- Email: [email protected]
Prostaglandin cytoprotection against ethanol-induced gastric injury in the rat
GASTROESTEROLOGY
1985;88:649-59
Prostaglandin Cytoprotection Against Ethanol-Induced Gastric Injury in the Rat A Histologic
and Cytologic
Study
KARMEN L. SCHMIDT, JULIA M. HENAGAN, GREGORY S. SMITH, PAMELA J. HILBURN, and THOMAS A. MILLER Department of Pathology of Texas Medical School
and Laboratory Medicine, and Department at Houston, Houston, Texas
Using macroscopic criteria for injury, prostaglandins have been alleged to possess potent antiulcer properties despite meager histologic evidence for this cytoprotective action. This time-sequence study used light, scanning, and transmission electron microscopy to evaluate the effects of 16,16-dimethyl prostaglandin Ez on gastric mucosal integrity after exposure to 100% ethanol. Macroscopically, virtually complete protection against injury to the glanduJar mucosa of the in vivo rat stomach was noted in animals receiving 10 pgikg body wt of prostaglandin subcutaneously before oral ethanol administration when killed at 5, 20, and 60 min after ethanol exposure compared with oral ethanol after saline injection. On light microscopy the length of injured epithelium in prostaglandinlethanoland saline1 ethanol-treated tissues was not significantly different at all time periods studied. Although the depth of injury extended into gastric glands in both groups killed at 5 min, the deep pit surface mucus cells in prostaglandiniethanol mucosa were less damaged and necrotic lesions were virtually absent when compared with saline/ethanol mucosa. At 20 and 60 min, cellular injury could still be identified in prostaglandiniethanol-treated mucosa but the depth of injury became even less pronounced over time in Received April 25. 1983. Accepted September 20, 1984. Address requests for reprints to: Thomas A. Miller, M.D.. Department of Surgery, The University of Texas Medical School, 6431 Fannin, Houston, Texas 77030. This work was supported by Research Grant AM 25838 [to Dr. Miller) from the National Institutes of Health. The authors thank Dr. John Pike of the Upjohn Company, Kalamazoo, Michigan for supplying the prostaglandin analogue. They also thank Inci Akkaya and Rebecca Fullen for secretarial assistance and Udayan Parikh for electron-microscopy technical . expertise. c 1985 by the American Gastroenterological Association 0016-5085/85/$3.30
of Surgery,
The University
contrast to mucosa exposed to ethanol without prostaglandin. Scanning electron microscopy and transmission electron microscopy confirmed these differences. Despite the macroscopic findings, these results indicate that prostaglandin does not prevent superficial surface mucus cell necrosis in ethanolexposed mucosa even though it does spare cells in the pit base. The reduction in damaged cells over time in prostagJandinlethanoJ-treated mucosa, in contrast to saline/ethanol-treated mucosa, supports the hypothesis that the reepithelialization of the Jamina propria is initiated by spared deep-lying pit cells. A substantial body of literature has accumulated in recent years suggesting that prostaglandins (PGs) are capable of preventing gastric mucosal ulceration induced by a large number of noxious agents and under a wide variety of experimental conditions independent of their known inhibitory effects on gastric acid secretion (1,Z). This unique property of these fatty acids has been called cytoprotection (1,~). Despite numerous reports (l-6) supporting a role for PGs as cytoprotective agents, histologic data verifying such protection are meager. Most studies evaluating this phenomenon have assessed protection in terms of the absence or reduction in macroscopically visible necrotic lesions in response to PGs. Thus, if a known damaging agent, such as ethanol.* induced the formation of necrotic lesions that could be obAbbreviations used in this paper: prostaglandin E2; LM, light microscopy; scanning electron microscopy: SMC. transmission electron microscopy. * Ethanol this paper.
and alcohol
will be used
dmPGE,. 16.16-dimethyl PG. prostaglandin; SEM. surface mucus cell: TEM,
interchangeably
throughout
GASTKOENTEROLOGY
650 SCHMIDT ET AL.
served macroscopically, and pretreatment with a given PG prevented such lesion formation, the PG being tested is said to be cytoprotective. A major problem in assessing cytoprotection in this fashion is that what might be perceived as damage or protection macroscopically may not correspond to what can actually be observed microscopically. In a recent study by Lacy and Ito (73, the first comprehensive histologic evaluation of cytoprotection was reported and the difficulty in judging the absence or presence of this phenomenon on purely macroscopic grounds was emphasized. In studies on the rat gastric mucosa treated with PG and subsequently challenged with ethanol, they observed on light microscopy, using semithin sections, that although PG was able to reduce the longitudinal red streaks and apparent necrotic lesions associated with alcohol-induced gastric injury when mucosa was visualized with the naked eye, areas of mucosa presumed to be normal macroscopically with PG treatment showed extensive cytologic disruption which was often indistinguishable from mucosa exposed to ethanol alone. In an effort to explore further the effects of prostaglandin on gastric epithelium when exposed to ethanol, and to correlate macroscopic findings with those observed microscopically, the present study was undertaken using light (LM), scanning (SEM), and transmission electron microscopic (TER4) techniques.
Materials and Methods Female Sprague-Dawley rats weighing -200 g were housed in cages with wide mesh bottoms to prevent coprophagia. The animals were fasted for 48 h and allowed free access to water during the fasting period. On the day of experimentation, animals were randomized into various groups and injected subcutaneously with 10 pgikg body wt of 16,16-dimethyl prostaglandin E2 (dmPGE2) or a comparable volume of saline. Thirty minutes later, each animal received an oral bolus of 1.0 ml of absolute ethanol or saline. Animals were killed at 5, 20, or 60 min after oral ethanol or saline administration to determine the cytologic and histologic effects of each treatment regimen over the course of time. Table
1. Criteria
Depth of damage
Type 0 Type 1 Type 2 Type 3 Type 4
for Depth
of Injury”
Damage None Luminal surface mucus cells only Luminal surface and upper gastric pit mucus cells Luminal surface and gastric pit mucus cells and some gastric gland cells Necrotic lesions: Severe mucosal injury to all surface and all or most glandular epithelium
’ Modification
of a scoring system devised by Lacy and Ito (7).
Vol. 88, No. 3
At all times of death after oral ethanol or saline exposure, rats were quickly anesthetized with ether and opened via a midline laparotomy. A silk ligature was tied around the lower esophagus and duodenum and 2.0 ml of halfstrength Karnovsky’s fixative (8) was injected into the gastric lumen through a puncture wound along the lesser curvature. The stomach was then excised and immersed in fixative. After fixation, the stomach was opened along the lesser curvature, laid flat, and examined for evidence of gross macroscopic damage by an observer unfamiliar with the experimental protocol. An ulcer index was assigned in which the percentage of glandular gastric mucosa injured was assessed. This was accomplished by measuring the width of each opened stomach and from this measurement visually dividing the exposed mucosa into two equal parts. As gross damage was confined to the glandular epithelium (oxyntic and antral mucosa; see Results), a visual estimate of the percentage surface area of the glandular stomach damaged in each part was made. The total surface area of glandular mucosa damaged in each stomach was calculated by summating the individual surface areas of each part. Samples for LM, SEM, and TEM were excised from the gastric glandular epithelium at a region located 2 mm below the limiting ridge (which separates the forestomach from the glandular epithelium) along the greater curvature of the stomach. Blocks were removed at right angles to the long axis of the stomach. A light microscopic sample measuring -1 mm x 10 mm was removed alongside a sample measuring 1 mm x 4 mm. The latter sample was for SEM subdivided into a l-mm x 3-mm block processed and a l-mm x l-mm block processed for TEM. The same procedure was performed on a second sample that was removed at -0.5 cm distal to the first sampling area. Thus, two LM, SEM, and TEM blocks were removed and processed from each stomach. Both samples were examined by the same individual (K. L. S.]. Routine paraffin-embedded hematoxylin and eosinstained sections and toluidine blue-stained, semithin plastic sections (0.25 Frn) were used for LM evaluation. Slides were viewed in a Zeiss photomicroscope (Carl Zeiss, Inc., Thornwood, N.Y.). Tissue blocks prepared for SEM were postfixed in aqueous 1.0% osmium tetroxide, dehydrated, and critical-point dried in a Bomar SPC 1500. The tissues were then gold-palladium-coated using a Denton Desk-I vacuum coater (Denton Vacuum Inc., Cherry Hill, N.J.) and viewed in a JEOL JSM-35 scanning electron microscope. Those tissues prepared for TEM were postfixed in cacodylate-buffered 2.0% osmium tetroxide, dehydrated, and embedded in epoxy resins. Semithin sections were made for orientation, followed by ultrathin sectioning using an LKB Ultratome III. Sections were viewed in a Siemens 102 electron microscope (Siemens Corp., Iselin, N.J.). Using hematoxylin and eosin-stained paraffin-embedded sections of tissue, the extent of mucosal injury was quantified. Slides were examined in a blinded fashion by coding them in such a way so that the examiner (K. L. S.) was unfamiliar with the experimental protocol. Only after examination was complete were they decoded. Using a graded micrometer eyepiece, within each tissue block pair from each stomach, the overall length of each tissue section was measured and the corresponding percentage of
HISTOLOGY AND CYTOLOGY OF CYTOPROTECTION
March 1985
Table
2. Percent Damage of Glandular Epithelium on GrossVisual Inspection at Various Time Periods of Death After 100% Ethanol Exposure’
Time of death (min)
100% Ethanol without PG 19.5 2 5.1 38.3 + 5.3 34.7 '- 3.7
5 20 60
100% Ethanol with PG 2.2 2 1.4[' 0" Oh
PG = 16,16-dimethyl prostaglandin EZ (10 &kg body wt subcutaneously). a n = 6 for each experimental group. ‘p C: 0.01 compared with 100% ethanol without prostaglandin. the length of mucosal surface injured was determined. The extent of injured mucosa from luminal surface to muscularis mucosa was further assessed using a modification of the criteria for depth of injury established previously by Lacy and Ito (7) and summarized in Table 1. Surface mucus cells were considered injured if one or more of the following criteria were present: marked cytoplasmic vacuolization, cytoplasmic swelling, nuclear pyknosis, or nuclear swelling with chromatin margination. Lucent cytoplasm and pyknotic nuclei indicated parietal cell injury. Dilation of glands, hyperemic vessels, and hemorrhage were additional determinants of glandular injury. In addition to these measurements, the left wall of every 10th gastric pit from each block was evaluated until a total of 10 pits per section had been assessed to ascertain subtle changes in the depth of gastric pit injury. This evaluation consisted of the micrometric measurement of the ratio of iniured versus normal pit mucus cells. Only those pits that were sectioned in the vertical plane with the entire lumen fully exposed were included in the determination. Paired measurements for length of mucosal surface injured as well as both the mucosal depth of injury and gastric pit depth of injury were averaged to obtain a single value for each of these indices of damage for each stomach, All values obtained were reported as means * SER?. Differences among the various experimental groups were evaluated statistically by analysis of variance. A p value of CO.05 was considered to represent statistical significance.
651
at all time periods of nonsecreting forestomach death. These lesions consisted of elongated red bands ranging from 2 to 10 mm in length and from 1 to 4 mm in width, generally parallel to the long axis of the stomach and mostly confined to the corpus (the portion of the stomach secreting acid and pepsin) and less commonly encountered in the antrum. In terms of total surface area, these lesions involved about 20% of the glandular epithelium in animals killed 5 min after alcohol exposure, and 35%38% of the mucosa in those killed at 20 and 60 min (Table 2). In animals receiving PG and ethanol (10 pgikg body wt & ‘dmPGE, inje%on followed by 100% ethanol administration orally), macroscopic evidence of damage was virtually absent regardless of the length of ethanol exposure. In the 2 animals in the 5-min sacrifice group in which some iniury was
visually present, it was in the form of petechiae small punctate lesions [Table 21.
Microscopic
and
Findings
Control tissues. (Saline injection followed by saline administration orally]: On LM at all time periods, the surface mucus cells (SMCs) formed a continuous epithelial sheet that lined the gastric
Results Macroscopic
Findings
No macroscopic
evidence
of mucosal
injury
was noted in any of the control stomachs (saline injection followed by saline administration orally) or
in stomachs exposed to PG alone (10 pgikg body wt of dmPGE2 injection followed by saline administration orally) regardless of the time of death. In contrast, stomachs exposed to ethanol alone (saline injection followed by lOOoh ethanol administration orally) showed obvious evidence of mucosal injury. In each of the stomachs treated with ethanol alone! necrotic hemorrhagic lesions were noted in the glandular portion of the stomach with sparing of the
Figure 1 Light microscopy of gastric mucosa from a salineinjected animal taken 5 min after oral saline. Luminal surface mucus cells and pit mucus cells (arrows) are intact, columnar or cuboidal in shape, and contain a single nucleus and varying amounts of mucin in the apical cytoplasm. Parietal cells (PC) are present in the glands. Plastic embedded, toluidine blue stain. (x600.)
652
Figure
SCHMIDT ET AL.
2. Scanning electron microscopy of gastric mucosa from a saline-injected animal taken 60 min after oral saline. The mucosal surface is formed by closely packed. polygonal surface mucus cells that are periodicall> interrupted by the round or slitlike openings of the gastric glands (arrows). Adherent granular or filmy material is mucus (double arrow,). (x280.)
surface and pits. At the tip of the mucosal surface, an occasional cell with a pyknotic nucleus and vacuolated cytoplasm was observed; such cells were presumed to be undergoing the normal exfoliative process. Most SMCs were intact and cuboidal or columnar in shape, with varying amounts of apical mucin granules. Mucin granules were visualized as a thin, dark-staining crescent in the apical surface of these cells (Figure 1). Gastric pits were of normal depth and gastric glands were formed of normal parietal, chief, endocrine, and mucus neck cells. On SEM, the luminal surface of the gastric mucosa was formed of closely packed SMCs and the round or slitlike openings of the gastric pits (Figure 2). On TEM, the SMCs were cuboidal to columnar in shape and contained numerous intensely stained, apical mucin granules in addition to the usual complement of cell organelles (Figure 3). No injury among control tissues obtained at the various times of death after oral saline administration could be detected by histologic examination except a rare, minute region where the surface epithelial cells were absent. These regions, when pres-
(:ASTKOENTEKOLO(;T
Figure
Vol. 88, No. 3
3. Transmission electron microscopy of gastric mucosa from a saline-injected animal taken 60 min after oral saline. Surface mucus cells are typically cuboidal to columnar in shape, contain a single nucleus (N) and no unusual organelles except numerous mucin granules (orrows) in the apical cytoplasm. (~5700.)
ent, generally involved -Cl?; of the surface epitheliurn (Table 3) and were graded as type I injury. Ethanol-exposed tissues. (Saline injection followed by 100% ethanol administration orally]: Extensive lesions of the gastric mucosa were noted on LM after alcohol exposure regardless of the time of death, with virtually the entire length of epitheliurn examined showing evidence of injury (Table 3).
Table
3. Percentage Length of Smface Damaged Microscopically” 5 min”
SAL-SAL PG-SAL SAL-&OH PG-EtOH
0.1 k 0.3 + 97.6 i 98.9 i
Epithelium
20 min” 0.1 0.2 1.4 0.5
1.5 i 1.0 r 94.9 i 85.3 +
0.6 0.6 2.2 5.8
60 min” 0.4 -t 0.2 0.9 + 0.9 97.1 2 1.3 91.7 -t 2.9
SAL-SAL, saline injection followed by oral saline. PG-SAL, 16.16dimethyl prostaglandin E2 (10 pg/kg body wt) injection followed by oral saline. SAL-EtOH. saline injection followed by 100% ethanol orally. PG.EtOH. 16.16-dimethyl prostaglandin E, (10 pgi kg body wt) injection followed by loo%, ethanol orally. ” n = 6 for each experimental group. “Time of death after 100% ethanol exposure.
March 1985
HISTOLOGY
SALIEIOH WOO25
PG/EIOH
p
5 min D
1
NOM
1
xl
2
p
1
C
p
---
2
p
p
T
40
7
s 20 mm
s 5
20
E x & a
p
NS
NS
P
.--
40
60 mm 20 NOW i
4 Category
Figure
of Injury
4. Effect of prostaglandin pretreatment on the depth of gastric mucosal injury. Animals were killed at 5, 20, and 60 min after ethanol injury. n = 6 animals for each group. NS, not significant: SALIEtOH, saline injection followed by ethanol orally; PGIEtOH. 16,16-dimethyl prostaglandin E, injection followed by ethanol orally.
AND
CYTOLOGY
OF CYTOPROTECTION
effects of ethanol on gastric epithelium are shown in Figures 6 and 7. This extensive erosive process was also evident on SEM regardless of the time of death. In many regions, the SMCs were sloughed, exposing the underlying lamina propria and the outlines of the denuded gastric foveolae. Masses of exfoliated cells and other debris were observed on the mucosal surface, obscuring the gastric pit openings and normal mucosal morphology [Figure 8). On TEM examination, the dense aggregation of luminal material included cells with organelles in various stages of degeneration along with extensive quantities of mucus, red blood cells, and other debris. Many of the adherent SMCs were injured to varying degrees as evidenced by pyknotic, crenated nuclei and concentric rings of rough endoplasmic reticulum [Figure 9), some of which exhibited dilated cisternae. High amplitude swelling and floccular densities in mitochondria were present in many of these cells. Prostaglandin-exposed tissues. [16,16-Dimethyl prostaglandin E, (10.0 pgikg body wt) injection followed by saline administration orally]: Stomachs of the dmPGEp-treated animals that received an oral bolus of physiologic saline were morphologically very similar to control tissues despite the time of death after saline exposure. Similar to control tissues, a rare, minute region where surface epithelial
SALlEtOH PG/EtOH p
100
p
In many regions, surface mucus cells were sloughed into the gastric lumen as an intact epithelial sheet. Exfoliated parietal cells and other debris also were observed within the gastric lumen. The depth of injury, however, varied with the length of alcohol exposure. At 5 min, after ethanol exposure, mucosal damage was characterized largely by type 3 and type 4 lesions (Figure 4). In addition, the average depth of gastric pit injury was -90% (Figure 5). By 20 min after ethanol administration, a slight shift in the depth of injury occurred with a reduction in type 3 lesions and an increase in type 2 (p < 0.05 compared with 5-min data) even though the percentage of type 4 lesions remained unchanged (Figure 4). A slight decrease in the depth of gastric pit injury (80%) was also observed at 20 min (Figure 5). Findings at 60 min were virtually unchanged from those observed at 20 min. At 60 min, >60% of damaged mucosa was still characterized by type 3 and type 4 lesions [Figure 4) and the average pit depth injury remained at 80% (Figure 5). Micrographs demonstrating the
653
‘1
80,
0
5
min Time
Figure
p
20
min
of
Sacrifice
30
mm
5. The depth of gastric pit injury in mucosa exposed to absolute ethanol with and without prostaglandin pretreatment. Animals were killed at 5, 20. and 60 min after ethanol exposure. n = 6 animals for each group. SALIEtOH, saline injection followed by ethanol orally; PGIEtOH, 16,16-dimethyl prostaglandin E2 injection followed by ethanol orally.
654
SCHMLDT
GASTKOENTEKOLOCY
ET Al,
cells were absent could be occasionally identified that involved 51% of the surface epithelium (Table 3). Such regions were very superficial and were graded as type 1 injury. Prostaglandiniethanol-exposed tissues. (16,16Dimethyl prostaglandin E, injection followed by loo%, ethanol administration orally): Animals receiving 10.0 pgikg body wt of dmPGEp subcutaneously before ethanol exposure were noted to have a different pattern of gastric mucosal injury when compared with those exposed to ethanol without dmPGEz pretreatment. Although the length of mucosal surface injured in animals receiving PG pretreatment before ethanol exposure was not significantly different from mucosa exposed to ethanol without PG pretreatment at any time of death (Table 31, the depth and type of injury were considerably altered if PG had been administered. More than 87% of the mucosal injury in the PC/ethanol-treated group was confined to types 2 and 3, with type 4 damage being totally absent at the 5-min time of death (Figure 4). Further, the mean depth of gastric pit injury was 60% in contrast to 90% in animals exposed to
‘V’ i
Figure
(I
Figure
6. Light microscopy ot gastric mucosa from a salineinjected animal taken 5 min after oral ethanol. Features of type 3 injury are represented. Surface mucus cells form an incomplete covering of the mucosal surface and those of the gastric pits demonstrate pyknotic nuclei and crowding (orrows). Although parietal cells deep in the glands appear normal, those at the surface (PC) have advanced toward the lumen. Plastic embedded, toluidine blue stain. (x800.)
Vol. 88. No. 3
7. Light microscopy of gastric mucosa from a salineinjected animal taken 5 min after oral ethanol showing the features of a type 4 or necrotic lesion as judged microscopically. Damage to cells is severe in the upper portion of the mucosa. Surface mucus cells are absent and parietal cells (PC) are sloughed into the lumen. Extensive hyperemia and hemorrhage are evident in the mucosal vessels and surface (arrorv). Plastic embedded, toluidine blue stain. (x800.)
ethanol without PG pretreatment (Figure 5). By 20 min, mucosal injury in the PGiethanol group was predominantly types 1 and 2 damage (73%). Again, type 4 or necrotic injury was virtually absent except for one focal region of necrotic tissue in one stomach sample (Figure 4). The mean gastric pit injury at 20 min was 30% (Figure 5). By 60 min after ethanol treatment, mucosal injury had shifted further toward types 1 and 2 injury (86%), whereas the mean gastric pit damage had fallen to 20% (Figures 4 and 5). Again, necrotic lesions were not observed at 60 min. Light micrographs depicting effects of PGiethanol on gastric mucosa are presented in Figures 10 and 11. Despite differences in the depth and severity of mucosal damage at the 5-min time of death, uninjured, minimally injured, severely injured, and necrotic SMCs and gland cells were observed in mucosa from both PGiethanoland the salineiethanol-treated groups on LM (Figures 6, 7, 10, and ll), SEM, and TEM evaluation. By SEM, the surface mucosa of PGiethanol-treated tissues at the 5-min time of death was morphologically similar to that
March
Figure
1985
8. Scanning electron microscopy of gastric mucosa from a saline-injected animal taken 60 min after oral ethanol. A portion of the lamina propria bearing outlines of gastric pit openings is exposed and largely devoid of epithelium (arroM.s). The remaining mucosal surface is covered by cells in various degrees of exfoliation as evidenced by obliteration of the gastric pit openings and loss of normal surface architecture. Compare with a similar region from saline/saline-exposed mucosa (control. Figure 2). (X 280.)
observed in the 5-min time of death in the saline/ ethanol-treated group. In some regions mucus cells were exfoliated from the interfoveolar surface, exposing the underlying lamina propria. In other areas, the exfoliated cells and other debris remained on the surface and were sufficiently dense to obscure the gastric pit openings. Minimally injured cells were difficult to identify on LM but on TEM were typified by swelling of mitochondrial matrices, margination of chromatin, and ring formation of rough endoplasmic reticulum. Severely injured cells could be identified on LM by the presence of pyknotic nuclei and vacuolated cytoplasm. On TEM, swollen mitochondria with floccular densities, in addition to the above cytologic alterations, marked the severely damaged cells. Despite these cytologic changes, junctional complexes between cells adherent to the basal lamina remained undissociated in PC/ethanoland saline/ethanol-treated groups. By 20 min after ethanol administration, many cells within the gastric pit and the mucosal surface ap-
HISTOLOGY
Figure
AN)
CYTOLOGY
OF (:1’TOPKOTECTIOS
655
9. Transmission electron microscopy elf gastric mucosa from a saline-injected animal taken 60 min after oral ethanol. A severely injured surface mucus vt!ll contains a pyknotic nucleus (N). In portrons of other cells. concentric rings of rough endoplasmi~ reticulum (orroTz,sJ are present. (X8800.)
peared normal or attenuated by LM in the samples treated with PGiethanol. This phenomenon was even more pronounced by 60 min, when rounded and attenuated SMCs were observed on the gastric surface (Figure 12). By SEM, these cells appeared to be emerging from the gastric pits and were intermingled with regions that were covered by attenuated cells or bare lamina propria. Although initiated, reepithelialization was not complete in the PGiethanol group by 60 min and focal zones of bare lamina propria indicative of type 3 injury were still present (Figures 13 and 14). By TEM, the adherent SMCs in regions of repair appeared almost normal (Figure 15). An occasional heterophagic vacuole or ring formation of rough endoplasmic reticulum was the only feature indicating cytologic alterations.
Discussion In an earlier study by Lacy and Ito (7), a clear disparity was demonstrated when assessing the cytoprotective effects of dmPGEz on ethanol-exposed mucosa as judged macroscopically with the naked eye and histologically with LM. It was noted in that study that all rats receiving absolute ethanol on examination of macroscopically nonnecrotic regions
Figure
Figu re
10. Light microscopy of gastric mucosa from a prostaglandin-injected animal taken at 5 min after oral ethanol showing focal regions of type I injury. Only luminal surface mucus cells are damaged in type I injury as evidenced by pyknotic nuclei and cytoplasmic vacuolization (arrow]; pit and parietal cells are normal appearing. A few surface mucus cells are exfoliating from the mucosal surface (double arrow). PC, parietal cells. Plastic embedded, toluidine blue stain. (X 800.)
Figure
11. Light microscopy of gastric mucosa from a prostaglandin-injected animal taken 5 min after oral ethanol and showing type z damage. Surface mucus cells with pyknotic nuclei and vacuolated cytoplasm are injured (arrows] to varying depths within the gastric pit. PC, parietal cells. Plastic embedded, toluidine blue stain. (X800.)
Light microscopy of gastric mucosa from a prostaglandin-injected animal taken 60 min after oral ethanol showing a region of epithelial repair. Attenuated surface mucus cells (double arrow) flatten to cover the underlying lamina propria (LP) adjacent to an area of connective tissue that appears to be exposed to the gastric lumen (arrow]. Surface mucus cells in m ost other regions appear normal in cytologic detail and distribution. PC, parietal cell. Plastic embedded, toluidine blue stain. I[x1100.)
March
1985
of the stomach, with or without PG treatment, had about 78% of the total area of the glandular portion of the gastric mucosal surface damaged when examined microscopically at 15 and 30 min after ethanol exposure. On macroscopic inspection, though, PG was felt to have rendered almost complete protection against alcohol injury at both time periods. Prostaglandin treatment, however, reduced the depth of all cells from damage by about 20% when evaluating the luminal surface through the gastric glands, and the necrotic lesions observed in mucosa exposed to ethanol alone were virtually eliminated. Wallace et al. (9) reported similar findings in a rat chambered stomach preparation when assessing the protective effects of the natural PC, PGE2, against gastric injury induced by 40% ethanol. We also observed in the present study that the presumed virtually complete protection against ethanol injury by PG when evaluated macroscopically could not be confirmed histologically at any of the three time periods examined after alcohol exposure. Such observations point out
Figure
13. Scanning electron microscopy of gastric mucosa from a prostaglandin-injected animal taken 60 min after oral ethanol. A focal region showing denuded lamina propria (L) with gastric pit openings is shown surrounded by a mucosal surface which is undergoing reepithelialization. Rounded surface mucus cells (white arrows) bulge into the lumen. Other surface mucus cells are flattened and attenuated (bluck orrow). a feature consistent with migrating cells or cells that extend their boundaries to cover underlying connective tissue components. (X280.)
HISTOLOGY
Figure
AS11 C:Y’f()LOG\r’ OF (:TI’TOPRO~E(:TION
657
14. Scanning electron microscopy of gastric mucosa from a prostaglandin-injected animal taken 60 min after oral ethanol showing the mouth of a gastric pit from a region of surface regeneration. Rounded surface mucus cells have emerged from the pit region although some still possess flattened cvtoplasmic surfaces common to migrating cells (nrrow). Cells in the background are attenuated (double orron) possibly to cover the underlying connective tissue stratum. GP. gastric pit. (X 3000.)
the importance of assessing the presence or absence of gastric injury on histologic grounds and emphasize the shortcomings of evaluating cytoprotection on purely macroscopic criteria. On histologic comparison of mucosa exposed to ethanol with and without PG pretreatment. we noted that the severe gastric hemorrhagic erosions (type 4 or necrotic lesions] so readily apparent microscopically in mucosa exposed to absolute ethanol alone were virtually eliminated in animals receiving PG pretreatment, confirming the previous findings of Lacy and Ito (7) and Wallace et al. (9). As much as SO%-40% of the gastric epithelium revealed evidence of necrotic lesions (type 4) in alcohol-exposed mucosa without PG pretreatment that was not altered over the 60 min of evaluation, whereas only one stomach (during the 20-min time of death) exposed to ethanol after PG pretreatment demonstrated evidence of one small necrotic lesion. Using the same PG (dmPGE2) as Lacy and Ito (7), we further observed no difference between the length of mucosal surface damaged in rats receiving absolute alcohol with and without PC pretreatment at 5, 20, or 60 min as quantified by LM, similar to what these investigators reported at 15 and 30 min after ethanol exposure. In contrast to their study, though, clear differences between the two experimental groups
658
Figure
SCHMIDT
ET AL.
15. Transmission electron microscopy of gastric mucosa from a prostaglandin-injected animal taken 60 min after oral ethanol. Several cells from a region of epithelial repair are either normal or show minor injury: concentric rings of rough endoplasmic reticulum (arrow) and one heterophagic vacuole (double arrow] are present in one cell. Minor margination and central clearing of chromatin is present within the nuclei (N). MG, mucin granules. (~5700.)
could be demonstrated with respect to the depth of mucosal injury over time. Although the distribution of type 1 and type 3 lesions was initially identical in animals killed at 5 min whether or not PG pretreatment was rendered, type 2 lesions were more prominent in PG-pretreated stomachs. In animals killed at 60 min, type 3 and type 4 injury still accounted for >60% of the lesions observed in alcohol-exposed mucosa without PG pretreatment, whereas type 1 and 2 lesions (>80%) were most prominent in animals killed at a similar time period in which PG had been rendered. The depth of pit injury was also different between the two experimental groups at all time periods studied. The explanation for the difference between the present study and that of Lacy and Ito (7) is uncertain. Although differences in the techniques used for specimen preparation may in part account for the discrepancy, it is unlikely that such differences alone are responsible for this disparity. Other considerations include the standardized area of epitheliurn that was evaluated microscopically in our studies in contrast to the separate analyses of necrotic
GASTROENTEROLOGY
Vol.
88, No. 3
and nonnecrotic regions evaluated by Lacy and Ito (7). Further, in this study the dose of PG used was 10 pg/kg body wt given parenterally in contrast 10 the 4pg/kg body wt dose given orally by these previous investigators. An additional possible explanation for our findings may be related to the time element involved. We sampled tissue at three different time periods from 5 to 60 min, whereas Lacy and Ito (7) examined only two time periods at 15 and 30 min. Although we saw differences in the magnitude of gastric damage by alcohol that was clearly less at 5 and 20 min if PG had been given, in contrast to the observations of Lacy and Ito at 15 and 30 min, these differences were most pronounced at the 60-min sampling time, a time period that Lacy and Ito did not evaluate. It is interesting to postulate that during an hour time period, spared pit mucus cells could move up the pit wall in an effort to reconstitute the exfoliated cell population, an event that would explain the reduced proportion of injured surface mucus cells observed with PG pretreatment over time. The observation in our studies that the depth of pit injury in response to ethanol exposure was significantly less in mucosa pretreated with PG supports this hypothesis. Further, the findings of Lacy and Ito, as well as our own findings, in which PGs were noted to reduce the depth of mucosal injury, also seem consistent with such an explanation. Thus, reduced depth of injury with sparing of deep-lying pit and mucus neck cells could serve as a cellular pool for rapid reconstitution by cell migration. It seems doubtful that an increased mitotic rate over an hour’s duration by the deeplying mucus neck cells could account for any significant reconstitution, although the present study did not address that consideration. The additional feature of absence of necrotic hemorrhagic erosions in ethanol-exposed mucosa after PG pretreatment may enhance the reconstitution process as it has previously been shown that an intact lamina propria and basal lamina would greatly expedite the attempt by pit cells to cover an epithelial defect (10). Experimental evidence for such a reconstitution process in ethanol-damaged mucosa has been reported. Using a rat chambered stomach preparation, Morris and Wallace (11) were the first to demonstrate epithelial reconstitution after exposure to 40% ethanol and noted cell migration as early as IO min after alcohol administration that was quite advanced by 1 h. They emphasized the importance of an intact basal lamina as scaffolding for such reepithelialization to occur. Using an in vivo rat stomach preparation in which the esophagus and duodenum were ligated, Lacy and Ito (12) exposed the gastric epithelium to absolute ethanol for 30-45 s via an injection in the nonglandular portion, after which the alcohol
March
1985
was removed. Stomachs examined microscopically immediately after this insult revealed damage to 98% of the glandular gastric surface. Examination of the stomach 30 min after exposure to ethanol revealed that >75% of the mucosal surface was reconstituted with surface mucus cells. Although it might be argued that our findings are in conflict with these observations, it must be stressed that our studies differ in a number of ways from the design of these previously published experiments. First, whereas the concentration of ethanol used in the present study was identical to that used by Lacy and Ito (1 Z), the mucosal exposure was considerably longer (even though the speed with which ethanol was emptied into the duodenum is not known) than 30-45 s, presumably resulting in more profound injury and thus requiring a longer period of time for reconstitution of the epithelium to occur. For similar reasons, and because of the higher concentration of alcohol used in the present investigation, the reconstitution observed by Morris and Wallace (11) over a 1-h time period with 40% ethanol is not comparable to our findings. Despite these differences, it is noteworthy that some reconstitution could even be identified in our studies in animals exposed to ethanol in which PG pretreatment was not rendered. By 60 min, for example, the percentage of type 3 lesions had significantly decreased and the percentage of type 2 lesions had significantly increased when compared with corresponding values at 5 min. It seems clear from our studies that two conclusions can be reached. First, although PG was unable to prevent injury to the gastric epithelium, at least in response to absolute ethanol, it did significantly reduce the depth of injury as exemplified by the absence of necrotic lesions and the prominence of type 2 injury at the 5-min time of death which was not observed in animals exposed to ethanol without receiving PG pretreatment. Thus, although the concept of cytoprotection, as originally defined (2-4), which refers to complete absence of cellular damage when exposed to an injurious agent, could not be supported in our studies, protection of deep layers by PG was clearly present. Whether one wishes to redefine cytoprotection as protection against deep injury, as Robert has done (l3), or use another term, seems to us to be of academic interest only. The fact remains that PG significantly reduced the depth of injury in gastric mucosa exposed to absolute ethanol and in that sense was clearly protective. Second, although mucosa exposed to absolute ethanol, with and without PG pretreatment, demonstrated evi-
HISTOLOGY
AND CYTOLOGY
OF CYTOPROTECTION
659
dence of healing over a 1-h time period, as demonstrated by a reduction in the depth of injury when compared with ti-min findings, this healing process was markedly accelerated in mucosa pretreated with PG. Whether this relates to the prevention of injury to pit mucus neck or mucus cells, or both, and thereby enhancement of reepithelialization, or to a direct effect of PG on cell mitosis or cell migration is unknown. The mechanisms responsible for these observations must await further study.
References 1. Miller TA. Protective effects of prostaglandins against gastric mucosal damage: current knowledge and proposed mechanisms. Am J Physiol 1983;245:G601-23. 2 Robert A. Cytoprotection by prostaglandins. Gastroenterology 1979;77:761-7. 3
4
5
6
7
8
Robert A. Antisecretory, antiulcer, cytoprotective and diarrheogenic properties of prostaglandins. In: Samuelsson B. Paoletti R, eds. Advances in prostaglandin and thromboxane research. New York: Raven, 1976:507-20. Robert A. Nezamis JE, Lancaster C, Hanchar AJ. Cytoprotection by prostaglandins in rats: prevention of gastric necrosis produced by alcohol, HCl, NaOH, hypertonic NaCl and thermal injury. Gastroenterology 1979:77:433-43. Bommelaer G, Guth PH. Protection by histamine receptor antagonists and prostaglandin against gastric mucosal barrier disruption in the rat. Gastroenterology 1979;77:303-8. Whittle BJR. Mechanisms underlying gastric mucosal damage induced by indomethacin and bile salts and the actions of prostaglandins. Br J Pharmacol 1977:60:455-60. Lacy ER, Ito S. Microscopic analysis of ethanol damage to rat gastric mucosa after treatment with a prostaglandin. Gastroenterology 1982;83:619-25. Karnovsky MJ. A formaldehyde-glutaraldehyde fixative of high osmolality for use in electron microscopy. J Cell Biol 1965;27:137A-EA.
9
10
11
12
13
Wallace JL. Morris GP, Krausse EJ, Greaves SE. Reduction by cytoprotective agents of ethanol-induced damage to the rat gastric mucosa: a correlated morphologic and physiologic study. Can J Physiol Pharmacol 1982;60:1686-99. Vracko R. Basal lamina scaffold: anatomy and significance for maintenance of orderly tissue structure: a review. Am J Pathol 1974;77:314-38. Morris GP, Wallace JL. The roles of ethanol and of acid in the production of gastric mucosal erosions in rats. Virchows Arch [Cell Pathol] 1981;38:23-38. Lacy ER, Ito S. Ethanol-induced insult to the superficial rat gastric epithelium: a study of damage and rapid repair. In: Allen A, Flemstrijm G. Garner A. Silen W. Turnberg LA, eds. Mechanisms of mucosal protection in upper gastrointestinal tract. New York: Raven, 1984:49-56. Robert A. Role of endogenous and exogenous prostaglandins in mucosal protection. In: Allen A, FlemstrBm G. Garner A, Silen W, Turnberg LA, eds. Mechanisms of mucosal protection in the upper gastrointestinal tract. New York: Raven, 1984:377-82.
1985;88:649-59
Prostaglandin Cytoprotection Against Ethanol-Induced Gastric Injury in the Rat A Histologic
and Cytologic
Study
KARMEN L. SCHMIDT, JULIA M. HENAGAN, GREGORY S. SMITH, PAMELA J. HILBURN, and THOMAS A. MILLER Department of Pathology of Texas Medical School
and Laboratory Medicine, and Department at Houston, Houston, Texas
Using macroscopic criteria for injury, prostaglandins have been alleged to possess potent antiulcer properties despite meager histologic evidence for this cytoprotective action. This time-sequence study used light, scanning, and transmission electron microscopy to evaluate the effects of 16,16-dimethyl prostaglandin Ez on gastric mucosal integrity after exposure to 100% ethanol. Macroscopically, virtually complete protection against injury to the glanduJar mucosa of the in vivo rat stomach was noted in animals receiving 10 pgikg body wt of prostaglandin subcutaneously before oral ethanol administration when killed at 5, 20, and 60 min after ethanol exposure compared with oral ethanol after saline injection. On light microscopy the length of injured epithelium in prostaglandinlethanoland saline1 ethanol-treated tissues was not significantly different at all time periods studied. Although the depth of injury extended into gastric glands in both groups killed at 5 min, the deep pit surface mucus cells in prostaglandiniethanol mucosa were less damaged and necrotic lesions were virtually absent when compared with saline/ethanol mucosa. At 20 and 60 min, cellular injury could still be identified in prostaglandiniethanol-treated mucosa but the depth of injury became even less pronounced over time in Received April 25. 1983. Accepted September 20, 1984. Address requests for reprints to: Thomas A. Miller, M.D.. Department of Surgery, The University of Texas Medical School, 6431 Fannin, Houston, Texas 77030. This work was supported by Research Grant AM 25838 [to Dr. Miller) from the National Institutes of Health. The authors thank Dr. John Pike of the Upjohn Company, Kalamazoo, Michigan for supplying the prostaglandin analogue. They also thank Inci Akkaya and Rebecca Fullen for secretarial assistance and Udayan Parikh for electron-microscopy technical . expertise. c 1985 by the American Gastroenterological Association 0016-5085/85/$3.30
of Surgery,
The University
contrast to mucosa exposed to ethanol without prostaglandin. Scanning electron microscopy and transmission electron microscopy confirmed these differences. Despite the macroscopic findings, these results indicate that prostaglandin does not prevent superficial surface mucus cell necrosis in ethanolexposed mucosa even though it does spare cells in the pit base. The reduction in damaged cells over time in prostagJandinlethanoJ-treated mucosa, in contrast to saline/ethanol-treated mucosa, supports the hypothesis that the reepithelialization of the Jamina propria is initiated by spared deep-lying pit cells. A substantial body of literature has accumulated in recent years suggesting that prostaglandins (PGs) are capable of preventing gastric mucosal ulceration induced by a large number of noxious agents and under a wide variety of experimental conditions independent of their known inhibitory effects on gastric acid secretion (1,Z). This unique property of these fatty acids has been called cytoprotection (1,~). Despite numerous reports (l-6) supporting a role for PGs as cytoprotective agents, histologic data verifying such protection are meager. Most studies evaluating this phenomenon have assessed protection in terms of the absence or reduction in macroscopically visible necrotic lesions in response to PGs. Thus, if a known damaging agent, such as ethanol.* induced the formation of necrotic lesions that could be obAbbreviations used in this paper: prostaglandin E2; LM, light microscopy; scanning electron microscopy: SMC. transmission electron microscopy. * Ethanol this paper.
and alcohol
will be used
dmPGE,. 16.16-dimethyl PG. prostaglandin; SEM. surface mucus cell: TEM,
interchangeably
throughout
GASTKOENTEROLOGY
650 SCHMIDT ET AL.
served macroscopically, and pretreatment with a given PG prevented such lesion formation, the PG being tested is said to be cytoprotective. A major problem in assessing cytoprotection in this fashion is that what might be perceived as damage or protection macroscopically may not correspond to what can actually be observed microscopically. In a recent study by Lacy and Ito (73, the first comprehensive histologic evaluation of cytoprotection was reported and the difficulty in judging the absence or presence of this phenomenon on purely macroscopic grounds was emphasized. In studies on the rat gastric mucosa treated with PG and subsequently challenged with ethanol, they observed on light microscopy, using semithin sections, that although PG was able to reduce the longitudinal red streaks and apparent necrotic lesions associated with alcohol-induced gastric injury when mucosa was visualized with the naked eye, areas of mucosa presumed to be normal macroscopically with PG treatment showed extensive cytologic disruption which was often indistinguishable from mucosa exposed to ethanol alone. In an effort to explore further the effects of prostaglandin on gastric epithelium when exposed to ethanol, and to correlate macroscopic findings with those observed microscopically, the present study was undertaken using light (LM), scanning (SEM), and transmission electron microscopic (TER4) techniques.
Materials and Methods Female Sprague-Dawley rats weighing -200 g were housed in cages with wide mesh bottoms to prevent coprophagia. The animals were fasted for 48 h and allowed free access to water during the fasting period. On the day of experimentation, animals were randomized into various groups and injected subcutaneously with 10 pgikg body wt of 16,16-dimethyl prostaglandin E2 (dmPGE2) or a comparable volume of saline. Thirty minutes later, each animal received an oral bolus of 1.0 ml of absolute ethanol or saline. Animals were killed at 5, 20, or 60 min after oral ethanol or saline administration to determine the cytologic and histologic effects of each treatment regimen over the course of time. Table
1. Criteria
Depth of damage
Type 0 Type 1 Type 2 Type 3 Type 4
for Depth
of Injury”
Damage None Luminal surface mucus cells only Luminal surface and upper gastric pit mucus cells Luminal surface and gastric pit mucus cells and some gastric gland cells Necrotic lesions: Severe mucosal injury to all surface and all or most glandular epithelium
’ Modification
of a scoring system devised by Lacy and Ito (7).
Vol. 88, No. 3
At all times of death after oral ethanol or saline exposure, rats were quickly anesthetized with ether and opened via a midline laparotomy. A silk ligature was tied around the lower esophagus and duodenum and 2.0 ml of halfstrength Karnovsky’s fixative (8) was injected into the gastric lumen through a puncture wound along the lesser curvature. The stomach was then excised and immersed in fixative. After fixation, the stomach was opened along the lesser curvature, laid flat, and examined for evidence of gross macroscopic damage by an observer unfamiliar with the experimental protocol. An ulcer index was assigned in which the percentage of glandular gastric mucosa injured was assessed. This was accomplished by measuring the width of each opened stomach and from this measurement visually dividing the exposed mucosa into two equal parts. As gross damage was confined to the glandular epithelium (oxyntic and antral mucosa; see Results), a visual estimate of the percentage surface area of the glandular stomach damaged in each part was made. The total surface area of glandular mucosa damaged in each stomach was calculated by summating the individual surface areas of each part. Samples for LM, SEM, and TEM were excised from the gastric glandular epithelium at a region located 2 mm below the limiting ridge (which separates the forestomach from the glandular epithelium) along the greater curvature of the stomach. Blocks were removed at right angles to the long axis of the stomach. A light microscopic sample measuring -1 mm x 10 mm was removed alongside a sample measuring 1 mm x 4 mm. The latter sample was for SEM subdivided into a l-mm x 3-mm block processed and a l-mm x l-mm block processed for TEM. The same procedure was performed on a second sample that was removed at -0.5 cm distal to the first sampling area. Thus, two LM, SEM, and TEM blocks were removed and processed from each stomach. Both samples were examined by the same individual (K. L. S.]. Routine paraffin-embedded hematoxylin and eosinstained sections and toluidine blue-stained, semithin plastic sections (0.25 Frn) were used for LM evaluation. Slides were viewed in a Zeiss photomicroscope (Carl Zeiss, Inc., Thornwood, N.Y.). Tissue blocks prepared for SEM were postfixed in aqueous 1.0% osmium tetroxide, dehydrated, and critical-point dried in a Bomar SPC 1500. The tissues were then gold-palladium-coated using a Denton Desk-I vacuum coater (Denton Vacuum Inc., Cherry Hill, N.J.) and viewed in a JEOL JSM-35 scanning electron microscope. Those tissues prepared for TEM were postfixed in cacodylate-buffered 2.0% osmium tetroxide, dehydrated, and embedded in epoxy resins. Semithin sections were made for orientation, followed by ultrathin sectioning using an LKB Ultratome III. Sections were viewed in a Siemens 102 electron microscope (Siemens Corp., Iselin, N.J.). Using hematoxylin and eosin-stained paraffin-embedded sections of tissue, the extent of mucosal injury was quantified. Slides were examined in a blinded fashion by coding them in such a way so that the examiner (K. L. S.) was unfamiliar with the experimental protocol. Only after examination was complete were they decoded. Using a graded micrometer eyepiece, within each tissue block pair from each stomach, the overall length of each tissue section was measured and the corresponding percentage of
HISTOLOGY AND CYTOLOGY OF CYTOPROTECTION
March 1985
Table
2. Percent Damage of Glandular Epithelium on GrossVisual Inspection at Various Time Periods of Death After 100% Ethanol Exposure’
Time of death (min)
100% Ethanol without PG 19.5 2 5.1 38.3 + 5.3 34.7 '- 3.7
5 20 60
100% Ethanol with PG 2.2 2 1.4[' 0" Oh
PG = 16,16-dimethyl prostaglandin EZ (10 &kg body wt subcutaneously). a n = 6 for each experimental group. ‘p C: 0.01 compared with 100% ethanol without prostaglandin. the length of mucosal surface injured was determined. The extent of injured mucosa from luminal surface to muscularis mucosa was further assessed using a modification of the criteria for depth of injury established previously by Lacy and Ito (7) and summarized in Table 1. Surface mucus cells were considered injured if one or more of the following criteria were present: marked cytoplasmic vacuolization, cytoplasmic swelling, nuclear pyknosis, or nuclear swelling with chromatin margination. Lucent cytoplasm and pyknotic nuclei indicated parietal cell injury. Dilation of glands, hyperemic vessels, and hemorrhage were additional determinants of glandular injury. In addition to these measurements, the left wall of every 10th gastric pit from each block was evaluated until a total of 10 pits per section had been assessed to ascertain subtle changes in the depth of gastric pit injury. This evaluation consisted of the micrometric measurement of the ratio of iniured versus normal pit mucus cells. Only those pits that were sectioned in the vertical plane with the entire lumen fully exposed were included in the determination. Paired measurements for length of mucosal surface injured as well as both the mucosal depth of injury and gastric pit depth of injury were averaged to obtain a single value for each of these indices of damage for each stomach, All values obtained were reported as means * SER?. Differences among the various experimental groups were evaluated statistically by analysis of variance. A p value of CO.05 was considered to represent statistical significance.
651
at all time periods of nonsecreting forestomach death. These lesions consisted of elongated red bands ranging from 2 to 10 mm in length and from 1 to 4 mm in width, generally parallel to the long axis of the stomach and mostly confined to the corpus (the portion of the stomach secreting acid and pepsin) and less commonly encountered in the antrum. In terms of total surface area, these lesions involved about 20% of the glandular epithelium in animals killed 5 min after alcohol exposure, and 35%38% of the mucosa in those killed at 20 and 60 min (Table 2). In animals receiving PG and ethanol (10 pgikg body wt & ‘dmPGE, inje%on followed by 100% ethanol administration orally), macroscopic evidence of damage was virtually absent regardless of the length of ethanol exposure. In the 2 animals in the 5-min sacrifice group in which some iniury was
visually present, it was in the form of petechiae small punctate lesions [Table 21.
Microscopic
and
Findings
Control tissues. (Saline injection followed by saline administration orally]: On LM at all time periods, the surface mucus cells (SMCs) formed a continuous epithelial sheet that lined the gastric
Results Macroscopic
Findings
No macroscopic
evidence
of mucosal
injury
was noted in any of the control stomachs (saline injection followed by saline administration orally) or
in stomachs exposed to PG alone (10 pgikg body wt of dmPGE2 injection followed by saline administration orally) regardless of the time of death. In contrast, stomachs exposed to ethanol alone (saline injection followed by lOOoh ethanol administration orally) showed obvious evidence of mucosal injury. In each of the stomachs treated with ethanol alone! necrotic hemorrhagic lesions were noted in the glandular portion of the stomach with sparing of the
Figure 1 Light microscopy of gastric mucosa from a salineinjected animal taken 5 min after oral saline. Luminal surface mucus cells and pit mucus cells (arrows) are intact, columnar or cuboidal in shape, and contain a single nucleus and varying amounts of mucin in the apical cytoplasm. Parietal cells (PC) are present in the glands. Plastic embedded, toluidine blue stain. (x600.)
652
Figure
SCHMIDT ET AL.
2. Scanning electron microscopy of gastric mucosa from a saline-injected animal taken 60 min after oral saline. The mucosal surface is formed by closely packed. polygonal surface mucus cells that are periodicall> interrupted by the round or slitlike openings of the gastric glands (arrows). Adherent granular or filmy material is mucus (double arrow,). (x280.)
surface and pits. At the tip of the mucosal surface, an occasional cell with a pyknotic nucleus and vacuolated cytoplasm was observed; such cells were presumed to be undergoing the normal exfoliative process. Most SMCs were intact and cuboidal or columnar in shape, with varying amounts of apical mucin granules. Mucin granules were visualized as a thin, dark-staining crescent in the apical surface of these cells (Figure 1). Gastric pits were of normal depth and gastric glands were formed of normal parietal, chief, endocrine, and mucus neck cells. On SEM, the luminal surface of the gastric mucosa was formed of closely packed SMCs and the round or slitlike openings of the gastric pits (Figure 2). On TEM, the SMCs were cuboidal to columnar in shape and contained numerous intensely stained, apical mucin granules in addition to the usual complement of cell organelles (Figure 3). No injury among control tissues obtained at the various times of death after oral saline administration could be detected by histologic examination except a rare, minute region where the surface epithelial cells were absent. These regions, when pres-
(:ASTKOENTEKOLO(;T
Figure
Vol. 88, No. 3
3. Transmission electron microscopy of gastric mucosa from a saline-injected animal taken 60 min after oral saline. Surface mucus cells are typically cuboidal to columnar in shape, contain a single nucleus (N) and no unusual organelles except numerous mucin granules (orrows) in the apical cytoplasm. (~5700.)
ent, generally involved -Cl?; of the surface epitheliurn (Table 3) and were graded as type I injury. Ethanol-exposed tissues. (Saline injection followed by 100% ethanol administration orally]: Extensive lesions of the gastric mucosa were noted on LM after alcohol exposure regardless of the time of death, with virtually the entire length of epitheliurn examined showing evidence of injury (Table 3).
Table
3. Percentage Length of Smface Damaged Microscopically” 5 min”
SAL-SAL PG-SAL SAL-&OH PG-EtOH
0.1 k 0.3 + 97.6 i 98.9 i
Epithelium
20 min” 0.1 0.2 1.4 0.5
1.5 i 1.0 r 94.9 i 85.3 +
0.6 0.6 2.2 5.8
60 min” 0.4 -t 0.2 0.9 + 0.9 97.1 2 1.3 91.7 -t 2.9
SAL-SAL, saline injection followed by oral saline. PG-SAL, 16.16dimethyl prostaglandin E2 (10 pg/kg body wt) injection followed by oral saline. SAL-EtOH. saline injection followed by 100% ethanol orally. PG.EtOH. 16.16-dimethyl prostaglandin E, (10 pgi kg body wt) injection followed by loo%, ethanol orally. ” n = 6 for each experimental group. “Time of death after 100% ethanol exposure.
March 1985
HISTOLOGY
SALIEIOH WOO25
PG/EIOH
p
5 min D
1
NOM
1
xl
2
p
1
C
p
---
2
p
p
T
40
7
s 20 mm
s 5
20
E x & a
p
NS
NS
P
.--
40
60 mm 20 NOW i
4 Category
Figure
of Injury
4. Effect of prostaglandin pretreatment on the depth of gastric mucosal injury. Animals were killed at 5, 20, and 60 min after ethanol injury. n = 6 animals for each group. NS, not significant: SALIEtOH, saline injection followed by ethanol orally; PGIEtOH. 16,16-dimethyl prostaglandin E, injection followed by ethanol orally.
AND
CYTOLOGY
OF CYTOPROTECTION
effects of ethanol on gastric epithelium are shown in Figures 6 and 7. This extensive erosive process was also evident on SEM regardless of the time of death. In many regions, the SMCs were sloughed, exposing the underlying lamina propria and the outlines of the denuded gastric foveolae. Masses of exfoliated cells and other debris were observed on the mucosal surface, obscuring the gastric pit openings and normal mucosal morphology [Figure 8). On TEM examination, the dense aggregation of luminal material included cells with organelles in various stages of degeneration along with extensive quantities of mucus, red blood cells, and other debris. Many of the adherent SMCs were injured to varying degrees as evidenced by pyknotic, crenated nuclei and concentric rings of rough endoplasmic reticulum [Figure 9), some of which exhibited dilated cisternae. High amplitude swelling and floccular densities in mitochondria were present in many of these cells. Prostaglandin-exposed tissues. [16,16-Dimethyl prostaglandin E, (10.0 pgikg body wt) injection followed by saline administration orally]: Stomachs of the dmPGEp-treated animals that received an oral bolus of physiologic saline were morphologically very similar to control tissues despite the time of death after saline exposure. Similar to control tissues, a rare, minute region where surface epithelial
SALlEtOH PG/EtOH p
100
p
In many regions, surface mucus cells were sloughed into the gastric lumen as an intact epithelial sheet. Exfoliated parietal cells and other debris also were observed within the gastric lumen. The depth of injury, however, varied with the length of alcohol exposure. At 5 min, after ethanol exposure, mucosal damage was characterized largely by type 3 and type 4 lesions (Figure 4). In addition, the average depth of gastric pit injury was -90% (Figure 5). By 20 min after ethanol administration, a slight shift in the depth of injury occurred with a reduction in type 3 lesions and an increase in type 2 (p < 0.05 compared with 5-min data) even though the percentage of type 4 lesions remained unchanged (Figure 4). A slight decrease in the depth of gastric pit injury (80%) was also observed at 20 min (Figure 5). Findings at 60 min were virtually unchanged from those observed at 20 min. At 60 min, >60% of damaged mucosa was still characterized by type 3 and type 4 lesions [Figure 4) and the average pit depth injury remained at 80% (Figure 5). Micrographs demonstrating the
653
‘1
80,
0
5
min Time
Figure
p
20
min
of
Sacrifice
30
mm
5. The depth of gastric pit injury in mucosa exposed to absolute ethanol with and without prostaglandin pretreatment. Animals were killed at 5, 20. and 60 min after ethanol exposure. n = 6 animals for each group. SALIEtOH, saline injection followed by ethanol orally; PGIEtOH, 16,16-dimethyl prostaglandin E2 injection followed by ethanol orally.
654
SCHMLDT
GASTKOENTEKOLOCY
ET Al,
cells were absent could be occasionally identified that involved 51% of the surface epithelium (Table 3). Such regions were very superficial and were graded as type 1 injury. Prostaglandiniethanol-exposed tissues. (16,16Dimethyl prostaglandin E, injection followed by loo%, ethanol administration orally): Animals receiving 10.0 pgikg body wt of dmPGEp subcutaneously before ethanol exposure were noted to have a different pattern of gastric mucosal injury when compared with those exposed to ethanol without dmPGEz pretreatment. Although the length of mucosal surface injured in animals receiving PG pretreatment before ethanol exposure was not significantly different from mucosa exposed to ethanol without PG pretreatment at any time of death (Table 31, the depth and type of injury were considerably altered if PG had been administered. More than 87% of the mucosal injury in the PC/ethanol-treated group was confined to types 2 and 3, with type 4 damage being totally absent at the 5-min time of death (Figure 4). Further, the mean depth of gastric pit injury was 60% in contrast to 90% in animals exposed to
‘V’ i
Figure
(I
Figure
6. Light microscopy ot gastric mucosa from a salineinjected animal taken 5 min after oral ethanol. Features of type 3 injury are represented. Surface mucus cells form an incomplete covering of the mucosal surface and those of the gastric pits demonstrate pyknotic nuclei and crowding (orrows). Although parietal cells deep in the glands appear normal, those at the surface (PC) have advanced toward the lumen. Plastic embedded, toluidine blue stain. (x800.)
Vol. 88. No. 3
7. Light microscopy of gastric mucosa from a salineinjected animal taken 5 min after oral ethanol showing the features of a type 4 or necrotic lesion as judged microscopically. Damage to cells is severe in the upper portion of the mucosa. Surface mucus cells are absent and parietal cells (PC) are sloughed into the lumen. Extensive hyperemia and hemorrhage are evident in the mucosal vessels and surface (arrorv). Plastic embedded, toluidine blue stain. (x800.)
ethanol without PG pretreatment (Figure 5). By 20 min, mucosal injury in the PGiethanol group was predominantly types 1 and 2 damage (73%). Again, type 4 or necrotic injury was virtually absent except for one focal region of necrotic tissue in one stomach sample (Figure 4). The mean gastric pit injury at 20 min was 30% (Figure 5). By 60 min after ethanol treatment, mucosal injury had shifted further toward types 1 and 2 injury (86%), whereas the mean gastric pit damage had fallen to 20% (Figures 4 and 5). Again, necrotic lesions were not observed at 60 min. Light micrographs depicting effects of PGiethanol on gastric mucosa are presented in Figures 10 and 11. Despite differences in the depth and severity of mucosal damage at the 5-min time of death, uninjured, minimally injured, severely injured, and necrotic SMCs and gland cells were observed in mucosa from both PGiethanoland the salineiethanol-treated groups on LM (Figures 6, 7, 10, and ll), SEM, and TEM evaluation. By SEM, the surface mucosa of PGiethanol-treated tissues at the 5-min time of death was morphologically similar to that
March
Figure
1985
8. Scanning electron microscopy of gastric mucosa from a saline-injected animal taken 60 min after oral ethanol. A portion of the lamina propria bearing outlines of gastric pit openings is exposed and largely devoid of epithelium (arroM.s). The remaining mucosal surface is covered by cells in various degrees of exfoliation as evidenced by obliteration of the gastric pit openings and loss of normal surface architecture. Compare with a similar region from saline/saline-exposed mucosa (control. Figure 2). (X 280.)
observed in the 5-min time of death in the saline/ ethanol-treated group. In some regions mucus cells were exfoliated from the interfoveolar surface, exposing the underlying lamina propria. In other areas, the exfoliated cells and other debris remained on the surface and were sufficiently dense to obscure the gastric pit openings. Minimally injured cells were difficult to identify on LM but on TEM were typified by swelling of mitochondrial matrices, margination of chromatin, and ring formation of rough endoplasmic reticulum. Severely injured cells could be identified on LM by the presence of pyknotic nuclei and vacuolated cytoplasm. On TEM, swollen mitochondria with floccular densities, in addition to the above cytologic alterations, marked the severely damaged cells. Despite these cytologic changes, junctional complexes between cells adherent to the basal lamina remained undissociated in PC/ethanoland saline/ethanol-treated groups. By 20 min after ethanol administration, many cells within the gastric pit and the mucosal surface ap-
HISTOLOGY
Figure
AN)
CYTOLOGY
OF (:1’TOPKOTECTIOS
655
9. Transmission electron microscopy elf gastric mucosa from a saline-injected animal taken 60 min after oral ethanol. A severely injured surface mucus vt!ll contains a pyknotic nucleus (N). In portrons of other cells. concentric rings of rough endoplasmi~ reticulum (orroTz,sJ are present. (X8800.)
peared normal or attenuated by LM in the samples treated with PGiethanol. This phenomenon was even more pronounced by 60 min, when rounded and attenuated SMCs were observed on the gastric surface (Figure 12). By SEM, these cells appeared to be emerging from the gastric pits and were intermingled with regions that were covered by attenuated cells or bare lamina propria. Although initiated, reepithelialization was not complete in the PGiethanol group by 60 min and focal zones of bare lamina propria indicative of type 3 injury were still present (Figures 13 and 14). By TEM, the adherent SMCs in regions of repair appeared almost normal (Figure 15). An occasional heterophagic vacuole or ring formation of rough endoplasmic reticulum was the only feature indicating cytologic alterations.
Discussion In an earlier study by Lacy and Ito (7), a clear disparity was demonstrated when assessing the cytoprotective effects of dmPGEz on ethanol-exposed mucosa as judged macroscopically with the naked eye and histologically with LM. It was noted in that study that all rats receiving absolute ethanol on examination of macroscopically nonnecrotic regions
Figure
Figu re
10. Light microscopy of gastric mucosa from a prostaglandin-injected animal taken at 5 min after oral ethanol showing focal regions of type I injury. Only luminal surface mucus cells are damaged in type I injury as evidenced by pyknotic nuclei and cytoplasmic vacuolization (arrow]; pit and parietal cells are normal appearing. A few surface mucus cells are exfoliating from the mucosal surface (double arrow). PC, parietal cells. Plastic embedded, toluidine blue stain. (X 800.)
Figure
11. Light microscopy of gastric mucosa from a prostaglandin-injected animal taken 5 min after oral ethanol and showing type z damage. Surface mucus cells with pyknotic nuclei and vacuolated cytoplasm are injured (arrows] to varying depths within the gastric pit. PC, parietal cells. Plastic embedded, toluidine blue stain. (X800.)
Light microscopy of gastric mucosa from a prostaglandin-injected animal taken 60 min after oral ethanol showing a region of epithelial repair. Attenuated surface mucus cells (double arrow) flatten to cover the underlying lamina propria (LP) adjacent to an area of connective tissue that appears to be exposed to the gastric lumen (arrow]. Surface mucus cells in m ost other regions appear normal in cytologic detail and distribution. PC, parietal cell. Plastic embedded, toluidine blue stain. I[x1100.)
March
1985
of the stomach, with or without PG treatment, had about 78% of the total area of the glandular portion of the gastric mucosal surface damaged when examined microscopically at 15 and 30 min after ethanol exposure. On macroscopic inspection, though, PG was felt to have rendered almost complete protection against alcohol injury at both time periods. Prostaglandin treatment, however, reduced the depth of all cells from damage by about 20% when evaluating the luminal surface through the gastric glands, and the necrotic lesions observed in mucosa exposed to ethanol alone were virtually eliminated. Wallace et al. (9) reported similar findings in a rat chambered stomach preparation when assessing the protective effects of the natural PC, PGE2, against gastric injury induced by 40% ethanol. We also observed in the present study that the presumed virtually complete protection against ethanol injury by PG when evaluated macroscopically could not be confirmed histologically at any of the three time periods examined after alcohol exposure. Such observations point out
Figure
13. Scanning electron microscopy of gastric mucosa from a prostaglandin-injected animal taken 60 min after oral ethanol. A focal region showing denuded lamina propria (L) with gastric pit openings is shown surrounded by a mucosal surface which is undergoing reepithelialization. Rounded surface mucus cells (white arrows) bulge into the lumen. Other surface mucus cells are flattened and attenuated (bluck orrow). a feature consistent with migrating cells or cells that extend their boundaries to cover underlying connective tissue components. (X280.)
HISTOLOGY
Figure
AS11 C:Y’f()LOG\r’ OF (:TI’TOPRO~E(:TION
657
14. Scanning electron microscopy of gastric mucosa from a prostaglandin-injected animal taken 60 min after oral ethanol showing the mouth of a gastric pit from a region of surface regeneration. Rounded surface mucus cells have emerged from the pit region although some still possess flattened cvtoplasmic surfaces common to migrating cells (nrrow). Cells in the background are attenuated (double orron) possibly to cover the underlying connective tissue stratum. GP. gastric pit. (X 3000.)
the importance of assessing the presence or absence of gastric injury on histologic grounds and emphasize the shortcomings of evaluating cytoprotection on purely macroscopic criteria. On histologic comparison of mucosa exposed to ethanol with and without PG pretreatment. we noted that the severe gastric hemorrhagic erosions (type 4 or necrotic lesions] so readily apparent microscopically in mucosa exposed to absolute ethanol alone were virtually eliminated in animals receiving PG pretreatment, confirming the previous findings of Lacy and Ito (7) and Wallace et al. (9). As much as SO%-40% of the gastric epithelium revealed evidence of necrotic lesions (type 4) in alcohol-exposed mucosa without PG pretreatment that was not altered over the 60 min of evaluation, whereas only one stomach (during the 20-min time of death) exposed to ethanol after PG pretreatment demonstrated evidence of one small necrotic lesion. Using the same PG (dmPGE2) as Lacy and Ito (7), we further observed no difference between the length of mucosal surface damaged in rats receiving absolute alcohol with and without PC pretreatment at 5, 20, or 60 min as quantified by LM, similar to what these investigators reported at 15 and 30 min after ethanol exposure. In contrast to their study, though, clear differences between the two experimental groups
658
Figure
SCHMIDT
ET AL.
15. Transmission electron microscopy of gastric mucosa from a prostaglandin-injected animal taken 60 min after oral ethanol. Several cells from a region of epithelial repair are either normal or show minor injury: concentric rings of rough endoplasmic reticulum (arrow) and one heterophagic vacuole (double arrow] are present in one cell. Minor margination and central clearing of chromatin is present within the nuclei (N). MG, mucin granules. (~5700.)
could be demonstrated with respect to the depth of mucosal injury over time. Although the distribution of type 1 and type 3 lesions was initially identical in animals killed at 5 min whether or not PG pretreatment was rendered, type 2 lesions were more prominent in PG-pretreated stomachs. In animals killed at 60 min, type 3 and type 4 injury still accounted for >60% of the lesions observed in alcohol-exposed mucosa without PG pretreatment, whereas type 1 and 2 lesions (>80%) were most prominent in animals killed at a similar time period in which PG had been rendered. The depth of pit injury was also different between the two experimental groups at all time periods studied. The explanation for the difference between the present study and that of Lacy and Ito (7) is uncertain. Although differences in the techniques used for specimen preparation may in part account for the discrepancy, it is unlikely that such differences alone are responsible for this disparity. Other considerations include the standardized area of epitheliurn that was evaluated microscopically in our studies in contrast to the separate analyses of necrotic
GASTROENTEROLOGY
Vol.
88, No. 3
and nonnecrotic regions evaluated by Lacy and Ito (7). Further, in this study the dose of PG used was 10 pg/kg body wt given parenterally in contrast 10 the 4pg/kg body wt dose given orally by these previous investigators. An additional possible explanation for our findings may be related to the time element involved. We sampled tissue at three different time periods from 5 to 60 min, whereas Lacy and Ito (7) examined only two time periods at 15 and 30 min. Although we saw differences in the magnitude of gastric damage by alcohol that was clearly less at 5 and 20 min if PG had been given, in contrast to the observations of Lacy and Ito at 15 and 30 min, these differences were most pronounced at the 60-min sampling time, a time period that Lacy and Ito did not evaluate. It is interesting to postulate that during an hour time period, spared pit mucus cells could move up the pit wall in an effort to reconstitute the exfoliated cell population, an event that would explain the reduced proportion of injured surface mucus cells observed with PG pretreatment over time. The observation in our studies that the depth of pit injury in response to ethanol exposure was significantly less in mucosa pretreated with PG supports this hypothesis. Further, the findings of Lacy and Ito, as well as our own findings, in which PGs were noted to reduce the depth of mucosal injury, also seem consistent with such an explanation. Thus, reduced depth of injury with sparing of deep-lying pit and mucus neck cells could serve as a cellular pool for rapid reconstitution by cell migration. It seems doubtful that an increased mitotic rate over an hour’s duration by the deeplying mucus neck cells could account for any significant reconstitution, although the present study did not address that consideration. The additional feature of absence of necrotic hemorrhagic erosions in ethanol-exposed mucosa after PG pretreatment may enhance the reconstitution process as it has previously been shown that an intact lamina propria and basal lamina would greatly expedite the attempt by pit cells to cover an epithelial defect (10). Experimental evidence for such a reconstitution process in ethanol-damaged mucosa has been reported. Using a rat chambered stomach preparation, Morris and Wallace (11) were the first to demonstrate epithelial reconstitution after exposure to 40% ethanol and noted cell migration as early as IO min after alcohol administration that was quite advanced by 1 h. They emphasized the importance of an intact basal lamina as scaffolding for such reepithelialization to occur. Using an in vivo rat stomach preparation in which the esophagus and duodenum were ligated, Lacy and Ito (12) exposed the gastric epithelium to absolute ethanol for 30-45 s via an injection in the nonglandular portion, after which the alcohol
March
1985
was removed. Stomachs examined microscopically immediately after this insult revealed damage to 98% of the glandular gastric surface. Examination of the stomach 30 min after exposure to ethanol revealed that >75% of the mucosal surface was reconstituted with surface mucus cells. Although it might be argued that our findings are in conflict with these observations, it must be stressed that our studies differ in a number of ways from the design of these previously published experiments. First, whereas the concentration of ethanol used in the present study was identical to that used by Lacy and Ito (1 Z), the mucosal exposure was considerably longer (even though the speed with which ethanol was emptied into the duodenum is not known) than 30-45 s, presumably resulting in more profound injury and thus requiring a longer period of time for reconstitution of the epithelium to occur. For similar reasons, and because of the higher concentration of alcohol used in the present investigation, the reconstitution observed by Morris and Wallace (11) over a 1-h time period with 40% ethanol is not comparable to our findings. Despite these differences, it is noteworthy that some reconstitution could even be identified in our studies in animals exposed to ethanol in which PG pretreatment was not rendered. By 60 min, for example, the percentage of type 3 lesions had significantly decreased and the percentage of type 2 lesions had significantly increased when compared with corresponding values at 5 min. It seems clear from our studies that two conclusions can be reached. First, although PG was unable to prevent injury to the gastric epithelium, at least in response to absolute ethanol, it did significantly reduce the depth of injury as exemplified by the absence of necrotic lesions and the prominence of type 2 injury at the 5-min time of death which was not observed in animals exposed to ethanol without receiving PG pretreatment. Thus, although the concept of cytoprotection, as originally defined (2-4), which refers to complete absence of cellular damage when exposed to an injurious agent, could not be supported in our studies, protection of deep layers by PG was clearly present. Whether one wishes to redefine cytoprotection as protection against deep injury, as Robert has done (l3), or use another term, seems to us to be of academic interest only. The fact remains that PG significantly reduced the depth of injury in gastric mucosa exposed to absolute ethanol and in that sense was clearly protective. Second, although mucosa exposed to absolute ethanol, with and without PG pretreatment, demonstrated evi-
HISTOLOGY
AND CYTOLOGY
OF CYTOPROTECTION
659
dence of healing over a 1-h time period, as demonstrated by a reduction in the depth of injury when compared with ti-min findings, this healing process was markedly accelerated in mucosa pretreated with PG. Whether this relates to the prevention of injury to pit mucus neck or mucus cells, or both, and thereby enhancement of reepithelialization, or to a direct effect of PG on cell mitosis or cell migration is unknown. The mechanisms responsible for these observations must await further study.
References 1. Miller TA. Protective effects of prostaglandins against gastric mucosal damage: current knowledge and proposed mechanisms. Am J Physiol 1983;245:G601-23. 2 Robert A. Cytoprotection by prostaglandins. Gastroenterology 1979;77:761-7. 3
4
5
6
7
8
Robert A. Antisecretory, antiulcer, cytoprotective and diarrheogenic properties of prostaglandins. In: Samuelsson B. Paoletti R, eds. Advances in prostaglandin and thromboxane research. New York: Raven, 1976:507-20. Robert A. Nezamis JE, Lancaster C, Hanchar AJ. Cytoprotection by prostaglandins in rats: prevention of gastric necrosis produced by alcohol, HCl, NaOH, hypertonic NaCl and thermal injury. Gastroenterology 1979:77:433-43. Bommelaer G, Guth PH. Protection by histamine receptor antagonists and prostaglandin against gastric mucosal barrier disruption in the rat. Gastroenterology 1979;77:303-8. Whittle BJR. Mechanisms underlying gastric mucosal damage induced by indomethacin and bile salts and the actions of prostaglandins. Br J Pharmacol 1977:60:455-60. Lacy ER, Ito S. Microscopic analysis of ethanol damage to rat gastric mucosa after treatment with a prostaglandin. Gastroenterology 1982;83:619-25. Karnovsky MJ. A formaldehyde-glutaraldehyde fixative of high osmolality for use in electron microscopy. J Cell Biol 1965;27:137A-EA.
9
10
11
12
13
Wallace JL. Morris GP, Krausse EJ, Greaves SE. Reduction by cytoprotective agents of ethanol-induced damage to the rat gastric mucosa: a correlated morphologic and physiologic study. Can J Physiol Pharmacol 1982;60:1686-99. Vracko R. Basal lamina scaffold: anatomy and significance for maintenance of orderly tissue structure: a review. Am J Pathol 1974;77:314-38. Morris GP, Wallace JL. The roles of ethanol and of acid in the production of gastric mucosal erosions in rats. Virchows Arch [Cell Pathol] 1981;38:23-38. Lacy ER, Ito S. Ethanol-induced insult to the superficial rat gastric epithelium: a study of damage and rapid repair. In: Allen A, Flemstrijm G. Garner A. Silen W. Turnberg LA, eds. Mechanisms of mucosal protection in upper gastrointestinal tract. New York: Raven, 1984:49-56. Robert A. Role of endogenous and exogenous prostaglandins in mucosal protection. In: Allen A, FlemstrBm G. Garner A, Silen W, Turnberg LA, eds. Mechanisms of mucosal protection in the upper gastrointestinal tract. New York: Raven, 1984:377-82.