INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES APPENDIX

THE COLONIC EPITHELIUM of mammals secretes fluid and mucus in response to various neurotransmitters and local mediators (9, 12, 13). Fluid secretion modifies luminal composition by increasing the aqueous volume and by adjusting concentrations of electrolytes such as K⁺ and H⁺. Low rates of fluid secretion could aid colonic motility and bacterial fermentation, whereas higher rates seen in pathophysiological states would serve to flush bacteria from the lumen. Mucus secretions contribute to the barrier function of the epithelium by protecting the epithelial cells from abrasion and also by slowing access of bacteria to those epithelial cells. Release of mucus from goblet cells generally is separate from sustained fluid secretion on the basis of sensitivity to specific secretagogues (35). A mucous substance distinct from goblet cell mucin is released from crypt columnar cells during stimulation of fluid secretion (14, 16). Control of secretion from goblet and columnar cells by separate secretagogues permits alteration of the relative proportion of mucus types released and the rate of fluid production.

Mucus secretion by goblet cells involves exocytotic release of granule contents at the apical membrane, as indicated by ultrastructural studies (51). Movement of granules within the cell is a microtubule-dependent process, and access of the granules to the apical membrane is restricted by the actin cytoskeleton. Further advances in examining dynamics of goblet cell secretion have been made using video-enhanced differential interference contrast (DIC) microscopy of living cells (4, 14, 20, 46, 52). Individual events of goblet granule release were visualized as changes in light intensity, indicating a rapid discharge (0.03-1.0 s) of mucus from granules (4, 20, 46). The time course of stimulation with mucus secretagogues consists of a rapid increase in the number of granule release events that reaches a maximum in 1-3 min and then subsides to a much slower sustained rate. Discharge of mucus from goblet cells in rabbit colonic crypts occurs via a similar exocytotic process (14, 52), and columnar cells also show stimulated emptying of contents from the apical pole of the cell (14). The results presented in this study were obtained with DIC microscopy of isolated human colonic crypts and indicate the presence of two morphologically distinguishable types of cells: goblet and columnar cells. Distinct secretagogues stimulated release of mucuslike material from apically stored granules in each of these cell types.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES APPENDIX

Specimens of human colon were obtained from surgical resection material by a procedure approved by the Institutional Review Board (Ohio State University and Wright State University). After release by the pathologist, specimens were refrigerated in HEPES-buffered Ringer solution until they were picked up and transported to the laboratory (generally ~30 min). The epithelium was removed from underlying muscle by blunt dissection and placed in ice-cold standard Ringer solution. Colonic mucosal biopsies from cotton-top tamarins were obtained during semiannual endoscopic screening of a colony maintained by Dr. J. D. Wood at Ohio State University. This procedure was approved by the Institutional Laboratory Animal Care and Use Committee at Ohio State University. The tamarin biopsies were transported to the laboratory in ice-cold HEPES-buffered Ringer solution within 20 min.

Human colonic crypts were isolated from resection material of 17 patients (9 women and 8 men, 15 Caucasians and 2 African-Americans; Ohio State University Hospitals) ranging in age from 35 to 69 yr. Resections were roughly divided between ascending or transverse sites (~55%) and descending or sigmoid sites (~45%); normal margins were released by the pathologist for experimental study. Resection had been performed for cancer (12 of 17 patients), diverticulitis (2 of 17 patients), and inflammatory bowel disease (IBD, 3 of 17 patients). All three patients with IBD were women: one had active ulcerative colitis, one had nonactive colitis (resected for metastatic cancer), and one had Crohn's disease. Information on patient medications was not available for comparison with experimental results. Cotton-top tamarins are a species of New World monkey that develops symptoms resembling ulcerative colitis when in captivity (30). The tamarins supplying colonic crypts for imaging (n = 4) were assessed for extent of disease by histological measures, including white blood cell infiltration: three had severe colitis, and one had moderate colitis (personal communication, K. S. Tefend and J. D. Wood). Differences among the human specimens were not readily discernible, except for IBD tissue, which could be distinguished by using objective measures presented in RESULTS. On the basis of this general similarity and the pathologist's assessment of normal tissue margins, tissues from patients with cancer or diverticulitis were grouped together as normal.

Individual colonic crypts were dissected from the colonic epithelium with use of fine forceps on the stage of a dissecting microscope (×40) in HEPES-buffered Ringer solution with 5% serum albumin added. The standard mammalian Ringer solution contained (in mM) 145 Na⁺, 5 K⁺, 2 Ca²⁺, 1.2 Mg²⁺, 125 Cl, 25 HCO₃, 4 PO^x₄, and 10 D-glucose. Solutions were continually gassed with 95% O₂-5% CO₂, which maintained solution pH at 7.4. HEPES-buffered Ringer solution contained 10 mM HEPES with HCO₃ replaced by Cl, and the pH was titrated to 7.4.

Crypts were imaged in a chamber mounted on the stage of an inverted microscope (Zeiss Axiovert); the bottom of the chamber was formed by a no. 1 coverslip (14). Isolated crypts were held by glass pipettes or adhered by a coat of polylysine to the chamber bottom coverslip. The glass pipettes were made to accommodate the relatively short length of colonic crypts and were manipulated with a system from Vestavia Scientific (Vestavia Hills, AL). For luminal perfusion, the blind end of the crypt was cut off with a sharpened needle before transfer to the imaging chamber. Luminal perfusate was HEPES-buffered Ringer solution. Luminal perfusion rate was estimated as 1-5 nl/min on the basis of perfusion pressure, pipette dimensions, and previous measurements with renal tubules (45). The bath was continuously perfused with standard Ringer solution with use of a peristaltic pump, and effluent was removed by suction tube; the solution was bubbled with 95% O₂-5% CO₂ and warmed by passage through a heated water jacket to maintain bath temperature at 37°C. Bath solution flow was 3 ml/min (~8 chamber vol/min).

Crypt images were formed with DIC optics: a ×40 oil immersion (1.4 NA) lens and a nonimmersion (0.65 NA) condenser (6). A video camera (model CCD300E, Videoscope International) was used to record images to videotape (Sony) and to computer disk with use of the Image-1 system (Universal Imaging). A ×2 coupling lens (Diagnostic Instruments) was used to fill the camera field.

Image analysis. Morphometric measurements of recorded crypt images were performed with Image-1 software. Quantification of volume was obtained with focus adjusted to the midline of the crypt, because this plane of section allowed ready measurement of crypt and lumen diameter as well as a complete profile of the epithelium from base to apex. Crypt diameter was taken at the base of epithelial cells on opposite sides of the crypt (inside the pericryptal sheath); lumen diameter was taken at the apexes of opposing cells. A consistent group of cells was monitored in successive images by marking the boundaries between distinct, identified cells near the lateral margins of the image frame. Diameters were measured at nine evenly spaced points along the crypt segment defined by the length markers. Segment length was determined at cell apex, cell base, and midway along cell height for both epithelial margins (6 values). Crypt (V_cr) and luminal volumes (V_lu) (23) were calculated on the basis of a piecewise cylindrical model with use of average crypt diameter (D), average lumen diameter (d), average section length (L), and standard deviations (_D and _d) of these values (Eq. 1)

(1a)

(1b)

This calculation assumes that cells seen in the plane of focus are representative of all cells in that annulus of the crypt. Epithelial volume was calculated as the difference between crypt and luminal volume. All volumes were normalized to epithelial volume obtained in the initial control image for each experiment to adjust for differences in defined segment lengths (cell number) between crypts. Generally, 300-400 crypt epithelial cells were represented by epithelial volume in each experiment. Repeated measures on the same image indicated that reproducibility was ±0.1 µm for diameter measurements and ±0.5 µm for length measurements. Propagation of these error estimates for volume calculations (Eq. 1) gave an error estimate for relative epithelial volume of ±0.6%.

(2)

(3a)

(3b)

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES APPENDIX

Isolated colonic crypts imaged with DIC microscopy (Fig. 1) exhibited the two predominant cell types associated with this epithelium: columnar and goblet cells (3, 14, 21, 40, 48). Focusing at the midline of the crypt showed the full height of the epithelium, with apically located granules above basally located nuclei. Surrounding the crypt epithelium was a pericryptal sheath of myoepithelial cells (36, 44), which was apparent as a fibrous layer with flattened nuclei. Goblet cells, in particular, were distinct because of a densely packed cluster of highly refractile granules stored in the apical pole, which had a characteristic ovate profile. Generally, goblet cells were separated by columnar cells having a more hourglass-shaped profile that arched over the neighboring goblet cells. Thus the luminal surface often was dominated by columnar apexes, even though goblet granules filled more than one-half of the apical epithelial volume.

View larger version (137K): [in this window] [in a new window]   Fig. 1.   Isolated colonic crypts. Differential interference contrast (DIC) microscopy was used to image individual dissected colonic crypts. Focus at midline of crypt tube provided a view of maximal crypt and lumen diameters. A full epithelial profile from apex to base can be seen. Pericryptal sheath is apparent at basal margin of crypt epithelium, with oval nuclear profiles scattered along crypt (a few are marked by *). Nuclei are apparent in basal pole of epithelial cells (a few are marked by arrows), and goblet granule clusters in apical pole (a few are marked by arrowheads). Columnar cell apexes generally appear as triangular regions between goblet granule clusters. Crypt orifices are to left. Scale bars, 20 µm. A: central portion of a normal human crypt shows goblet and columnar cells. B: crypt from a patient with active ulcerative colitis had a dilated lumen and an outside diameter (~110 µm) larger than most normal crypts. Epithelial height varied considerably. Goblet granule clusters appeared narrower and shorter than typical in normal crypts. C: crypt from a patient with nonactive colitis appeared similar to normal crypts, except lumen was relatively dilated and epithelial height varied. D: crypt from a patient with Crohn's colitis appeared normal. E: crypt from a tamarin with severe colitis had goblet granule clusters that were smaller than in tamarin with moderate colitis or normal human crypts. Epithelial height varied considerably.

Crypts also were obtained from colonic epithelium of patients with IBD (ulcerative colitis and Crohn's disease; Fig. 1). During dissection ulcerated regions of colon with active colitis had a fragile surface epithelium and a sparse number of recognizable crypts (4 specimens with active colitis were examined by dissection microscope; data not shown). Those crypts present (Fig. 1B) were generally large in diameter (>100 µm). Goblet cells appeared less full of mucous granules than did normal crypts. Colonic epithelium from patients with nonactive ulcerative colitis (Fig. 1C) or Crohn's disease (Fig. 1D) was indistinguishable on dissection from normal epithelium in crypt density or crypt size. Crypts from all IBD patients had more focal spots of short epithelial height than did normal crypts. Colonic epithelium of tamarins had densely packed crypts, even in specimens with severe colitis. As for crypts from patients with IBD, the tamarin epithelium had focal spots of short cell height (Fig. 1E) and relatively depleted granule stores in goblet cells. Results from tamarin crypts are specifically indicated; all other results are for human crypts.

Crypt dimensions were measured and assigned to three groups (Table 1): normal (see METHODS), IBD, and tamarin. A distinction also was made for crypts from the normal group that were opened for luminal perfusion. Differences in size among these groups were not apparent statistically. The focal spots of short epithelial height seen in crypts from patients with IBD and in tamarin crypts were not extensive enough to cause a decrease in average epithelial height; however, the lumen diameter of perfused crypts and crypts from patients with ulcerative colitis had a tendency to be larger.

Secretory stimulation. Addition of secretagogues that stimulate mucus and fluid secretion produced several changes in crypt geometry, as well as epithelial cell morphology. The cholinergic agent carbachol (CCh) stimulated mucus release from goblet cells, seen as shrinking of apical granule clusters and as light flashes, presumably the result of refractile contents exiting individual granules (4, 20, 46). In some cases, evanescent plumes could be seen forming above the sites of flashes, which then drifted away from the cell surface. Flashes ceased within a few seconds of CCh removal. Concurrent with goblet granule release, crypt lumens dilated (Fig. 2). Histamine also stimulated mucus release from goblet cells, as indicated by numerous apical flashes. Stimulation with prostaglandin E₂ (PGE₂) or adenosine produced fluid secretion and recession of columnar cell apexes without any apparent release from goblet cells (Fig. 3). Occasional light flashes were seen at columnar cell apexes, but these events were not consistently observed. Fluid flow toward the crypt orifice was visualized by the rapid movement of refractile particles along the lumen. The cholinergic response of tamarin crypts was dramatic, with nearly complete disgorging of mucous granules and large dilation of the lumen (Fig. 4). Large plumes of mucus stayed tethered to the cell of origin (Fig. 4C). Subsequent stimulation with adenosine (Fig. 4D) further shortened the epithelium and produced rapid fluid flow along the lumen that eventually ripped mucous plumes from goblet cells.

View larger version (140K): [in this window] [in a new window]   Fig. 2.   Cholinergic stimulation. Goblet granule clusters (arrowheads) responded to a cholinergic agonist. Individual granules are discernible within some of these clusters. Crypt orifice is to right. Scale bars, 10 µm. A: midline DIC image from central portion of a normal crypt in control condition. B: crypt in A 5 min after beginning of cholinergic stimulation [100 µM carbachol (CCh)]. Lumen diameter increased concurrent with loss of apically stored material from goblet cells. Release of mucin granules produced craters in some goblet cells.

View larger version (132K): [in this window] [in a new window]   Fig. 3.   Prostaglandin E2 (PGE2) stimulation. Response of columnar cells (arrowheads) to a fluid secretagogue. Crypt orifice is to left. Scale bars, 10 µm. A: midline DIC image from central portion of a normal crypt in control condition (after return from stimulation with 2 µM CCh). A large globule of goblet mucus (~10 µm diameter) can be seen in lumen (under 2nd downward arrowhead from left), remaining from earlier release event. B: crypt in A 5 min after beginning of stimulation with PGE2 (2 µM). Lumen diameter increased through a recession of columnar apexes without any apparent change in goblet cells.

View larger version (183K): [in this window] [in a new window]   Fig. 4.   Secretory response of tamarin crypt. Hypersensitivity of goblet cells to cholinergic stimulation was apparent in a crypt from a tamarin with moderate colitis. Crypt orifice is to left. Scale bars, 10 µm. A: control goblet cells had apical poles with moderately large granule clusters (arrowheads). Individual granules were apparent within goblet clusters. B: during stimulation (5 min) with CCh (100 µM), almost all goblet clusters of granules were extruded into lumen. Individual granules were still apparent within extruded luminal mucous plumes. C: after removal of CCh, further release of granule contents was not apparent, and lumen remained full of goblet mucus. Lysis of extruded granules became evident (10 min) from a uniform evanescent appearance of mucous plumes above goblet cells. (Top center and bottom right plumes in B could not be clearly identified.) D: subsequent stimulation of fluid secretion by adenosine (10 µM) for 10 min pushed mucous plumes down lumen toward crypt orifice, at left ~100 µm. (Only top right and bottom center plumes in B could still be identified.)

View larger version (60K): [in this window] [in a new window]   Fig. 5.   Crypt dimensions. Lumen and crypt diameters of individual crypts are shown. Dashed line, lumen diameter expected for case of epithelial cells with zero height; dotted line, lumen diameter expected for case of constant epithelial volume (with assumption of constant cell width), fit to mean of normal control group. A: crypt dimensions before any stimulation. , Normal; , normal-perfused; , ulcerative colitis (N, nonactive); , Crohn's disease; , tamarin. B: changes in diameter induced by CCh (10-min stimulation). Open symbols, control; filled symbols, stimulated; circles, normal; squares, normal-perfused; triangles, ulcerative colitis; inverted triangles, tamarin. Some control points are return to control (~10 min) from another secretagogue. C: changes in diameter induced (10-min stimulation) by a fluid secretagogue, PGE2 or adenosine (A); symbols as in B. D: diameters after return to control (10-15 min) from sequential stimulation by CCh and a fluid secretagogue. Symbols as in A.

Secretagogue-induced changes in epithelial volume. Epithelial volume of crypts was monitored during secretagogue stimulation by picking a consistent segment of crypt length over which to measure crypt diameter and lumen diameter. Segment length was defined from the boundaries between identifiable cells positioned near the edges of the recorded image fields. A representative experiment from the central portion of a normal crypt is shown in Fig. 6. Volumes (Fig. 6B) were calculated from the measured crypt dimensions (Fig. 6A; see METHODS). Epithelial volume decreased with CCh addition and did not return in the subsequent control period. With addition of PGE₂, a decrease in epithelial volume that was not recovered on return to control conditions also occurred.

View larger version (33K): [in this window] [in a new window]   Fig. 6.   Epithelial volume changes. A: a representative time course of dimensions from central portion of a normal crypt (~300 cells) during stimulation by CCh (100 µM) and PGE2 (5 µM). Open symbols, equal to prior measured value, indicate addition of CCh and PGE2. Segment length represents distance between distinct, identified cell boundaries positioned at either margin of recorded image field. Epithelial height (hep) was calculated from crypt diameter (D) and lumen diameter (d): hep = (D  d)/2. Error bars, SE. B: crypt and lumen volumes calculated using crypt and lumen diameters together with segment lengths (see METHODS). Epithelial volume was calculated as difference between crypt and lumen volumes.

View larger version (80K): [in this window] [in a new window]   Fig. 7.   Epithelial volume loss. Relative epithelial volume changes were normalized to control value preceding stimulation. , Normal; , normal-perfused; , ulcerative colitis (N, nonactive); , Crohn's disease; , tamarin. Last point generally was return to control (*). A: CCh (maximal stimulation, 2-100 µM). B: histamine (10 µM). C: PGE2 (2-5 µM). D: adenosine (100 µM).

View larger version (65K): [in this window] [in a new window]   Fig. 8.   Relative volume loss of apical granule storage zone. Time courses in Fig. 7 were normalized to apical zone volume fraction (~0.35) for each crypt. , Normal; , normal-perfused; , ulcerative colitis (N, nonactive); , Crohn's disease; , tamarin. Experiments without clearly sustained volume decreases in Fig. 7 were not included. * Return to control. A: CCh. B: PGE2. C: adenosine.

View larger version (15K): [in this window] [in a new window]   Fig. 9.   Depletion of apically stored materials. Proportion of epithelial volume lost after sequential stimulation by CCh and a fluid secretagogue (PGE2 or adenosine) is shown in relation to measured initial contribution of apical zone to total epithelial volume. , Normal; , normal-perfused; , ulcerative colitis (N, nonactive); , Crohn's disease; , tamarin.

Luminal fluid flow. Fluid flow along the lumen was apparent from movement of refractile particles. Sustained movement of these objects in the lumen was seen only during secretagogue stimulation. Presumably, the objects were made up of released mucus and other cellular debris. In some crypts, particles were small (~3 µm diameter) and seen infrequently, whereas in other crypts the lumen was crowded with globular objects. Stimulation by CCh of a crypt with globular objects already present in the initial control condition (Fig. 10A) produced a measurable particle flow toward the crypt orifice. Removal of the stimulus stopped flow with a lag of ~1 min, although the flashes indicative of granule release stopped within a few seconds. PGE₂ addition immediately produced flow toward the crypt orifice. Movement was episodic, with particles slowing down as larger globular objects squeezed past each other. Dilation of the lumen alleviated this congestion and progressively led to faster particle speeds. Addition of CCh together with PGE₂ led to reduced speed, consistent with inhibition of fluid secretion. Particle speeds were highest for smaller objects (Fig. 10B). All moving particles were tracked during stimulation, but blurred streaks expected for very fast particles were not observed. In crypts with relatively unobstructed lumens, PGE₂ produced particle speeds of ~10 µm/s throughout the period of stimulation, whereas CCh stimulation resulted in particle speeds comparable to PGE₂ for only 2-3 min, after which movement slowed and became undetectable.

View larger version (25K): [in this window] [in a new window]   Fig. 10.   Luminal fluid flow. Speeds of refractile particles moving in lumen of a normal crypt (shown in Fig. 6) measured during stimulation with secretagogues. A: time course of particle speeds. Particle size was measured to compare cross-sectional area with luminal cross-sectional area. , Particles with area <5% that of lumen; , particles with area >20% that of lumen; , particles of intermediate cross section. Large particle monitored at ~25 min squeezed past several other objects at a relatively high speed, propelled by full force of stream. B: variation of particle speed with particle volume during PGE2 stimulation. One large particle occluded lumen during speed measurement (*).

Stimulated release of cellular contents. Loss of epithelial volume during stimulation by goblet- and fluid-type secretagogues appeared to be restricted to specific cell types. Goblet cells were the site of cholinergic volume loss. Release of mucin from apical granules began at the luminal margin, occasionally progressing to make deep craters within the apex of goblet cells, presumably by fusion of individual pits (Fig. 11). Columnar cell apexes continued to arch over goblet cells during CCh stimulation, suggesting a lack of volume release from this cell type. An en face view of the epithelium (Fig. 12) shows the craters centered within the goblets and the spatial arrangement of goblet and columnar cells within the epithelium. Craters generally did not persist, with goblet granule clusters instead rearranging to form smaller spheroids.

View larger version (143K): [in this window] [in a new window]   Fig. 11.   Cellular response to cholinergic stimulation. Addition of CCh stimulated progressive release of mucin granules, producing apical craters in goblet cells (arrowheads), without evidence of release from columnar cells (bars with *). Scale bars, 5 µm. Three pairs of images are shown for control and stimulated conditions (A1 and B1, A2 and B2, and A3 and B3). A: control goblet cells had broad clusters of granules. Columnar cells arched over goblet cells, such that only goblet apexes were in contact with lumen. B: during CCh stimulation (100 µM for ~10 min), pits formed in apexes of goblets, with some becoming deep invaginations into apical granule clusters. C: goblet during CCh stimulation (~10 min) shows an apical crater connected to lumen (apical margin is along lower edge of goblet, with lumen below). Individual ~0.9-µm-diameter granules (arrowhead) also are visible. Scale bar, 2 µm.

View larger version (127K): [in this window] [in a new window]   Fig. 12.   Goblet response to CCh. A deeper focal plane (below midline), along crypt center line, provides an optical section through lower apical region of crypt cells. This en face image plane shows goblet cells as circular cross sections (arrowheads) and columnar cells as polygonal shapes filling intervening spaces. Focus was ~16 µm above base of cells along center line, which was determined from crypt width (w) in these images compared with crypt diameter (Dcr) and with assumption of a cylindrical crypt shape: h = (Dcr/2) {1  [1  (w/Dcr)2]1/2}. Longitudinal axis of crypt runs from bottom to top of these images. Scale bars, 5 µm. A: control condition. B: during CCh response (~5 min), goblet granule clusters showed rearrangement and pitting. Deformation of goblet cluster at top of field is readily apparent from convex edges of cell.

View larger version (180K): [in this window] [in a new window]   Fig. 13.   Cholinergic response in ulcerative colitis. CCh addition stimulated mucus release in a crypt from a patient with active ulcerative colitis (Fig. 1B). Scale bars, 5 µm. A: control goblet cells had relatively depleted stores of granules. This crypt had been stimulated previously by 5 µM PGE2. B: during CCh stimulation (100 µM for ~10 min) mucus was released and remained adherent to epithelium.

View larger version (165K): [in this window] [in a new window]   Fig. 14.   Cellular response to PGE2 stimulation. PGE2 addition produced recession of columnar cell apical margins (bars with *), without evidence of release from goblet cells (arrowheads). Scale bars, 5 µm. Two pairs of images are shown for control and stimulated conditions (A1 and B1, and A2 and B2). A: control columnar cells had narrow profiles that fanned out near luminal margin. Lack of sharp contrast at lumen edge suggests a lower density of cell contents than in neighboring goblet granules. B: during PGE2 stimulation (5 µM for ~10 min), luminal margins between goblet cells (presumptive columnar cells) receded.

View larger version (14K): [in this window] [in a new window]   Fig. 15.   Size and shape of apical goblets. Goblet granule clusters were measured for individual goblet cells in control conditions. Volume and eccentricity (mean ± SE) were calculated for each crypt (number of goblet cells measured ranged from 9 to 23/crypt). , Normal; , normal-perfused; , ulcerative colitis (N, nonactive); , Crohn's disease; , tamarin. Positive eccentricities indicate a prolate shape; negative values indicate an oblate shape (see METHODS).

View larger version (25K): [in this window] [in a new window]   Fig. 16.   Volume loss from apical goblets. Volume released from goblet granule clusters is shown in relation to volume stored before stimulation (mean ± SE). Dashed line, total release. A: volumes released with CCh stimulation [100 or 2 µM (2) or 10 µM (10)]. , Normal; , normal-perfused; , ulcerative colitis (N, nonactive); , Crohn's disease; , tamarin. Dotted line, 25% release. Some crypts had been stimulated by histamine before stimulation by CCh (*). B: volumes released with histamine stimulation (10 µM). Symbols as in A. Dotted line, 30% release.

View larger version (15K): [in this window] [in a new window]   Fig. 17.   Comparison of volume release measurements. Average relative volume change of individual goblet granule clusters (Vgob) in crypts produced by CCh stimulation (mean ± SE) is shown in relation to relative volume change of apical goblet zone [goblet fraction of epithelial volume (Vep)]. , Normal; , normal-perfused; , ulcerative colitis (N, nonactive); , tamarin. Dashed line, line of identity. Ulcerative colitis crypts are significantly off line of identity; all other values are not significantly different from line of identity.

Responsiveness of individual goblet cells. The distribution of CCh-induced fractional goblet volume release for individual cells is shown in Fig. 18A. Two distinct peaks occurred: one near zero volume release and one at higher release. This bimodal distribution is consistent with a nonresponding group of goblet cells dispersed among responding goblet cells. Comparison of the fraction of goblet cells responding within a crypt to the average initial goblet volume (Fig. 18B) indicated a greater proportion of nonresponding cells in crypts with larger goblet clusters. A similar relation for histamine stimulation (Fig. 18C) suggested that the histamine response was not related to goblet size. In those crypts exposed to both CCh and histamine (3 crypts, 27 goblets), 43% of goblet cells responded (defined as in Fig. 18) to both agents, but the CCh-nonresponsive cells responded to histamine and the histamine-nonresponsive cells responded to CCh. Apparently all these goblet cells were capable of stimulated release of granule contents.

View larger version (77K): [in this window] [in a new window]   Fig. 18.   Responsiveness of apical goblets. A: relative frequencies of CCh-induced fractional goblet volume release for all goblet cells in normal crypts (n = 74). Fitted curve is sum of 2 Gaussian distributions: peak at small fractional release contains 23% of goblets with a mean of 0.04 ± 0.01 and peak at higher fractional release contains 77% of goblets with a mean of 0.29 ± 0.02. B: fraction of CCh-responsive goblet cells for each crypt relative to average control goblet volume (mean ± SE). , Normal; , normal-perfused; , ulcerative colitis (N, nonactive); , Crohn's disease; , tamarin. Responding goblet cells were defined as those with a volume release fraction 0.20. C: fraction of histamine-responsive goblet cells for each crypt relative to average control goblet volume (mean ± SE). Symbols as in B. Responding goblet cells were defined as those with a fractional goblet volume release 0.20.

View larger version (48K): [in this window] [in a new window]   Fig. 19.   Stimulation of goblet cell and columnar cell release. A: time course of changes in goblet volume and eccentricity induced by adenosine (100 µM), histamine (10 µM), and CCh (100 µM) for responding goblet cells (defined as in Fig. 18) within a single crypt (mean ± SE, n = 10). Goblet volume was normalized to initial control level. Positive eccentricities indicate a prolate shape, and negative values indicate an oblate shape. Open symbols indicate addition of agonist. B: relative goblet volume (mean ± SE) for responding goblet cells during CCh stimulation of normal crypts (; 6 crypts, 41 goblets). Histamine stimulation was averaged from crypts in Fig. 16B (; 5 crypts, 32 goblets). Average CCh response of goblet cells from tamarin crypts [; 2 crypts (1 severely affected and 1 moderately affected), 26 goblets] is also shown. Apical goblet zone volume changes during CCh stimulation (, 7 crypts) were obtained from epithelial volume by using goblet volume fraction, including fraction of goblets responding (0.77, from Fig. 18). Dashed line, exponential decline with a half-time of 3.7 min, which fits all but earliest time point in response. C: volume loss (mean ± SE) from nongoblet (columnar cell) apical fraction of epithelial volume for stimulation by fluid secretagogues PGE2 and adenosine (8 crypts). Half-time obtained from exponential fit was 3.3 min.

View larger version (53K): [in this window] [in a new window]   Fig. 20.   Stimulated rearrangement of goblet volume. A: goblet eccentricity in relation to goblet volume (mean ± SE) during stimulation by CCh for responding goblet cells (defined as in Fig. 18). Open symbols, control; filled symbols, stimulated; circles, normal; triangles, ulcerative colitis; diamonds, Crohn's disease; inverted triangles, tamarin. Some crypts had been stimulated by histamine before stimulation by CCh (*). B: goblet eccentricity (for all goblet cells) in relation to goblet volume before (#), during, and after stimulation by a fluid secretagogue, PGE2 or adenosine. Symbols as in A. C: changes in goblet dimensions for CCh-responsive goblet cells of each crypt (, , ), for average of CCh-nonresponsive goblet cells from all crypts (, volume fraction released <0.10), and for all goblet cells during fluid secretagogue stimulation of each crypt (, , ).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES APPENDIX

A major task of the colonic epithelium as the most distal site along the alimentary tract is to reabsorb fluid. In contrast to this conservation function, the colonic epithelium also secretes fluid. This fluid has a distinct electrolyte and macromolecular composition. The largest macromolecule secreted is mucus (9), which serves to buffer cells against abrasion by luminal contents and to create a microclimate (unstirred layer) near the epithelial surface. Mucus released in crypts of Lieberkühn is cleared into the colonic lumen by fluid secretion. As shown in RESULTS, crypts from human colon release apically stored material (presumably mucus) from goblet and columnar cells. Separate control of these mucus secretions occurs through distinct types of secretagogues.

Epithelial cell types in colonic crypts. The colonic epithelium is comprised of several cell types that form the two major epithelial structures: the surface epithelium and the crypts of Lieberkühn. Similar to other mammals (3, 14), human colonic epithelium contains two predominant cell types: columnar and goblet cells (21, 40, 48). Columnar cells have been distinguished further on the basis of ultrastructural features associated with the degree of differentiation. Enteroendocrine cells, which release signaling molecules, generally make up <5% of the cells. Goblet cells are distinguished by the large number of mucous granules in the apical pole. Although in crypts these cells lack the characteristic narrow basal pole seen in the surface epithelium, crypt goblet cells have a rigid cytoskeletal arrangement that maintains the round profile of the granule mass (43, 51). In living crypts, with use of DIC microscopy, goblet cells can be recognized readily by the ovate shape of this granule cluster because of the highly refractile contents of the granules (Figs. 1-4 and 11-14). Columnar cells fill the spaces between goblet cells; distinctions among these nongoblet cells were not apparent. Deformation by the more rigid goblet cell neighbors produced a narrow waist between basal and apical poles, giving columnar cells an hourglass appearance. The apical poles of columnar cells fan out over goblet cells, such that only the small apexes of goblet cells contact the lumen. Whereas in rabbits and mice the apical poles of columnar cells are filled with large vacuoles (3, 14), columnar cells in human colon have small apical vesicles (21, 40, 48). Thus goblet and columnar cells have apically stored products positioned for release into the lumen. The general epithelial appearance is of goblet ovals pointing toward the lumen and fan-shaped deltas of columnar cells with a wide luminal extent.

Mucus secretion by goblet and columnar cells. Secretory control of columnar and goblet cells in colonic crypts is distinct. Release of mucus from goblet cells is stimulated by cholinergic agonists and by histamine (35); goblet cells do not release mucus in response to vasoactive intestinal peptide (VIP) or cAMP and theophylline (35), agents that stimulate fluid secretion (13). Responsiveness was assessed in fixed specimens by the presence of apical cavitation in goblet cells, a procedure that would miss modest stimulation of granule release. A morphometric assessment of cholinergic stimulation indicated that goblet cells that did not cavitate, particularly in the surface epithelium and crypt base, did show a decrease in mucous granules (37, 39). From these studies of fixed colonic tissue, the strongest goblet secretagogues are cholinergic agonists and histamine, but other agents are not conclusively ruled out as minor secretagogues. In living human colonic crypts, with use of morphometric assessment, it was found that CCh and histamine stimulated goblet granule release (Figs. 11 and 16),