GASTROENTEROLOGY

ALIMENTARY

77:433-443,1979

TRACT

Cytoprotection by Prostaglandins in Rats

Prevention of Gastric Necrosis Produced by
Alcohol, HCl, NaOH, Hypertonic NaCl, and
Thermal Injury
ANDRk
and

ROBERT,

ALEXANDER

Department

of Experimental

JAMES

E. NEZAMIS,

Biology,

The Upjohn

(PG) have been shown to inin animals-

and humans.4-6

Received October 13,1978. Accepted

February
27.1979.
Address
requests for reprints to: Andre Robert, M.D., Ph.D, Department
of Experimental
Biology, The Upjohn Company,
Kalamazoo, Michigan 49001.
The authors
thank Dr. Eugene D. Jacobson,
University
of Cincinnati College of Medicine,
for his valuable suggestions,
and for
proposing
the term cytoprotection
after examining
our data.
0 1979 by the American
Gastroenterological
Association.
0016-5085/79/090433-11/$02.00

LANCASTER,

J. HANCHAR

Oral administration
to fasted rats of either absolute
ethanol,
0.6 N hydrochloric
acid, 0.2 N sodium hydroxide, 25% sodium chloride,
or boiling water produced extensive
necrosis
of the gastric mucosa.
Pretreatment
with several prostaglandins
of the A, E, or
F type, either orally or subcutaneously,
prevented
such necrosis,
and the effect was dose-dependent.
This property
of prostaglandins
is called
cytoprotection.
The protective
effect against
oral administration
of absolute
ethanol
was already
maximal 1 min after PGE, given orally, and 15-30 min
after PGE, given subcutaneously.
Cytoprotection
by
prostaglandins
is unrelated
to the inhibition
of gastric acid secretion
since, (a) it is maximal
at doses
that have no effect on gastric secretion,
and (b) antisecretory
compounds
(cimetidine,
methscopolamine
bromide)
and antacids
are not cytoprotective.
AIthough
the mechanism
of gastric cytoprotection
is
unknown,
prostaglandins
appear to increase
the resistance
of gastric mucosal
cells to the necrotizing
effect of strong irritants.
These results suggest that
certain prostaglandins,
by a mechanism
other than
the inhibition
of gastric acid secretion,
maintain
the
cellular integrity
of the gastric mucosa, and might be
beneficial
in the treatment
of a variety of diseases in
which gastric mucosal injury is present.

Several prostaglandins
hibit gastric secretion

CLEO

Company,

Kalamazoo,

Michigan

They also prevent ulcer formation

in animals,, and
accelerate
ulcer healing in humans.*. It was usually
assumed
that the antiulcer
effect was due to the inhibition
of gastric acid secretion.
We now present
evidence
that PG can prevent
the formation
of gastric lesions, even mucosal
necrosis,
produced
in rats
by a variety of necrotizing
agents, without
reducing
gastric acid secretion.
This property
of prostaglandins is called cytoprotection.
An abstract has been
published
earlier.

Methods
In all the studies reported
here, female rats (Upjohn rats derived
from the Sprague-Dawley
strain) of a
body weight of 205-215 g were used. Food, but not water,
was removed
in the morning.
At ~:OO pm, water was also
withheld,
and the animals were placed in individual
cylindrical
stainless-steel
cages with flat bottoms
and with
perforations
to allow ventilation.
These cages limited their
movements
and thus prevented
the ingestion
of hair and
feces. This procedure,
as described
earlier, did not appear
to be stressful;
actually, the animals were often observed
to be asleep in such cages. On the following morning, various necrotizing
agents (listed below) were given orally,
and the animals
were killed 1 hr later. Their stomachs
were dissected
out, opened
along the greater curvature,
and the mucosa was examined
by an observer,
who was
unaware
of the treatment
given, with a 2~ binocular
magnifier for the presence
of necrotic
lesions
which
were
counted.
The prostaglandins
were administered
at multiple dose
levels. The results were plotted on semilog paper. The ordinate shows
either the average
number
of lesions
per
stomach as such or as a percent of the control values. The
abscissa
shows the doses on a log scale. The ED,,, is the
dose of PG reducing
the average number
of lesions
per
stomach by half in comparison
with control values, as determined
by graphic interpolation
on a plot of the dose on
a log scale. The number of animals per group is indicated
in the tables and legends for figures. In all studies, the Stu-

434

ROBERT ET AL.

GASTROENTEROLOGY Vol. 77. No. 3

dents t-test was used to determine

Gastric Lesions Produced

Necrotizing
Agents

statistical

significance.

by Various

One of the following agents was given orally in a

volume of 1 ml: absolute ethanol, 0.6 N HCl, 0.2 N NaOH,
25% NaCl, and boiling water. In the case of boiling water,
the animals were first anesthetized
with sodium methoxital (Brevital Sodium, Eli Lilly and Co.), 12 mg in 0.6 ml intraperitoneally,
to prevent pain. Anesthesia was found not
to interfere with the production of a gastric burn.

Effect of PG on Gastric Lesions

Several Necrotizing
Agents

Produced

by

Several PG were administered

at various doses, either orally or subcutaneously,
30 min before administration of each of the necrotizing agents mentioned
above.
These PG were: PGE,, 16,16-dimethyl
PGE,, 15(S)-15methyl PGF,,, 15(R)-15-methyl PGF,,, and 16,18dimethyl
PGA,. The PG were first diluted with a few drops of 95%
ethanol and then made up to a volume of 1 ml with either
water (oral) or saline (subcutaneous).
The animals were
killed 1 hr after receiving the necrotizing agents.

Comparison
Routes

of Oral

and

Subcutaneous

Both PGE, and 16,16-dimethyl PGE, were given at

various dose levels, either orally or subcutaneously,
30
min before oral administration
of 1 ml absolute ethanol.
The animals were killed 1 hr after receiving ethanol.

Onset

of Action

and Duration

of Action

PGE, (500 pg/kg) was given orally or subcutaneously at various times up to 5 hr before oral administration of 1 ml of absolute ethanol. Control animals received
1 ml of vehicle at the same time intervals before administration of ethanol. The animals were killed 1 hr after receiving ethanol.

Effect of the Volume

of Diluent

To find out whether, when the PG were given

orally, the critical factor was the total amount of PG or the
concentration
of PG, three doses of PGE, were administered in various volumes of diluent, namely 0.5, 1.0, and
2.0 ml. The doses of PGE, were 10, 25, and 100 pg/kg. Absolute ethanol (1 ml) was given 30 min later, and the animals were killed 1 hr after receiving ethanol.

16,26-Dimethyl
PGE, and Absolute
both Given after Pylorus Ligation

Ethanol,

In a first study, both the necrotizing agent (absolute ethanol) and the cytoprotective
agent (16,16-dimethyl
PGE,) were administered
orally after ligation of the py-

lorus, to see whether the PG could protect by direct contact with the gastric mucosa. The pylorus was first ligated
under ether anesthesia.
16,16-Dimethyl
PGE, was then
given orally 2 min later at 1.5 and 5.0 pg/kg. Ten minutes
after the PG, 1 ml of absolute ethanol was given orally.
The animals were killed 1 hr later.
In another study, 16,16-dimethyl PGE, was injected subcutaneously
immediately after pylorus ligation, at 1.5 and
5.0 pg/kg. Thirty minutes later, in order to allow enough
time for absorption of the PG from the injection site, 1 ml
of absolute ethanol was administered
orally. The animals
were killed 1 hr after receiving ethanol.

Effect of Cytoprotective
Secretion

PG on Gastric

To find out whether cytoprotective

PG affected
gastric acid secretion, several PG were tested in two different preparations:
(a) Shay rats (pylorus ligated), and (b)
intact, unoperated
rats. The advantage of the Shay rat is
that it permits collection of all gastric juice that accumulates during a given time interval. The advantage of the intact, unoperated rat is that the secretion obtained is influenced only by the treatment given, in this case a PG, and is
not altered by pylorus ligation, a procedure known to
stimulate gastric acid secretion.
Shay Rats. After a 24 hr fast, the animals received
orally either 1 ml of water (controls) or of PG. The PGs
used and their doses are shown in Table 2. One hour later,
the pylorus was ligated under ether anesthesia. The animals were killed 4 hr after pylorus ligation, and gastric
juice was collected. The volume was measured, and the
acid content was titrated with NaOH 0.1 N to pH 7 with a
glass electrode, using an automatic titrator (Copenhagen
Radiometer).
The results were expressed
in meq/liter
(concentration)
and meq/4 hr (output). The doses of PG
were chosen to be from 40-500 times greater than the cytoprotective doses in order to find out whether these prostaglandins affected gastric secretion even when given at
doses much in excess of the cytoprotective dose.
Intact, Unoperated Rats. The animals were fasted
for 24 hr as above and then given 16,16-dimethyl PGE,, 1
pg/kg orally in 1 ml. This dose is 20 times higher than the
cytoprotective
ED,,. Groups of 12 rats were killed at the
following intervals after treatment: 1, 5, 15, and 30 min.
Controls received 1 ml of water orally and were killed at
the same time intervals. At autopsy, the esophagus and the
pylorus were clamped, and gastric juice was collected into
a test tube. Since there was too little secretion for direct
titration, the volume was first measured, and then 1 ml of
water was added to the test tube. The acidity of the diluted juice was determined by titrating 0.5 ml with 0.1 N
NaOH. The amount of acid in the stomach was calculated
as follows:
Vol. of juice
before dilution

Vol. of juice
x

after dilution

Vol. of juice after dilution

peq/liter of
diluted juice

~2

minus 1 ml X 1000
Acid in
= stomach
in peq

September

GASTRIC CYTOPROTECTION

1979

Example:

x 1.4X 6200 X 2
= 17.36peq/stomach
0.4
[1.4- 1.01x 1000

This study was done to find out whether gastric acid secretion was decreased at any time during the 30 min after
treatment with a PG, since, in other experiments reported
here, necrotizing agents were administered
30 min after a
PG.

Effect of Antisecretory
Drugs and of Antacids
on Gastric Necrosis Produced
by Absolute
Ethanol
Two gastric antisecretory
drugs, cimetidine
(a histamine H, receptor antagonist)
50 mg/kg intraperitoneally,
and methscopolamine
bromide
or Pamine@
(an anticholinergic,
Upjohn Co.) 10 mg/kg subcutaneously,
were
given either 30 min (cimetidine)
or 1 hr (methscopolamine
bromide)
before oral administration
of 1 ml of absolute
ethanol.
Control
animals
received
saline. Two antacids,
sodium bicarbonate
0.15 M and a Tris buffer at pH 7 (0.15
M), were given orally in 2 ml to other groups, 1 min before
oral administration
of 1 ml of absolute
ethanol.
Control
animals received 2 ml of water. The animals were killed 1
hr after receiving absolute ethanol.

Figure

BY PROSTAGLANDINS

435

Results
Gastric Lesions Produced
Necrotizing
Agents

by Various

Each of these five agents produced

severe gastric damage visible from the outside of the stomach
as thick black or red lines. After opening
the stomach, lesions were found in the mucosa and consisted
of elongated
bands, l-10 mm long by l-3 mm wide,
usually parallel to the long axis of the stomach
(Figure 1). Their color varied with the agent: yellow or
red (ethanol),
red (25% NaCl, boiling water), black
(0.6 N HCl, 0.2 N NaOH).
Usually,
15-20 lesions
could be counted.
They were located mostly in the
corpus
(the portion
of the stomach
secreting
acid
and pepsin); the antrum
was less affected. No gross
lesion
developed
in the forestomach
(the nonsecreting
part covered with a squamous
epithelium).
Histologically,
the lesions
consisted
of necrosis
usually extending
down through about two-thirds
of
the mucosa (involving
the surface epithelium,
the region of mucus neck cells, and of parietal cells); occasionally,
the necrosis
involved
the full mucosal
thickness.
Necrotic patches were also present in the

1. Gastric lesions produced

by five necrotizing
agents. All agents were given orally in 1 ml, and the animals were killed 1 hr later.
Stomachs
were opened along the greater curvature.
A. Control. B. Absolute ethanol. C. 0.6 N HCl. D. 0.2 N NaOH. E. 25% NaCl.
F. Boiling water. In all cases: extensive
necrotic lesions of the corpus. The antrum is almost intact grossly, although
lesions
were present histologically.
The forestomach
(upper white portion) is intact.

436

Figure

ROBERT

ET AL.

GASTROENTEROLOGY

Vol. 77, No. 3

2. Gastric cytoprotection
of prostaglandins
against absolute
ethanol.
One ml of absolute
ethanol was given orally and the animals were killed 1 hr later. A. Control vehicle given orally 36 min before ethanol. Multiple and severe necrotic lesions of the
was administered
30 min before ethanol. B. PGE, 500 pg/kg
corpus caused by absolute
ethanol. In B, C and D, a prostaglandin
subcutaneously.
C. PGE, 150 pg/kg orally. D. 16,16Dimethyl
PGA, 50 pgg/kg orally. These three prostaglandins
prevented
the
formation
of gastric lesions by absolute ethanol.

antral mucosa. Mucus cells of the corpus were depleted of granules

(as shown by PAS staining),
sometimes
even in areas that were not necrotic.
Large globs of mucus were pften found in the lumen
and were particularly abundant after administration
of boiling water; sometimes these masses of mucus
were still attached to the mucosa and contained mucosal cells, either isolated or as detached fragments
of the mucosa. The muscularis mucosae was intact.
The submucosa of the corpus, the antrum, and the
forestomach
was markedly thickened by edema, but
was not necrotic. The edematous submucosa did not
contain polymorphonuclear
leucocytes 1 hr after ad-

ministration of the necrotizing agents. In other studies in which the animals were killed 24 hr after the
necrotizing
agents,
the submucosa
was still
markedly
edematous
and, in addition, was infiltrated with polymorphonuclear
leucocytes.
At that
time (24 hr), most of the mucosa had become necrotic. In animals pretreated with a cytoprotective
PG and in which no gross lesions were visible, the
histology of the stomach showed a normal mucosa
(no necrosis, intact mucosal cells) and a normal submucosa (no edema, no infiltration with polymorphonuclear leucocytes).
These histological findings will
be reported in detail elsewhere.

September

Figure

GASTRIC

1979

CYTOPROTECTION

BY PROSTAGLANDINS

437

3. Gastric cytoprotection
by prostaglandins
against boiling water. Rats whose stomachs
are shown in B, C, and D received 1 ml of
boiling water orally and were killed 1 hr later. A. Untreated
control. B. Control vehicle given orally 30 min before boiling water. Massive necrosis of the corpus caused by boiling water. C and D. X,16-Dimethyl
PGE, 5 pg/kg either orally (C) or subcutaneously (D), 30 min before boiling water. Complete protection.

Effect of PG on Gastric
Several Agents

Necrosis Produced by

The PGs used prevented the formation of gastric lesions, regardless of the necrotizing agent used
(Figures 2 and 3). The only difference among the PGs
was their potency: some were active at less than 1

pg/kg, whereas others required up to 500 pg/kg to inhibit the lesions. Figure 4 shows the cytoprotective
effect of five PG, given orally, against gastric mucosal necrosis produced by absolute ethanol. The
protection was dose-dependent.
The most potent PG
was 16,16-dimethyl
PGE,. Figure 5 shows the effect
of l6,16-dimethyl
PGE,, given orally, on gastric mu-

438

ROBERT

ET AL.

GASTROENTEROLOGY

Vol. 77, No. 3

GASTRIC LESIONS

GASTRIC LESIONS
100

56
B
5
0.025

Figure

10 15 25

pglhg ORALLY

50

1 150
100

( 500
300

4. Gastric
cytoprotection
by five prostaglandins
against
absolute
ethanol.
The prostaglandins
were administered orally 30 min before giving 1 ml of absolute
ethanol orally. Animals
were killed 1 hr after ethanol. The
vertical axis shows the average
number
of lesions per
stomach
in each group, expressed
as percent
of controls (control
animals
received
ethanol
but no prostaglandin).
Each prostaglandin
reduced the number of lesions per stomach
in a dose-dependent
manner.
The
most potent
prostaglandin
was 16,16-dimethyl
PGE,.
Ten rats per point.

z4
2
-iy/ II\

.;'..
.
.
.;....
.. "0
.-s
0.05
I
0.1 0.15 0.3 0.5
1.5
,

0.025 0.03
16.WOIMETHYL PGE2:pg/kg ORALLY, 30 MIN
BEFORE NECROTIZING AGENT
Figure

5. Cytoprotection
of 16,16-dimethyl
PGE, against five different
necrotizing
agents.
16,16-Dimetbyl
PGE, was
given orally 30 min before the necrotizing
agents (1 ml
of either absolute ethanol, 0.6 N HCl, 0.2 N NaOH, 25%
NaCl, or boiling water). The vertical axis shows the average number of lesions per stomach in each group. For
each necrotizing
agent, 16,16-dimethyl
PGE, reduced
the number of lesions and the degree of protection
was
dose dependent.
Ten to sixteen rats per point.

GASTRIC

LESIONS

cosal lesions produced by five different agents. In all

cases, the cytoprotective
effect was dose-dependent.
The cytoprotective
ED,, of 16,16-dimethyl
PGE,
ranged from 0.03 to 0.15 pg/kg, depending on the
necrotizing agent.
100

Comparison
Routes.

16.16.LItMETHYL PGE2

of Oral and Subcutaneous

PGEz

PGE, and 16,16-dimethyl PGE, were about 3-4

times more cytoprotective
when given orally than
when given subcutaneously
(Figure 6). Their cytoprotective
ED,, were as follows: PGE,, 25 pg/kg
orally and 100 pg/kg subcutaneously;
16,16-dimethyl
PGE,, 0.05 pg/kg orally and 0.15 pg/kg subcutaneously. By either route, 16,16-dimethyl PGE, was approximately 500 times more potent than PGE,.
Onset of Action and Duration

of Action

When PGE, was given at the same time as absolute ethanol, the number of gastric lesions was reduced by 34% (14.5 lesions in controls, 9.6 lesions in
PGE,-treated
animals, just below statistical significance) (Figure 7). When PGE, was given from 1 to 30
min before ethanol, the inhibition was nearly complete (85-97%). The onset of cytoprotection
after
subcutaneous
administration
of PGE, was longer:
The PC had to be given 2.5 min before ethanol to reduce the number of gastric lesions by half. Near total inhibition (88-100%) was achieved when it was

pdkg
Figure

6. Gastric
cytoprotection
of PGE, and 16,16-dimethyl
PGE, against absolute
ethanol: comparison
of oral and
subcutaneous
administration.
The prostaglandins
were
given 30 min before oral administration
of 1 ml of absolute ethanol. The vertical axis shows the average number of lesions per stomach in each group, expressed as
percent of controls (control animals received ethanol
but no prostaglandin).
Each prostaglandin
was 3-4
times more cytoprotective after oral than after subcutaneous administration. Ten animals per point.

September

GASTRIC

1979

GASTRIC

CYTOPROTECTION

GASTRIC

LESIONS
P

100

80

20

439

BY PROSTAGLANDINS

LESIONS

100 -

60 -

3, PGE2
\ Subcutaneously

5 60 ii

It

40-

i!i

20 -

2.5

MINUTES

10

BEFORE

15

30

60

ETHANOL

CONTROL

120180 300
90 150240

DOSEOFPGE:,

Figure
Figure

7. Time of onset and duration

of action of cytoprotection
by PGE,. PGE,, 500 &kg,
was given either orally or
subcutaneously,
either at the same time (Time 0) as absolute ethanol (I ml orally) or at various time intervals
before absolute ethanol; these time intervals
are shown
on the horizontal
axis. After oral administration,
cytoprotection
was almost immediate.
After subcutaneous
administration,
cytoprotection
was apparent
at -2.5
min and maximal
at -30 min. The duration
of action
was about twice as long after oral than after subcutaneous administration.
Ten to twenty-five
rats per point.

injected
15-60 min before ethanol
(Figure 7). Cytoprotection
diminished
when PGE, was given longer
than 1 hr before
ethanol.
Gastric
lesions
had returned to 50% of control values 4 hr after oral administration
of PGE,, and about 2 hr after subcutaneous
administration,
as determined
by graphic
interpolation on a plot of the time on a log scale,

Effect

Table

of the Volume

at a low dose (10 pg/kg)

1. Cytoprotection

by a Prostaglandin,

was more
After

Pyiorus

0.5

25 p9/k9

16,16-Dimethyl
PGE, and Absolute
Given After Pylorus Ligation.

duced

16,16-Dimethyl
gastric necrosis

100 p9lk9

PGE, inhibited
even when both

Ethanol

ethanol-inthe PC and

Ligation

PGE,
-

1.5 pg/kg

ml

0.5
1
2
ML OF DILUENT

8 Effect of the volume of diluent on the cytoprotective

activity of PGE,. PGE, (at three dose levels) was given
orally in either 0.5,1.0, or 2.0 ml, 30 min before absolute
ethanol.
The dilution
factor played a role only at the
lowest dose. NS = Not statistically
significant
when
compared
to control
group. *= P < 0.01 when compared with 0.5 and 1.0 ml at a dose of 10 &kg.
At 25
and 100 pg/kg, none of the differences
between
volumes of diluent was statistically
significant.
However,
at those doses (25 and 100 pg/kg), the degree of inhibition was statistically
different
from the control group
regardless
of the volume of diluent. Twelve
rats per
point.

Vehicle

to p9lk9

Oral
16,16-dimethyl

cytoprotective
when diluted
in a small volume (0.5
and 1.0 ml) than in a larger volume (2.0 ml) (Figure
8). At higher doses, 25 and 100 pg/kg, that were 7085% cytoprotective,
respectively,
the volume
of
diluent played no role (Figure 8).

of Diluent

PGE, given

0.5

5.0 &kg

Vehicle

Subcutaneous
-___
16,16-dimethyl
~I__
1.5 pg/kg

PGE,
5.0 pg/kh

___No. of animals
Gastric Lesions
No. per stomach
Change

from controls

P
16,Wdimethyl
PGE,: given
min after oral prostaglandin,

10

10

10

11.7 -e 3.04

3.3 -t 0.90

0.3 +- 0.16

13.8 f 0.98

4.6 + 1.42

3.1 + 0.70

-72%

-97%

-677 0

-78%

<O.Ol

<O.Ol

to.01

<O.Ol

in 1 ml immediately
after pylorus ligation, either orally
and 30 min after subcutaneous
prostaglandin.
Animals

or subcutaneously.
Absolute
ethanol: 1 ml orally,
killed 1 hr after absolute ethanol. fl SEM.

10

440

Table

ROBERT

ET AL.

2. Effect

of

Gastric

GASTROENTEROLOGY

Cytoprotective
Prostaglandins
Secretion: Pylorus Ligation
Acid output
(meq/4

Controls
PGE, 500 &kg
Controls
16,16-dimethyl
Controls
15(S)-15-methyl
Controls
15(R)-15-methyl
Controls
16,16-dimethyl

PGF,,

1000 pg/kg

PGF,@ 5000 pg/kg

PGA,

5000 pg/kg

Effect of Antisecretory
Drugs and of Antacids
on Gastric Lesions Produced
by Absolute
Ethanol

Cytoprotective

hr1

ED,, (pg/kgI

0.78
0.82
0.65
0.67
0.68
0.69
0.87
0.81
0.88
0.93

PGE, 3 &kg

on

Vol. 77, No. 3

25
0.05
25
100
10

Each prostaglandin
was given orally, at a dose 20-60 times higher
than the ED,, for cytoprotection.
Gastric juice collected 4 hr after
pylorus
ligation.
Six rats per group. None of the differences
was
statistically
significant.
For comparison,
the oral cytoprotective
ED,, is given in the column on the right.

absolute
ethanol were given orally after pylorus ligation, and the degree of protection
was dose dependent (Table 1). When 16,16-dimethyl
PGE, was given
subcutaneously
after pylorus
ligation
and absolute
ethanol was administered
orally 30 min later, the PG
was also cytoprotective,
and in a dose-dependent
manner
(Table 1).

To ascertain
whether
or not gastric acid, even
in normal
amounts,
was necessary
for the development of lesions produced
by the necrotizing
agents,
as it is the case for gastric lesions produced
by nonsteroidal
antiinflammatory
compounds,*
gastric
acidity was eliminated
by administration
of either:
(a) antisecretory
drugs such as cimetidine
(a histamine H, receptor
antagonist)
and methscopolamine
bromide
(an anticholinergic
agent), or (b) antacids
such as sodium bicarbonate
0.15 M, and a Tris buffer at pH 7 (0.15 M). None of these compounds
prevented the development
of gastric necrosis produced
by absolute
ethanol (Table 4). Methscopolamine
bromide reduced
the number
of lesions by only 17%, a
borderline
decrease,
although
the dose used had
been found in other studies to produce near-total
inhibition
of gastric acid secretion.
In the same experiment, 16,16-dimethyl
PGE, given orally at 3 pg/kg (a
dose that does not affect acid secretion)
completely
prevented
the formation
of ethanol-induced
gastric
lesions.

Discussion
Effect of Cytoprotective
Secretion

PG on Gastric

To find out whether

gastric
cytoprotection
was related
to inhibition
of gastric acid secretion,
the cytoprotective
prostaglandins
listed in Figure 4
were administered
orally to Shay rats 1 hr before
pylorus
ligation,
at a dose found
earlier
to completely
prevent
the formation
of gastric lesions
by
absolute
ethanol.
In fact, the doses chosen were supramaximal,
40-500 times in excess
of the cytoprotective
ED,,. As shown in Table 2, none of the PG
inhibited
gastric
acid secretion.
Similarly,
16,16dimethyl
PGE, given orally to intact, unoperated
rats
at 1 pg/kg (i.e., 20 times the cytoprotective
EDm),
failed to affect gastric acid secretion
at any time during the 30 min after treatment
(Table 3).

Table

3. Effect

of 16,16-dimethyl

PGE, on Gastric

Acid

Secretion:

1 min
Vehicle
No. of animals
Acid content of
stomach,
PEq
PG: 16,16-dimethyl
min after treatment,
nificant.

12
9.4
PGE2, 1 pg/kg in 1 ml, given
and acid output determined.

These studies describe

a model to produce extensive
gastric
necrosis
in rats within
an hour. It
consists
of administering
orally, after a 24 hr fast, 1
ml of a variety of damaging
agents, such as absolute
ethanol,
a strong acid (0.6 N HCl), a strong base (0.2
N NaOH), a hypertonic
solution
(25% NaCl), or boiling water. We found that several prostaglandins
protect the stomach
against
the necrotizing
effect proincluding
a gastric
burn
duced
by these
agents,
produced
by boiling
water.
The forestomach
(the
nonglandular
anterior
portion
of the rat stomach,
lined by a squamous
epithelium)
did not become necrotic. This may be due to the thick layer of cornified epithelium
that covers its surface.
This phenomenon
is called gastric cytoprotection,
since the gastric mucosal
cells are protected
against

Unoperated

Rats
15 min

5 min
PG

Vehicle

PG

Vehicle

30 min
PG

Vehicle

PG

12

10

12

10

10

12

12

12.8

14.9

11.3

15.2

14.6

10.4

13.7

orally. Corresponding
controls received 1 ml of vehicle. Animals were killed from
+:l SEM. None of the differences
between vehicle and PG groups was statistically

l-30
sig-

September

GASTRIC

1979

the effect of necrotizing

agents. Cytoprotection,
gastric and intestinal,
is defined
as the property
of
many PG to protect the mucosa of the stomach
and
the intestine
from becoming
inflamed
and necrotic,
when
this mucosa
is exposed
to noxious
agents.
Cytoprotection
is separate
from and unrelated
to inhibition
of gastric secretion.
This property
of PG appears to be distinct from other known properties
described
for these compounds.
Other examples
of
cytoprotection
by PG were reported
earlier for the
intestine,
and consist of the prevention
of intestinal
lesions produced
in rats either by nonsteroidal
antiinflammatory
compounds,7.1-15
such as indomethatin and flufenamic
acid, or by prolonged
treatment
with high doses of prednisolone..7
These intestinal
lesions,
especially
those produced
by nonsteroidal
antiinflammatory
compounds,
lead to fatal peritonitis. In rats, 15(R)-15-methyl
PGE, methyl ester was
found
to reduce
by half gastric
mucosal
hemorrhages produced
by oral administration
of taurocholate plus HCl and aspirin plus HCl. In other studies, using the pylorus ligated rat, both 16,l&dimethyl
PGE, and cimetidine
inhibited
gastric lesions
produced by aspirin plus HCl.
The mechanism
of gastric cytoprotection
by PG is
unknown.
Several possibilities
have been discussed
in an excellent
review
by Miller and Jacobson.
There is no apparent
structure-activity
relationship
for cytoprotection;
cytoprotective
PG have no common structural
configuration.
Thus, PG of the A, E,
and F type share that property.
The rapidity
with
which the effect takes place (within I min after oral
administration
of a PG) suggests that PG may stimulate gastric cells to secrete some product(s)
acting as
a shield for the mucosa. Such a shield is so effective
that it prevents
damage
produced
by strong chemicals, and it even protects
against the physical
burn
due to boiling water. The following
possible mechanisms can be considered.
Table

4. Effect of Antisecretory
Protection

Mcthscopolamint:

bromide

10 mg/kg
Cimctidinr:
50 m&kg
NaHCO,, 0.15 M 2 ml
pH 7 Tris buffer 2 ml
16,16-Dimethyl
PGE, 3 pg/kg

Drugs

and Antacids

CYTOPROTECTION

BY PROSTAGLANDINS

441

Effect of PG on Mucus
A sudden
secretion
of mucus,
which would
provide a viscous physical
barrier between
the damaging agents and the surface
epithelium,
could explain the protection.
Although
mucus secretion
after
administration
of PG has not been studied
quantitatively,
gastric juice of rats treated
with certain
PG was found to bind more alcian blue than juice of
untreated
rats.22 Alcian
blue is commonly
used to
stain acid mucopolysaccharides
of mucus granules.
Also, biopsies
of human
gastric mucosa
obtained
1
hr after oral administration
of 15(R)-15-methyl
PGE,
methyl
ester showed
an increase
in intracellular
mucus granules.Z3

Effect of PG on Gastric

Acid

Secretion

Cytoprotection
by PG is independent
of inhibition of gastric secretion
for the following
reasons:
(a) The cytoprotective
dose is a small fraction
of the
antisecretory
dose, in the case of PG that possess
antisecretory
activity.
The antisecretory
ED,,, are
more than 100 times higher than the cytoprotective
doses. (b) Certain
cytoprotective
PG are not antisecretory
at any dose when given orally to rats (e.g.,
16,16-dimethyl
PGA,, 15(R)-15-methyl
PGF,,).
(c)
Other antisecretory
agents such as cimetidine
and
methscopolamine
bromide,
as well as antacids,
are
not cytoprotective
in our models
(Table 4). Moreover, in the case of intestinal
cytoprotection,
the lesions produced
by nonsteroidal
antiinflammatory
compounds
(e.g., indomethacin)
and by prednisolone
are located in the jejunum
and the ileum,
regions deprived
of gastric juice and whose luminal
pH is 7 or greater. These intestinal
lesions are also
prevented
by cytoprotective
PG.7 I:7L1

on Gastric

Lesions

Produced

by Absolute

No. of Lesions per Stomach

___.____
_______
Control
Treatment
_~__
17.7 -t 2.05 (72)
14.8 + 1.03 (6)

Ethanol:

Lack of

Statistical
significance
P < 0.05

17.1 +- 1.2 (12)

17.7 * 1.7 (12)

NS

11.1 + 2.03 (8)

10.5 + 2.58 (8)
17.3 +- 1.21 (8)

7.8 f 1.83 (8)

7.1 +- 2.15 (9)

NS
NS
P < 0.01

0 (10)

All animals received 1 ml of absolute ethanol orally and were killed 1 hr later. tJ Mcthscopolamine
bromide: given subcutaneously
1 hr
before absolute ethanol. Cimetidine:
given intraperitoncally
30 min before absolute ethanol. d NaHCO,,
pH 7 Tris buffer: given orally I
min before absolute
ethanol. Control animals received 2 ml of water at the same time; therefore
the absolute ethanol became diluted in
the stomach,
a fact that explains the smaller number of lesions by comparison
to the other control groups. e 16,16-Dimethyl
PGE,: given
orally 30 min bcfow absolute ethanol. f Students t-test. ( ) = number of animals.

442

ROBERT

Local

ET AL.

Versus

Systemic

GASTROENTEROLOGY

Action

of PG

Regardless
of the mechanism
of action, PG do
not need to be placed in direct contact
with gastric
mucosal
cells to exert their action since PG were
cytoprotective
after subcutaneous
as well as after
oral administration.
However,
only one-third
to onefourth as much PG was required
by oral as by subcutaneous
administration.
This may be due to the
high concentration
of PG near gastric cells after oral
administration.
Such high concentration
is not obtained
after parenteral
administration
because,
in
the latter case, the compound
is distributed
throughout the body before reaching
the stomach,
and is
also partially
degraded
by certain organs; therefore,
only a fraction
of the injected dose reaches the stomach. However,
this difference
in potency
between
oral and subcutaneous
routes suggests
that PG can
exert local cytoprotective
action on gastric cells. The
fact that cytoprotection
can take place within 1 min
after a PG has been introduced
in the stomach
also
points
to a local action.
Therefore,
PG are cytoprotective
after reaching
their target cells from either the luminal
or the serosal
side, the luminal
pathway
being more efficient. It is remarkable
that a
PG is cytoprotective
even when placed together with
a necrotizing
agent into a gastric pouch (i.e., in a pylorus ligated stomach)
(Table 1). In that particular
study, although
absolute
ethanol
was trapped
in the
stomach,
and therefore
remained
in contact with the
gastric mucosa for 1 hr, it failed to damage the mucosa when 16,16-dimethyl
PGE, (as little as 1.5 pLg/
kg, corresponding
to a total amount
of 0.3 pg) was
also introduced
in the lumen.
Whether
PGE, is introduced
in the stomach
in a
small (0.5 ml) or large (2.0 ml) volume makes a difference
only at a low cytoprotective
dose. When
such a low dose (10 pg/kg) is diluted
in 2.0 ml, it
loses its cytoprotective
activity,
probably
because
the surface cells are in contact
with a low PG concentration.
At higher
doses,
however,
PGE, is
equally cytoprotective
regardless
of the dilution;
the
effect is probably
too pronounced
to be influenced
by the PG concentration.

Effect

of PC on the Sodium

Pump

16,16-Dimethyl
PGE, was found to stimulate
the sodium
pump of the canine gastric mucosa,
an
effect leading
to an increase
in sodium
transport
from the mucosal to the serosal side. Indomethacin,
given at concentrations
that break the gastric mucosal barrier, reversed
the direction
of sodium transport; administration
of 16,16-dimethyl
PGE, prevented
that
reversal
of sodium
transport
by
indomethacin.25~2
These
results
suggest
that indomethacin,
by depleting
the stomach
of PG, may

Vol. 77, No. 3

damage
the gastric mucosa
by interfering
with sodium transport,
and that, conversely,
PG may be
cytoprotective
by keeping
the sodium
pump intact.
Whether
gastric cytoprotection
against the necrotizing agents used in the present
study is due to the
same mechanism
is unknown.
Direct

Effect of Cells

Instead
of stimulating
the secretion
of a protective substance
from the mucosal cells, or of acting
on the sodium pump, PC may increase
the intrinsic
resistance
of these cells to injury. Although
such a
hypothesis
is speculative
at the moment,
it could be
that cytoprotective
PG act directly on the cell membrane or on some intracellular
structure
that in turn
makes the cell stronger.
In this regard, cyclic AMP
may be the mediator
of cytoprotection,
since several
PG were found to markedly
increase
adenylyl
cyclase activity
in the gastric mucosa
of several
spePGE, was found to double
cies.27.2H 16,16-Dimethyl
the levels of cyclic AMP in the isolated
dog gastric
mucosa.2.y If an increase
in cyclic AMP plays a role
in cytoprotection,
the effect may take place in nonparietal cells since PG were found to increase
cyclic
AMP content
in such cells whereas
they decrease
histamine-stimulated
accumulation
of cyclic AMP in
parietal cells.
Effect on Gastric

Circulation

It is unlikely
that PG exert cytoprotection
by
affecting
gastric
circulation
for the following
reasons: (a) Although
E-type PG were found to decrease
gastric mucosal
blood flow (although
not reducing
the ratio of blood flow to gastric secretion),
this effect was shown only at doses inhibiting
gastric secretion31; in our studies,
16,16-dimethyl
PGE, was
cytoprotective
at about
1/500th
its antisecretory
dose, whereas
other PG were not antisecretory
at all;
(b) PC shown to exert opposite
effects on the vascular system (PGE, is a vasodilator,
PGF,, is a vasoconstrictor) are cytoprotective
for the intestine.3.4 However,
even
if a change
in the gastric
mucosal
circulation
is unlikely
to explain
the cytoprotective
property
of PG, the possibility
should be tested especially since accurate
and sensitive
methods
to measure gastric mucosal blood flow are available.
Some of the implications
of our results are as follows: (a) Certain PG may have a high specificity
for
cytoprotection
without
displaying
any of the other
biologic actions common to many PG, such as stimulation of smooth muscle, inhibition
of gastric secretion, inhibition
or stimulation
of platelet aggregation,
increase
or decrease
of blood pressure,
etc. Such
seemingly
inert compounds
still retain their cytoprotective
property.
Even if a PG exerted some other

September

1979

effects, it is likely that these would not be exhibited

since doses needed for cytoprotection
are so low. (b)
Cytoprotective
PG may have therapeutic
applications
They may be beneficial
in the treatment
of: (1) acute gastritis,
such as caused
by irritants
(alcohol,
aspirin,
and other
nonsteroidal
antiinflammatory
compounds)
and biliary
reflux;
(2)
stress
ulcers
often associated
with sepsis,
severe
trauma,
prolonged
surgical
procedures,
renal transplantation,
extensive
burns; (3) the gastritis
usually
associated
with gastric ulcers; (4) gastric infections
(such as gastroenteritis
of infants);
(5) gastritis of unknown
origin;
and (6) reflux
esophagitis.
Cytoprotective
PG may also accelerate
the healing
of
peptic ulcers and prevent their recurrence.
In conclusion,
several PG protect the gastric mucosa against necrosis
otherwise
produced
by agents
such as absolute
ethanol,
strong acid and base, a hypertonic
solution,
and boiling water. This property
is
called gastric cytoprotection.
Since PG are natural
compounds
present
in most cells, including
gastric
mucosal
cells, it appears
that PG are necessary
for
maintaining
cellular integrity
of the gastric mucosa.
By raising the tissue levels of PG through exogenous
administration,
the cellular resistance
is enhanced
to
the point that gastric cells remain indifferent
to the
presence
of strong irritants
in the lumen.
Endogenous PG may play an important
role in keeping gastric cells intact and functional
in spite of the hostile
environment
of the stomach
which contains
acid at
a pH of 2 or less, pepsin, indiscriminate
foodstuffs,
spices, hypertonic
fluids, and liquids ingested
at a
wide temperature
range.

References
Robert A, Nezamis JE, Phillips JP: Inhibition
of gastric secretion by prostaglandins.
Am J Dig Dis 12:1073-1076,1967
Robert A, Nczamis
JE, Phillips JP: Effect of prostaglandin
E,
on gastric secretion
and ulcer formation
in the rat. Gastroenterology 55:481-487,1968
Robert A, Schultz JR, Nezamis JE, Lancaster
C: Gastric antisecretory
and antiulcer
properties
of PGE,, and 16,16-dimethy1 PGE,. lntravcnous,
oral and intrajejunal
administration.
Gastroenterology
70:359-370,1976
Wilson DE, Phillips C, Levine RA: Inhibition
of gastric secretion in man by prostaglandin
A,. Gastroenterology
61:201206,197l
Karim SMM, Carter
DC, Bhana D, Ganesan
PA: Effect of
orally and intravenously
administered
prostaglandin
15(R)15methyl
E, on gastric secretion
in man. Adv Bios 9255-264,
1973
Robert A, Nylander
B, Andersson
S: Marked inhibition
of gastric secretion
by two prostaglandin
analogs
given orally to
man. Life Sci [I] 14:533-538. 1974
Robert
A: Antisecretory,
antiulcer.
cytoprotective
and
diarrheogenic
properties
of prostaglandins.
In: Advances
in
Prostaglandin
and Thromboxane
Research.
Edited by B Samu&son,
R Paoletti. Raven Press, New York, 1976, p 507-520
Fung WP, Karim SMM, Tye CY: Effect of 15(R)-15-methyl-

GASTRIC

CYTOPROTECTION

BY PROSTAGLANDINS

443

ulcers.
prostaglandin
E, methyl
ester on healing
of gastric
Controlled
endoscopic
study. Lancet Z:lO-12,1974
K, Rybicka J, Mikos E, Nowak A: Double-blind
clini9. Gibinski
cal trial on gastroduodenal
ulcer healing with prostaglandin
E, analogs. Gut 12:636-639,1977
10. Robert A, Nezamis
JE, Lancaster
C, Hanchar
AJ: Gastric cytoprotective
property
of prostaglandins.
Gastroenterology
72:1121. 1977
of gastric hyperacidity
produced
11. Brodie DA: The mechanism
by pylorus ligation in the rat. Am J Dig Dis 11:231-241,1966
acid in aspirin-induced
12. Brodie DA, Chase BJ: Role of gastric
gastric irritation
in the rat. Gastroenterology
53:604-610, 1967
on the stomach
and the
13. Robert A: Effects of prostaglandins
intestine.
Prostaglandins
6:523-532.1974
disease produced
experimentally
by a
14. Robert A: An intestinal
Gastroenterology
69:1045-1047,
prostaglandin
deficiency.
1975
A, Jacobson
ED: Indomethacin-induced
15. Fang WF, Broughton
intestinal
inflammation.
Am J Dig Dis 22:749-760,1977
C, Robert A: Intestinal
lesions produced
by pred16. Lancaster
nisolone: prevention
by 16,16-dimethyl
prostaglandin
E,. Fed
Proc 36:1020,1977
C, Robert A: Intestinal
lesions produced
by pred17. Lancaster
nisolone: prevention
(cytoprotection)
by 16,16-dimethyl
prostaglandin
E,. Am J Physiol235:E703-E708,1978
18. Carmichael
HD, Nelson L, Russell RI, Lyon A, Chandra
V:
The effect of the synthetic
prostaglandin
analogs
15(R)-15
methyl PGE, methyl ester on gastric mucosal hemorrhage
induced in rats by taurocholic
acid and hydrochloric
acid. Am J
Dig Dis 22:411-414.1977
HA, Nelson LM, Russell RI: Cimetidine
and pros19. Carmichael
taglandin:
evidence
for different
modes of action on the rat
gastric mucosa. Gastroenterology
74:1229-1232.
1978
20. Guth PH, Aures D, Paulsen G: Topical aspirin plus HCl gastric lesions in the rat. Cytoprotective
effect of prostaglandin,
cimctidine
and prohanthine.
Gastrocnterology
76:88-93,1979
ED: Gastrointestinal
cytoprotection
by
21. Miller TA, Jacobson
prostaglandins.
Gut 20:75-87, 1979
22. Bolton JP, Palmer D, Cohen MM: Effect of the E, prostaglandins on gastric mucus production
in rats. Surg Forum 27:402403,1976
23 Fung WP, Lee SK, Karim SMM: Effect of prostaglandin
15(R)l5-methyl
E, methyl
ester. An endoscopic
and histological
study. Prostaglandins
10:465-472,1974
24 Bowen JC, Kuo Y-Y, Pawlik W, Williams
D, Shanbour
LL, Jacobson
ED: Electrophysiological
effects of burimamide
and
16,16-dimethyl
prostaglandin
E, on the canine
gastric
mucosa. Gastroenterology
68:1480-1484,1975
25 Jacobson
ED, Chaudhury
TK, Thompson
WL: Mechanism
of
gastric mucosal
cytoprotection
by prostaglandins.
Gastroenterology 70:897,1976
26. Chaudhury
TK, Jacobson
ED: Prostaglandin
cytoprotection
of
gastric mucosa. Gastroenterology
74:58-63.1978
27 Wollins A, Code CF, Dousa TP: Interaction
of prostaglandins
and histamine
with enzymes of cyclic AMP metabolism
from
guinea pig gastric mucosa. J Clin Invest 57:1548,1976
WJ, Chang LK, Jacobson
ED: Rat gastric mucosal
28. Thompson
Adenylyl
cyclase. Gastroenterology
72:244,1977
29. Chaudhury
TK: Mechanism
of the cytoprotective
action of
prostaglandins
in gastric mucosa. Ph.D. Thesis, University
of
Texas, Houston,
1976
30. Sol1 AH: Prostaglandin
inhibition
of histamine-stimulated
aminopyrine
uptake and cyclic AMP generation
by isolated
canine parietal ce11s. Gastroenterology
74:1146, 1978
31. Jacobson
ED: Comparison
of prostaglandin
E, and norepinephrine
on the gastric
mucosal
circulation.
Proc Sot Exp
Biol Med 133:516-519.1970