Rapid epithelial restitution of human and rabbit colonic mucosa

GASTROENTEROLOGY

1989:97:685-701

Rapid Epithelial Restitution of Human and Rabbit Colonic Mucosa WOLFGANG FEIL, ERIC R. LACY, YU-MAN MATTHEW WONG, DORIS BURGER, ETIENNE WENZL, MICHAEL STARLINGER, and RUDOLF SCHIESSEL Department of Anatomy and Cell Biology, Medical University of South Carolina, Charleston, South Carolina and University Clinic of Surgery I, Vienna General Hospital, Vienna, Austria

Rapid epithelial restitution is now considered one of the primary defense mechanisms of the stomach and duodenum. Because there is currently no evidence as to whether restitution occurs in human tissue, this study examined human and rabbit coionic mucosa after superficial injury and monitored the potential difference, alkaline flux, and speed and mechanisms of mucosal restitution as observed with light and electron microscopy. Luminal exposure of the in vivo rabbit colon to 100 mM HCl for 5 min or the in vitro human colon to 10 mM HCl for 10 min caused superficial mucosal injury to 76% of the epithelial surface in the rabbit and 95% in the human. The necrotic epithelial cells detached in sheets from the intact basal lamina and formed a protective mucoid layer. Morphologic evidence of restitution occurred within 15 min after injury in the rabbit and 30 min in the human, as viable nongoblet cells projected lamellipodia and migrated over the denuded basal lamina at a speed of -2 pm/min. One hour after damage 61% of the mucosal surface was still damaged in the rabbit, and 86% of the human mucosal surface was damaged after 2 h. In the following 60 min restitution progressed rapidly, so that only 10% of the surface remained unrepaired in the rabbit after 2 h and 19% in the human after 3 h. Small areas with deeper injury did not repair until 5 h after damage. The potential difference dropped after mucosal injury and did not recover despite morphologic repair. Rapid epithelial restitution is considered to be a basic defense mechanism of the gastrointestinal mucosa that is obviously not necessarily related to the presence of an acidic environment in the stomach or duodenum.

N

umerous studies have attempted to create a model for experimental colitis by using ricinoleic acid (I), various bile acids (2), deoxycholic acid

(3,4), carrageenan (5), melphalan (6), formaldehyde (7,8), alcohol (9), or acetic acid (10,ll). Nonetheless, little is known about the etiology of colitis and even less about the cellular mechanisms of colonic epithelial repair. Most studies of colonic and rectal wound repair have involved acute trauma that destroys large masses of the tissue (12-17). Those investigations provided knowledge concerning the long-term repair process of the colonic mucosa, but it is doubtful that they reflect usual pathophysiologic conditions, as the colon is not normally subjected to acute trauma nor does the luminal fluid contain potentially necrotizing agents such as the acid and pepsin found in the stomach and duodenum. Although some ultrastructural observations of superficial wound healing have been made in the rectum (18), it was not until recently that cell migration was shown to play an important role in the restitution of the epithelial barrier in the in vivo porcine colon after deoxycholate exposure (19). The primary objective of the present study was to develop a model for superficial mucosal damage in the colonic mucosa and to determine how and under what conditions the damage was repaired. Rabbit colon (in vivo) and human colon (in vitro) were investigated simultaneously. Quantitative morphologic and physiologic parameters were correlated. Materials

and

Methods

Tissue Preparation Rabbit colon [in vivo). Experiments were performed on white, nonfasted, female New Zealand rabbits (Ivanovas, Kissleg, F.R.G.) with an average weight of 3 kg.

Abbreviation used in this paper: AF, alkaline flux. 0 1989 by the American Gastroenterological Association 0016-5085/89/$3.50

GASTROENTEROLOGYVol. 97,No. 3

686 FEIL ET AL.

Arterial blood pressure and acid-base balance were monitored as previously reported (20). After sedation with Vetanarcol* (0.3 mg/kg i.v.), the abdomen was opened under local anesthesia consisting of l-2 ml of 1% Scandicain+ (Astra, Linz, Austria) given subcutaneously. A 6-cm-long segment of the first portion of the proximal colon (l-1.5 cm from the cecal valve) was isolated between ligations without impairing the blood supply. A surgical incision was made at each end of the segment and the tips of two polyvinyl catheters (Tygon, inside diameter 2.4 mm, outside diameter 3.1 mm; Norton Performance Plastics, Akron, Ohio) were introduced into the colonic lumen. The cecal segment was perfused in situ with isotonic NaCl solution for the baseline measurements. A Buchler multistaltic pump (Buchler Inc., Fort Lee, Va.) assured a constant perfusion rate of 10 ml/min, recirculating -45 ml of fluid through the colon and the reservoir. The perfusate was kept at 37°C and gassed with 100% oxygen, which was prewashed in 1 M KOH to avoid possible traces of CO,. Transmucosal potential difference (PD) was monitored with KC1 agar-filled polyethylene tubes (Portex 3 FG O/O1 x 0.2 mm; Portex Ltd., Hythe, Kent, U.K.); one positioned in the lumen of the perfused segment via a small incision and the other tube placed in the ear vein. Both bridges were connected via saturated KC1 solutions and calomel electrodes to a voltmeter (pHM 82; Radiometer, Copenhagen, Denmark). Correction for junction potentials was not necessary. The values were given in millivolts and recorded every 10 min. Alkaline flux (AF) was measured with the pH-stat method at pH = 7.4 by automatic titration with 0.1 M HCl in the reservoir (pHM 82 standard pH-meter, TTT 80 titrator, ABU 80 autoburette; Radiometer] and was expressed as microequivalents per square centimeter per 10 min. Human colon (in vitro). The tissues were obtained from patients who underwent surgery for colorectal cancer. The sigmoid and descending colon were mobilized without impairing the blood supply. The colon segment (sigma) was removed quickly and put into an oxygenated Euro-Collins solution* (Fresenius AG, Bad Homburg, F.R.G.) at 4°C for transport to the laboratory (5-10 min). The seromuscular layer was removed by scalpel dissection and placed in Euro-Collins solution (4%). This process took -2-3 min. The resulting mucosal sheets were mounted horizontally in modified Ussing chambers (20) with the luminal side facing upwards (2-cm* surface area). Isotonic NaCl solution was kept at 37°C and gassed with 100% oxygen [prewashed in 1 M KOH to remove CO,) and placed on the luminal side. The nutrient solution (37°C) contained the following (mM/L): 122.0 NaCl, 5.0 KCl, 1.3 MgSO,, 2.0 CaCl,, 20.0 glucose; it was gassed with 95%

* One hundred milliliters Vetanarcol contains 0.162g pentobarbitalum natricum, 20 mg benzylic alcohol, and 500 mg propylene glycol. ’ One milliliter 1% Scandicain contains 10 mg mepivacaine hydrochloride. * One thousand milliliters Euro-Collins solution contains 115 mEq Na+, 15 mEq Cl-, 10 mEq HCO,-, 15 mEq H,PO,-, 85 mEq HPO,‘-, 35 g glucose (375mosmol/L at pH = 7.5,4'C).

O,-5% CO, and kept at pH, = 7.4 using an automatic titration system (Radiometer). Transmucosal PD and AF were measured at the luminal side as described above.

Experimental

Groups

Rabbit colon (in vivo). CONTROLS. For histologic controls, colonic segments were perfused with isotonic NaCl solution at pH, = 7.4 parallel to acid-treated segments from the same animal for 90 min (n = 3), 3 h (n = 6), or 6 h (n = 6). Potential difference was measured as described above in the 6-h series. In addition, unperfused colonic segments of each animal were excised at the end of each experiment and fixed for microscopic examination. EXPERIMENTAL.After 20 min of baseline measurements (PD and AF), the colonic segment for each experiment was luminally perfused with either 100 or 200 mM HCl for 5 or 30 min followed by an acid washout period for 10 min and a postinjury perfusion with isotonic saline (pH, = 7.4) for up to 10 h. NaCl was added to hypotonic HCl to obtain a total ionic strength of 308 mosmol/L. Hydrochloric acid was used for superficial mucosal injury to compare the effects with acid-induced mucosal damage in the stomach (21,221 and rabbit duodenum (20,23). Furthermore, this agent produces uniform superficial damage when administered under appropriate conditions, as opposed to components such as aspirin that form focal hemorrhagic lesions (24). Different acid concentrations and exposure times were tested to develop a model for superficial damage: (a) 200 mM HCl for 30 min, postinjury perfusion for 90 min (n = 3). 3 h (n = 8), or 6 h (n = 15); (b) 100 mM HCl for 30 min, postinjury perfusion for 30 min (n = 3), 90 min (n = 5), 6 h (n = 5), or 10 h (n = 5). (c) 100 mM HCl for 5 min, fixation immediately afterwards (n = 3), postinjury perfusion for 3 min (n = 3), 7 min (n = 3), 15 min (n = 3), 30 min (n = 6), 1 h (n = 7), 2 h (n = 6). or 5 h (n = 6). Potential difference was measured during the postinjury period in all experiments, AF in one series of trials (100 mM HCl for 5 min, postinjury perfusion for 5 h). At the end of the experiments the tissues were fixed in situ by injection of 5% buffered formalin or Ito’s fixative (25) containing 4% formaldehyde, 5% glutaraldehyde, and 0.2% picric acid buffered with 0.2 M cacodylate buffer. Human colon (in vitro). CONTROLS. Mucosal sheets (n = 5) remained in the Ussing chamber for 5 h bathed with isotonic NaCl solution (pH, = 7.4). Potential difference and AF were measured as described above. EXPERIMENTAL.After IO-20 min of baseline perfusion, the luminal solution was changed to 10 mM HCl for 10 min. The remaining acid was washed out with 10 quick changes of isotonic NaCl solution. The tissues remained in the chambers (isotonic NaCl solution, pH, = 7.4) for 15 min (n = 3), 30 min (n = 3), 2 h (n = 7), 3 h (n = 7), or 5 h [n = 7) after the end of acid exposure. Potential difference and AF were monitored and recorded every 10 min in all experiments; AF was measured in one series (5 h). Untreated mucosal preparations from each experiment were fixed immediately after stripping to serve as additional morphologic controls. Experimental and control

September 1989

tissues were fixed for light and electron microscopy at the end of the experiments in Ito’s fixative (251, as described above.

Microscopy The fixed specimens from all described experiments and controls were pinned out on wax plates with the mucosal side up. Three strips of tissue, each -3 mm wide and representing the complete circumference of the colon, were excised from the middle, proximal, and distal end (each 1 cm apart] of the fixed specimen with a razor blade and embedded in paraffin. Sections were cut about 3 pm thick, stained with hematoxylin and eosin, mounted on glass slides, coded, and used for morphometric analysis of damage and restitution. Tissues were also embedded in epoxy resin for qualitative light and electron microscopic observation. Fixed colonic segments were rinsed with 0.2 M cacodylate buffer, then postfixed in 1% 0~0, in 0.1 M cacodylate buffer at pH = 7.4 for at least 2 h at room temperature. After washing in maleate buffer (0.2 M at pH = 5.21, tissues were treated with 1% uranyl acetate in maleate buffer (pH = 5.2) in three changes for 1 h. After a 2-h wash in the same buffer, the specimens were dehydrated in ethanol and embedded in Epon-araldite. Semithin sections (0.5 pm thick) were stained with toluidine blue. Thin sections (0.2 pm) were stained with a fresh mixture of saturated uranyl acetate followed by lead citrate and examined with an Hitachi HU12 electron microscope. Selected fixed tissues were prepared for scanning electron microscopy by postfixation in osmium tetroxide and subsequent dehydration in graded concentrations of ethanol. Specimens were then critical point dried, sputter-coated with gold-palladium, and examined in a JEOL 35 at accelerating voltages of 15 kV.

Morphometry Time-course of mucosal restitution. To assess the extent of mucosal injury and to monitor the progress of epithelial restitution, morphometry was performed on the paraffin-embedded sectioned tissue from all experiments. The three sections, each of which represented the complete circumference of the colonic segment at a different defined location, were examined. Measurements were made using a Zeiss IM 35 research microscope (Carl Zeiss, Oberkochen, F.R.G.; objective magnification, x6.3, X10, x 40; ocular magnification, x 8). Pictures were transmitted using a MT1 series 68 videocamera (NC68DX; Dage-MTI, Inc., Michigan City, Ind.) to a monitor screen (model PVM-1271Q, superfine pitch; Sony, Japan] and the computer system (IBM-PC AT, model 5170; IBM, Armonk, N.Y.). A special morphometry programs (MIPSY, “The Micro-based Image Processing System”] was downloaded

s”The Micro-based Image Processing System” (MIPSY) is a custom made morphometry program edited by Yu-Man Matthew Wong and John Lindroth, Medical University of South Carolina, Charleston, South Carolina.

RAPID RESTITUTION OF COLONIC MUCOSA 687

to the CPU and followed on a second monitor (Magnavox Professional RGB Monitor 80; N.A.P. Consumer Electronics Corp., Knoxville, Tenn.). The system was calibrated to each objective using a micrometer slide (2 mm in 200 equal parts; Wild, Heerbrugg, Switzerland). Calibration values could be stored as a file in the CPU and downloaded for the elected objective magnification. Additionally, the signal from a graphic tablet (HIPAD Digitizer DT 114; Houston Instruments, Austin, Tex.) overlayed the tissue image on the first screen. The equipment allowed the investigator to draw contours of the mucosa on the screen with the “puck” of the digitizer in a live mode by following the superficial epithelial layer, including the crypts, with the cursor. By this method at first the length of undamaged epithelial lining was added to the variable “0,” then the damaged areas in between were added to the variable “1.” Double measurements were avoided, as the drawing of the 0 sections remained displayed on the screen, thus marking the points of demarcation between damaged and undamaged mucosa. To facilitate the decision on the exact margins of the two areas (damaged and undamaged), the slide could be viewed on the microscope at any time throughout the procedure. The drawing overlay was then discarded and the slide moved to the next area of investigation. The results were coded for each tissue and stored on back-up copies for further statistical calculation. Values give the percentage of mucosal length damaged by acid exposure. Absolute values were measured in microns to the next 0.01 pm and usually given in millimeters to the next 0.01 mm. Migration speed. The initial lo-cm-long portion of the rabbit colon shows wartlike protrusions, which characterize the mucosal surface topography of this species (26,27). Mucosal damage was confined to the apical parts of these villuslike warts and the migrating cells protruded from both sides out from the huge crypts over the protrusions. The surface of the mucosal protrusions was measured by following the mucosa with the puck on the screen of the morphometry equipment. The intercrypt distance was defined as the length of mucosal surface lining between the deepest points of the crypts surrounding welloriented protrusions. The intercrypt distance in randomly chosen untreated control tissues was 539.37 t 11.47 pm (n = 60 measurements in 10 tissues). Six measurements were performed at randomly chosen areas in each of the seven tissues fixed 1 h after mucosal damage for comparison of the intercrypt distance. Additionally, damaged and undamaged mucosal surface was measured in the l-h and 2-h series as described above. Thus it was possible to estimate the percentage of mucosal length that should have been covered with intact epithelial cells in the ensuing 1 h in the 1-h series. The number of mucosal protrusions in the 1-h series was calculated by dividing the total mucosal length by the mean intercrypt distance. By knowing the mucosal surface (in microns), which had to restitute in 1 h, and the number of mucosal protrusions, it became possible to estimate the speed of the migrating cells in the rabbit colon.

686

FEIL ET AL.

GASTROENTEROLOGY Vol. 97, No. 3

Statistics All values are expressed as mean 2 SEM. Probabilities of ~0.05 were regarded as significant. Statistical differences were calculated from the t-test, Wilcoxon twosample test, and Kruskal-Wallis test (2 approximation) using a modified SAS program (SAS Institute, Inc., Cary, NC.). Except when specially indicated, p values are from the Kruskal-Wallis test.

Results Effect of Mucosal Damage on Potential Difference

01

Rabbit colon. CONTROLS. Baseline PD was -12.38 (kO.95) mV (n = 6) and remained stable throughout the experiments (-11.82 ? 0.82 mV after 6 h, n = 6 and -11.04 ?z 0.62 mV after 10 h, n = 3) (Figure 1). EXPERIMENTAL. HCl (200 mM, 30 min): PD dropped after acid exposure and reversed polarity from -12.43 (kO.95) mV to +5.07 (20.55) mV 10 min after the end of acid injury (Figure 1). Potential difference started to recover after 1 h and reached +0.93 (kO.55) mV 6 h after acid injury. HCl (100 mM, 30 min): PD changed from a baseline value of -9.20 (kO.62) mV to +2.80 (-+l.ll) mV after damage. This experiment was continued for 10 h to determine if the PD could recover. Although there was a progressive return of the PD toward preinjury levels, at 10 h after insult the PD was only -4.50 (t 0.39) mV. HCl (100 mM, 5 min) was the only group in which

I-, -c

5minlOOmMHCl 301nin1OOmMHCl

al

4

30min2OOmMHCI

•j

+

Control

??

-1

.

@ 1

0

.

1

1

.

1

2

.

1

.

3

1

4

Figure

1 2 3 4 5 6 7 8 HOURS AFTER ACID EXPOSURE

0

I. Potential

10

of the in vivo rabbit colon before exposure to luminal HCl. Symbols indicate means and bars indicate SEM. Columns depict concentration of acid, moment of administration, and exposure time. and

after

difference

9

luminal

f

5

HOURS AFTER ACID EXPOSURE Figure 2. Potential difference of the in vitro human colon before and after exposure to luminal HCI. Symbols indicate means and bars indicate SEM. Column depicts moment of luminal exposure (10 mM HCI for 10 min).

acid exposure did not cause a reversal in the PD values (Figure 1). From a baseline value of -14.0 (k1.20) mV, the PD fell to -5.50 (k1.20) mV 10 min after luminal exposure to acid. Within 30 min after acid damage, the PD fell further to -4.71 (20.20) mV. The PD rose slightly thereafter but remained steady until the end of the experiments 5 h after injury at -5.50 (k1.17) mV. In all three experimental series, the PD values at the end of the experiments were significantly different (0.020 2 p 2 0.003) from the respective baseline values. Human colon. There was no statistically significant difference between the baseline PD of the untreated control group (-16.80 ? 0.06 mV) and the preinjury levels of the acid-treated tissue (-18.13 +0.23 mV) (Figure 2). In the control group (isotonic NaCl, pH, = 7.4), PD remained constant at -17.20 (20.20) mV after 5 h. Ten minutes after damage with acid, PD dropped to -5.25 (21.30) mV. There was no significant recovery of PD during the ensuing 5 h of the experiment (-4.80 t 0.80 mV). Effect of Mucosal Damage on Alkaline

-1

-

Flux

Rabbit colon. Luminal exposure to the weakest concentration of acid (100 mM HCl, 5 min) caused a significant drop in AF from 1.12 (20.04) pEq/cm’ . 10 min during the baseline period to 0.19 (kO.08) pEq/cm” 10 min 10 min after injury (Figure 3). Alkaline flux recovered to 0.84 (kO.21) pEq/ cm’ . 10 min after 30 min, reached baseline values after 1 h, and remained relatively constant (1.20 + 0.06) during the remaining 2 h of the experiment.

September

RAPID RESTITUTION OF COLONIC MUCOSA

1989

f

689

10 min caused a sharp drop of AF from the baseline value of 0.55 (+O.lO) to zero, as measured after the lo-min washout period. In the following 20 min, AF recovered to -50% (0.31 ? 0.17) of the baseline values. Complete recovery was achieved after 1 h (0.50 2 0.15). Alkaline flux remained constant for the succeeding 4 h (0.45 + 0.14) to the end of the experiment. There was no significant difference in AF between the control and the acid injury group except in the first hour after damage.

Human (10 min 10 mMHCl)

1.5 3 2 % 3” 1.0 x’ 2 z 0.5

Microscopy

2 a 0 -1

0

1 2 3 4 HOURS AFl-ER ACID EXPOSURE

5

Figure 3. Alkaline flux measured in rabbit and human colon before and after exposure to luminal HCl. Symbols are means with corresponding SEM. Column depicts concentration of acid and moment of luminal administration.

Human colon. Basal AF was approximately half of that measured in the in vivo rabbit colon (0.52 + 0.05 PEqlcm’ 10 min). These values remained constant up to 5 h (0.51 ? 0.06) in the untreated control group. Luminal exposure to 100 mM HCl for

Rabbit colon. CONTROLS. Perfused or unperfused colonic segments showed normal morphology [Figure 4). EXPERIMENTAL. HCl (100 mM or 200 mM, 30 min): exposure to either acid concentration produced nearly identical results. Almost the entire mucosa from the epithelium to the muscularis mucosae was severely damaged. Vascular congestion and thrombosis were consistently observed in the mucosal, submucosal, and occasionally the subserosal vessels. Large areas of parenchymal hemorrhage were present. In these regions, the necrotic tissue did not slough off into the lumen but remained in situ for the duration of the experiments. Within 10 h postinjury, scanning electron and light microscopy re-

Figure 4. Scanning electron micrograph of the rabbit colon 6 h after isotonic NaCl perfusion shows normal morphology. This view reveals the typical anatomy of the proximal rabbit colon with wartlike protrusions (black asterisk) as well as deep crypts (white asterisk). Magnification, x70.

690

GASTROENTEROLOGY

FEIL ET AL.

Vol. 97. No. 3

vealed heavy bacterial overgrowth adjacent to the luminal border [Figure 5). We could not detect any signs of repair in these necrotic regions during the experiments. Contrary to the condition of almost all of the mucosa, small and patchy areas escaped deep necrosis. In these restricted areas, there was either no injury or it was limited to the epithelium and was not accompanied by thrombosis and vascular congestion. Histologic examination at several postinjury intervals showed that only these superficially damaged regions underwent a cycle of rapid restitution in which necrotic surface cells were sloughed and the viable, deeper cells migrated to cover the defect, as described in more detail in the next section. In these superficially damaged areas, restitution occasionally could be detected by light microscopy as early as 30 min but was unequivocally present 3 h after damage and within 6 h there was nearly complete repair. HCl (100mM, 5 min): in tissues fixed immediately after luminal acid exposure, light and electron microscopy revealed an absence of the deep necrotic damage present in the tissues exposed to 100 mM Figure

Figure

5. Light micrograph of the superficial rabbit colon from an area deeply damaged 10 h after exposure to 100 mM HCl for 30 min. Superficial necrotic epithelium remains attached to the mucosa along the basal lamina (BL) and is overgrown with bacteria on the luminal surface (arrows]. LP, damaged lamina propria; M, mucous granules of necrotic goblet cells. Magnification, X63.

6. Scanning electron micrograph of rabbit colon 15 min after exposure to 100 mM HCl for 5 min. Damage is evident only at the tips of the protruding mucosal “warts” (arrows) and not deeper toward the crypts. Magnification, x250.

HCl for 30 min, as described above. Even at 7 min after injury, scanning electron microscopy did not reveal epithelial damage. Areas showing epithelial injury could be detected distinctly by electron microscopy at 15 min postinjury. Damage was first detected in the superficial epithelial cells as their apical plasma membranes were ruptured and cell contents started to erupt (Figure 6). Sheets of necrotic epithelial cells began to lift from the intact underlying basal lamina (Figure 7) and form subepithelial blebs. These subepithelial blebs were often filled with a fine granular precipitate similar to that found in the underlying capillaries. The epithelia in these regions developed loss in staining intensity, acquired swollen cytoplasm with rich granularity, and showed signs of karyolysis. There was minimal damage to the components of the lamina propria (Figure 7). There were no signs of mucosal repair by light microscopy, but transmission electron microscopy showed that in regions where the necrotic superficial epithelium had exfoliated there were viable nongoblet cells forming lamellipodia and beginning to migrate across the

September

Figure

1989

RAPID RESTITUTION OF COLONIC MUCOSA

691

7. Transmission electron micrograph of the superficial rabbit colon 15 min after exposure to 100 mM HCl for 5 min. Necrotic epithelial cells (NL) are just beginning to exfoliate off the basal lamina (BL) as a single layer. Asterisk denotes extravasated red blood cell In lamina propria (LPI. Magnification, X4800

basal lamina. This was the first sign of epithelial restitution. At 30 min, most of the necrotic epithelial layer remained attached to the viable mucosa, although it still formed blebs [Figure 8). In some areas, the necrotic epithelial layer started to detach from the viable cells along the basal lamina. In these regions, distinctively flattened nongoblet epithelial cells could be detected slipping over the basal lamina under detached necrotic epithelial lining, providing reepithelialization (Figure 9). Occasionally the migrating enterocytes extended lamellipodia away from the basal lamina into the lumen. At 1 h, as the necrotic tissue continued to detach from the basal lamina, the flattened intact epithelial cells started to bridge the defect and to further

accomplish restitution. Detached necrotic tissue usually remained embedded in the mucus, thus forming a layer composed of epithelial cell debris and formed blood elements. This layer was periodically attached to the mucosa at points where the viable cells projected into the lumen (Figures 8 and 10). At 2 h, as epithelial restitution proceeded, the basal lamina was covered with a new epithelial layer formed by the viable migrating enterocytes. The necrotic layer usually remained covering the mucosal surface even after complete epithelial restitution. In some areas, the necrotic tissue was directly adherent to the mucosa without any sign of restitution. Epithelial restitution was nearly accomplished at 5 h (Figure 10). Only very small areas showed

692

FEIL ET AL.

GASTROENTEROLOGY Vol. 97, No. 3

Figure 8. Light micrograph of histologic section from rabbit colon 30 min after 100 mM HCl damage for 5 min. A segment of superficial necrotic epithelial cells (NL] has exfoliated as a single sheet from the basal lamina (arrows] but remains attached to the adjacent viable epithelial cells (E) at positions denoted by arrowheads. LP, lamina propria. Magnification, X121.

incomplete repair. Nearly the entire surface of the rabbit colonic mucosa was covered with squamous to cuboidal cells largely consisting of absorptive cells resting on the basal lamina. Absorptive cells had a midcentrally located nucleus and abundant mitochondria, among other organelles, and had repolarized as shown by apical microvilli and tight junctions (Figures 11 and 12). Goblet cells were occasionally present within the luminal epithelium but showed signs of damage such as vacuoles, swollen organelles, and a bulging apical plasma membrane. They were obviously not actively contributing to the mechanisms of restitution. There were small areas where necrotic cells remained attached to the mucosa. In other regions, damaged cells indicated by numerous vacuoles, dilated mitochondria and endoplasmic reticulum, and broken organelle plasma membranes, were present among the viable cells. In addition, some cells were still squamous with lamellipodia, suggesting they had migrated across the basal lamina. The necrotic layer was mostly detached and sloughed off. Human colon. CONTROLS. Tissues taken immediately after mucosal stripping or after 5 h in the Ussing chamber retained normal morphology. EXPERIMENTAL. Luminal acid exposure to IO mM HCl for 10 min caused nearly uniform mucosal damage confined to the superficial epithelial layer. The crypts remained intact. The most impressive feature at 15 min postinjury was the formation of

subepithelial blebs or blisters at the mucosal tips (Figure 13). Epithelial damage became more obvious by light microscopy after 30 min, showing swelling of the cytoplasm as well as granulation and loss of staining intensity. The basal lamina generally remained undamaged as did the underlying propria tissue. Only small patchy areas showed deeper injury. There was no evidence of epithelial restitution at this time. However, by transmission electron microscopy, rudimentary lamellipodia were detected in the crypt necks at 30 min postinjury. Two hours after epithelial injury, most of the superficial epithelial sheets of necrotic tissue were exfoliated. These sloughed cells were only loosely attached to the basal lamina, which itself was often completely denuded to the luminal side. At the demarcation line between intact and damaged mucosa, which was usually located at the necks of the crypts, flattened epithelial cells could be detected migrating over the denuded basal lamina. This was the first sign of incipient epithelial restitution with light microscopy, but scanning electron microscopy showed numerous flattened cells with lamellipodia in these positions (Figure 14). After 5 h the necrotic layer had nearly completely lifted off [Figure 15) and mucosal restitution proceeded continuously as the basal lamina was nearly completely covered with flattened epithelial cells. However, transmission electron microscopy of the

September

1989

RAPID RESTITUTION OF COLONIC MUCOSA

693

Figure 9. Transmission electron micrograph of rabbit colon 30 min after 100 mM HCl damage for 5 min. Exfoliated epithelial cells have exposed the basal lamina (BL) over which an enterocyte (E) has extended a lamellipodium and begun to migrate. The underlying capillary (C) remains patent and the endothelium and cells in the lamina propria (F, fibroblast) show mild membrane blebbing (asterisk) indicative of injury. Magnification, x8640.

human colon 5 h after acid exposure showed considerably more damage than the comparable time periods in rabbit tissue. Figure 16 shows that, although this portion of the mucosa is covered with epithelium, the cell appears to be actively engaged in restitution and reorganization. Viable cells extended lamellipodia between the basal lamina and the adjacent necrotic cells. Numerous cells were exfoliating into the lumen and had vacuoles and breaks in the plasma membranes. Both viable and necrotic-appearing cells were held together by tight junctions (Figure 17). Those areas, where the cellular constituents of the lamina propria also showed more extensive damage, remained unrepaired at that time. Morphometry Rabbit colon. Histologic sections from tissue exposed to 100 mM luminal acid for 5 min were quantitatively evaluated for damage and restitution. A total of 2985.05 mm of mucosal surface length was

examined (n = 37) at all time periods in these tissues. Immediately after damage, 79.52% ? 1.44% of the mucosa was damaged (Figure 18). Values were unchanged up to 30 min (76.41%+ 1.07%),but by 1 h the damaged surface had decreased to 61.35% * 1.47%. During the ensuing 60 min, there was a significant decrease to only 10.43% k 0.87% of the mucosa damaged. Within 5 h after acid, only 2.36% 2 0.34% of the colonic mucosa was damaged. The remaining islands of damage were not expected to repair by rapid restitution, as damage extended beyond the basal lamina and destroyed lamina propria tissue. One hour after injury, 61.35% of the total mucosal surface length of 668.45 mm [n = 7) was classified as necrotic. During the next 60 min, the damaged tissue was reduced by 50.92%, so that the total mucosal surface remaining damaged was 10.43%. The average intercrypt distance was 534.85 ? 14.44 pm (n = 42) in the 1-h series. Thus, 1249.74 protrusions (total

694 FEILET AL.

GASTROENTEROLOGY Vol. 97. No. 3

Figure

mucosal length divided by mean intercrypt distance] could be expected for the total mucosal length of 668.45 mm. As the remaining viable epithelial cells started to migrate from both sides over the denuded protrusions, we calculated that 2499.48 cells starting at the leading edge had to bridge 340.36 mm (50.92% of total mucosal length) in 1 h. Therefore, the calculated migration speed of the epithelial cells was -2.3 pmimin. Human colon. A total of 1008.39 mm of mucosal surface was measured in 27 slides. Fifteen minutes after luminal acid exposure, 95.95% * 1.06% of the surface was judged necrotic and after 30 min this value was virtually unchanged (96.13% ? 0.17%) (Figure 18).The injured surface was reduced to 85.70% ? 2.27% at 2 h. Mucosal restitution proceeded rapidly within the following 60 min so that 19.27% -C 1.28% of the mucosa remained damaged at 3 h and 16.96% -+ 1.18% after 5 h. Because the human colonic mucosa lacks the unique structure of the proximal rabbit colon (no mucosal protrusions), it was not possible to calculate the migration speed as in the rabbit colon.

10. Light micrograph of histologic section from rabbit colon 5 h after 100 mM HCl damage for 5 min. Exfoliated necrotic superficial epithelium (NL) remains as a layer attached at a few points (arrowhead] to the underlying mucosa. Note the restituted superficial epithelium (arrows) consists of irregularly low cuboida1 enterocytes overlying the lamina propria (LP). Magnification, x 121.

Discussion The present study shows that the colonic mucosa of both human and rabbit respond to superficial injury by reestablishing the damaged surface epithelium through a process of cell migration termed “rapid epithelial restitution” (20,28,29). Although numerous studies have documented that cell migration is an integral component of wound repair (18,30,31), it was not until recently that its importance in the repair of superficial lesions in the gastrointestinal tract was recognized. This process was first described in the stomach (28,29,32), followed by its detection in the duodenum (20) and colon (19,33). It is now considered to be an important defense mechanism against luminal challenge (20,34). The present study is the first to our knowledge that determined the response of the human colon to superficial injury, a condition more closely reflecting the in vivo conditions than acute trauma studies previously reported (12-18). Rapid epithelial restitution occurs only under conditions in which damage is confined to the superfi-

September

1989

RAPID RESTITUTION OF COLONIC MUCOSA

695

Figm -e 11. Transmission electron micrograph showing a portion of the restituted rabbit colonic surface 5 h after acid exposure. ALsingle damaged goblet cell with swollen vacuole (V) is situated between columnar absorptive cells. Arrow indicates a portit on of a migrating low cuboidal epithelial cell. Asterisks show two undifferentiated cells. C, capillary; BL, basal lamina (arrow rhead]. Magnification, x3600.

cial mucosa. Regions of the mucosa with gross hemorrhagic lesions heal by a lengthy process of tissue Coreplacement involving cell mitosis (20,21,34). ionic tissue exposed to the stronger concentrations of acid in the present study had deep hemorrhagic lesions in vivo and extensive mucosal necrosis in vitro. This type of damage was not repaired by epithelial restitution but, had our observations been carried out longer, would probably have been subject to either scarring or total parenchymal replacement by processes involving mitosis. In the first series of experiments, 200 and 100 mM HCl were administered to the luminal surface of the rabbit colon in vivo for 30 min, because these concentrations were used to produce superficial injury in our previous studies in the rabbit duodenum (XI). In contrast to the duodenum, these concentrations of acid caused deep injury to the mucosa. It only became visible in the long-term experiments after 6 or 10 h. We hypothesize that this type of deep injury consists of two different components: an initial damage to the superficial epithelial cells, which is clearly visible at the light microscopic level immediately after the acid exposure and a delayed indirect, secondary injury affecting the entire mucosa. The first

yigure 12. Transmission electron micrograph showing the apically positioned zonula occludens (tight junction) between two restituted rabbit absorptive cells. Arrows show points where the outer leaflet of the adjacent plasma membranes have fused. Magnification, x180.000.

696

FEIL ET AL.

GASTROENTEROLOGY

Vol. 97. No. 3

‘igu1 *e 13. Light micrograph of histologic section from human colonic mucosa 30 min after exposure to 100 mM HCl for 10 min. Damage has been restricted to the superficial epithelium at the villar tip, which has blistered into the lumen forming subepithelial blebs (asterisks) above the basal lamina (arrows]. MM, muscularis mucosae. Magnification, x 77.

component could be explained by direct cellular injury due to acid. The latter, indirect damage may be due to stimulation of mediators such as oxygen free radicals, proteases, amines, or other products derived from cells in the lamina propria or vasculature. Regardless of the potential mediator, the result was vascular congestion in addition to thrombosis and complete stasis of blood flow. This more severe injury is associated with the ischemic lesion due to acid-induced vascular injury. Indeed, our observations in the present investigation show that vascular thromboses were consistently present in tissues hav-

ing severe, deep damage. Vascular occlusion was absent only in those small, patchy areas that escaped severe injury and thus retained their epithelial repair facilities. The important role of regional blood flow for the prevention or development of lesions caused by luminal agents is generally accepted in the stomach (35) and has also been suggested in the duodenum (23,36). Luminal exposure to 100 mM HCl for 5 min produced uniform superficial mucosal injury in the rabbit colon. However, some patchy areas were damaged more deeply and the propria tissue was also

September

1989

RAPID RESTITUTION

OF COLONIC

MUCOSA

697

affected by the injury. As shown here, there was no restitution in these regions up to the end of our experiments in vivo. Reproducible and comparable superficial mucosal damage of the human colonic tissue in vitro could be produced by exposure to 10 mM isotonic HCl for 10 min. This damage model has been used previously in the rabbit duodenum in vitro (20). Rapid restitution observed in the colon shares many similarities with this process in the stomach and duodenum (33). As a first sign of incipient repair, the remaining viable epithelial cells in the vicinity of the necrosis are transformed into rapidly migrating cells (20,28,29,32). When the necrotic epithelial cell linings detached from the basal lamina, broad flat extensions (lamellipodia) protruded along the denuded basal lamina as the cells began to migrate and reepithelialize the damaged mucosal surface. This mechanism is very similar to the stomach, but somewhat different from the duodenum (20). During mucosal repair in the duodenum, lamellipodia are formed preceding the exfoliation of the

Figure

15. Scanning electron micrograph showing nearly completely restituted superficial mucosal surface from human colon 5 h after damage. Asterisks indicate some of the necrotic epithelial cells still attached to the viable-appearing surface cells. Arrowheads indicate migrating cells with lamellipodia. Magnification, X255.

Figure

14. Scanning electron micrograph of the human colonic mucosa 2 h after damage. Asterisks indicate lamellipodia extending from cells migrating toward the lumen. One necrotic epithelial cell (NC) is being extruded by the lamellipodium of a migrating cell. Magnification, x2600.

necrotic tissue. Cell extensions appeared to actively extrude the damaged tissue (20). However, mucosal restitution in the duodenum takes exceedingly longer, because the requisite basal lamina for the orientation of the rapidly crawling epithelia is broken (20,34). Cell flattening as the first sign of repair could be detected in the electron microscope 15 min after injury in the rabbit colon and after 30 min in the human colon. Light microscopy showed occasional cell flattening 30 min after injury in the rabbit and 2 h later in the human. This difference is assumedly due to the better fixation using glutaraldehyde-formaldehyde, the less disruptive embedding processes, and thinner sections in epoxy resin. Physiologic studies have demonstrated important interspecies variations in the physiology of the large intestine (37) that may also reflect a difference in the speed of repair of the mucosa. It is also well known from previous studies that there is a broad interanima1 variation in the time-course of rapid epithelial restitution in the stomach, even when the same damaging agents were used (28,29,32,38,39). Fur-

698

GASTROENTEROLOGY

FEIL ET AL.

Figure

thermore, there is also a difference in the time-course of restitution when in vivo and in vitro experiments of the rabbit duodenum were compared in the same species (20). Nevertheless, the “S” shape of the restitution curve of the colon is similar to that described in the in vivo rat stomach after ethanol injury (29). In both cases, there is an initial period when restitution begins, slowly followed by a phase in which most of the restitution takes place, after which little if any further reepithelialization occurs. Despite these similarities, the in vivo rat stomach restitutes about 80% of the superficially damaged mucosa in -15 min, whereas in the rabbit and human colon, 75% of the surface is restituted in -1.5 h. In addition, the onset of restitution in the colon is delayed when compared with the stomach. Using electron microscopy, we were able to recognize cell flattening and lamellipodial formation 7 min after damage in the stomach (29), but 15 min and 30 min after injury in the rabbit and human colon, respectively. In the rat stomach, -90% of the damaged mucosa was restituted at the end of 60 min, whereas in the rabbit colon it took 2 h and in the human colon 3 h to achieve 90% and 80% restitution, respectively. Furthermore, even after 5 h the human colon had achieved only 80% epithelial restitution, whereas in the rabbit 95% of the surface was reestablished by 5 h. Whether these findings reflect intrinsic cellular differences between these two species or the absence of an intact blood supply in the human tissue cannot

Figure

Vol. 97. No. 3

16. Transmission electron showing micrograph mucosal surface of the human colon 5 h after acid exposure. Asterisk shows degenerative goblet cell next to two viable cells that have extended lamellipodia (arrows) along the basal lamina (BL) Other (arrowhead). regions cells show filled with vacuoles (V) and broken plasma membranes. Magnification, x3800.

17. Transmission electron micrograph showing zonula occludens (arrows) between two adjacent human enterocytes on mucosal surface 5 h after acid damage. Note membrane-bound vacuoles (V). Magnification, X80,000.

September

1989

RAPID RESTITUTION OF COLONIC MUCOSA

0 0 Figure

1

5 2 3 4 Hours after acid exposure

18. Graph showing the time sequence of morphologic restitution of the superficial mucosa after luminal exposure to 100 mM HCl for 5 min in the human colon in vitro and 10 mM HCl for 10 min in the rabbit colon in vivo. Values are means. SEM values were so small they could not be graphically represented here.

be determined but may provide an insight into possible mechanisms for the etiology of colonic erosions. Another component of rapid epithelial restitution is the formation of a mucoid layer over the damaged mucosa. As the necrotic epithelial cells exfoliate from the mucosa, they form a single layer composed of mucus released from epithelial cells, as well as cell debris, fibrin, and other components of the plasma (20,28,29,34,39-42). Once thought to be of little importance, it is now clear that this mucoid layer plays an important role in protecting the injured mucosa from luminal agents and probably aids the restitution process by trapping alkaline tissue exudate, including plasma, next to the superficial mucosa (20,40-43). Similar to the damage and repair process in the stomach and duodenum, a necrotic layer is formed in the colon. This layer remained in position throughout the time frame of colonic restitution observed here but then lifted off. The observed delay in colonic restitution when compared with the stomach may be due in part to a failure of the necrotic tissue to lift off the mucosa. Shedding of the necrosis is likely to be a prerequisite for restitution to occur. Nevertheless, this mucoid layer is considered to be an important protective factor for the gastroduodenal mucosa (20,34,41-44) and appears to be conducive to the repair process in the colon as well. Epithelial restitution in the colon has to be called “rapid” because the cells migrate with a speed of about 2 pm/min. This value is comparable to that estimated for the stomach. Lacy and Ito (29,34) calculated a migration velocity of 1-2 pm/min for the rapid repair of epithelial cells in the stomach,

699

whereas Grant (45) estimated a migration speed of 3.3-6.6 pm/min in the cat superficial mucosa. Mucosal restitution supposedly is a prerequisite for the restoration of normal electrical parameters. Damage of colonic mucosa (rabbit and human) was invariably associated with a fall in PD. The amount of decrease was correlated with the morphologic extent of damage. Thus PD seems to be a good indicator of damage in this model, as in the stomach and duodenum (20,22,23,28,35). However, tissues showing substantial morphologic repair did not have an accompanying return of the PD from the initial drop to the end of the experiment. Potential difference does not reflect mucosal restitution in vivo or in vitro in the present investigation. Therefore, restitution could only be judged morphologically (20). We measured alkaline flux in the present investigation because of its importance in mucosal defense in both the duodenum and stomach. The basal alkaline flux of the in vivo rabbit colon (1.12 pEq/ cm2 10 min) is similar to that in the distal duodenum in vivo (1.4 pEq/cm’ 10 min) or in the proximal duodenum in vitro (0.95 pEq/cm’ 10 min) (20,23). Alkaline flux in the in vitro human colon was about half these values. This difference seems to be more likely due to the lack of additional supply of bicarbonate from an intact circulation in the in vitro model rather than to interspecies difference. This suggestion becomes reasonable regarding the difference in the basal AF between the proximal rabbit duodenum in vivo (3.6 pEq/cm’ .10 min) and in vitro (20,23). The influence of blood flow on AF has also been reported recently (23,361. In the present study, luminal acid exposure caused a significant initial drop of AF and recovery to baseline values after 1 h comparable in the rabbit and human. Bicarbonate supposedly is trapped in the necrotic layer that obviously formed in the first hour after acid injury. It is assumed that there is no net alkaline output into the lumen in that period until the necrotic layer, which forms a diffusion barrier, is saturated with alkali. This might be an explanation for the drop and the recovery of AF in the rabbit colon. One of the postulates concerning the function of the necrotic mucoid layer is that it serves to trap alkaline tissue exudate next to the damaged mucosa, thus facilitating epithelial restitution (2942,431. There is experimental evidence from both the gastric and duodenal mucosa that supports this hypothesis (23,38,43). However, in the present investigation, there is no evidence that AF was significantly important for restitution. Rapid restitution of the colonic mucosa after superficial damage caused by luminal acid occurs in the rabbit in vivo and human in vitro. Rapid epithelial restitution of the gastroduodenal mucosa is gen-

700

GASTROENTEROLOGY

FEIL ET AL.

erally considered to be an important defense mechanism against luminal challenge by acid. The colon lacks the presence of an acidic environment, which is common in the stomach and duodenum. Nevertheless, rapid epithelial restitution appears to be a basic protective mechanism of the gastrointestinal mucosa that is not strictly related to the presence of an acidic luminal environment.

References

19

20

1982:27:865-g. 5. Fath RB, Deschner EE, Winawer SJ, Dworkin BM. Degraded carrageenan-induced colitis in CF, mice. A clinical, histopathological and kinetic analysis. Digestion 1984:29:197-203. 6. Robinson BA. Clutterbuck RD, Millar JL, McElwain TJ. Epidermal growth factor (hEGF) has no effect on murine intestine epithelial damage and regeneration after melphalan. Br J Cancer 1985;52:733-7. 7. Cohen JD, Kao HW, Tan ST, Lechago J. Snape WJ Jr. Effect of acute experimental colitis on rabbit colonic smooth muscle. Am J Physiol 1986;251:G538-45. 8. Seidman EG, Hanson DG, Walker WA. Increased permeability

9.

10. 11.

12. 13. 14.

15.

16. 17.

18.

to polyethylene glycol 4000 in rabbits with experimental colitis. Gastroenterology 1986:90:120-6. Psaila JV, Myers B, Jones IR, Rhodes J. Effect of prostaglandin PGE, on alcohol-induced ulceration in the rat colon. Digestion 1986;35:224-8. Hagihara PF. Experimental colitis as a promotor in largebowel tumorigenesis. Arch Surg 1982:117:1304-7. Davey PG, McGuinness EE, Nicol S, Wormsley KG. Experimental colitis and colonic cancer in the rat. Mt Sinai J Med 1985;52:209-11. O’Connor RJ. The healing of localised areas of necrosis in the colon of the mouse. Br J Exp Path01 1954;35:545-50. O’Connor RJ. Mucosal repair in the rectum. Br J Surg 1956; 44:93-7. Braucher RE, Kirsner JB. Regeneration of colon mucosa: a morphologic and histochemical study. Gastroenterology 1962;42:706-17. Braucher RE, Kirsner JB. Regeneration of colon mucosa in large lesions. A morphological and histochemical study. Gastroenterology 1966;50:372-5. Reeve DRE. An electron microscopical study of chronic ulcers of the colon in rats. J Exp Path01 1977;58:63-77. Bank S, Novis BH, Burns DG, Marks IN. Dissecting microscopy of the rectal mucosa in health and disease. Dis Colon Rectum 1973:16:459-67. Buck RC. Ultrastructural features of rectal epithelium of the

mouse during the early phases of migration to repair a defect. Virch Arch [B] 1986:51:331-40. Argenzio RA, Henrikson CK, Liacos JA. Restitution of barrier and transport function of porcine colon after acute mucosal injury. Am J Physiol 1988;255:G62-71. Feil W. Wenzl E. Vattay P. Starlinger M, Sogukoglu T, Schiessel R. Repair of rabbit duodenal mucosa after acid injury in vivo and in vitro. Gastroenterology 1987;92:197386.

21

Morris GP, Wallace JL. The roles of ethanol and of acid in the production of gastric mucosal erosions in rats. Virchows Arch [B] 1981;38:23-38.

22

Rutten MJ. Ito S. Luminal acid effects on reconstitution damaged guinea pig gastric mucosa in vitro. In: Allen

1. Gaginella TS, Chadwick VS, Debongnie JC. Lewis JC. Phillips SF. Perfusion of rabbit colon with ricinoleic acid: doserelated mucosal injury, fluid secretion, and increased permeability. Gastroenterology 1977;73:95-101. 2. Chadwick VS, Gaginella TS, Carlson GL, Debongnie JC, Phillips SF, Hofman AF. Effect of molecular structure on bile acid-induced alterations in absorption. function, permeability, and morphology in the perfused rabbit colon. J Lab Clin Med 1979;94:661-74. 3. Rafter JJ, Eng VWS, Furrer R, Medline A, Bruce WR. Effects of calcium and pH on the mucosal damage produced by deoxycholic acid in the rat colon. Gut 1986;27:1320-9. 4. Camilleri M, Murphy R. Chadwick VS. Pharmacological inhibition of chenodeoxycholate-induced fluid and mucus secretion and mucosal injury in the rabbit colon. Dig Dis Sci

Vol. 97. No. 3

of A,

Flemstrom G, Garner A, Silen W, Turnberg LA, eds. Mechanisms of mucosal protection in the upper gastrointestinal tract. New York: Raven, 1984:41-7. Wenzl E, Feil W, Starlinger M, Schiessel R. Alkaline secretion A protective mechanism against acid injury in rabbit duodenum. Gastroenterology 1987;92:709-15. 24. Hingson DJ, Ito S. Effect of aspirin and related compounds on the fine structure of mouse gastric mucosa. Gastroenterology 1971:61:156-77. 25. Ito S. Karnovsky MJ. Formaldehyde-glutaraldehyde fixatives containing tri-nitro compounds. J Cell Biol 1968:39:168a. 26. Snipes RL. Clauss W, Weber A, Hornicke H. Structural and 23

functional differences in various divisions Cell Tissue Res 1982;225:331-46.

of the rabbit

colon.

27. Clauss W. Schafer H, Horch I, Hornicke H. Segmental differences in electrical properties and Na-transport of rabbit caecum, proximal and distal colon in vitro. Pflugers Arch 1985; 403:278-82. 28. Svanes K. Ito S, Takeuchi K. Silen W. Restitution of the surface epithelium of the in vitro frog gastric mucosa after damage with hyperosmolar sodium chloride. Gastroenterology 1982;82:1409-26. 29. Lacy ER, Ito S. Rapid epithelial restitution of the rat gastric mucosa after ethanol injury. Lab Invest 1984;51:573-83. 30. Buck RC. Cell migration in repair of mouse cornea1 epithehum. Invest Ophthalmol Vis Sci 1979;18:767-81. 31. Radice GP. The spreading of epithelial cells during wound closure in Xenopus larvae. Dev Biol 1980;76:26-46. 32. Rutten MJ, Ito S. Morphology and electrophysiology of guinea pig gastric mucosal repair in vitro. Am J Physiol 1983: 244:G171-82. 33. Feil W. Lacy ER, Wong M, et al. Rapid restitution of the rabbit and human colonic mucosa after superficial damage (abstr). Gastroenterology 1988:94:A124. 34. Lacy ER. Gastric mucosal defense after superficial injury. Clin Invest Med 1987;10:189-200. 35. Dzienis H, Grenbech JE, Varhaug JE, Lekven J, Svanes K. Regional blood flow and acid secretion associated with damage and restitution of the gastric surface epithelium in cats. Eur Surg Res 1987;19:98-112. 36. Starlinger M, Matthews J, Yoon Ch, Wenzl E, Feil W, Schiesse1 R. The effect of acid perfusion on mucosal blood flow and intramural pH of rabbit duodenum. Surgery 1987;101:433-8. 37. Von Engelhardt W, Rechkemmer G, Luciano L, Reale E. The hindgut: lessons for man from other species. In: Barbara L Miglioli M, Phillips SF, eds. New trends in pathophysiology and therapy of the large bowel. Amsterdam: Elsevier Science Publishers, 1983:3-17. K, Ito S, Silen W. Effect of luminal pH 38. Svanes K, Takeuchi and nutrient bicarbonate concentration on restitution after gastric surface cell injury. Surgery 1983;94:494-500. 39. Ito S, Lacy ER. Morphology of rat gastric mucosal damage,

September

1989

defense,

and

RAPID RESTITUTION

restitution

Gastroenterology 40.

Morris

GP, Harding

extracellular

in the presence

of luminal

ethanol

1985;88:250-60. RK, Wallace mucus

gastric

JL. A functional

in the

rat.

model

for

Arch

(B]

Virchows

1984:46:239-51. 41. Lacy

ER. Gastric

insult. 42.

Stand

Wallace

JL, Whittle

epithelial

damage

by mucolytic 43.

mucosal

Vattay

BJR. Role of mucus

agents.

Gastroenterology

P, Feil W, Klimesch is passive

to a repeated

1985;2O(Suppl

in the rat: inhibition

se1 R. Acid-stimulated num

resistance

J Gastroenterol

and

S. Wenzl

ethanol

110):63-72.

in the repair

of gastric

of epithelial

recovery

1986;91:603-11. E, Starlinger

alkaline

secretion

correlates

with

M, Schies-

in the rabbit mucosal

duode-

damage.

Gut

1988;29:284-90. 44. Wallace JL. Increased resistance of the gastric mucosa to hemorrhagic damage after exposure to an irritant. Gastroenterology

1988;94:22-32.

OF COLONIC

MUCOSA

45. Grant R. Rate of replacement of the surface epithelial the gastric mucosa. Anat Ret 1945;91:175-85.

701

cells of

Received June 30. 1988. Accepted April 10, 1989. Address requests for reprints to: Dr. Wolfgang Feil, University Clinic of Surgery I, Vienna General Hospital, Alser Strasse 4, A-1090 Vienna. Austria. The experimental work was supported by a grant from the “Anton Dreher-Gedlchtnisschenkung des Medizinischen Dekanats der Universitat Wien.” Wolfgang Feil is an international research fellow partially supported by a grant from the Austrian “Bundesministerium fur Wissenschaft und Forschung.” The authors gratefully acknowledge the excellent technical assistance of Kathy Cowart and Jan King and the secretarial work of Marion Hinson. Preliminary results of this work were published in abstract form (Gastroenterology 1988;94:A124).