Both crypt and villus intestinal goblet cells secrete mucin in response to cholinergic stimulation THOMAS E. PHILLIPS Division of Biological Sciences, University of Missouri, Columbia, Missouri 65211 Phillips, Thomas E. Both crypt and villus intestinal goblet cells secrete mucin in response to cholinergic stimulation. Am. J. Physiol. 262 (Gczstrointest. Liver Physiol. 25): G327-G331, 1992.-Computer-assisted morphometric analysis was used to quantify the effects of cholinergic stimulation on intestinal goblet cells. Within 5 min of stimulation (250 pg/kg carbachol SC), many crypt goblet cells were depleted of mucin secretory granules and their apical membranes had the deep cavitation that accompanies recent compound exocytotic activity. The percentage of crypt epithelial volume occupied by mucin secretory granules was decreased by 58.4% at 5 min and 45.9% at 60 min. Although villus goblet cells never showed signs of recent compound exocytosis, morphometric analysis revealed a 22.4% decrease in the percentage of villus epithelial volume occupied by mucin secretory granules within 5 min of stimulation and a 32.4% decrease by 60 min. The decrease in villus mucin stores was due to both a reduction in the volume of mucin in an average villus goblet cell and a drop in the number of recognizable goblet cells per square micrometer of villus epithelium. Mucin stores in both the crypt and villus regions were largely replenished by 4 h poststimulation. mucus; exocytosis;
ileum; morphometry
of cholinergic stimulation, both crypt and villus intestinal goblet cells secrete mucin at a slow baseline rate using simple exocytosis. Cholinergic stimulation causes crypt goblet cells to rapidly accelerate their discharge of mucin by compound exocytosis (4-8). Compound exocytosis of mucin secretory granules is easily recognized in light-microscopic semithin sections or in transmission electron microscopy by the deep cavitation of the apical cell surface. Previous studies established that induction of exocytosis results from a direct action of cholinomimetics on the epithelium (4) and that the same effect could be obtained by the electrical stimulation of mucosal nerves (5). In contrast, neither cholinergic agonists (4, 6-8) nor nonspecific electrical stimulation of mucosal nerves (5) could elicit any sign of compound exocytotic activity by villus goblet cells. It was unknown, however, whether villus cells respond to stimulation by cholinomimetics or other agents by increasing their mucin discharge by a mechanism that does not result in the dramatic compound exocytotic cavitation of their apical membrane, such as acceleration of simple exocytosis or selective exfoliation of goblet cells. The purpose of this study was to address this possibility using computer-assisted morphometric analysis. IN THE ABSENCE
MATERIALS
AND
METHODS
Sprague-Dawley rats, 44-54 days old and weighing between 220 and 255 g, were used for this study. Animals were lightly anesthetized with methoxyflurane, weighed, and injected subcutaneously with 250 pg/kg carbachol (carbamylcholine, Sigma) (6) in phosphate-buffered saline (PBS) and stimulated for 5 (n = 3), 15 (n = 5), 60 (n = 5), or 240 (n = 5) min before 0193-1857/92
$2.00
Copyright
being killed by pentobarbital overdose. Mock-treated control rats (n = 5) were injected with PBS 5 min before being killed. All protocols were approved by the University of Missouri Animal Use Committee and followed National Institutes of Health guidelines. The abdominal cavity was opened with a midline laparotomy and a small slice of tissue was removed from the ileum (- 3 cm proximal to the ileocecal junction), immediately placed in cold PFG fixative [ 2.5% glutaraldehyde, 2% freshly depolymerized paraformaldehyde, 70 mM NaCl, 30 mM N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid (HEPES), 2 mM CaClz, pH 7.41, and cut into l- to &mm3 cubes. After 2 h of fixation in fresh PFG at 4”C, the tissues were rinsed in the HEPES buffer (70 mM NaCl + 30 mM HEPES, pH 7.4) and postfixed in 1% 0~0~ in HEPES buffer. After rinses in HEPES buffer and deionized water (dHzO), tissues were stained en bloc with 0.5% uranyl magnesium acetate overnight at 4OC, washed in dHzO, dehydrated in a series of ethanol solutions, and embedded in epoxy resin (EmBed 812, EMS, Fort Washington, PA). For computer-assisted morphometric analysis (American Innovision, San Diego, CA), a color image of a O.&pm toluidine blue-stained section was displayed on a monitor at ~1,364 magnification. A mouse-driven cursor was used to outline total crypt or villus epithelial surface area or the accumulations of mucin granules contained in individual goblet cells. All slides were randomized and coded for analysis by the investigator. Three randomly chosen blocks of tissue from each animal were examined; a slide with 5-10 sequential sections was produced from each tissue block. One randomly chosen section on each slide was used for analysis; within the section, all complete villi (i.e., not including those on margins of sections in which one side was not present) and all crypt epithelia directly beneath these villi were analyzed. Typically, a section would include three to four villi. A two-group unpaired Mann-Whitney U test was performed on each set on data using the StatView SE+ computer program (Abacus Concepts, Berkeley, CA). For photomicrography, 0.9-pm sections were stained with hematoxylin and 1% safranin 0 and photographed using a Wratten no. 44 filter with a Zeiss Axiophot microscope. RESULTS
Goblet cells in the mock-stimulated rat ileum contained a complete complement of secretory granules (Fig. I&. Mucin secretory granules occupied 5.5 t 0.3% (SE) of the epithelial area in ileal crypts of mock-stimulated rats. The average profile of secretory granules in the crypt zone was 21.1 t 1.2 pm’. The number of goblet cells per square micrometer, estimated by the number of granule profiles, averaged 2.65 x 10B3 & 1.58 X 10v4 per ,um2of crypt epithelium. As expected from previous studies (e.g., Ref. 6), goblet cells in ileal crypts showed a loss of secretory granules and a deep cavitation of their apical surface within 5 min of cholinergic stimulation (Fig. 1B). At 5 min poststimulation, the area occupied by secretory granules decreased to 41.5 t 2.6% (P 5 0.0001) of the control value (Fig. 2A). This decrease in crypt mucin stores resulted
0 1992 the American
Physiological
Society
G327
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
G328
INTESTINAL
GOBLET
CELL MUCIN
SECRETION
Fig. 1. Time course of ileal response to cholinergic stimulation. All micrographs are at same magnification (bar = 50 pm). A: in mock-stimulated ileum, all goblet cells in crypt (arrows) and on villi (arrowheads) appeared packed with mucin secretory granules. B: 5 min after cholinergic stimulation, many crypt goblet cells (small arrows) are depleted of secretory granules and their apical membranes showed deep cavitation indicative of recent compound exocytotic activity. Exfoliated goblet cells (large arrowheads) were occasionally observed trapped in extracellular mucous coat. Some crypt goblet cells (large arrows) and all villus goblet cells (small arrowheads) showed no sign of cavitation and appeared filled with granules. C: by 60 min poststimulation, crypt goblet cells (arrows) no longer showed apical membrane cavitation but intracellular mucin stores remained partially depleted. Villus goblet cells (arrowheads) continue to show no sign of cavitation. D: by 4 h poststimulation, both crypt (arrows) and villus (arrowheads) goblet cells were difficult to distinguish from mock-stimulated controls.
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
INTESTINAL
r
GOBLET
CELL
MUCIN
SECRETION
G329
r
5
15
60
240
C
5
15
60
240
D
5
15
60
240
60
240 minutes of cholinergic
5
15
60
240
60
240
Fig. 2. A and B: percentage of total epithelial area occupied by mucin secretory granules for crypt (A) and villus (B) regions expressed as percentage of control value and plotted against minutes of cholinergic stimulation. C and D: average accumulation of mucin secretory granules (total pm2 surface area occupied by secretory granules divided by number of profiles) expressed as percentage of control values and plotted against min of cholinergic stimulation for crypt (C) and villus (D) cells. E and F: no. of cells with mucin secretory granules per ,um2 of epithelial surface area expressed as percentage of control value in crypt (E) and villus (F) regions plotted against min of cholinergic stimulation. In all graphs, each bar represents average response of 5 rats each at 0, 15, 60, and 240 min and 3 rats at 5 min; error bars indicate SE of mean. P values determined using a 2-group unpaired Mann-Whitney U test are indicated above each bar.
stimulation
largely from an all-or-none secretory response, since the number of crypt cells containing recognizable secretory granules cells decreased to a similar 44.4 t 4.5% (P = 0.0002) of the control value at 5 min (Fig. 2E), whereas the average accumulation of mucin in cells still containing secretory granules did not change (99.0 t 7.6% of control value, P = 0.98) at this time (Fig. 2C). Cavitation was still prominent at 15 min, but by 60 min poststimulation cells with deeply cavitated apical membranes were observed less frequently (Fig. 1C). Shedding of “membrane tags” (7) and endocytotic retrieval apparently removed the excess granule membrane from the apical surface by 60 min. At 15 and 60 min, crypt mucin stores remained significantly depleted from the control level (P -< 0.0001) but were not significantly changed (P = 0.30) from the 5-min level (Fig. 2A). Although mucin stores had not increased significantly by 15 min, replenishment had begun, since the number of cells containing secretory granules had rebounded to 59.7 t 7.2% of the control value (P = 0.0007), which is significantly higher than the 5-min value (P = 0.0165). As would be predicted, these refilling cells caused a small drop in the size of the average granule accumulation (Fig. 2C). By 240 min poststimulation (Fig. lo), the crypt mucin stores had refilled to 89.7 t 8.2% of the control value (P = 0.40).
At 240 min, the number of granule profiles was 106.5 t 7.6% (P = 0.58) of the control value, but this value may not accurately reflect the number of goblet cells, since at this time point active synthesis resulted in distinct accumulations of mucin-filled granules in the Golgi region of some cells. Because Golgi accumulations of granules could not be discerned from true thecal accumulations except in fortuitously sectioned cells, all accumulations were counted. Because both the Golgi and thecal accumulations in some cells may have been individually counted, the number of granule profiles per square micrometer might overestimate the number of goblet cells at 240 min poststimulation. In the villus region of mock-stimulated rats, mucin secretory granules occupied 7.62 t 0.52% of the epithelial area. The number of goblet cells per square micrometer of control villi (1.40 X lo-” t 7.3 X lo-“) was 47.2% less than in the crypt zone, but the average profile of mucin granules was 2.57 times larger (54.1 t 2.6 pm’). Less than 1 of every 500 goblet cells on villi above cholinergically responsive crypts showed any sign of the apical membrane cavitation indicative of recent compound exocytotic activity (Fig. 1, B-D). Electron-microscopic examination also failed to reveal evidence of compound exocytosis from villus goblet cells. Morphometric analy-
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
G330
INTESTINAL
GOBLET
CELL
sis, however, revealed that villus mucin stores decreased to 77.6 t 10.2% of the control value (P = 0.089) by 5 min and to 67.5 t 5.1% (P = 0.002) by 60 min (Fig. ZB). The average profile of secretory granules was also decreased by cholinergic stimulation, with a maximal decrease occurring at 60 min (78.8 t 5.7% of control value, P = 0.015; Fig. 2D). Decreases in the number of granulecontaining cells occurred at 15 min (88.0 t 4.8% of control value, P = 0.085) and 60 min (88.7 t 6.5%, P = 0.120; Fig. 2F). Occasional examples of exfoliated intact goblet cells trapped in the luminal mucous blanket were observed (Fig. 1B). By 240 min poststimulation, villus mucin stores recovered to 93.8 t 8.1% (P = 0.237) of the control value and the average accumulation of granules was 81.4 t 3.0% of control (P = 0.009). At 240 min, the number of granule-containing cells was 115.0 t 7.7% (P = 0.19) of the mock-stimulated value, but as in the case of the crypt cells accumulations of granules in the Golgi region probably led to an overestimation of the number of goblet cells at this time point. In mock-stimulated rats, the ratio of villus epithelial area to the underlying crypt epithelial area was 0.97 t 0.02. Large-scale exfoliation of villus cells in response to stimulation would be expected to reduce this ratio. Only two rats, both at 60 min poststimulation, showed significant decreases from the control value (ratios: 0.47 t 0.02, P = 0.013; 0.63 t 0.04, P = 0.038). DISCUSSION
This study demonstrated that cholinergic stimulation affects mucin secretion from villus as well as crypt goblet cells. In the rat ileum, the villus epithelium underwent reproducible reductions in intracellular mucin stores that were not previously detected. Although crypt goblet cells showed the deep cavitation of their apical membranes indicative of recent compound exocytotic activity, only rare profiles of cavitation of villus goblet cells were ever observed. The cholinergically evoked decrease in the average accumulation of secretory granules in villus goblet cells presumably reflects an acceleration of simple exocytosis (i.e., fusion of single granules with the plasmalemmal membrane) but without the granule-granule fusion mechanism seen as compound exocytosis. This variation in the rate of simple exocytosis is consistent with earlier autoradiographic studies in which the rate of granule transport during unstimulated baseline release of mucin varied widely between individual cells (2). Accelerated simple exocytosis could also conceivably account for the decrease in the number of villus cells with secretory granules. Without the apical cavitation due to compound exocytosis, goblet cells depleted of mucin granules would have been difficult to distinguish from the predominate columnar cells at the light-microscopic level. Alternatively, villus goblet cells could have fashion in which there was responded in a “holocrine” an acceleration of the normal exfoliation of entire goblet cells from villus tips. Because the villus epithelium is constantly being replaced by crypt cells, our finding that there are fewer goblet cells per square micrometer on villi of mock-stimulated rats than in the crypt region is
MUCIN
SECRETION
evidence that goblet cells must normally be exfoliated from the villus at a faster rate than columnar enterocytes. We also observed evidence for a large-scale nonselective exfoliation of villus epithelial cells in two rats at the 60min time point. Tissue from these two rats had been prepared on different days, and it is unlikely that an error in processing had caused an artifactual loss of cells. Except for the obvi .ous truncation of all villi, these tissues looked healthy and well preserved. It is unclear why only these two animals showed this response. There are approximately twice as many goblet cells in the crypt zone as on the overlying villi, but because the average villus goblet cell contains 2.5 times as much mucin as the typical crypt goblet cell, the villus region accounts for 55% of the total ileal mucin store. Although the decrease in intracellular stores of mucin in crypt cells at 5 min is 1.42 times the volume missing from villus cells at 60 min, these measurements cannot be corrected for any replenishment of granules during the stimulation period. It is therefore not possible to calculate the relative contribution of the crypt and villus stores of mucin to the final luminal mucous product. Nevertheless, the present study clearly demonstrates a significant contribution of villus-derived mucin to the luminal mucous blanket formed in response to cholinergic stimulation. The spatial localization of cells responsible for the ultimate secretory product is of importance because it has previously been shown that glycosylation of the mucin protein backbone varies along the crypt-villus axis (10). Previous observations that villus goblet cells did not respond to cholinergic stimulation had always seemed inconsistent with other lines of evidence that suggested mucin secretion from villi could be induced. It is known that villus cells, like crypt cells, continuously secrete mucin at a slow baseline rate (8) and that both villus and crypt cells can be totally depleted of mucin (without any evidence of compound exocytotic activity) when exposed to bacterial enterotoxins from Vibrio cholerae (1, 9) or Escherichia coli (1). Furthermore, radioreceptor binding assays have previously demonstrated that rat ileal villus and crypt cells have equal numbers of highaffinity cholinergic receptors (11). Quantitative analysis of the secretory response of individual goblet cells has proved to be a formidable task. Mucin is relatively insoluble and designed to adhere to mucosal surfaces. Biochemical assay of mucin discharge requires complete removal of previously discharged mutin before the assay period and quantitative recovery of mucin after stimulation. Biochemical assays can underestimate or overestimate the secretion of mucin if secreted mucins are not fully recovered or if vigorous washing causes an artifactual stimulation of mucin discharge (8). More importantly, biochemical assays are incapable of discerning differences in the secretory responses of the crypt and villus populations of goblet cells. For these reasons, our previous analyses of mucin secretion relied on the obvious and highly reproducible cavitation of the crypt goblet cell’s apical membrane, a phenomenon that allowed simple morphological assessment of mucin discharge. This led us to fail to detect the more subtle but quantitatively significant response of the villus goblet cells. In the present study, it has been
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
INTESTINAL
GOBLET
demonstrated that mucin secretion can be accelerated even in the absence of compound exocytotic cavitation but that the phenomenon can be detected by morphometric analysis. The drawback in this approach as in any morphological method is that it only allows assessment of intracellular stores at a single moment in time. If an agent accelerates mucin synthesis to the same degree as mucin discharge, morphometric analysis would be incapable of detecting secretion. In the present study, morphometric evaluation of the effect of a secretagogue at multiple time points minimized this pitfall. Alternatively, a balanced stimulation of synthesis and secretion might be morphologically detected using autoradiographic visualization of the transport and release of pulse-radiolabeled mucin glycoproteins, but this appreach also has limitations. Mucin granules located on the periphery of the goblet cell theta are radiolabeled and transported more rapidly than centrally located granules (2). If, as with crypt goblet cells (7), cholinergic stimulation were to preferentially accelerate the secretion of the more centrally located granules in villus cells, changes in the secretory rate might be missed or misinterpreted. Furthermore, variations in the granule transport rate in the unstimulated animal (2) would require large numbers of cells sectioned along the central axis to be analyzed if subtle changes in secretory rates were to be detected. Finally, the cost of pulse-radiolabeling mutin glycoproteins would limit the ability to routinely screen secretagogues in vivo. The present results therefore establish that morphometric analysis is the method of choice for quantitatively demonstrating the action of a mucous secretagogue on individual cells in the intestinal mucosa. The expert assistance of Jan Wilson is gratefully acknowledged. This research was supported by grants from the Cystic Fibrosis Foundation and the Crohn’s and Colitis Foundation of America. The computerized morphometric analysis system was purchased with the assistance of the University of Missouri Institutional Biomedical Re-
CELL
MUCIN
G331
SECRETION
search Support Grant RR-07053 from the National Institutes of Health. Address for reprint requests: T. E. Phillips, Div. of Biological Sciences, Tucker Hall, University of Missouri, Columbia, MO 65211. Received
12 April
1991; accepted
in final
form
17 September
1991.
REFERENCES
1.
Moon, H. W., S. C. Whipp, and A. L. Baetz. Comparative effects of enterotoxins from Escherichia coZi and Vibrio cholerae on rabbit and swine small intestine. Lab. Invest. 25: 133-140, 1975. M. R., R. J. Grand, and J. S. Trier. Glycoprotein 2* Neutra, synthesis, transport, and secretion by epithelial cells of human rectal mucosa. Lab. Invest. 36: 535-546, 1977. 3. Neutra, M. R., L. J. O’Malley, and R. D. Specian. Regulation of intestinal goblet cells. I. A survey of potential secretagogues. Am. J. Physiol. 242 (Gastrointest. Liver Physiol. 5): G380-G387, 1982. 4 Phillips, T. E., T. L. Phillips, and M. R. Neutra. Regulation * of intestinal goblet cell secretion. III. Isolated intestinal epithelium. Am. J. Physiol. 247 (Gastrointest. Liver Physiol. 10): G674-G681, 1984. Phillips, T. E., T. L. Phillips, and M. R. Neutra. Regulation 5* of intestinal goblet cell secretion. IV. Electrical field stimulation in vitro. Am. J. Physiol. 247 (Gastrointest. Liver Physiol. 10): G682G687, 1984. T. E., T. L. Phillips, and M. R. Neutra. Cholinergic 6. Phillips, responsiveness of goblet cells during intestinal maturation. Biol. Neonate 55: 197-203, 1989. 7 Specian, R. D., and M. R. Neutra. Mechanism of rapid mucus * secretion in goblet cells stimulated by acetylcholine. J. CeZZ Biol. 85: 626-640,198O. R. D., and M. R. Neutra. Regulation of intestinal $0 Specian, goblet cell secretion. I. Role of parasympathetic stimulation. Am. J. Physiol. 242 (Gastrointest. Liver Physiol. 5): G370-G379, 1982. 9. Steinberg, S. E., J. G. Banwell, J. H. Yardley, G. T. Keusch, and T. R. Hendrix. Comparison of secretory and histological effects of shigella and cholera enterotoxins in rabbit jejunum. Gastroenterology 68: 309-317, 1975. ho. Vecchi, M., G. Torgano, M. Monti, E. Berti, D. Agape, M. Primignani, G. Ronchi, and R. deFranchis. Evaluation of structural and secretory glycoconjugates in normal human jejunum by means of lectin histochemistry. Histochemistry 86: 359-364, 1987. 11. Wahawisan, R., L. J. Wallace, and T. S. Gaginella. Muscarinic receptors exist on ileal crypt and villus cells of the rat (Abstract). Federation Proc. 12: 761, 1983.
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
ileum; morphometry
of cholinergic stimulation, both crypt and villus intestinal goblet cells secrete mucin at a slow baseline rate using simple exocytosis. Cholinergic stimulation causes crypt goblet cells to rapidly accelerate their discharge of mucin by compound exocytosis (4-8). Compound exocytosis of mucin secretory granules is easily recognized in light-microscopic semithin sections or in transmission electron microscopy by the deep cavitation of the apical cell surface. Previous studies established that induction of exocytosis results from a direct action of cholinomimetics on the epithelium (4) and that the same effect could be obtained by the electrical stimulation of mucosal nerves (5). In contrast, neither cholinergic agonists (4, 6-8) nor nonspecific electrical stimulation of mucosal nerves (5) could elicit any sign of compound exocytotic activity by villus goblet cells. It was unknown, however, whether villus cells respond to stimulation by cholinomimetics or other agents by increasing their mucin discharge by a mechanism that does not result in the dramatic compound exocytotic cavitation of their apical membrane, such as acceleration of simple exocytosis or selective exfoliation of goblet cells. The purpose of this study was to address this possibility using computer-assisted morphometric analysis. IN THE ABSENCE
MATERIALS
AND
METHODS
Sprague-Dawley rats, 44-54 days old and weighing between 220 and 255 g, were used for this study. Animals were lightly anesthetized with methoxyflurane, weighed, and injected subcutaneously with 250 pg/kg carbachol (carbamylcholine, Sigma) (6) in phosphate-buffered saline (PBS) and stimulated for 5 (n = 3), 15 (n = 5), 60 (n = 5), or 240 (n = 5) min before 0193-1857/92
$2.00
Copyright
being killed by pentobarbital overdose. Mock-treated control rats (n = 5) were injected with PBS 5 min before being killed. All protocols were approved by the University of Missouri Animal Use Committee and followed National Institutes of Health guidelines. The abdominal cavity was opened with a midline laparotomy and a small slice of tissue was removed from the ileum (- 3 cm proximal to the ileocecal junction), immediately placed in cold PFG fixative [ 2.5% glutaraldehyde, 2% freshly depolymerized paraformaldehyde, 70 mM NaCl, 30 mM N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid (HEPES), 2 mM CaClz, pH 7.41, and cut into l- to &mm3 cubes. After 2 h of fixation in fresh PFG at 4”C, the tissues were rinsed in the HEPES buffer (70 mM NaCl + 30 mM HEPES, pH 7.4) and postfixed in 1% 0~0~ in HEPES buffer. After rinses in HEPES buffer and deionized water (dHzO), tissues were stained en bloc with 0.5% uranyl magnesium acetate overnight at 4OC, washed in dHzO, dehydrated in a series of ethanol solutions, and embedded in epoxy resin (EmBed 812, EMS, Fort Washington, PA). For computer-assisted morphometric analysis (American Innovision, San Diego, CA), a color image of a O.&pm toluidine blue-stained section was displayed on a monitor at ~1,364 magnification. A mouse-driven cursor was used to outline total crypt or villus epithelial surface area or the accumulations of mucin granules contained in individual goblet cells. All slides were randomized and coded for analysis by the investigator. Three randomly chosen blocks of tissue from each animal were examined; a slide with 5-10 sequential sections was produced from each tissue block. One randomly chosen section on each slide was used for analysis; within the section, all complete villi (i.e., not including those on margins of sections in which one side was not present) and all crypt epithelia directly beneath these villi were analyzed. Typically, a section would include three to four villi. A two-group unpaired Mann-Whitney U test was performed on each set on data using the StatView SE+ computer program (Abacus Concepts, Berkeley, CA). For photomicrography, 0.9-pm sections were stained with hematoxylin and 1% safranin 0 and photographed using a Wratten no. 44 filter with a Zeiss Axiophot microscope. RESULTS
Goblet cells in the mock-stimulated rat ileum contained a complete complement of secretory granules (Fig. I&. Mucin secretory granules occupied 5.5 t 0.3% (SE) of the epithelial area in ileal crypts of mock-stimulated rats. The average profile of secretory granules in the crypt zone was 21.1 t 1.2 pm’. The number of goblet cells per square micrometer, estimated by the number of granule profiles, averaged 2.65 x 10B3 & 1.58 X 10v4 per ,um2of crypt epithelium. As expected from previous studies (e.g., Ref. 6), goblet cells in ileal crypts showed a loss of secretory granules and a deep cavitation of their apical surface within 5 min of cholinergic stimulation (Fig. 1B). At 5 min poststimulation, the area occupied by secretory granules decreased to 41.5 t 2.6% (P 5 0.0001) of the control value (Fig. 2A). This decrease in crypt mucin stores resulted
0 1992 the American
Physiological
Society
G327
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
G328
INTESTINAL
GOBLET
CELL MUCIN
SECRETION
Fig. 1. Time course of ileal response to cholinergic stimulation. All micrographs are at same magnification (bar = 50 pm). A: in mock-stimulated ileum, all goblet cells in crypt (arrows) and on villi (arrowheads) appeared packed with mucin secretory granules. B: 5 min after cholinergic stimulation, many crypt goblet cells (small arrows) are depleted of secretory granules and their apical membranes showed deep cavitation indicative of recent compound exocytotic activity. Exfoliated goblet cells (large arrowheads) were occasionally observed trapped in extracellular mucous coat. Some crypt goblet cells (large arrows) and all villus goblet cells (small arrowheads) showed no sign of cavitation and appeared filled with granules. C: by 60 min poststimulation, crypt goblet cells (arrows) no longer showed apical membrane cavitation but intracellular mucin stores remained partially depleted. Villus goblet cells (arrowheads) continue to show no sign of cavitation. D: by 4 h poststimulation, both crypt (arrows) and villus (arrowheads) goblet cells were difficult to distinguish from mock-stimulated controls.
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
INTESTINAL
r
GOBLET
CELL
MUCIN
SECRETION
G329
r
5
15
60
240
C
5
15
60
240
D
5
15
60
240
60
240 minutes of cholinergic
5
15
60
240
60
240
Fig. 2. A and B: percentage of total epithelial area occupied by mucin secretory granules for crypt (A) and villus (B) regions expressed as percentage of control value and plotted against minutes of cholinergic stimulation. C and D: average accumulation of mucin secretory granules (total pm2 surface area occupied by secretory granules divided by number of profiles) expressed as percentage of control values and plotted against min of cholinergic stimulation for crypt (C) and villus (D) cells. E and F: no. of cells with mucin secretory granules per ,um2 of epithelial surface area expressed as percentage of control value in crypt (E) and villus (F) regions plotted against min of cholinergic stimulation. In all graphs, each bar represents average response of 5 rats each at 0, 15, 60, and 240 min and 3 rats at 5 min; error bars indicate SE of mean. P values determined using a 2-group unpaired Mann-Whitney U test are indicated above each bar.
stimulation
largely from an all-or-none secretory response, since the number of crypt cells containing recognizable secretory granules cells decreased to a similar 44.4 t 4.5% (P = 0.0002) of the control value at 5 min (Fig. 2E), whereas the average accumulation of mucin in cells still containing secretory granules did not change (99.0 t 7.6% of control value, P = 0.98) at this time (Fig. 2C). Cavitation was still prominent at 15 min, but by 60 min poststimulation cells with deeply cavitated apical membranes were observed less frequently (Fig. 1C). Shedding of “membrane tags” (7) and endocytotic retrieval apparently removed the excess granule membrane from the apical surface by 60 min. At 15 and 60 min, crypt mucin stores remained significantly depleted from the control level (P -< 0.0001) but were not significantly changed (P = 0.30) from the 5-min level (Fig. 2A). Although mucin stores had not increased significantly by 15 min, replenishment had begun, since the number of cells containing secretory granules had rebounded to 59.7 t 7.2% of the control value (P = 0.0007), which is significantly higher than the 5-min value (P = 0.0165). As would be predicted, these refilling cells caused a small drop in the size of the average granule accumulation (Fig. 2C). By 240 min poststimulation (Fig. lo), the crypt mucin stores had refilled to 89.7 t 8.2% of the control value (P = 0.40).
At 240 min, the number of granule profiles was 106.5 t 7.6% (P = 0.58) of the control value, but this value may not accurately reflect the number of goblet cells, since at this time point active synthesis resulted in distinct accumulations of mucin-filled granules in the Golgi region of some cells. Because Golgi accumulations of granules could not be discerned from true thecal accumulations except in fortuitously sectioned cells, all accumulations were counted. Because both the Golgi and thecal accumulations in some cells may have been individually counted, the number of granule profiles per square micrometer might overestimate the number of goblet cells at 240 min poststimulation. In the villus region of mock-stimulated rats, mucin secretory granules occupied 7.62 t 0.52% of the epithelial area. The number of goblet cells per square micrometer of control villi (1.40 X lo-” t 7.3 X lo-“) was 47.2% less than in the crypt zone, but the average profile of mucin granules was 2.57 times larger (54.1 t 2.6 pm’). Less than 1 of every 500 goblet cells on villi above cholinergically responsive crypts showed any sign of the apical membrane cavitation indicative of recent compound exocytotic activity (Fig. 1, B-D). Electron-microscopic examination also failed to reveal evidence of compound exocytosis from villus goblet cells. Morphometric analy-
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
G330
INTESTINAL
GOBLET
CELL
sis, however, revealed that villus mucin stores decreased to 77.6 t 10.2% of the control value (P = 0.089) by 5 min and to 67.5 t 5.1% (P = 0.002) by 60 min (Fig. ZB). The average profile of secretory granules was also decreased by cholinergic stimulation, with a maximal decrease occurring at 60 min (78.8 t 5.7% of control value, P = 0.015; Fig. 2D). Decreases in the number of granulecontaining cells occurred at 15 min (88.0 t 4.8% of control value, P = 0.085) and 60 min (88.7 t 6.5%, P = 0.120; Fig. 2F). Occasional examples of exfoliated intact goblet cells trapped in the luminal mucous blanket were observed (Fig. 1B). By 240 min poststimulation, villus mucin stores recovered to 93.8 t 8.1% (P = 0.237) of the control value and the average accumulation of granules was 81.4 t 3.0% of control (P = 0.009). At 240 min, the number of granule-containing cells was 115.0 t 7.7% (P = 0.19) of the mock-stimulated value, but as in the case of the crypt cells accumulations of granules in the Golgi region probably led to an overestimation of the number of goblet cells at this time point. In mock-stimulated rats, the ratio of villus epithelial area to the underlying crypt epithelial area was 0.97 t 0.02. Large-scale exfoliation of villus cells in response to stimulation would be expected to reduce this ratio. Only two rats, both at 60 min poststimulation, showed significant decreases from the control value (ratios: 0.47 t 0.02, P = 0.013; 0.63 t 0.04, P = 0.038). DISCUSSION
This study demonstrated that cholinergic stimulation affects mucin secretion from villus as well as crypt goblet cells. In the rat ileum, the villus epithelium underwent reproducible reductions in intracellular mucin stores that were not previously detected. Although crypt goblet cells showed the deep cavitation of their apical membranes indicative of recent compound exocytotic activity, only rare profiles of cavitation of villus goblet cells were ever observed. The cholinergically evoked decrease in the average accumulation of secretory granules in villus goblet cells presumably reflects an acceleration of simple exocytosis (i.e., fusion of single granules with the plasmalemmal membrane) but without the granule-granule fusion mechanism seen as compound exocytosis. This variation in the rate of simple exocytosis is consistent with earlier autoradiographic studies in which the rate of granule transport during unstimulated baseline release of mucin varied widely between individual cells (2). Accelerated simple exocytosis could also conceivably account for the decrease in the number of villus cells with secretory granules. Without the apical cavitation due to compound exocytosis, goblet cells depleted of mucin granules would have been difficult to distinguish from the predominate columnar cells at the light-microscopic level. Alternatively, villus goblet cells could have fashion in which there was responded in a “holocrine” an acceleration of the normal exfoliation of entire goblet cells from villus tips. Because the villus epithelium is constantly being replaced by crypt cells, our finding that there are fewer goblet cells per square micrometer on villi of mock-stimulated rats than in the crypt region is
MUCIN
SECRETION
evidence that goblet cells must normally be exfoliated from the villus at a faster rate than columnar enterocytes. We also observed evidence for a large-scale nonselective exfoliation of villus epithelial cells in two rats at the 60min time point. Tissue from these two rats had been prepared on different days, and it is unlikely that an error in processing had caused an artifactual loss of cells. Except for the obvi .ous truncation of all villi, these tissues looked healthy and well preserved. It is unclear why only these two animals showed this response. There are approximately twice as many goblet cells in the crypt zone as on the overlying villi, but because the average villus goblet cell contains 2.5 times as much mucin as the typical crypt goblet cell, the villus region accounts for 55% of the total ileal mucin store. Although the decrease in intracellular stores of mucin in crypt cells at 5 min is 1.42 times the volume missing from villus cells at 60 min, these measurements cannot be corrected for any replenishment of granules during the stimulation period. It is therefore not possible to calculate the relative contribution of the crypt and villus stores of mucin to the final luminal mucous product. Nevertheless, the present study clearly demonstrates a significant contribution of villus-derived mucin to the luminal mucous blanket formed in response to cholinergic stimulation. The spatial localization of cells responsible for the ultimate secretory product is of importance because it has previously been shown that glycosylation of the mucin protein backbone varies along the crypt-villus axis (10). Previous observations that villus goblet cells did not respond to cholinergic stimulation had always seemed inconsistent with other lines of evidence that suggested mucin secretion from villi could be induced. It is known that villus cells, like crypt cells, continuously secrete mucin at a slow baseline rate (8) and that both villus and crypt cells can be totally depleted of mucin (without any evidence of compound exocytotic activity) when exposed to bacterial enterotoxins from Vibrio cholerae (1, 9) or Escherichia coli (1). Furthermore, radioreceptor binding assays have previously demonstrated that rat ileal villus and crypt cells have equal numbers of highaffinity cholinergic receptors (11). Quantitative analysis of the secretory response of individual goblet cells has proved to be a formidable task. Mucin is relatively insoluble and designed to adhere to mucosal surfaces. Biochemical assay of mucin discharge requires complete removal of previously discharged mutin before the assay period and quantitative recovery of mucin after stimulation. Biochemical assays can underestimate or overestimate the secretion of mucin if secreted mucins are not fully recovered or if vigorous washing causes an artifactual stimulation of mucin discharge (8). More importantly, biochemical assays are incapable of discerning differences in the secretory responses of the crypt and villus populations of goblet cells. For these reasons, our previous analyses of mucin secretion relied on the obvious and highly reproducible cavitation of the crypt goblet cell’s apical membrane, a phenomenon that allowed simple morphological assessment of mucin discharge. This led us to fail to detect the more subtle but quantitatively significant response of the villus goblet cells. In the present study, it has been
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
INTESTINAL
GOBLET
demonstrated that mucin secretion can be accelerated even in the absence of compound exocytotic cavitation but that the phenomenon can be detected by morphometric analysis. The drawback in this approach as in any morphological method is that it only allows assessment of intracellular stores at a single moment in time. If an agent accelerates mucin synthesis to the same degree as mucin discharge, morphometric analysis would be incapable of detecting secretion. In the present study, morphometric evaluation of the effect of a secretagogue at multiple time points minimized this pitfall. Alternatively, a balanced stimulation of synthesis and secretion might be morphologically detected using autoradiographic visualization of the transport and release of pulse-radiolabeled mucin glycoproteins, but this appreach also has limitations. Mucin granules located on the periphery of the goblet cell theta are radiolabeled and transported more rapidly than centrally located granules (2). If, as with crypt goblet cells (7), cholinergic stimulation were to preferentially accelerate the secretion of the more centrally located granules in villus cells, changes in the secretory rate might be missed or misinterpreted. Furthermore, variations in the granule transport rate in the unstimulated animal (2) would require large numbers of cells sectioned along the central axis to be analyzed if subtle changes in secretory rates were to be detected. Finally, the cost of pulse-radiolabeling mutin glycoproteins would limit the ability to routinely screen secretagogues in vivo. The present results therefore establish that morphometric analysis is the method of choice for quantitatively demonstrating the action of a mucous secretagogue on individual cells in the intestinal mucosa. The expert assistance of Jan Wilson is gratefully acknowledged. This research was supported by grants from the Cystic Fibrosis Foundation and the Crohn’s and Colitis Foundation of America. The computerized morphometric analysis system was purchased with the assistance of the University of Missouri Institutional Biomedical Re-
CELL
MUCIN
G331
SECRETION
search Support Grant RR-07053 from the National Institutes of Health. Address for reprint requests: T. E. Phillips, Div. of Biological Sciences, Tucker Hall, University of Missouri, Columbia, MO 65211. Received
12 April
1991; accepted
in final
form
17 September
1991.
REFERENCES
1.
Moon, H. W., S. C. Whipp, and A. L. Baetz. Comparative effects of enterotoxins from Escherichia coZi and Vibrio cholerae on rabbit and swine small intestine. Lab. Invest. 25: 133-140, 1975. M. R., R. J. Grand, and J. S. Trier. Glycoprotein 2* Neutra, synthesis, transport, and secretion by epithelial cells of human rectal mucosa. Lab. Invest. 36: 535-546, 1977. 3. Neutra, M. R., L. J. O’Malley, and R. D. Specian. Regulation of intestinal goblet cells. I. A survey of potential secretagogues. Am. J. Physiol. 242 (Gastrointest. Liver Physiol. 5): G380-G387, 1982. 4 Phillips, T. E., T. L. Phillips, and M. R. Neutra. Regulation * of intestinal goblet cell secretion. III. Isolated intestinal epithelium. Am. J. Physiol. 247 (Gastrointest. Liver Physiol. 10): G674-G681, 1984. Phillips, T. E., T. L. Phillips, and M. R. Neutra. Regulation 5* of intestinal goblet cell secretion. IV. Electrical field stimulation in vitro. Am. J. Physiol. 247 (Gastrointest. Liver Physiol. 10): G682G687, 1984. T. E., T. L. Phillips, and M. R. Neutra. Cholinergic 6. Phillips, responsiveness of goblet cells during intestinal maturation. Biol. Neonate 55: 197-203, 1989. 7 Specian, R. D., and M. R. Neutra. Mechanism of rapid mucus * secretion in goblet cells stimulated by acetylcholine. J. CeZZ Biol. 85: 626-640,198O. R. D., and M. R. Neutra. Regulation of intestinal $0 Specian, goblet cell secretion. I. Role of parasympathetic stimulation. Am. J. Physiol. 242 (Gastrointest. Liver Physiol. 5): G370-G379, 1982. 9. Steinberg, S. E., J. G. Banwell, J. H. Yardley, G. T. Keusch, and T. R. Hendrix. Comparison of secretory and histological effects of shigella and cholera enterotoxins in rabbit jejunum. Gastroenterology 68: 309-317, 1975. ho. Vecchi, M., G. Torgano, M. Monti, E. Berti, D. Agape, M. Primignani, G. Ronchi, and R. deFranchis. Evaluation of structural and secretory glycoconjugates in normal human jejunum by means of lectin histochemistry. Histochemistry 86: 359-364, 1987. 11. Wahawisan, R., L. J. Wallace, and T. S. Gaginella. Muscarinic receptors exist on ileal crypt and villus cells of the rat (Abstract). Federation Proc. 12: 761, 1983.
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 10, 2019.
