke inositol lipid hydrolysis have been shown to induce oscillations in [Ca²⁺]_i in some, but not all, nonexcitable cell types, but the physiological significance of these oscillations is not clearly established. For example, carbachol induces sustained oscillations in [Ca²⁺]_i in mouse pancreatic -cells exposed to 20 mM (but not in such cells exposed to 3 mM) of glucose (12). In rabbit cultured gastric parietal cells (20), carbachol induced apparently rhythmic oscillations but only at concentrations below the half-maximal concentration for initiation of acid secretion. Moreover, higher concentrations of carbachol (>1 µM), which did initiate a secretory response, induced only a transient spike followed by a sustained elevation of [Ca²⁺]_i in these cells.

In contrast to the cholinergic signaling pathway, other agonists that stimulate duodenal mucosal HCO₃ secretion have been shown to increase cAMP production. For example, dopamine D₁-receptor agonists, which stimulate HCO₃ secretion in rat (9) and human (18) duodenum in vivo, increased cAMP production in fractions of crypt as well as in fractions of villous duodenocytes from the rat (26). In addition, VIP, which is a potent secretory stimulant in all mammalian species tested (8), also increased cAMP production in crypt and villus duodenocytes from the rat (26) and guinea pig (25). Not unexpectedly, this peptide was not found to increase [Ca²⁺]_i in the present study. It would thus seem highly likely that both the cAMP and the [Ca²⁺]_i signaling pathways are important modulators of duodenocyte function(s). It has also recently been shown (11) that 8-bromo-cGMP as well as cGMP stimulate HCO₃ in rat duodenum in vitro.

With respect to the specific intracellular events activated by the cAMP and Ca²⁺ signaling pathways, studies of amphibian mucosa in vitro, of membrane vesicles, and of acutely isolated rat and rabbit duodenocytes (1, 8, 16) have indicated that duodenocytes import HCO₃ at the serosal membrane by NaHCO₃ cotransport and export Cl/HCO₃ exchange via an apical anion conductive pathway, recently suggested to be the cystic fibrosis transmembrane conductance regulator channel (11, 14). The cellular location (villus or crypt) within the duodenal mucosa of the secretory processes and their relation to the mode of intracellular signaling have, however, not yet been clarified. Electrogenic secretion of HCO₃ mediated by cAMP may, as suggested for secretion of Cl by more distal small intestine, be a property of crypt cells. In contrast, villus cells may export HCO₃ by Cl/HCO₃ exchange (8). However, the results of the present study suggest that villus as well as crypt cells respond to muscarinic M₃ receptor stimulation with an increase in [Ca²⁺]_i.

CCK released postprandially from the duodenal mucosa is an interesting potential stimulus of duodenal mucosal HCO₃ secretion, but contraction of the smooth muscular layers induced by CCK and obstructing luminal perfusion makes it experimentally difficult to interpret actions of this hormone on intestinal electrolyte transport in intact animals. The presence of CCK receptors at vagal afferent fibers may pose a further problem in elucidating results. Studies of distal ileum, stripped of the longitudinal muscle layer and mounted in an in vitro chamber (17), have suggested that CCK acts at enteric neuronal receptors to stimulate mucosal short-circuit current (anion secretion), in part by release of ACh. Actions of CCK at the enterocyte level have not been reported before. The results of the present study indicate that, similar to carbachol, CCK-8 increases [Ca²⁺]_i in isolated duodenocytes. Thus it would seem most likely that CCK exerts a direct stimulatory effect on the duodenocyte in vivo. Interestingly, the duodenocyte [Ca²⁺]_i response seems rapidly downregulated as was previously observed with respect to the response to CCK and gastrin in the gastric parietal cell (5).

In summary, the results of the present study indicate that muscarinic receptors are present on both human and rat duodenocytes and are M₃ in nature. Because these duodenocytes respond not only to the cholinergic agonist carbachol but also to the peptide hormone CCK-8 with an increase in [Ca²⁺]_i, our results further suggest that postprandially released CCK, similar to cholinergic stimuli, may be involved in stimulating duodenocyte HCO₃ secretion and protecting the duodenum against acid discharged from the stomach.