Transforming Growth Factor Alpha and Epidermal Growth Factor in Protection and Healing of Gastric Mucosal Injury S. J. KONTUREK, T. BRZOZOWSKI, J. MAJKA, A . DEMBINSKI, A. SLOMIANY & B. L. SLOMIANY Institute of Physiology, University Medical School, Cracow, Poland, and Research Center New Jersey, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, USA
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
Konturek SJ, Brzozowski T, Majka J, Dembinski A, Slomiany A, Slomiany BL. Transforming growth factor alpha and epidermal growth factor in protection and healing of gastric mucosal injury. Scand J Gastroenterol 1992;27:649-655.
r
Transforming growth factor alpha (TGF) and epidermal growth factor (EGF) present in the gastric mucosa are polypeptides with similar biologic activity. This study compares the activity of TGF and EGF in the protection against injury by 100% ethanol and stress and in healing of acute gastric ulcerations. TGF and EGF (12.5-100 pg/kg-h) infused subcutaneously 30 min before and during ethanol or stress decreased mucosal lesions dose-dependently. The IDSofor ethanol- and stress-induced lesions after TGF were 40 and 70pg/kg-h and after EGF 60 and lOOpg/kg-h. TGF and EGF infused subcutaneously into intact rats inhibited gastric acid secretion but did not affect the gastric blood flow or mucosal generation of prostaglandin E, (PGE2). Both TGF and EGF also significantly enhanced the healing of stress-induced lesions and the restoration of DNA synthesis. Ethanol and stress reduced blood flow in the oxyntic mucosa by 68% and 51%, respectively, and this effect was partially reversed by EGF and TGF. Pretreatment with indomethacin (5 rng/kg intraperitoneally), which reduced mucosal generation of PGEz by 85%, decreased in part the protection by TGF and EGF against ethanol-induced damage and virtually abolished the protective action of these peptides against stress-induced injury. We conclude therefore that 1) TGF and EGF show similar and comparable gastroprotective activity against ethanol- and stress-induced damage; 2) the protection by TGF and EGF is accompanied by an increase in gastric blood flow which appears to be an essential factor in gastroprotection; 3) mucosal PG is necessary for manifestation of the protective activity of TGF and EGF against acute gastric damage; and 4) TGF and EGF enhance the healing of gastric lesions, possibly via stimulation of DNA synthesis and cell proliferation. Key words: Cytoprotection; ethanol; growth factor; prostaglandins; stress Prof. S. J . Konturek, M . D., Institute of Physiology, 31-5.31 Krakow, ul.Grzegorzecka 16, Poland
Transforming growth factor alpha (TGF) and epidermal growth factor (EGF) (1, 2 ) are homologous peptides produced in the gastrointestinal tract (3-5) and show a similar spectrum of biologic activities (1-3). Their action appears to be mediated via binding to the common EGF/TGF, receptors, which have been identified in the gastric mucosa ( 5 , 6). Although both EGF and TGF inhibit gastric acid secretion, the gastroprotective effect occurs at doses that are too low to affect the gastric secretion, and therefore the mechanism underlying their antiulcer action remains an area of considerable interest. Similar gastroprotective activity is also displayed by prostaglandins (PG) of the E and I series, but little is known about the interdependence of growth factors and mucosal PGs. This study was designed 1) to compare the protective and healing effects of TGF and EGF o n acute gastric lesions; 2)
to determine the role of endogenous PG in the action of T G F and EGF on the gastric mucosa; and 3) t o evaluate the influence of these peptides on gastric blood flow. MATERIALS A N D METHODS
Animals Male Wistar rats, weighing about 300g, were equipped with a chronic gastric fistula ( G F ) and used about 2 weeks later in studies o n gastric secretion. The animals were starved for about 18 h and placed in Bollman-type cages to maintain the minimum restraint necessary. The fistula was opened, and the stomach was rinsed gently with 5-10 ml of tap water at 37°C. Basal gastric secretion was collected for 60 min, and then T G F or EGF was infused subcutaneously by means of a peristaltic pump in gradually increasing doses ranging from 12.5 to 100 pg/kg-h. Each dose was infused for 60 min and then doubled. In control tests, vehicle (saline) was infused
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
650
S. J . Kontutek e1 al.
at the same constant rate of about 4 ml/h. I n separate tests T G F or EGF was given intragastrically in a dose of 100 yg/ kg. Acid and pepsin concentrations were determined in each 30-min collected sample and expressed as mean outputs per 30 min. Acute gastric lesions were produced by 100% ethanol or water immersion and restraint stress as described previously (8). Male Wistar rats weighing 180-200g and starved for 20 h were used. They were divided into several groups (AD). Animals in group A, divided into 6 subgroups of 8-10 rats, received 1.5 ml of 100% ethanol intragastrically and were then infused either with vehicle (saline) or with T G F or E G F at doses of 12.5,25,50, or I00 yg/kg-h. The infusion started 30 min before the administration of ethanol and continued for next 60min. In a separate series of experiments, T G F or E G F was applied intragastrically in a dose of 100 pg/kg, and then 30min later 1.5 ml 100% ethanol was administered. For comparison, 16,16-demethyl PGE2 (dniPGEz) was given intragastrically at a dose of 10 yg/kg 30 min before the application of ethanol. Animals in group B were divided into 6 subgroups of 810 rats and also received 1.5ml 100% ethanol. The first subgroup (control) was given 1 ml of saline intraperitoneally and then 1.5 ml of 100% ethanol; the second subgroup was given indomcthacin, 5 mg/kg, intraperitoneally and 90 rnin later 100% ethanol; the third and fourth subgroups received intraperitoncal injection of saline and 60 min later the infusion of TGF and E G F (50 pg/kg-h), respectively, and, tinally, 90min later 100% ethanol; and the fifth and sixth subgroups were given indomethacin ( 5 mg/kg) intraperitoneally and then 60 niin later infusion of T G F and E G F (50 pg/kg-h), respectively, and, finally, 90 min later 100% ethnnol. In tests with T G F and E G F , each peptide was infused 30 rnin before and 60 min after the administration of e t hand. Rats i n C group were exposed to 3 h of stress by immobilizing them in stress cages and immersing them in 23°C water to their xyphoid process, as described previously (8). In this
series of experiments, the first subgroup (control) received saline or indomethacin (5 mg/kg) intraperitoneally and then 90 rnin later was exposed t o 3 h of stress; the second subgroup was also given indomethacin 90min before the stress and then infused with TGF (25 pg/kg-h) for 30 rnin before and during 3 h of stress; the third subgroup was given indomethacin 90min before the stress and then infused with E G F (25 yg/kg-h) for 30 rnin before and during the 3-h stress. The fourth group was given indomethacin intraperitoneally 90min before the stress, and then dmPGEz (25 pg/kg) was administered intragastrically 30 min before the stress. Group D was used for studies on healing of stress lesions. All rats were exposed to 3 h of stress and then transferred from the stress cages to normal cages t o permit recovery from the stress for 6, 12, or 24 h. Subgroups were infused with vehicle saline (4ml/h), T G F (25 yg/kg-h), or EGF (25 yg/kg-h) starting immediately after the stress and continuing throughout the recovery period in rats without o r with pretreatment with indomethacin ( 5 mg/kg intraperitoneally).
Measurement of gustric lesions One hour after the administration of 100% ethanol and 3 h after the start of water immersion and restraint stress, the animals were anesthetized with ether, the abdomen was opened, and the stomach was exposed for measurement of gastric blood flow. Then the stomach was removed and examined grossly for the area and number o f gastric lesions, using computerized planimetry (Morphomat 10, Zeiss, Berlin, Germany). Measurement of gustric mucosal blood pow Gastric mucosal blood flow was measured by the laser Doppler flowmeter (LDF) technique (9), using a Laserflow, model BPM 403A, Blood Perfusion Monitor (Vasamedics Inc., St. Paul, Minn., USA). Blood flow was measured in three areas o f the oxyntic and antral portions of the stomach
Tab!e 1. Effccts of various doscs of transforming growth factor ( T G F ) alpha or cpidcrrnal growth factor (EGF) on basal gastric acid and pepsin sccrction in rats treated with vehiclc (saline) and T G F o r EGF infused subcutaneously in gradcd doscs (12.5-100 pg/kg-h) or givcn intragastric;illy in a dosc of 100 pg/kg
HCI (pmoI/3O min) Vehicle (control) 'TGF alpha subcutaneously, 12.5 pg/kg-h 'I'GF alpha subcutaneously, 25 pg/kg-h TGF alpha subcutaneously, 50 pg/kg-h TGF alpha subcutaneously, 100 pg/kg-h 'IGF alpha intragastrically, 100 pg/kg E G F subcutaneously, 12.5 pg/kg-h EGF subcutaneously, 25 pg/kg-h E G F subcutaneously, SO pg/kg-h E G F subcutaneously, 100 pg/kg-h EGF intragastrically. 100 kg/kg ~~
~
~
* Significant decrease below the vehicle (control) value
115 % 9 95 8 78 5* 18 2 2* 14 2 4* 92 k 12 110 2 12 98 9 61 +- 7* 28 5* 98' 12
' *
' *
Pepsin (mg/30 min) 0.72 2 0.09 0.69 5 0.07 0.46 t 0.03* 0.12 2 0.01* 0.10 -t 0.01* 0.66 0.08 0.71 2 0.09 0.68 5 0.08 0.43 -+ 0.05* 0.31 -t 0.04* 0.74 2 0.10
*
Growth Factors in Gastric Protection
I I O O ~ /ETHANOL ~ I
Fig. I . Mean area of gastric lesions induced by 100% ethanol in rats infused subcutaneously with vehicle (saline) or various doses of transforming growth factor alpha (TGF,), epidermal growth factor ( E G F ) , or 16,16-dimethyl prostaglandin E2 (dmPGE,). Means t SEM of 8-12 rats. Asterisk indicates significant ( P < 0.05) decrease below the value obtained with vehicle.
Determination of mucosal generation of PGE, In some tests the mucosal samples of the oxyntic gland area were taken by biopsy (about 50 mg) immediately after the animals were killed to determine the mucosal generation of PGE, by means of specific radioimmunoassay. The mucosal sample was placed in preweighed Eppendorf vials, and 1 ml of Tris buffer (50 mM, p H 3.4) was added to each vial. The samples were finally minced (about 15 sec) with scissors, washed, and centrifuged for 10 sec, the pellet being resuspended again in 1 ml of Tris. Then each sample was incubated on a Vortex mixer for 1 min and centrifuged for 15sec. The pellet was weighed, and the supernatant was transferred to a second Eppendorf vial containing indomethacin (10 mM) and kept at -20°C until radioimmunoassay as described previously (11, 12). P G E 2 was measured in duplicate using radioimmunoassay kits (New England Nuclear, Munich, Germany). T h e capability of the mucosa to generate P G E z was expressed in nanograms of wet tissue weight.
on its anterior and posterior wall. The zero value was defined by placing the laser probe against a white board. The area of laser emission of the probe was 1 mm2. Since the depth of measurement by LDF is about 6 mm, the technique used determined total blood flow of the stomach wall. The gastric blood flow measured by LDF was expressed in millimeters per minute per 100 g tissue, as displayed on the digital panel meter and the trend recorder. Previous studies (9) confirmed that total blood How correlates significantly with the gastric mucosal blood flow and the validity of laser Doppler flowmetry for measurement of gastric mucosal blood How.
Determination of DNA synthesis In tests with healing of stress ulcerations, the D N A synthesis was measured in the oxyntic mucosa at 0, 6, 12, and 24 h after stress in tests with or without infusion of TGF o r E G F as described previously (11).In brief, the mucosa was scraped from the oxyntic area and incubated for 30min in culture medium containing 2 yCi/ml of [3H]thymidine (Amersham, U.K.). With thymus D N A as a standard, the D N A content of the samples was determined by the incorporation of [3HH]thymidineinto D N A , and D N A synthesis was expressed as disintegration per minute (dpm) per 1 pg DNA. Results are reported as means 2 SEM. Statistical analyses
-
100
-
7s
0VEHICLE or dm PGE2
N
a
25
50
Y
a
25
W
SI
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
65 1
50 10.0 dmPGE2 VEHICLE 12.5 25 EGF or TGFd(pg/kgh SLI * (lOJJg/kgid
Table 11. Mucosal generation of 16,16-dimethyl prostaglandin E z (drnPGE2) and gastric blood flow measured by laser Doppler flowmetry in intact oxyntic and antral areas and 1 h after intragastric instillation of 1 S m l 100% ethanol with or without pretreatment with indomcthacin (5 mg/kg) intraperitoneally and transforming growth factor (TGF) or epidermal growth factor ( E G F ) (50 pg/kg-h) infused 30 min before and 60 min after administration of ethanol or dmPGEz (10 pg/kg) given intragastrically 30 min before ethanol. Means t SEM of 8-10 determinations in 8-10 rats Blood flow lm1/100 e m i n ) Oxyntic area
Antral area
60.4 t 1.3 62.3 2 2.9 59.7 2 3.1 64.3 t 4.2 19.3 2 1.1* 26.6 t 2.l't 29.1 & 2.2*t 40.2 t 3.2*t 16.8 -+ 1.8' 21.4 ? 2.2* 20.3 t 1.8* 38.4 2 2.7*t
54.2 -+ 1.2 53.1 2 2.8 52.6 t 2.9 59.4 t 3.8 26.5 t 1.S* 29.2 t 1.8' 28.1 t 1.6* 41.5 2 3.2*t 24.5 -+ 2.1* 26.3 t 2.8* 24.1 t 3.3* 39.4 2.9*t
~
Intact TGF EGF dmPGE, Ethanol T G F + ethanol E G F ethanol dmPGE, t ethanol Indomethacin Indomethacin + TGF + ethanol Indomethacin E G F + ethanol Indomethacin dmPGE, + ethanol
+
+ +
420 t 38 392 & 42 448 t 53 418 -t 58 452 i_ 64 410 -t 43 68 zk 12* 14 ? 20* 79 t t6*
-
* Significant (P < 0.05) decrease below the value obtained in intact rats. t Significant ( P < 0.05) increase above the value obtained with 100% ethanol.
*
652
S.
J. Konturek et al.
1-1
0LESION AREA
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
ULESION NUMBER
Fig. 2. Mean area and number of gastric lesions induced by 100% ethanol in rats pretreated with indomethacin, transforming growth factor alpha (TGF,), and epidermal growth factor (EGF), or their combination. Means SEM of %lo rats. Asterisk indicates significant decrease below the value obtained with vchicle (saline).
*
were determined by analysis of variance and, when appropriate, by the unpaired Student’s t test; a Pvalue of less than 0.05 was considered significant. RESULTS The effects of graded doses of TGF and EGF on gastric acid and pepsin secretion from the chronic gastric fistulas in rats are shown on Table I. In control rats infused with vehicle (saline), basal acid output averaged 115 2 9 pmol/30 min, and pepsin output 0.72 i 0.09 mg/30 min. T G F resulted in significant inhibition of acid and pepsin secretion at a dose of 25 pg/kg-h or higher, and EGF was effective in gastric inhibition at a dose of 50 pg/kg-h or higher. TGF and EGF given intragastrically at a dose of 100 pg/kg did not affect gastric secretion of acid or pepsin. The effects of graded doses of TGF and EGF on the formation of acute gastric lesions induced by 100% ethanol are shown in Fig. 1. Both peptides reduced dose-dependently the area of gastric lesions caused by ethanol. The reduction was significant at the dose of 25 pg/kg-h of TGF and of SO pg/kg-h of EGF, and the dose reducing the area of these lesions by SO% (IDso) averaged about 25 yg/kg-h for TGF and about 40 pg/kg-h for EGF. At a dose of 100 pg/kg-h each of these peptides prevented the formation of gastric lesions to a similar extent as dmPGE, (lOyg/kg) given intragastrically. TGF or EGF applied intragastrically at a dose of 100 pg/kg did not affect significantly the area or the numbcr of ethanol-induced gastric lesions, and these results were not included, for the sake of clarity. Pretreatment with indomethacin, which reduced mucosal generation of PGE2 by about 85% (Table II), tended to increase ethanol-induced damage, but this was not significant (Fig. 2). Infusion of TGF or EGF at a dose of SOpg/kg-h caused a reduction in gastric lesion area similar to that observed in previous series of experiments, but the pre-
treatment with indomethacin significantly attenuated the reduction in mucosal damage obtained with infusion of T G F or EGF (Fig. 2). Pretreatment with indomethacin did not affect significantly the reduction in mucosal damage obtained with dmPGEZ,and these results have not been included. Gastric blood flow as measured by LDF in the oxyntic and antral portions of the intact stomach averaged 60.4 2 1.3 and 54.2 2 1.2 ml/min-100 g tissue, and infusion of TGF or EGF at a dose of SO pg/kg-h did not affect this blood flow significantly (Table 11). Administration of 1.5 ml of 100% ethanol, which produced widespread mucosal lesions, resulted in about 68% and 51% reduction of the blood flow in the oxyntic and antral area, respectively. Infusion of TGF or E G F in ethanol-treated rats resulted in a small but significant increase in the mucosal blood flow in tests without pretreatment with indomethacin. After pretreatment with indomethacin the mucosal blood flow was reduced to a similar extent as in tests with ethanol alone, but the addition of TGF or EGF did not affect this reduced blood flow. In contrast, the pretreatment with dmPGE2 caused a marked increase in the blood flow as compared with that recorded in the ethanol-treated stomach both in rats without and with the pretreatment with indomethacin (Table 11). The exposure of rats to 3 h of water immersion and restraint stress caused visible lesions in the oxyntic gland mucosa in all animals tested. The damage appeared as elongated bands of bleeding erosions (about 12 _t 0.6 lesions per rat) following the long axis of the mucosal folds. lnfusion of TGF or EGF (25 yg/kg-h) 30 min before and during 3 h of stress significantly reduced (by about 60%) the number of gastric lesions, and this effect was similar to that attained
I WATER IMMERSION+RESTRAINTSTRESS ( 3hI I 0WITHOUT
INDOMEMACIN INDOMETHACIN
Fig. 3. Mean number of gastric lesions induced by 3 h of water immersion and restraint stress in rats without or with pretreatment with indomethacin and with infusion of transforming growth factor alpha (TGF,), epidermal growth factor (EGF), or Ih,lh-dimethyl prostaglandin E2 (dmPGE,). Means 2 SEW of 8-10 rats. Single asterisk indicates significant increase above the value obtained in corresponding rats without pretreatment with indomethacin. Double asterisks indicate significant decrease below the valuc obtained with vehicle control.
Growth Factors in Gastric Protection
653
Table Ill. Mucosal generation of 16,16-dimethyl prostaglandin E, (dmPGE,) and gastric blood flow measured by laser Doppler flowmetry in oxyntic and antral mucosa in rats, measured after 3 h of water immersion and restraint stress with and without pretreatment with intraperitoneal indomethacin (5 mg/kg) and/or transforming growth factor (TGF) or epidermal growth factor (EGF) (25 pg/kg-h) infused subcutaneously 30 rnin before and during 3 h of stress. Means 5 SEM of 8-10 determinations in 8-10 rats
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
Blood flow (rnl/min-100 g)
Intact Stress TGF + stress E G F + stress dmPGE2 + stress Indomethacin Indomethacin + TGF + stress E G F + stress Indomethacin Indomethacin + dmPGE, + stress
+
PGE, w g j
Oxyntic area
Antral area
480 ? 56 289 t 37" 271 ? 48* 238 5 52* 54 -t 7* 62 ? 8* 75 t 12* -
56.8 t 5.7 27.9 t 3.8* 38.1 t 3.2*t 36.2 t 2.4*1 41.8 t 3.8*t 19.6 ? 2.4* 2 1 . 2 t 1.8* 20.1 ? 2.4* 37.2 t 2.7*t
43.4 t 5.9 30.3 2 4.6* 37.2 ? 5.6 35.3 t 4.2 39.2 ? 3.8*t 20.7 t 1.5* 22.6 ? 2.4* 23.8 2 1.9* 36.4 -+ 2.3*t
* Significant ( P < 0.05) decrease below the value obtained in intact rats. t Significant ( P < 0.05) increase above the value obtained with stress.
after intragastric administration of dmPGEz (25 pg/kg-h). Pretreatment with indomethacin ( 5 mg/kg intraperitoneally) significantly increased (by about 26%) the number of gastric lesions and almost completely eliminated the reduction in lesion number achieved with infusion of TGF or EGF. In contrast, indomethacin failed to affect the reduction in stressinduced gastric damage obtained with dmPGEz (Fig. 3). Gastric mucosal generation of PGEz was decreased by about 40% after 3 h of stress, and pretreatment with indomethacin caused further decline in mucosal PGEz to about 10% of the normal value (Table 111). Infusion of TGF or EGF before and during the stress did not significantly affect the generation of PGEz in the oxyntic mucosa of rats without or with pretreatment with indomethacin (Table 111). After 3 h of stress the gastric blood flow was reduced significantly as compared with that in the intact stomach. Pretreatment with indomethacin ( 5 mg/kg intraperitoneally) resulted in a further significant decrease in the gastric blood flow. Infusion of TGF or EGF (25 bg/kg-h) caused a small but significant increase in the blood flow in rats without
indomethacin but not in those pretreated with indomethacin (Table 111). In tests with intragastric administration of dmPGEz the mucosal blood flow was significantly increased (over that observed in the stomach exposed to 3 h of stress alone) in rats with or without pretreatment with indomethacin. The time course of gastric lesions observed at 0 , 6, 12, and 24 h after withdrawal of stress without or with administration of TGF or EGF is shown in Fig. 4. In rats treated with TGF or E G F the disappearance of gastric lesions was significantly accelerated at 6, 12, and 24 h after stress. Mucosal generation of PGEz started to increase after 12 h and was almost completely restored after 24h of stress (Table IV). The incorporation of labeled thymidine into DNA in the intact oxyntic mucosa was about 38.0 1.6 dpm/ pg DNA. Significant increase in DNA synthesis started 12 h after stress, and almost complete restoration of DNA synthesis occurred after 24 h. Mucosal blood flow was almost completely restored to control value already 6 h after stress. The infusion of TGF or EGF (25 pg/kg-h) caused significant
*
Table IV. Generation of prostaglandin E 2 (PGEJ and DNA synthesis in oxyntic mucosa and blood flow in the oxyntic and antral area of intact stomach at 0, 6, 12, and 24 h after water immersion and restraint stress. Means SEM of &lo rats
*
Blood flow (ml/min-100 g)
Intact stomach 0 h after stress 6 h after stress 12 h after stress 24 h after stress TGF, 25 pg/kg-h, for 12 h after stress EGF, 25 pg/kg-h, for 12 h after stress ~
PGE, (ng/d
DNA synthesis, dpm/CLg DNA
Oxyntic area
Antral area
420 ? 58 231 2 48* 218 t 42* 264 t 47* 382 ? 50t 254 t 36*
38.0 +- 1.6 25.1 t 2.6* 27.2 ? 1.9* 30.5 2.0t 34.6 t 4.3t 34.2 t 3.2t
*
58.6 f 6.2 31.2 t 2.7* 67.8 f 6.5t 61.9 f 8.37 64.7 f 4.2t 64.7 f 5.8t
54.2 t 4.8 34.8 2 3.6* 64.9 t 5.5* 63.7 5 4.2t 59.3 t 5.2t 58.2 ? 4.2t
268 ? 42*
36.7 t 2.8t
62.8 f 7.4t
53.4 t 4.7t
~
TGF = transforming growth factor; EGF = epidermal growth factor. * Significant ( P < 0.05) decrease below the value recorded in intact stomach. t Significant increase above the value recorded immediately (at 0 h) after stress
S. J . Konturek et al.
6.54
[HEALING OF STRESS LESIONS]
15 & W
I
m
3 z
z
OVEHICLE TGFd (ZM91kg-h s.c.)
10
0
ezll
5
a w Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
r
HOURS AFTER STRESS Fig. 4. Mean number of stress-induced gastric lesions at 0, 6 , 12, or 24 h after stress in rats treated with vehicle (saline), transforming growth factor alpha (TGF,), or epidermal growth factor (EGF). Means * SEM of 8-12 rats. Asterisk indicates significant decrease below the value obtained in rats treated with vehicle.
increase in DNA synthesis but did not influence mucosal generation of PGE2 or gastric blood flow as measured 12 h after stress (Table IV). DISCUSSION The results demonstrate that T G F and EGF show similar gastroprotective activity against acute damage induced by ethanol or stress and have similar healing effect on acute gastric ulcerations induced by stress. Prcviouc studies showed that the major source of EGF is the salivary glands ( 3 , 12, 13), but small amounts of the peptide have also been detected in gastric niucosa (3,4, 12), where specific binding of E G F receptors has been observed (6). Removal of salivary glands was reported to exacerbate the rnucosal lesions induced by various topical irritants and ulcerogens, including stress ( 3 , 12, 13), and to abolish the adaptive cytoprotection (14). Furthermore, sialoadenectomy has been shown to inhibit the healing of chronic gastroduodenal ulcerations (15, 16). Exogenous EGF is also known to protect the gastric mucosa against acute damage provoked by both acid-dependent (stress, aspirin, bile salts) and acid-independent (ethanol) mucosal lesions ( 3 ) . These results could be interpreted as showing that functional salivary glands and constant generation of E G F are essential for both gastroprotection and mucosal repair processes (11, 12, 14). Our data presented here confirm the protective activity of E G F against ethanol- and stress-induced damage and show that this protection is mediated by mucosal PG. The suppression of P G generation by indomethacin (to about 15% of normal value) reduced the protective effect of EGF on ethanol-induced damage and completely reversed its protective action against stress-induced gastric lesions. In con-
trast, dmPGE2-;nduced gastroprotection against ethanol- or stress-induced lesions was not affected by pretreatment with indomethacin. It has been suggested that the interaction between EGF and PG may involve the maintenance of the mucosal blood flow (17). Indeed, our results showed that E G F attenuated the reduction in mucosal blood flow caused by ethanol o r stress and that indomethacin abolished this effect. Exogenous drnPGEz improved the mucosal blood flow in rats exposed to ethanol or stress to a similar extent in tests without and with indoniethacin pretreatment, whereas exogenous EGF was effective only in the animals without administration of indomethacin when mucosal PG biosynthesis was preserved even at a reduced level. This indicates that mucosal PG is necessary for the maintenance of gastric blood flow by EGF. T G F was originally suggested to function only during embryonic development and to be abnormally expressed by tumor cells, resulting in autocrine secretion and tumor growth (18). Cartlidge & Elder (4) recently demonstrated the presence of immunoreactive TGF in human mucosa throughout the gastrointestinal tract and especially in the stomach. Beauchamp et al. (5) reported the expression of TGFmRNA and TGFa/EGF receptor m R N A in normal gastric mucosa of the adult guinea pig, rat, and dog. They concluded that the local production of TGF may play an important role in the regulation of acid secretion and mucosal cell proliferation. Our study demonstrates that TGF is about twice as potent as EGF in the inhibition of gastric secretion and in the protection of gastric mucosal integrity against ethanol- and stress-induced damage. Thus T G F , like E G F , protects the gastric mucosa against the damage by both acid-independent (ethanol) and acid-dependent (stress) irritants. As in the case of E G F , the protection afforded by TGF against ethanol or stress-induced mucosal damage apparently requires the preservation of mucosal PG biosynthesis and blood flow. Since the protective effects of TGF or EGF against stressinduced lesions occurred at doses that caused partial inhibition of gastric acid secretion, the possibility could not be excluded that this factor also plays some role in gastroprotection by these peptides. One of the important considerations is whether TGF and E G F play any physiologic role in the maintenance of the integrity of gastrointestinal mucosa. Studies show that stress augments the release of E G F and increases its content in the gastric mucosa, whereas sialoadenectomy greatly augments the stress-induced gastric lesions, probably due to the deficiency of EGF (12). O n the other hand, TGF and E G F applied intragastrically failed to affect gastric acid secretion or to prevent the mucosal damage, suggesting that these peptides released into the gut lumen even in excessive amounts are unable to influence the mucosal integrity. Studies are required to determine whether alterations in T G F and E G F levels in gastric mucosa under various phy-
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
Growth Factors in Gastric Protection
siopathologic conditions correlate with the degree of mucosal damage. Another aspect of our study was the role of TGF and EGF in the recovery of gastric mucosa from the damage provoked by stress. Our results confirmed the findings of Wang & Johnson (18) that the healing of acute mucosal lesions is significant 12 h after the withdrawal of the stress and that this is accompanied by a marked increase in mucosal blood flow already evident 6 h after the stress and a significant increase in DNA synthesis and mucosal generation of PGE2 12 and 24 h after the stress. Both TGF and EGF were found to enhance significantly the healing of stress ulcerations, which was accompanied by a significant stimulation in DNA synthesis but not PGE2generation. Although the mechanism of the observed healing effects by TGF and EGF has not been investigated, our contention is that this is probably due to the stimulation by the peptides of the process of early epithelization or mucosal restitution (3) and of the cell proliferation in the gastric mucosa. Indeed, both EGF and TGF were shown to stimulate the proliferation of mucosal cells in the stomach (19-21), and our results are in keeping with these findings.
ACKNOWLEDGEMENTS This work was supported in part by grants 403989101 from the National Research Committee, Warsaw, Poland, USPHS grantsfrom the National Institute of Diabetes and Digestive and Kidney Diseasesfrom the National Institute of Alcohol Abuse and Alcoholism, and-rom the National Institute of Dental Research, NIH.
655
epidermal growth factor levels in normal gastrointestine mucosa. Br J Cancer 1989;60:657-60. 5. Beauchamp RD, Barnard JA, McCutchen CM, Cherner JA, Coffey RJ. Localization of transforming growth factor a and its receptor in gastric mucosa cells. J Clin Invest 1989;84:1017-23. 6. Slomiany BL, Lin J , Yao P, et al. Characterization of the epidermal growth factor receptor in gastric mucosa. Digestion 1990~47:181-90. 7. Lane A, Ivy AC. Response to the chronic gastric fistula rat to histamine Am J Physiol 1957;190:221-6. 8. Konturek SJ, Brzozowski T, Drozdowicz D, Beck G. Role of leukotrienes in acute gastric lesions induced by ethanol, taurocholate, aspirin, platelet-activating factor and stress in rats. Dig Dis Sci 1989;33:806-13. 9. Ahu H, Ivarsson LE, Johansson K, Lindhagen J , Lundgren 0. Assessment of gastric blood flow with laser Doppler flowmetry. Scand J Gastroenterol 1988;23:1203-10. 10. Chen BW, Hui WM, Lamm SK, Cho CH, Ng MMT, Luk CT. Effect of sucralfate on gastric blood flow in rats. Gut 1989;30:1544-51. 11. Konturek PK, Brzozowski T, Konturek SJ, Dembinski A. Role of epidermal growth factor, prostaglandin and sulfhydryls in stress-induced gastric lesions. Gastroenterology 1990;YY:160715. 12. Konturek SJ, BrzozowskiT, Konturek PK, Majka J , Dembinski A. Role of salivary glands and epidermal growth factor in gastric secretion and mucosal integrity in rats exposed to stress. Reg Pept 1991;32:203-15. 13. Olsen PS, Poulsen SS, Kirkegaard P, Nexo. Role of submandibular saliva and epidermal growth factor in gastric cytoprotection. Gastroenterology 1984;87:103-6. 14. Tepperman BL, Soper BD, Morris GP. Effect of sialoadenectomy on adaptive cytoprotection in the rat. Gastroenterology 1989;97:123-9. 15. Olsen SP, Poulsen SS, Therkelsen K, Nexo E . Effect of sialoadenectomy and synthetic human urogastrone on healing of chronic gastric ulcers in rats. Gut 1986;27:1443-9. 16. Konturek SJ, Dembinski A, Warzecha Z, Brzozowski T, Gregory H . Role of epidermal growth factor in healing of chronic gastroduodenal ulcers in rats. Gastroenterology 1988;94:130% I. 17. Maramatsu I , Hollenberg MD, Lederis K. Vascular actions of epidermal growth factor urogastrone; possible relation to prostaglandin production. Can J Physiol Pharmacol 1985;63: 994-5.
REFERENCES 1. Waterfield MD. Epidermal growth factors and related molecules. Lancet 1989;l: 1243-6. 2. Gregory H. Isolation and structure of urogastrone and its relationship to epidermal growth factor. Nature 1975;257:3257. 3. Konturek SJ. Role of growth factors in gastroduodenal protection and healing of peptic ulcers. Gastroenterol Clin N Am 1990;1Y:41-65. 4. Cartlidge SA, Elder JB. Transforming growth factor a and Received 9 October 1991 Accepted 23 March 1992
18. Wang J-Y, Johnson LR. Luminal polyamines stimulate repair of gastric mucosal stress ulcers. Am J Physiol 1990;259:G58492. 19. Dembinski A , Gregory H, Konturek SJ, Polanski M. Trophic action of epidermal growth factor on the pancreas and gastrointestinal mucosa in rats. J Physiol 1982;325:35-42. 20. Derynek R. Transforming growth factor a structure and biological activities. J Cell Biol 1986;32:293-304. 21. Schreiber HB, Winkler MF, Derynck R. Transforming growth factors a more potent angiogenic mediator than epidermal growth factor. Science 1986;232:1250-3.
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
Konturek SJ, Brzozowski T, Majka J, Dembinski A, Slomiany A, Slomiany BL. Transforming growth factor alpha and epidermal growth factor in protection and healing of gastric mucosal injury. Scand J Gastroenterol 1992;27:649-655.
r
Transforming growth factor alpha (TGF) and epidermal growth factor (EGF) present in the gastric mucosa are polypeptides with similar biologic activity. This study compares the activity of TGF and EGF in the protection against injury by 100% ethanol and stress and in healing of acute gastric ulcerations. TGF and EGF (12.5-100 pg/kg-h) infused subcutaneously 30 min before and during ethanol or stress decreased mucosal lesions dose-dependently. The IDSofor ethanol- and stress-induced lesions after TGF were 40 and 70pg/kg-h and after EGF 60 and lOOpg/kg-h. TGF and EGF infused subcutaneously into intact rats inhibited gastric acid secretion but did not affect the gastric blood flow or mucosal generation of prostaglandin E, (PGE2). Both TGF and EGF also significantly enhanced the healing of stress-induced lesions and the restoration of DNA synthesis. Ethanol and stress reduced blood flow in the oxyntic mucosa by 68% and 51%, respectively, and this effect was partially reversed by EGF and TGF. Pretreatment with indomethacin (5 rng/kg intraperitoneally), which reduced mucosal generation of PGEz by 85%, decreased in part the protection by TGF and EGF against ethanol-induced damage and virtually abolished the protective action of these peptides against stress-induced injury. We conclude therefore that 1) TGF and EGF show similar and comparable gastroprotective activity against ethanol- and stress-induced damage; 2) the protection by TGF and EGF is accompanied by an increase in gastric blood flow which appears to be an essential factor in gastroprotection; 3) mucosal PG is necessary for manifestation of the protective activity of TGF and EGF against acute gastric damage; and 4) TGF and EGF enhance the healing of gastric lesions, possibly via stimulation of DNA synthesis and cell proliferation. Key words: Cytoprotection; ethanol; growth factor; prostaglandins; stress Prof. S. J . Konturek, M . D., Institute of Physiology, 31-5.31 Krakow, ul.Grzegorzecka 16, Poland
Transforming growth factor alpha (TGF) and epidermal growth factor (EGF) (1, 2 ) are homologous peptides produced in the gastrointestinal tract (3-5) and show a similar spectrum of biologic activities (1-3). Their action appears to be mediated via binding to the common EGF/TGF, receptors, which have been identified in the gastric mucosa ( 5 , 6). Although both EGF and TGF inhibit gastric acid secretion, the gastroprotective effect occurs at doses that are too low to affect the gastric secretion, and therefore the mechanism underlying their antiulcer action remains an area of considerable interest. Similar gastroprotective activity is also displayed by prostaglandins (PG) of the E and I series, but little is known about the interdependence of growth factors and mucosal PGs. This study was designed 1) to compare the protective and healing effects of TGF and EGF o n acute gastric lesions; 2)
to determine the role of endogenous PG in the action of T G F and EGF on the gastric mucosa; and 3) t o evaluate the influence of these peptides on gastric blood flow. MATERIALS A N D METHODS
Animals Male Wistar rats, weighing about 300g, were equipped with a chronic gastric fistula ( G F ) and used about 2 weeks later in studies o n gastric secretion. The animals were starved for about 18 h and placed in Bollman-type cages to maintain the minimum restraint necessary. The fistula was opened, and the stomach was rinsed gently with 5-10 ml of tap water at 37°C. Basal gastric secretion was collected for 60 min, and then T G F or EGF was infused subcutaneously by means of a peristaltic pump in gradually increasing doses ranging from 12.5 to 100 pg/kg-h. Each dose was infused for 60 min and then doubled. In control tests, vehicle (saline) was infused
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
650
S. J . Kontutek e1 al.
at the same constant rate of about 4 ml/h. I n separate tests T G F or EGF was given intragastrically in a dose of 100 yg/ kg. Acid and pepsin concentrations were determined in each 30-min collected sample and expressed as mean outputs per 30 min. Acute gastric lesions were produced by 100% ethanol or water immersion and restraint stress as described previously (8). Male Wistar rats weighing 180-200g and starved for 20 h were used. They were divided into several groups (AD). Animals in group A, divided into 6 subgroups of 8-10 rats, received 1.5 ml of 100% ethanol intragastrically and were then infused either with vehicle (saline) or with T G F or E G F at doses of 12.5,25,50, or I00 yg/kg-h. The infusion started 30 min before the administration of ethanol and continued for next 60min. In a separate series of experiments, T G F or E G F was applied intragastrically in a dose of 100 pg/kg, and then 30min later 1.5 ml 100% ethanol was administered. For comparison, 16,16-demethyl PGE2 (dniPGEz) was given intragastrically at a dose of 10 yg/kg 30 min before the application of ethanol. Animals in group B were divided into 6 subgroups of 810 rats and also received 1.5ml 100% ethanol. The first subgroup (control) was given 1 ml of saline intraperitoneally and then 1.5 ml of 100% ethanol; the second subgroup was given indomcthacin, 5 mg/kg, intraperitoneally and 90 rnin later 100% ethanol; the third and fourth subgroups received intraperitoncal injection of saline and 60 min later the infusion of TGF and E G F (50 pg/kg-h), respectively, and, tinally, 90min later 100% ethanol; and the fifth and sixth subgroups were given indomethacin ( 5 mg/kg) intraperitoneally and then 60 niin later infusion of T G F and E G F (50 pg/kg-h), respectively, and, finally, 90 min later 100% ethnnol. In tests with T G F and E G F , each peptide was infused 30 rnin before and 60 min after the administration of e t hand. Rats i n C group were exposed to 3 h of stress by immobilizing them in stress cages and immersing them in 23°C water to their xyphoid process, as described previously (8). In this
series of experiments, the first subgroup (control) received saline or indomethacin (5 mg/kg) intraperitoneally and then 90 rnin later was exposed t o 3 h of stress; the second subgroup was also given indomethacin 90min before the stress and then infused with TGF (25 pg/kg-h) for 30 rnin before and during 3 h of stress; the third subgroup was given indomethacin 90min before the stress and then infused with E G F (25 yg/kg-h) for 30 rnin before and during the 3-h stress. The fourth group was given indomethacin intraperitoneally 90min before the stress, and then dmPGEz (25 pg/kg) was administered intragastrically 30 min before the stress. Group D was used for studies on healing of stress lesions. All rats were exposed to 3 h of stress and then transferred from the stress cages to normal cages t o permit recovery from the stress for 6, 12, or 24 h. Subgroups were infused with vehicle saline (4ml/h), T G F (25 yg/kg-h), or EGF (25 yg/kg-h) starting immediately after the stress and continuing throughout the recovery period in rats without o r with pretreatment with indomethacin ( 5 mg/kg intraperitoneally).
Measurement of gustric lesions One hour after the administration of 100% ethanol and 3 h after the start of water immersion and restraint stress, the animals were anesthetized with ether, the abdomen was opened, and the stomach was exposed for measurement of gastric blood flow. Then the stomach was removed and examined grossly for the area and number o f gastric lesions, using computerized planimetry (Morphomat 10, Zeiss, Berlin, Germany). Measurement of gustric mucosal blood pow Gastric mucosal blood flow was measured by the laser Doppler flowmeter (LDF) technique (9), using a Laserflow, model BPM 403A, Blood Perfusion Monitor (Vasamedics Inc., St. Paul, Minn., USA). Blood flow was measured in three areas o f the oxyntic and antral portions of the stomach
Tab!e 1. Effccts of various doscs of transforming growth factor ( T G F ) alpha or cpidcrrnal growth factor (EGF) on basal gastric acid and pepsin sccrction in rats treated with vehiclc (saline) and T G F o r EGF infused subcutaneously in gradcd doscs (12.5-100 pg/kg-h) or givcn intragastric;illy in a dosc of 100 pg/kg
HCI (pmoI/3O min) Vehicle (control) 'TGF alpha subcutaneously, 12.5 pg/kg-h 'I'GF alpha subcutaneously, 25 pg/kg-h TGF alpha subcutaneously, 50 pg/kg-h TGF alpha subcutaneously, 100 pg/kg-h 'IGF alpha intragastrically, 100 pg/kg E G F subcutaneously, 12.5 pg/kg-h EGF subcutaneously, 25 pg/kg-h E G F subcutaneously, SO pg/kg-h E G F subcutaneously, 100 pg/kg-h EGF intragastrically. 100 kg/kg ~~
~
~
* Significant decrease below the vehicle (control) value
115 % 9 95 8 78 5* 18 2 2* 14 2 4* 92 k 12 110 2 12 98 9 61 +- 7* 28 5* 98' 12
' *
' *
Pepsin (mg/30 min) 0.72 2 0.09 0.69 5 0.07 0.46 t 0.03* 0.12 2 0.01* 0.10 -t 0.01* 0.66 0.08 0.71 2 0.09 0.68 5 0.08 0.43 -+ 0.05* 0.31 -t 0.04* 0.74 2 0.10
*
Growth Factors in Gastric Protection
I I O O ~ /ETHANOL ~ I
Fig. I . Mean area of gastric lesions induced by 100% ethanol in rats infused subcutaneously with vehicle (saline) or various doses of transforming growth factor alpha (TGF,), epidermal growth factor ( E G F ) , or 16,16-dimethyl prostaglandin E2 (dmPGE,). Means t SEM of 8-12 rats. Asterisk indicates significant ( P < 0.05) decrease below the value obtained with vehicle.
Determination of mucosal generation of PGE, In some tests the mucosal samples of the oxyntic gland area were taken by biopsy (about 50 mg) immediately after the animals were killed to determine the mucosal generation of PGE, by means of specific radioimmunoassay. The mucosal sample was placed in preweighed Eppendorf vials, and 1 ml of Tris buffer (50 mM, p H 3.4) was added to each vial. The samples were finally minced (about 15 sec) with scissors, washed, and centrifuged for 10 sec, the pellet being resuspended again in 1 ml of Tris. Then each sample was incubated on a Vortex mixer for 1 min and centrifuged for 15sec. The pellet was weighed, and the supernatant was transferred to a second Eppendorf vial containing indomethacin (10 mM) and kept at -20°C until radioimmunoassay as described previously (11, 12). P G E 2 was measured in duplicate using radioimmunoassay kits (New England Nuclear, Munich, Germany). T h e capability of the mucosa to generate P G E z was expressed in nanograms of wet tissue weight.
on its anterior and posterior wall. The zero value was defined by placing the laser probe against a white board. The area of laser emission of the probe was 1 mm2. Since the depth of measurement by LDF is about 6 mm, the technique used determined total blood flow of the stomach wall. The gastric blood flow measured by LDF was expressed in millimeters per minute per 100 g tissue, as displayed on the digital panel meter and the trend recorder. Previous studies (9) confirmed that total blood How correlates significantly with the gastric mucosal blood flow and the validity of laser Doppler flowmetry for measurement of gastric mucosal blood How.
Determination of DNA synthesis In tests with healing of stress ulcerations, the D N A synthesis was measured in the oxyntic mucosa at 0, 6, 12, and 24 h after stress in tests with or without infusion of TGF o r E G F as described previously (11).In brief, the mucosa was scraped from the oxyntic area and incubated for 30min in culture medium containing 2 yCi/ml of [3H]thymidine (Amersham, U.K.). With thymus D N A as a standard, the D N A content of the samples was determined by the incorporation of [3HH]thymidineinto D N A , and D N A synthesis was expressed as disintegration per minute (dpm) per 1 pg DNA. Results are reported as means 2 SEM. Statistical analyses
-
100
-
7s
0VEHICLE or dm PGE2
N
a
25
50
Y
a
25
W
SI
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
65 1
50 10.0 dmPGE2 VEHICLE 12.5 25 EGF or TGFd(pg/kgh SLI * (lOJJg/kgid
Table 11. Mucosal generation of 16,16-dimethyl prostaglandin E z (drnPGE2) and gastric blood flow measured by laser Doppler flowmetry in intact oxyntic and antral areas and 1 h after intragastric instillation of 1 S m l 100% ethanol with or without pretreatment with indomcthacin (5 mg/kg) intraperitoneally and transforming growth factor (TGF) or epidermal growth factor ( E G F ) (50 pg/kg-h) infused 30 min before and 60 min after administration of ethanol or dmPGEz (10 pg/kg) given intragastrically 30 min before ethanol. Means t SEM of 8-10 determinations in 8-10 rats Blood flow lm1/100 e m i n ) Oxyntic area
Antral area
60.4 t 1.3 62.3 2 2.9 59.7 2 3.1 64.3 t 4.2 19.3 2 1.1* 26.6 t 2.l't 29.1 & 2.2*t 40.2 t 3.2*t 16.8 -+ 1.8' 21.4 ? 2.2* 20.3 t 1.8* 38.4 2 2.7*t
54.2 -+ 1.2 53.1 2 2.8 52.6 t 2.9 59.4 t 3.8 26.5 t 1.S* 29.2 t 1.8' 28.1 t 1.6* 41.5 2 3.2*t 24.5 -+ 2.1* 26.3 t 2.8* 24.1 t 3.3* 39.4 2.9*t
~
Intact TGF EGF dmPGE, Ethanol T G F + ethanol E G F ethanol dmPGE, t ethanol Indomethacin Indomethacin + TGF + ethanol Indomethacin E G F + ethanol Indomethacin dmPGE, + ethanol
+
+ +
420 t 38 392 & 42 448 t 53 418 -t 58 452 i_ 64 410 -t 43 68 zk 12* 14 ? 20* 79 t t6*
-
* Significant (P < 0.05) decrease below the value obtained in intact rats. t Significant ( P < 0.05) increase above the value obtained with 100% ethanol.
*
652
S.
J. Konturek et al.
1-1
0LESION AREA
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
ULESION NUMBER
Fig. 2. Mean area and number of gastric lesions induced by 100% ethanol in rats pretreated with indomethacin, transforming growth factor alpha (TGF,), and epidermal growth factor (EGF), or their combination. Means SEM of %lo rats. Asterisk indicates significant decrease below the value obtained with vchicle (saline).
*
were determined by analysis of variance and, when appropriate, by the unpaired Student’s t test; a Pvalue of less than 0.05 was considered significant. RESULTS The effects of graded doses of TGF and EGF on gastric acid and pepsin secretion from the chronic gastric fistulas in rats are shown on Table I. In control rats infused with vehicle (saline), basal acid output averaged 115 2 9 pmol/30 min, and pepsin output 0.72 i 0.09 mg/30 min. T G F resulted in significant inhibition of acid and pepsin secretion at a dose of 25 pg/kg-h or higher, and EGF was effective in gastric inhibition at a dose of 50 pg/kg-h or higher. TGF and EGF given intragastrically at a dose of 100 pg/kg did not affect gastric secretion of acid or pepsin. The effects of graded doses of TGF and EGF on the formation of acute gastric lesions induced by 100% ethanol are shown in Fig. 1. Both peptides reduced dose-dependently the area of gastric lesions caused by ethanol. The reduction was significant at the dose of 25 pg/kg-h of TGF and of SO pg/kg-h of EGF, and the dose reducing the area of these lesions by SO% (IDso) averaged about 25 yg/kg-h for TGF and about 40 pg/kg-h for EGF. At a dose of 100 pg/kg-h each of these peptides prevented the formation of gastric lesions to a similar extent as dmPGE, (lOyg/kg) given intragastrically. TGF or EGF applied intragastrically at a dose of 100 pg/kg did not affect significantly the area or the numbcr of ethanol-induced gastric lesions, and these results were not included, for the sake of clarity. Pretreatment with indomethacin, which reduced mucosal generation of PGE2 by about 85% (Table II), tended to increase ethanol-induced damage, but this was not significant (Fig. 2). Infusion of TGF or EGF at a dose of SOpg/kg-h caused a reduction in gastric lesion area similar to that observed in previous series of experiments, but the pre-
treatment with indomethacin significantly attenuated the reduction in mucosal damage obtained with infusion of T G F or EGF (Fig. 2). Pretreatment with indomethacin did not affect significantly the reduction in mucosal damage obtained with dmPGEZ,and these results have not been included. Gastric blood flow as measured by LDF in the oxyntic and antral portions of the intact stomach averaged 60.4 2 1.3 and 54.2 2 1.2 ml/min-100 g tissue, and infusion of TGF or EGF at a dose of SO pg/kg-h did not affect this blood flow significantly (Table 11). Administration of 1.5 ml of 100% ethanol, which produced widespread mucosal lesions, resulted in about 68% and 51% reduction of the blood flow in the oxyntic and antral area, respectively. Infusion of TGF or E G F in ethanol-treated rats resulted in a small but significant increase in the mucosal blood flow in tests without pretreatment with indomethacin. After pretreatment with indomethacin the mucosal blood flow was reduced to a similar extent as in tests with ethanol alone, but the addition of TGF or EGF did not affect this reduced blood flow. In contrast, the pretreatment with dmPGE2 caused a marked increase in the blood flow as compared with that recorded in the ethanol-treated stomach both in rats without and with the pretreatment with indomethacin (Table 11). The exposure of rats to 3 h of water immersion and restraint stress caused visible lesions in the oxyntic gland mucosa in all animals tested. The damage appeared as elongated bands of bleeding erosions (about 12 _t 0.6 lesions per rat) following the long axis of the mucosal folds. lnfusion of TGF or EGF (25 yg/kg-h) 30 min before and during 3 h of stress significantly reduced (by about 60%) the number of gastric lesions, and this effect was similar to that attained
I WATER IMMERSION+RESTRAINTSTRESS ( 3hI I 0WITHOUT
INDOMEMACIN INDOMETHACIN
Fig. 3. Mean number of gastric lesions induced by 3 h of water immersion and restraint stress in rats without or with pretreatment with indomethacin and with infusion of transforming growth factor alpha (TGF,), epidermal growth factor (EGF), or Ih,lh-dimethyl prostaglandin E2 (dmPGE,). Means 2 SEW of 8-10 rats. Single asterisk indicates significant increase above the value obtained in corresponding rats without pretreatment with indomethacin. Double asterisks indicate significant decrease below the valuc obtained with vehicle control.
Growth Factors in Gastric Protection
653
Table Ill. Mucosal generation of 16,16-dimethyl prostaglandin E, (dmPGE,) and gastric blood flow measured by laser Doppler flowmetry in oxyntic and antral mucosa in rats, measured after 3 h of water immersion and restraint stress with and without pretreatment with intraperitoneal indomethacin (5 mg/kg) and/or transforming growth factor (TGF) or epidermal growth factor (EGF) (25 pg/kg-h) infused subcutaneously 30 rnin before and during 3 h of stress. Means 5 SEM of 8-10 determinations in 8-10 rats
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
Blood flow (rnl/min-100 g)
Intact Stress TGF + stress E G F + stress dmPGE2 + stress Indomethacin Indomethacin + TGF + stress E G F + stress Indomethacin Indomethacin + dmPGE, + stress
+
PGE, w g j
Oxyntic area
Antral area
480 ? 56 289 t 37" 271 ? 48* 238 5 52* 54 -t 7* 62 ? 8* 75 t 12* -
56.8 t 5.7 27.9 t 3.8* 38.1 t 3.2*t 36.2 t 2.4*1 41.8 t 3.8*t 19.6 ? 2.4* 2 1 . 2 t 1.8* 20.1 ? 2.4* 37.2 t 2.7*t
43.4 t 5.9 30.3 2 4.6* 37.2 ? 5.6 35.3 t 4.2 39.2 ? 3.8*t 20.7 t 1.5* 22.6 ? 2.4* 23.8 2 1.9* 36.4 -+ 2.3*t
* Significant ( P < 0.05) decrease below the value obtained in intact rats. t Significant ( P < 0.05) increase above the value obtained with stress.
after intragastric administration of dmPGEz (25 pg/kg-h). Pretreatment with indomethacin ( 5 mg/kg intraperitoneally) significantly increased (by about 26%) the number of gastric lesions and almost completely eliminated the reduction in lesion number achieved with infusion of TGF or EGF. In contrast, indomethacin failed to affect the reduction in stressinduced gastric damage obtained with dmPGEz (Fig. 3). Gastric mucosal generation of PGEz was decreased by about 40% after 3 h of stress, and pretreatment with indomethacin caused further decline in mucosal PGEz to about 10% of the normal value (Table 111). Infusion of TGF or EGF before and during the stress did not significantly affect the generation of PGEz in the oxyntic mucosa of rats without or with pretreatment with indomethacin (Table 111). After 3 h of stress the gastric blood flow was reduced significantly as compared with that in the intact stomach. Pretreatment with indomethacin ( 5 mg/kg intraperitoneally) resulted in a further significant decrease in the gastric blood flow. Infusion of TGF or EGF (25 bg/kg-h) caused a small but significant increase in the blood flow in rats without
indomethacin but not in those pretreated with indomethacin (Table 111). In tests with intragastric administration of dmPGEz the mucosal blood flow was significantly increased (over that observed in the stomach exposed to 3 h of stress alone) in rats with or without pretreatment with indomethacin. The time course of gastric lesions observed at 0 , 6, 12, and 24 h after withdrawal of stress without or with administration of TGF or EGF is shown in Fig. 4. In rats treated with TGF or E G F the disappearance of gastric lesions was significantly accelerated at 6, 12, and 24 h after stress. Mucosal generation of PGEz started to increase after 12 h and was almost completely restored after 24h of stress (Table IV). The incorporation of labeled thymidine into DNA in the intact oxyntic mucosa was about 38.0 1.6 dpm/ pg DNA. Significant increase in DNA synthesis started 12 h after stress, and almost complete restoration of DNA synthesis occurred after 24 h. Mucosal blood flow was almost completely restored to control value already 6 h after stress. The infusion of TGF or EGF (25 pg/kg-h) caused significant
*
Table IV. Generation of prostaglandin E 2 (PGEJ and DNA synthesis in oxyntic mucosa and blood flow in the oxyntic and antral area of intact stomach at 0, 6, 12, and 24 h after water immersion and restraint stress. Means SEM of &lo rats
*
Blood flow (ml/min-100 g)
Intact stomach 0 h after stress 6 h after stress 12 h after stress 24 h after stress TGF, 25 pg/kg-h, for 12 h after stress EGF, 25 pg/kg-h, for 12 h after stress ~
PGE, (ng/d
DNA synthesis, dpm/CLg DNA
Oxyntic area
Antral area
420 ? 58 231 2 48* 218 t 42* 264 t 47* 382 ? 50t 254 t 36*
38.0 +- 1.6 25.1 t 2.6* 27.2 ? 1.9* 30.5 2.0t 34.6 t 4.3t 34.2 t 3.2t
*
58.6 f 6.2 31.2 t 2.7* 67.8 f 6.5t 61.9 f 8.37 64.7 f 4.2t 64.7 f 5.8t
54.2 t 4.8 34.8 2 3.6* 64.9 t 5.5* 63.7 5 4.2t 59.3 t 5.2t 58.2 ? 4.2t
268 ? 42*
36.7 t 2.8t
62.8 f 7.4t
53.4 t 4.7t
~
TGF = transforming growth factor; EGF = epidermal growth factor. * Significant ( P < 0.05) decrease below the value recorded in intact stomach. t Significant increase above the value recorded immediately (at 0 h) after stress
S. J . Konturek et al.
6.54
[HEALING OF STRESS LESIONS]
15 & W
I
m
3 z
z
OVEHICLE TGFd (ZM91kg-h s.c.)
10
0
ezll
5
a w Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
r
HOURS AFTER STRESS Fig. 4. Mean number of stress-induced gastric lesions at 0, 6 , 12, or 24 h after stress in rats treated with vehicle (saline), transforming growth factor alpha (TGF,), or epidermal growth factor (EGF). Means * SEM of 8-12 rats. Asterisk indicates significant decrease below the value obtained in rats treated with vehicle.
increase in DNA synthesis but did not influence mucosal generation of PGE2 or gastric blood flow as measured 12 h after stress (Table IV). DISCUSSION The results demonstrate that T G F and EGF show similar gastroprotective activity against acute damage induced by ethanol or stress and have similar healing effect on acute gastric ulcerations induced by stress. Prcviouc studies showed that the major source of EGF is the salivary glands ( 3 , 12, 13), but small amounts of the peptide have also been detected in gastric niucosa (3,4, 12), where specific binding of E G F receptors has been observed (6). Removal of salivary glands was reported to exacerbate the rnucosal lesions induced by various topical irritants and ulcerogens, including stress ( 3 , 12, 13), and to abolish the adaptive cytoprotection (14). Furthermore, sialoadenectomy has been shown to inhibit the healing of chronic gastroduodenal ulcerations (15, 16). Exogenous EGF is also known to protect the gastric mucosa against acute damage provoked by both acid-dependent (stress, aspirin, bile salts) and acid-independent (ethanol) mucosal lesions ( 3 ) . These results could be interpreted as showing that functional salivary glands and constant generation of E G F are essential for both gastroprotection and mucosal repair processes (11, 12, 14). Our data presented here confirm the protective activity of E G F against ethanol- and stress-induced damage and show that this protection is mediated by mucosal PG. The suppression of P G generation by indomethacin (to about 15% of normal value) reduced the protective effect of EGF on ethanol-induced damage and completely reversed its protective action against stress-induced gastric lesions. In con-
trast, dmPGE2-;nduced gastroprotection against ethanol- or stress-induced lesions was not affected by pretreatment with indomethacin. It has been suggested that the interaction between EGF and PG may involve the maintenance of the mucosal blood flow (17). Indeed, our results showed that E G F attenuated the reduction in mucosal blood flow caused by ethanol o r stress and that indomethacin abolished this effect. Exogenous drnPGEz improved the mucosal blood flow in rats exposed to ethanol or stress to a similar extent in tests without and with indoniethacin pretreatment, whereas exogenous EGF was effective only in the animals without administration of indomethacin when mucosal PG biosynthesis was preserved even at a reduced level. This indicates that mucosal PG is necessary for the maintenance of gastric blood flow by EGF. T G F was originally suggested to function only during embryonic development and to be abnormally expressed by tumor cells, resulting in autocrine secretion and tumor growth (18). Cartlidge & Elder (4) recently demonstrated the presence of immunoreactive TGF in human mucosa throughout the gastrointestinal tract and especially in the stomach. Beauchamp et al. (5) reported the expression of TGFmRNA and TGFa/EGF receptor m R N A in normal gastric mucosa of the adult guinea pig, rat, and dog. They concluded that the local production of TGF may play an important role in the regulation of acid secretion and mucosal cell proliferation. Our study demonstrates that TGF is about twice as potent as EGF in the inhibition of gastric secretion and in the protection of gastric mucosal integrity against ethanol- and stress-induced damage. Thus T G F , like E G F , protects the gastric mucosa against the damage by both acid-independent (ethanol) and acid-dependent (stress) irritants. As in the case of E G F , the protection afforded by TGF against ethanol or stress-induced mucosal damage apparently requires the preservation of mucosal PG biosynthesis and blood flow. Since the protective effects of TGF or EGF against stressinduced lesions occurred at doses that caused partial inhibition of gastric acid secretion, the possibility could not be excluded that this factor also plays some role in gastroprotection by these peptides. One of the important considerations is whether TGF and E G F play any physiologic role in the maintenance of the integrity of gastrointestinal mucosa. Studies show that stress augments the release of E G F and increases its content in the gastric mucosa, whereas sialoadenectomy greatly augments the stress-induced gastric lesions, probably due to the deficiency of EGF (12). O n the other hand, TGF and E G F applied intragastrically failed to affect gastric acid secretion or to prevent the mucosal damage, suggesting that these peptides released into the gut lumen even in excessive amounts are unable to influence the mucosal integrity. Studies are required to determine whether alterations in T G F and E G F levels in gastric mucosa under various phy-
Scand J Gastroenterol Downloaded from informahealthcare.com by Chulalongkorn University on 01/02/15 For personal use only.
Growth Factors in Gastric Protection
siopathologic conditions correlate with the degree of mucosal damage. Another aspect of our study was the role of TGF and EGF in the recovery of gastric mucosa from the damage provoked by stress. Our results confirmed the findings of Wang & Johnson (18) that the healing of acute mucosal lesions is significant 12 h after the withdrawal of the stress and that this is accompanied by a marked increase in mucosal blood flow already evident 6 h after the stress and a significant increase in DNA synthesis and mucosal generation of PGE2 12 and 24 h after the stress. Both TGF and EGF were found to enhance significantly the healing of stress ulcerations, which was accompanied by a significant stimulation in DNA synthesis but not PGE2generation. Although the mechanism of the observed healing effects by TGF and EGF has not been investigated, our contention is that this is probably due to the stimulation by the peptides of the process of early epithelization or mucosal restitution (3) and of the cell proliferation in the gastric mucosa. Indeed, both EGF and TGF were shown to stimulate the proliferation of mucosal cells in the stomach (19-21), and our results are in keeping with these findings.
ACKNOWLEDGEMENTS This work was supported in part by grants 403989101 from the National Research Committee, Warsaw, Poland, USPHS grantsfrom the National Institute of Diabetes and Digestive and Kidney Diseasesfrom the National Institute of Alcohol Abuse and Alcoholism, and-rom the National Institute of Dental Research, NIH.
655
epidermal growth factor levels in normal gastrointestine mucosa. Br J Cancer 1989;60:657-60. 5. Beauchamp RD, Barnard JA, McCutchen CM, Cherner JA, Coffey RJ. Localization of transforming growth factor a and its receptor in gastric mucosa cells. J Clin Invest 1989;84:1017-23. 6. Slomiany BL, Lin J , Yao P, et al. Characterization of the epidermal growth factor receptor in gastric mucosa. Digestion 1990~47:181-90. 7. Lane A, Ivy AC. Response to the chronic gastric fistula rat to histamine Am J Physiol 1957;190:221-6. 8. Konturek SJ, Brzozowski T, Drozdowicz D, Beck G. Role of leukotrienes in acute gastric lesions induced by ethanol, taurocholate, aspirin, platelet-activating factor and stress in rats. Dig Dis Sci 1989;33:806-13. 9. Ahu H, Ivarsson LE, Johansson K, Lindhagen J , Lundgren 0. Assessment of gastric blood flow with laser Doppler flowmetry. Scand J Gastroenterol 1988;23:1203-10. 10. Chen BW, Hui WM, Lamm SK, Cho CH, Ng MMT, Luk CT. Effect of sucralfate on gastric blood flow in rats. Gut 1989;30:1544-51. 11. Konturek PK, Brzozowski T, Konturek SJ, Dembinski A. Role of epidermal growth factor, prostaglandin and sulfhydryls in stress-induced gastric lesions. Gastroenterology 1990;YY:160715. 12. Konturek SJ, BrzozowskiT, Konturek PK, Majka J , Dembinski A. Role of salivary glands and epidermal growth factor in gastric secretion and mucosal integrity in rats exposed to stress. Reg Pept 1991;32:203-15. 13. Olsen PS, Poulsen SS, Kirkegaard P, Nexo. Role of submandibular saliva and epidermal growth factor in gastric cytoprotection. Gastroenterology 1984;87:103-6. 14. Tepperman BL, Soper BD, Morris GP. Effect of sialoadenectomy on adaptive cytoprotection in the rat. Gastroenterology 1989;97:123-9. 15. Olsen SP, Poulsen SS, Therkelsen K, Nexo E . Effect of sialoadenectomy and synthetic human urogastrone on healing of chronic gastric ulcers in rats. Gut 1986;27:1443-9. 16. Konturek SJ, Dembinski A, Warzecha Z, Brzozowski T, Gregory H . Role of epidermal growth factor in healing of chronic gastroduodenal ulcers in rats. Gastroenterology 1988;94:130% I. 17. Maramatsu I , Hollenberg MD, Lederis K. Vascular actions of epidermal growth factor urogastrone; possible relation to prostaglandin production. Can J Physiol Pharmacol 1985;63: 994-5.
REFERENCES 1. Waterfield MD. Epidermal growth factors and related molecules. Lancet 1989;l: 1243-6. 2. Gregory H. Isolation and structure of urogastrone and its relationship to epidermal growth factor. Nature 1975;257:3257. 3. Konturek SJ. Role of growth factors in gastroduodenal protection and healing of peptic ulcers. Gastroenterol Clin N Am 1990;1Y:41-65. 4. Cartlidge SA, Elder JB. Transforming growth factor a and Received 9 October 1991 Accepted 23 March 1992
18. Wang J-Y, Johnson LR. Luminal polyamines stimulate repair of gastric mucosal stress ulcers. Am J Physiol 1990;259:G58492. 19. Dembinski A , Gregory H, Konturek SJ, Polanski M. Trophic action of epidermal growth factor on the pancreas and gastrointestinal mucosa in rats. J Physiol 1982;325:35-42. 20. Derynek R. Transforming growth factor a structure and biological activities. J Cell Biol 1986;32:293-304. 21. Schreiber HB, Winkler MF, Derynck R. Transforming growth factors a more potent angiogenic mediator than epidermal growth factor. Science 1986;232:1250-3.
