- Email: [email protected]
Gastric mucosal defense mechanisms: Effects of salicylate and histamine
Gastric Mucosal Defense Mechanisms: Effects of Salicylate and Histamine David Fromm, MD,* Boston, Massachusetts
The gastric mucosa generally is not harmed by the presence of large concentrations of acid in its lumen. Teorell [I] in 1933 postulated that this phenomenon was due to the presence of a mucosal barrier to acid. The clinical importance of this barrier was not fully appreciated until the early 1960s when Davenport [2] and Davenport, Warner, and Code [3] showed that diffusion of abnormally large amounts of acid from the gastric lumen back into the mucosa may be associated with bleeding. Since then many chemical agents have been shown to increase the back diffusion of acid and cause mucosal damage [4]. The ability of several drugs to alter back diffusion of acid is important clinically; for example, the ingestion of salicylates is associated with the development of gastritis, ulceration, and bleeding. A clearer understanding of the mechanisms by which the stomach protects itself against its own secreted acid also has implications for stress ulcers and peptic ulcers of the stomach not related to drugs, in which barrier defects are believed to be involved in the pathogenesis. Although the gastric mucosal barrier to acid has not been anatomically defined, one physiologic component is related to the permeability of the mucosa to acid. However, to think of the gastric mucosal barrier only in terms of permeability to acid is unduly restrictive, because it does not allow consideration of other possible mechanisms of protection. For example, the metabolic processes of the stomach may be important in protecting against acid back diffusion; the stomach could become very permeable to acid, but this might cause little damage if hydrogen ions are buffered within the mucosa. Perhaps the time has come to replace the term “gastric mucosal barrier” with “gastric mucosal defense mechanism.” From the Department of Surgery, Harvard Medical School, and Beth Israel Hospital, Boston, Massachusetts. This work was supported by United States Public Health Service Grant AM17919. Reprint requests should be addressed to David Fromm, MD, Department Of Surgery, Harvard Medical School and Beth Israel Hospital, 330 Brookline Avenue, Boston, Massachusetts 02215. ‘Recipient of Career Development Award AM 00053.
Volume 135, March 1979
Experimental studies of the effects of salicylate and histamine on the stomach have raised a number of questions about the actions of these agents on the gastric mucosal defense mechanisms and the nature of these mechanisms in general. Prevailing Concepts of Action Salicyhte. It has been postulated that the mechanism by which salicylate alters mucosal permeability is related to its lipid solubility. When the pH of a solution containing salicylate is lower than the log of the dissociation constant (pKa, approximately 3.5) of the drug, most of the salicylate is unionized and therefore soluble in fat. As a result, salicylate can permeate the lipoprotein mosaic membrane facing the lumen [2]. At higher pH values, salicylate is primarily ionized and insoluble in fat, and therefore less salicylate permeates the luminal membrane. Davenport [2,5] believed that this accounts for the damaging effects of salicylate at low but not high pH values of the gastric juice. He suggested that as unionized salicylate passes through the mucosa, alterations in permeability occur which allow enhanced tissue entry of acid. Although acid per se is not necessary for a change in permeability [6,7], it is generally agreed that acid is necessary for the tissues to become damaged. Thus, acid is believed to play a dual role in the damaging effects of salicylate; it is necessary for the drug to exist in a form which increases mucosal permeability, and it is the increased tissue entry of acid which causes damage. Mucus, once regarded as important in the gastric defense mechanism against damage by acid, is now discounted as a protective factor [7]. Histamine. Histamine is released during gastric mucosal damage [8-101 and is believed to stimulate acid secretion by the abnormally permeable mucosa [101. This effect is potentially harmful, since it would presumably increase the hydrogen ion gradient across the more permeable mucosa. However, the
379
Fromm
effects of histamine on mucosal permeability are controversial, particularly when mucosal permeability is unaltered [11-241. Although histamine stimulates mucosal metabolism., the effects of this alteration on mucosal defense mechanisms until recently have been largely ignored. In Vivo and In Vitro Studies Problems of Interpretation
In vivo studies have contributed greatly to the understanding of the nature of the gastric mucosal defense mechanisms and the harmful effects of back diffusion of acid. These studies are limited, however, in that it is difficult to differentiate between the unidirectional flows of various ions. In the presence of a semipermeable membrane such as the gastric mucosa, ions can move out of the lumen (that is, from the luminal to the serosal side of the mucosa) or into the lumen (from the serosal to the luminal side of the mucosa) by diffusional as well as by active transport processes. The former are largely concentrationdependent and therefore passive, whereas the latter are energy-consuming, allowing transfer of ions against their concentration gradients. An example of passive ion transport is the back diffusion of acid, and an example of active ion transport is acid secretion by the stomach. Most in vivo studies have investigated net ionic flow, which is the arithmetic difference between the unidirectional luminal to serosal and unidirectional serosal to luminal flows. Unidirectional ion flows have been studied in a few in vivo investigations, but no distinction between active and passive components of such ion flows has been made. The importance of this distinction is apparent when considering acid back diffusion, which is the net flux dependent on the luminal to serosal (passive) and serosal to luminal (active) hydrogen ion fluxes. Thus, inhibition of acid secretion would increase or unmask back diffusion, but this does not necessarily imply mucosal damage. Normal gastric mucosa usually has a slight permeability to luminal hydrogen ions. The small amount of acid that spontaneously diffuses back into the tissue does not, in health, appear to cause damage. Flux of bicarbonate can influence the net hydrogen ion flux, especially when permeability is increased. Back diffusion of acid generally amounts to a fraction of a milliequivalent, whereas the concentration of bicarbonate in blood is much larger. When permeability is increased, the concentration gradient of bicarbonate favors passage toward the gastric lumen. The influence of bicarbonate on the actual luminal loss of acid is difficult to quantify as is the amount of
380
buffer entering the mucosa from the capillaries. It is not unreasonable to assume that the latter may contribute to the tissue defense against acid, especially since blood flow to the gastric mucosa has been shown to increase during enhanced back diffusion of acid [IS]. Although most in vivo studies dealing with gastric mucosal permeability have focused on the flux of hydrogen ions, permeability of the mucosa is not restricted to hydrogen ions. Increased back diffusion of acid is associated with increased passage of sodium into the lumen when solutions in the gastric lumen contain high concentrations of hydrogen and low concentrations of sodium. Because a difference in sodium concentrations across the mucosa favors diffusion into the lumen, it has been assumed that increased sodium flux into the lumen is the result of increased (passive) permeability of the mucosa. However, the gastric mucosa of most mammals contains an active unidirectional absorptive (that is, luminal-to-serosal) sodium ion pump. Inhibition of active sodium absorption could result in an apparent increase in the unidirectional entry of sodium into the lumen (that is, serosal-to-luminal flux). If this is accompanied by inhibition of the active hydrogen ion secretory pump, which may result in an apparent increase in back diffusion of acid, the cause of these changes might be erroneously attributed to an increase in permeability rather than to inhibition of active transport. Thus, inhibition of active transport may unmask normal passive fluxes, and the resulting changes in ion fluxes could be the same as those seen with increased permeability alone. Net potassium fluxes also have been measured by several investigators, but the data are difficult to interpret. Because its intracellular concentrations are high, potassium may leak from damaged cells; therefore, the luminal concentration of potassium may not accurately reflect permeability. Few investigators have reported changes in chloride fluxes after exposure of the gastric mucosa to injurious agents. Many of the interpretive problems in alterations of ion flows across the gastric mucosa caused by salicylate and histamine can be resolved by in vitro studies in which a distinction between active and passive and unidirectional and net transport processes can be made more readily. Although it is recognized that in vitro experiments involve some artificial circumstances, several of the basic concepts of ion transport across the gastrointestinal tract have been derived primarily from them. These concepts have helped to explain many in vivo observations, and studies both in vivo and in vitro have been mutually complementary.
The American Journal of Surgery
Gastric Mucosal Defense Mechanisms
In Vltfo
Mothod8
In vitro studies of gastric physiology in our laboratory have been performed with an Ussing chamber which allows separate perfusion of the luminal (mucosal) and serosal surfaces of the tissue. (Figure 1.1 The tissue, either gastric fundus or antrum, is stripped of serosa and muscularis propria. The thin muscularis mucosa is placed between two Lucite@ half chambers attached to individual gas-lift circulating reservoirs containing Ringer’s solution kept at 38°C. Salt-agar bridges placed near the center and periphery of the Lucite chambers permit electrophysiologic measurements and manipulations which are useful in the interpretation of ion transport data. The rates of acid secretion and luminal acid loss (back diffusion) are measured with an automatic titrating system (pH stat) which maintains the bathing solution at any desired pH. Thus, acid secretion can be continuously measured under conditions that minimize back diffusion. Unidirectional ion fluxes can also be measured by placing isotopes in one reservoir and periodically sampling the opposite reservoir. Ion fluxes are calculated from the rates of isotopic transfer across the tissue or titration of acid entering or being lost from the luminal side per unit time per area of mucosa. In the absence of electrical and ion concentration gradients across the mucosa, the occurrence of a net ion flux is thought to be the result of an active process. This can usually be verified by inhibiting the flux with agents that interfere with cellular metabolism or receptor sites. Rabbit stomach has been used in most of our studies. The ion transport properties of both fundus and antrum under isolated conditions have been well characterized [14,16-221. Rabbit stomach is unique because it does not require exogenous carbon dioxide or bicarbonate to maintain acid secretion, and the antral mucosa does not secrete alkali in the absence of exogenous bicarbonate. Furthermore, since rabbit stomachs are more permeable than those of most other mammalian species, extremely low luminal pH levels are not required for back diffusion to occur and consequently, titration of luminal loss of acid is more accurate. Effects on H+ Transport Salicylate. There is nearly uniform agreement that salicylate increases gastric mucosal permeability to acid. However, the effect of this agent on acid secretion has been in doubt, partly because of the difficulty of measuring acid secretion at a low pH when permeability is increased. Studies using the isolated mucosa, however, indicate that in the presence of
Vohma 135, Much 1878
Ftgure 1. -tk d&am 0f.k v/ho pe&skn lyyrsralus. 77~ mucosa (dark sttppled bar) divktes the system Into mucosal (M), or turnhal, and serosal (S) halves. The two central salt-agar bridges comnwnkate wffh reference eh3ct~connectedtoavoHmetar(mV)krmesswement of the transmural etectrical potent/al dtfference. The two peripheral sail-agar bridges comnnmkate wffh Ag-AgC12 e/e&&es connected to a dhuct wtrent source @A). The arrows point to the gas inlets. 77~ gas IM syatem circulates the pe&skn ftu/d. The apparatus on the top of tha luminal perfusion chamber represents a ptt stat system.
small pH gradients across the tissue, the effect of salicylate is concentration-dependent [17]; high concentrations of salicylate (20 mM) decrease acid secretion, but low concentrations (3 to 5 mM) have a negligible effect. Despite these differences, both low and high concentrations of salicylate increase permeability of the fundic mucosa to acid. Mucosal permeability changes to acid also are dependent upon the continued presence of salicylate [18]. The rate of luminal acid loss returns to normal when salicylate is washed free. Similar effects on acid permeability occur in antral mucosa treated with salicylate. Although the rate of acid secretion is not significantly altered in the presence of low salicylate concentrations, the mucosa is resistant to known stimulants with high concentrations [18,19]. Under isolated conditions, exogenous cyclic adenosine monophosphate (AMP), theophylline (which increases cyclic AMP activity by inhibiting phosphodiester-
381
FUNDUS
UBaral ~Hirtamme,9x10e5M
-41
n=IO
n=14
Figure 2. Effects of hlstamlne and sallcylate on acid secret/on by koiated fundk mucosa. Each bar represents the mean of a 30 to 45 minute Interval of measurement. The f/d set of bars represents the controls. The second set shows histamine stimulation. The third set shows that a low concentratkn of salkyiate does not s/gnifkant& alter secretion. 77~ fourth set shows that histamine does not stimulate in the presence of salicylate. The last set shows that when sallcylate /s removed (clear bar), subsequent add/tkn of histamine stlnndates acid secret/on.
ase), and histamine (which increases cyclic AMP activity by stimulating adenylate cyclase) stimulate acid secretion in rabbit mucosa [16]. In the presence of salicylate, however, these agents have no effect. An example of this situation is shown in Figure 2. When salicylate is washed out of the solution bathing the lumen, cyclic AMP, theophylline, and histamine once again stimulate acid secretion. Histamine. Since histamine does not stimulate acid secretion in the presence of salicylate, a better estimate of the effects of histamine on H+ permeability can be made in this circumstance [19]. Histamine causes a small but significant decrease in the .rate of luminal acid loss by fundic mucosa treated with salicylate. (Figure 3.) Histamine has a similar but more pronounced effect on antral mucosa in both the presence and absence of salicylate. (Figure 3.) In addition to decreasing H+ permeability, histamine also decreases the permeability of a noncharged probe molecule, erythritol, in both fundus and antrum. The Hs receptor antagonists do not alter mucosal permeability of either fundus treated with salicylate or antrum treated or not treated with salicylate. However, Hz receptor antagonists prevent the effects of histamine on H+ permeability in fundus and antrum. Thus, in addition to the well known effect of Hs receptor antagonists on acid secretion, these agents inhibit an effect of histamine separate from acid secretion [19].
382
n=,o
Il.12
“.I0
Il.10
“40
“.I0
F/gure 3. EHects of hktam/ne on ksnbtal at/d kss. Each bar represents the mean of a 30 to 40 minute /nterva/ of measurement. The first set of bars represents fundk controls, indkating that the increased rate of krmlnai at/d kss in the presence of se/icy/ate does not change sign/f/cant/y over the two Intervals of measurement. The second set shows that histamine s/gn/fkantiy decreases /um/nai acid kss by fundus exposed to saikyiate. The third set rqnwen@ antral conbv/s, /ndkat/ng that iuminal acid loss in the absence of salicylate does not change s/gn/ficantiy over the two intervals of measurement. The same is true in the presence of se/icy/ate (fifth set). l?m rate of luminal at/d /ass by antrum exposed to saiky/ate (fifth set) is s/gn/fkantiy greater (p <0.001) than that observed in the absence of sa//cy/ate (third set). 77n?fourth and sixth sets show that histamine sign/f/cant/y decreases lumlnai acid loss by antrum in the absence and presence of saikyiate.
When capable of stimulating acid secretion, histamine appears to influence an additional mechanism for mucosal protection against acid [21]. As previously indicated, rabbit stomach has an unusual spontaneous permeability to acid. At pH 1 diffuse mucosal ulceration and histologic changes indicative of mucosal damage occur in vivo. This effect, however, does not occur in histamine-stimulated mucosa. Recent microelectrode studies before and after administration of histamine suggest that tissue pH is important in this mucosal defense mechanism. A decrease in tissue pH is associated with extensive mucosal damage, but maintenance of a near normal tissue pH by administration of histamine is associated with minimal damage. To what extent histamine affects mucosal permeability in these in vivo studies is unclear. However, it is likely that increased formation of alkali during acid secretion contributes to the protection associated with histamine stimulation. Support for this concept is provided by studies of isolated bullfrog mucosa [21]. The permeability of spontaneously secreting bullfrog stomach is not adversely affected by a luminal pH of approximately 1. However, at this pH inhibition of acid secretion by burimamide (an Hz receptor antagonist) is followed by tissue damage; at higher luminal pH values, bur-
lho Amwlcan Journalof Surgwy
Gastric
imamide does not adversely affect the mucosa. Inhibition of acid secretion by high concentrations of salicylate may perpetuate damage induced by the back diffusion of acid, but the above studies provide little information on how salicylate increases mucosal permeability.
Mucosal
I
NO AC/D
Defense
Mechanisms
SALICYLATE
I
/Al LUMEN
Mechanism of Salicylate on Mucosal Permeability
Studies of ion fluxes using isotopes provide some insight into the mechanism of salicylate’s action on permeability. Salicylate at both low and high concentrations increases fundic mucosal permeability to cations, but not to anions, in the absence of acid in the lumen [I&?].The selective effect of salicylate on cations is not unique to isolated gastric mucosa of the rabbit, since similar changes have been observed in red blood cells [23], urinary bladder of the turtle [24], and buccal ganglion of the marine mollusk [25]. In contrast to the changes occurring at neutral pH, in the presence of acid in the fundus sahcylate increases cation and anion permeability. These changes occur relatively rapidly, so that it is difficult to observe the sequence of events resulting in this nonspecific increase in permeability. This difficulty may be due to a modifying effect of the acid microenvironment provided by the parietal cells. However, ion flux studies on antral mucosa further clarify the mechanism of salicylate on altering mucosal permeability. In the absence of acid, salicylate added to antral mucosa also increases cation permeability and, in contrast to fundus, decreases anion permeability [20]. Similar effects occur in the presence of acid, but eventually these ion-selective effects give way to a nonspecific increase in permeability. (Figure 4.) The changes in permeability observed for fundus and especially for antrum are consistent with the concept that salicylate, by its selective increase in cation permeability, allows enhanced diffusion of hydrogen ions into the tissue. As the tissue becomes overwhelmed with acid, damage occurs, which results in a nonspecific, or diffuse, increase in ion permeability. This concept is supported by flux studies of noncharged permeability probe molecules. At neutral pH, salicylate does not alter mucosal permeability, (to erythritol for example), but in the presence of acid, the permeability (to erythritol) eventually increases. The observation that salicylate increases cation but decreases anion permeability by antral mucosa in the absence of acid, and initially in the presence of acid, is consistent with the salicylate anion exerting a negative electrical charge on or altering the electrical field of the mucosa. By making the luminal side
volume 139. March 1979
-Cation,
-Anion
Figure
4. Schematic representation of effects of salkylate on cation and anion permeability of an&al mucosa. In absence of lumlnal acid, salkylate increases cation and decreases anion permeability. Ihe same is lnlt/al~ true in the presence of acid, but subsequently anion permeablflty increases.
of the mucosa more negative, cations will be attracted into the mucosa and anions will initially be repelled. The electrical charge concept of salicylate on permeability suggests that alterations of cellular metabolism need not necessarily contribute to the initial mode of action. The ion-selective effects of salicylate persist when active transport processes are inhibited by pretreatment with ouabain, a metabolic poison [18,20]. However, salicylate does appear to have some metabolic effects on active transport [18,20]. Perhaps one way that the gastric mucosa protects itself against luminal acid ‘is by maintaining positive charges along its mucosal border; that is, these electrical charges may govern permeability and salicylate may affect this property. Studies of artificial membranes [26-281 and data from other laboratories [23,29] support this concept.
Summary
Some of the recent concepts about the gastric mucosal defense mechanisms against damage by luminal acid and the effects of histamine and salicylate on these mechanisms are reviewed. The mucosal barrier to acid appears to consist of at least two physiologic components: a permeability mechanism and a metabolic mechanism related to cellular bicarbonate production as a result of acid secretion. In
393
Fromm
the absence of salicylate, histamine appears to exert some protection by affecting both mechanisms, but in the presence of salicylate, histamine’s protective effect is limited to altering mucosal permeability. The actions of salicylate on the gastric mucosa are complex, related in part to the concentration of salicylate and the pH of the luminal fluid. The damaging effects of salicylate appear to be related more to the concentration of acid in the lumen than to the lipid solubility of the drug. Salicylate increases permeability regardless of pH; the increase is initially selective for cations and subsequently becomes nonselective, involving both cations and anions. Although both low and high concentrations of salicylate increase mucosal permeability to hydrogen ions, only high concentrations of salicylate affect cellular bicarbonate production.
12. 13.
14.
15.
16.
17.
16.
19.
References
20.
1. Teorell T: Untersuchungen uber die magensaftsekretion. &and Arch Physiol66: 225, 1933. 2. Davenport HW: Gastric mucosal injury by fatty and acetylsalicylic acids. Gastroenterology 46: 245, 1964. 3. Davenport HW. Warner HA, Code CF: Functional significance of gastric mucosal barrier to sodium. Gastroenterology 47: 142, 1964. 4. lvey KJ: Gastric mucosal barrier. Gastroenrerology 61: 247, 1971. 5. Davenport HW: Gastric mucosal hemorrhage in dogs. Gastroenterology 56: 439, 1969. 6. Black RB, Hole D, Rhodes J: Bile damage to the gastric mucosal barrier: the influence of pH and bile acid concentration. Gastroenterology61: 176, 1971. 7. Davenport HW: Protein-losing gastropathy produced by sulfhydryl reagents. Gastroenterology60: 870, 1971. 6. Davenport HW, Cohen BJ, Bree M, et al: Damage to the gastric mucosa: effects of salicylates and stimulation. Gastroenterolcgy 49: 169, 1965. 9. Johnson LR: Histamine liberation by gastric mucosa of pylorus-ligated rats damaged by acetic or salicylic acids. Proc Sot Exp Biol Med 121: 364, 1966. IO. Johnson LR. Overholt BF: Release of histamine into gastric venous blood following injury by acetic or salicylic acid. Gastroenterology 52: 505, 1967. II. Wlodek GK, Leach RK: Effects of histamine, feeding and insulin
21.
384
22.
23.
24.
25. 26.
27.
28. 29.
hypoglycemia on net ionic fluxes in gastric pouches. Arci Surg 93: 175, 1966. Altamirano M: Back diffusion of H+ during gastric secretion Am J Physiol216: I, 1970. Moody FG, Davis WL: Hydrogen and sodium permeation o canine gastric mucosa during histamine and sodium thio cyanate administration. Gasfroentero/ogy 59: 350, 1970. Fromm D, Schwartz JH, Quijano R: Transport of H+ and othe electrolytes across isolated gastric mucoSa of the rabbit. An J Physiol226: 166, 1975. Cheung LY, Moody FG, Reese RS: Effect of aspirin, bile salt an ethanol on gastric mucosal blood flow. Surgery 77: 766 1975. Fromm D, Schwartz JH, Quijano R: The effects of cyclic AMF and related agents on acid secretion by isolated rabbit gastric mucosa. Gastroenferology 69: 453, 1975. Fromm D, Robertson R: Effects of alcohol on ion transport b) isolated gastric and esophageal mucosa. Gastroenterolog; 70: 220, 1976. Fromm D, Schwartz JH, Quijano R: Effects of salicylate and hilt salt on ion transport by isolated gastric mucosa of the rabbi Am J Physioi 230: 319, 1976. Fromm D, Silen M, Robertson R: Histamine effects on H’ permeability by isolated gastric mucosa. Gastroenterolog~ 70: 1076, 1976. Fromm D: Ion selective effects of salicylate on antral mucosa Gastroenterology 71: 743, 1976. Smith P. O’Brien P, Fromm D, et al: Secretory state of gastric mucosa and resistance to injury by exogenous acid. Am, Surg 133: 61, 1977. Fromm D, Schwartz JH, Robertson R, et al: Ion transport across isolated antral mucosa of the rabbit. Am J Pnysio/ 23 1: 1763 1976. Wieth JO: Effect of some monovalent anions on chloride ant sulphate permeability of human red cells. J b%ysio/ (Lo@ 207: 581. 1970. Schwartz JH. Fromm D. Conner T, et al: Effect of salicylatea on ion fluxes across the turtle urinary bladder. fed Proc 32: 217, 1973. Levitan H, Barker JL: Membrane permeability: cation selectivib reversibly altered by salicylate. Science 178: 63, 1972. llani A: Ion discrimination by “millipore” filters saturated with organic solvents. I. Cation selectivity, mobility, and relative permeability of membranes saturated with bromobenzene. Biochim Biophys Acta 94: 405, 1965. llani A: Ion discrimination by “millipore” filters saturated with organic solvents. II. The significance of the hydrophobic medium. Biochim Biophys Acta 94: 415. 1965. llani A, Tzivoni D: Hydrogen ions in hydrophobic membranes. Biochim Biophys Acta 163: 429, 1968. Wieth JO: Paradoxical temperature dependence of sodium and potassium fluxes in human red cells. J Physio/ 207: 563, 1970.
The American Journal of Surgery
The gastric mucosa generally is not harmed by the presence of large concentrations of acid in its lumen. Teorell [I] in 1933 postulated that this phenomenon was due to the presence of a mucosal barrier to acid. The clinical importance of this barrier was not fully appreciated until the early 1960s when Davenport [2] and Davenport, Warner, and Code [3] showed that diffusion of abnormally large amounts of acid from the gastric lumen back into the mucosa may be associated with bleeding. Since then many chemical agents have been shown to increase the back diffusion of acid and cause mucosal damage [4]. The ability of several drugs to alter back diffusion of acid is important clinically; for example, the ingestion of salicylates is associated with the development of gastritis, ulceration, and bleeding. A clearer understanding of the mechanisms by which the stomach protects itself against its own secreted acid also has implications for stress ulcers and peptic ulcers of the stomach not related to drugs, in which barrier defects are believed to be involved in the pathogenesis. Although the gastric mucosal barrier to acid has not been anatomically defined, one physiologic component is related to the permeability of the mucosa to acid. However, to think of the gastric mucosal barrier only in terms of permeability to acid is unduly restrictive, because it does not allow consideration of other possible mechanisms of protection. For example, the metabolic processes of the stomach may be important in protecting against acid back diffusion; the stomach could become very permeable to acid, but this might cause little damage if hydrogen ions are buffered within the mucosa. Perhaps the time has come to replace the term “gastric mucosal barrier” with “gastric mucosal defense mechanism.” From the Department of Surgery, Harvard Medical School, and Beth Israel Hospital, Boston, Massachusetts. This work was supported by United States Public Health Service Grant AM17919. Reprint requests should be addressed to David Fromm, MD, Department Of Surgery, Harvard Medical School and Beth Israel Hospital, 330 Brookline Avenue, Boston, Massachusetts 02215. ‘Recipient of Career Development Award AM 00053.
Volume 135, March 1979
Experimental studies of the effects of salicylate and histamine on the stomach have raised a number of questions about the actions of these agents on the gastric mucosal defense mechanisms and the nature of these mechanisms in general. Prevailing Concepts of Action Salicyhte. It has been postulated that the mechanism by which salicylate alters mucosal permeability is related to its lipid solubility. When the pH of a solution containing salicylate is lower than the log of the dissociation constant (pKa, approximately 3.5) of the drug, most of the salicylate is unionized and therefore soluble in fat. As a result, salicylate can permeate the lipoprotein mosaic membrane facing the lumen [2]. At higher pH values, salicylate is primarily ionized and insoluble in fat, and therefore less salicylate permeates the luminal membrane. Davenport [2,5] believed that this accounts for the damaging effects of salicylate at low but not high pH values of the gastric juice. He suggested that as unionized salicylate passes through the mucosa, alterations in permeability occur which allow enhanced tissue entry of acid. Although acid per se is not necessary for a change in permeability [6,7], it is generally agreed that acid is necessary for the tissues to become damaged. Thus, acid is believed to play a dual role in the damaging effects of salicylate; it is necessary for the drug to exist in a form which increases mucosal permeability, and it is the increased tissue entry of acid which causes damage. Mucus, once regarded as important in the gastric defense mechanism against damage by acid, is now discounted as a protective factor [7]. Histamine. Histamine is released during gastric mucosal damage [8-101 and is believed to stimulate acid secretion by the abnormally permeable mucosa [101. This effect is potentially harmful, since it would presumably increase the hydrogen ion gradient across the more permeable mucosa. However, the
379
Fromm
effects of histamine on mucosal permeability are controversial, particularly when mucosal permeability is unaltered [11-241. Although histamine stimulates mucosal metabolism., the effects of this alteration on mucosal defense mechanisms until recently have been largely ignored. In Vivo and In Vitro Studies Problems of Interpretation
In vivo studies have contributed greatly to the understanding of the nature of the gastric mucosal defense mechanisms and the harmful effects of back diffusion of acid. These studies are limited, however, in that it is difficult to differentiate between the unidirectional flows of various ions. In the presence of a semipermeable membrane such as the gastric mucosa, ions can move out of the lumen (that is, from the luminal to the serosal side of the mucosa) or into the lumen (from the serosal to the luminal side of the mucosa) by diffusional as well as by active transport processes. The former are largely concentrationdependent and therefore passive, whereas the latter are energy-consuming, allowing transfer of ions against their concentration gradients. An example of passive ion transport is the back diffusion of acid, and an example of active ion transport is acid secretion by the stomach. Most in vivo studies have investigated net ionic flow, which is the arithmetic difference between the unidirectional luminal to serosal and unidirectional serosal to luminal flows. Unidirectional ion flows have been studied in a few in vivo investigations, but no distinction between active and passive components of such ion flows has been made. The importance of this distinction is apparent when considering acid back diffusion, which is the net flux dependent on the luminal to serosal (passive) and serosal to luminal (active) hydrogen ion fluxes. Thus, inhibition of acid secretion would increase or unmask back diffusion, but this does not necessarily imply mucosal damage. Normal gastric mucosa usually has a slight permeability to luminal hydrogen ions. The small amount of acid that spontaneously diffuses back into the tissue does not, in health, appear to cause damage. Flux of bicarbonate can influence the net hydrogen ion flux, especially when permeability is increased. Back diffusion of acid generally amounts to a fraction of a milliequivalent, whereas the concentration of bicarbonate in blood is much larger. When permeability is increased, the concentration gradient of bicarbonate favors passage toward the gastric lumen. The influence of bicarbonate on the actual luminal loss of acid is difficult to quantify as is the amount of
380
buffer entering the mucosa from the capillaries. It is not unreasonable to assume that the latter may contribute to the tissue defense against acid, especially since blood flow to the gastric mucosa has been shown to increase during enhanced back diffusion of acid [IS]. Although most in vivo studies dealing with gastric mucosal permeability have focused on the flux of hydrogen ions, permeability of the mucosa is not restricted to hydrogen ions. Increased back diffusion of acid is associated with increased passage of sodium into the lumen when solutions in the gastric lumen contain high concentrations of hydrogen and low concentrations of sodium. Because a difference in sodium concentrations across the mucosa favors diffusion into the lumen, it has been assumed that increased sodium flux into the lumen is the result of increased (passive) permeability of the mucosa. However, the gastric mucosa of most mammals contains an active unidirectional absorptive (that is, luminal-to-serosal) sodium ion pump. Inhibition of active sodium absorption could result in an apparent increase in the unidirectional entry of sodium into the lumen (that is, serosal-to-luminal flux). If this is accompanied by inhibition of the active hydrogen ion secretory pump, which may result in an apparent increase in back diffusion of acid, the cause of these changes might be erroneously attributed to an increase in permeability rather than to inhibition of active transport. Thus, inhibition of active transport may unmask normal passive fluxes, and the resulting changes in ion fluxes could be the same as those seen with increased permeability alone. Net potassium fluxes also have been measured by several investigators, but the data are difficult to interpret. Because its intracellular concentrations are high, potassium may leak from damaged cells; therefore, the luminal concentration of potassium may not accurately reflect permeability. Few investigators have reported changes in chloride fluxes after exposure of the gastric mucosa to injurious agents. Many of the interpretive problems in alterations of ion flows across the gastric mucosa caused by salicylate and histamine can be resolved by in vitro studies in which a distinction between active and passive and unidirectional and net transport processes can be made more readily. Although it is recognized that in vitro experiments involve some artificial circumstances, several of the basic concepts of ion transport across the gastrointestinal tract have been derived primarily from them. These concepts have helped to explain many in vivo observations, and studies both in vivo and in vitro have been mutually complementary.
The American Journal of Surgery
Gastric Mucosal Defense Mechanisms
In Vltfo
Mothod8
In vitro studies of gastric physiology in our laboratory have been performed with an Ussing chamber which allows separate perfusion of the luminal (mucosal) and serosal surfaces of the tissue. (Figure 1.1 The tissue, either gastric fundus or antrum, is stripped of serosa and muscularis propria. The thin muscularis mucosa is placed between two Lucite@ half chambers attached to individual gas-lift circulating reservoirs containing Ringer’s solution kept at 38°C. Salt-agar bridges placed near the center and periphery of the Lucite chambers permit electrophysiologic measurements and manipulations which are useful in the interpretation of ion transport data. The rates of acid secretion and luminal acid loss (back diffusion) are measured with an automatic titrating system (pH stat) which maintains the bathing solution at any desired pH. Thus, acid secretion can be continuously measured under conditions that minimize back diffusion. Unidirectional ion fluxes can also be measured by placing isotopes in one reservoir and periodically sampling the opposite reservoir. Ion fluxes are calculated from the rates of isotopic transfer across the tissue or titration of acid entering or being lost from the luminal side per unit time per area of mucosa. In the absence of electrical and ion concentration gradients across the mucosa, the occurrence of a net ion flux is thought to be the result of an active process. This can usually be verified by inhibiting the flux with agents that interfere with cellular metabolism or receptor sites. Rabbit stomach has been used in most of our studies. The ion transport properties of both fundus and antrum under isolated conditions have been well characterized [14,16-221. Rabbit stomach is unique because it does not require exogenous carbon dioxide or bicarbonate to maintain acid secretion, and the antral mucosa does not secrete alkali in the absence of exogenous bicarbonate. Furthermore, since rabbit stomachs are more permeable than those of most other mammalian species, extremely low luminal pH levels are not required for back diffusion to occur and consequently, titration of luminal loss of acid is more accurate. Effects on H+ Transport Salicylate. There is nearly uniform agreement that salicylate increases gastric mucosal permeability to acid. However, the effect of this agent on acid secretion has been in doubt, partly because of the difficulty of measuring acid secretion at a low pH when permeability is increased. Studies using the isolated mucosa, however, indicate that in the presence of
Vohma 135, Much 1878
Ftgure 1. -tk d&am 0f.k v/ho pe&skn lyyrsralus. 77~ mucosa (dark sttppled bar) divktes the system Into mucosal (M), or turnhal, and serosal (S) halves. The two central salt-agar bridges comnwnkate wffh reference eh3ct~connectedtoavoHmetar(mV)krmesswement of the transmural etectrical potent/al dtfference. The two peripheral sail-agar bridges comnnmkate wffh Ag-AgC12 e/e&&es connected to a dhuct wtrent source @A). The arrows point to the gas inlets. 77~ gas IM syatem circulates the pe&skn ftu/d. The apparatus on the top of tha luminal perfusion chamber represents a ptt stat system.
small pH gradients across the tissue, the effect of salicylate is concentration-dependent [17]; high concentrations of salicylate (20 mM) decrease acid secretion, but low concentrations (3 to 5 mM) have a negligible effect. Despite these differences, both low and high concentrations of salicylate increase permeability of the fundic mucosa to acid. Mucosal permeability changes to acid also are dependent upon the continued presence of salicylate [18]. The rate of luminal acid loss returns to normal when salicylate is washed free. Similar effects on acid permeability occur in antral mucosa treated with salicylate. Although the rate of acid secretion is not significantly altered in the presence of low salicylate concentrations, the mucosa is resistant to known stimulants with high concentrations [18,19]. Under isolated conditions, exogenous cyclic adenosine monophosphate (AMP), theophylline (which increases cyclic AMP activity by inhibiting phosphodiester-
381
FUNDUS
UBaral ~Hirtamme,9x10e5M
-41
n=IO
n=14
Figure 2. Effects of hlstamlne and sallcylate on acid secret/on by koiated fundk mucosa. Each bar represents the mean of a 30 to 45 minute Interval of measurement. The f/d set of bars represents the controls. The second set shows histamine stimulation. The third set shows that a low concentratkn of salkyiate does not s/gnifkant& alter secretion. 77~ fourth set shows that histamine does not stimulate in the presence of salicylate. The last set shows that when sallcylate /s removed (clear bar), subsequent add/tkn of histamine stlnndates acid secret/on.
ase), and histamine (which increases cyclic AMP activity by stimulating adenylate cyclase) stimulate acid secretion in rabbit mucosa [16]. In the presence of salicylate, however, these agents have no effect. An example of this situation is shown in Figure 2. When salicylate is washed out of the solution bathing the lumen, cyclic AMP, theophylline, and histamine once again stimulate acid secretion. Histamine. Since histamine does not stimulate acid secretion in the presence of salicylate, a better estimate of the effects of histamine on H+ permeability can be made in this circumstance [19]. Histamine causes a small but significant decrease in the .rate of luminal acid loss by fundic mucosa treated with salicylate. (Figure 3.) Histamine has a similar but more pronounced effect on antral mucosa in both the presence and absence of salicylate. (Figure 3.) In addition to decreasing H+ permeability, histamine also decreases the permeability of a noncharged probe molecule, erythritol, in both fundus and antrum. The Hs receptor antagonists do not alter mucosal permeability of either fundus treated with salicylate or antrum treated or not treated with salicylate. However, Hz receptor antagonists prevent the effects of histamine on H+ permeability in fundus and antrum. Thus, in addition to the well known effect of Hs receptor antagonists on acid secretion, these agents inhibit an effect of histamine separate from acid secretion [19].
382
n=,o
Il.12
“.I0
Il.10
“40
“.I0
F/gure 3. EHects of hktam/ne on ksnbtal at/d kss. Each bar represents the mean of a 30 to 40 minute /nterva/ of measurement. The first set of bars represents fundk controls, indkating that the increased rate of krmlnai at/d kss in the presence of se/icy/ate does not change sign/f/cant/y over the two Intervals of measurement. The second set shows that histamine s/gn/fkantiy decreases /um/nai acid kss by fundus exposed to saikyiate. The third set rqnwen@ antral conbv/s, /ndkat/ng that iuminal acid loss in the absence of salicylate does not change s/gn/ficantiy over the two intervals of measurement. The same is true in the presence of se/icy/ate (fifth set). l?m rate of luminal at/d /ass by antrum exposed to saiky/ate (fifth set) is s/gn/fkantiy greater (p <0.001) than that observed in the absence of sa//cy/ate (third set). 77n?fourth and sixth sets show that histamine sign/f/cant/y decreases lumlnai acid loss by antrum in the absence and presence of saikyiate.
When capable of stimulating acid secretion, histamine appears to influence an additional mechanism for mucosal protection against acid [21]. As previously indicated, rabbit stomach has an unusual spontaneous permeability to acid. At pH 1 diffuse mucosal ulceration and histologic changes indicative of mucosal damage occur in vivo. This effect, however, does not occur in histamine-stimulated mucosa. Recent microelectrode studies before and after administration of histamine suggest that tissue pH is important in this mucosal defense mechanism. A decrease in tissue pH is associated with extensive mucosal damage, but maintenance of a near normal tissue pH by administration of histamine is associated with minimal damage. To what extent histamine affects mucosal permeability in these in vivo studies is unclear. However, it is likely that increased formation of alkali during acid secretion contributes to the protection associated with histamine stimulation. Support for this concept is provided by studies of isolated bullfrog mucosa [21]. The permeability of spontaneously secreting bullfrog stomach is not adversely affected by a luminal pH of approximately 1. However, at this pH inhibition of acid secretion by burimamide (an Hz receptor antagonist) is followed by tissue damage; at higher luminal pH values, bur-
lho Amwlcan Journalof Surgwy
Gastric
imamide does not adversely affect the mucosa. Inhibition of acid secretion by high concentrations of salicylate may perpetuate damage induced by the back diffusion of acid, but the above studies provide little information on how salicylate increases mucosal permeability.
Mucosal
I
NO AC/D
Defense
Mechanisms
SALICYLATE
I
/Al LUMEN
Mechanism of Salicylate on Mucosal Permeability
Studies of ion fluxes using isotopes provide some insight into the mechanism of salicylate’s action on permeability. Salicylate at both low and high concentrations increases fundic mucosal permeability to cations, but not to anions, in the absence of acid in the lumen [I&?].The selective effect of salicylate on cations is not unique to isolated gastric mucosa of the rabbit, since similar changes have been observed in red blood cells [23], urinary bladder of the turtle [24], and buccal ganglion of the marine mollusk [25]. In contrast to the changes occurring at neutral pH, in the presence of acid in the fundus sahcylate increases cation and anion permeability. These changes occur relatively rapidly, so that it is difficult to observe the sequence of events resulting in this nonspecific increase in permeability. This difficulty may be due to a modifying effect of the acid microenvironment provided by the parietal cells. However, ion flux studies on antral mucosa further clarify the mechanism of salicylate on altering mucosal permeability. In the absence of acid, salicylate added to antral mucosa also increases cation permeability and, in contrast to fundus, decreases anion permeability [20]. Similar effects occur in the presence of acid, but eventually these ion-selective effects give way to a nonspecific increase in permeability. (Figure 4.) The changes in permeability observed for fundus and especially for antrum are consistent with the concept that salicylate, by its selective increase in cation permeability, allows enhanced diffusion of hydrogen ions into the tissue. As the tissue becomes overwhelmed with acid, damage occurs, which results in a nonspecific, or diffuse, increase in ion permeability. This concept is supported by flux studies of noncharged permeability probe molecules. At neutral pH, salicylate does not alter mucosal permeability, (to erythritol for example), but in the presence of acid, the permeability (to erythritol) eventually increases. The observation that salicylate increases cation but decreases anion permeability by antral mucosa in the absence of acid, and initially in the presence of acid, is consistent with the salicylate anion exerting a negative electrical charge on or altering the electrical field of the mucosa. By making the luminal side
volume 139. March 1979
-Cation,
-Anion
Figure
4. Schematic representation of effects of salkylate on cation and anion permeability of an&al mucosa. In absence of lumlnal acid, salkylate increases cation and decreases anion permeability. Ihe same is lnlt/al~ true in the presence of acid, but subsequently anion permeablflty increases.
of the mucosa more negative, cations will be attracted into the mucosa and anions will initially be repelled. The electrical charge concept of salicylate on permeability suggests that alterations of cellular metabolism need not necessarily contribute to the initial mode of action. The ion-selective effects of salicylate persist when active transport processes are inhibited by pretreatment with ouabain, a metabolic poison [18,20]. However, salicylate does appear to have some metabolic effects on active transport [18,20]. Perhaps one way that the gastric mucosa protects itself against luminal acid ‘is by maintaining positive charges along its mucosal border; that is, these electrical charges may govern permeability and salicylate may affect this property. Studies of artificial membranes [26-281 and data from other laboratories [23,29] support this concept.
Summary
Some of the recent concepts about the gastric mucosal defense mechanisms against damage by luminal acid and the effects of histamine and salicylate on these mechanisms are reviewed. The mucosal barrier to acid appears to consist of at least two physiologic components: a permeability mechanism and a metabolic mechanism related to cellular bicarbonate production as a result of acid secretion. In
393
Fromm
the absence of salicylate, histamine appears to exert some protection by affecting both mechanisms, but in the presence of salicylate, histamine’s protective effect is limited to altering mucosal permeability. The actions of salicylate on the gastric mucosa are complex, related in part to the concentration of salicylate and the pH of the luminal fluid. The damaging effects of salicylate appear to be related more to the concentration of acid in the lumen than to the lipid solubility of the drug. Salicylate increases permeability regardless of pH; the increase is initially selective for cations and subsequently becomes nonselective, involving both cations and anions. Although both low and high concentrations of salicylate increase mucosal permeability to hydrogen ions, only high concentrations of salicylate affect cellular bicarbonate production.
12. 13.
14.
15.
16.
17.
16.
19.
References
20.
1. Teorell T: Untersuchungen uber die magensaftsekretion. &and Arch Physiol66: 225, 1933. 2. Davenport HW: Gastric mucosal injury by fatty and acetylsalicylic acids. Gastroenterology 46: 245, 1964. 3. Davenport HW. Warner HA, Code CF: Functional significance of gastric mucosal barrier to sodium. Gastroenterology 47: 142, 1964. 4. lvey KJ: Gastric mucosal barrier. Gastroenrerology 61: 247, 1971. 5. Davenport HW: Gastric mucosal hemorrhage in dogs. Gastroenterology 56: 439, 1969. 6. Black RB, Hole D, Rhodes J: Bile damage to the gastric mucosal barrier: the influence of pH and bile acid concentration. Gastroenterology61: 176, 1971. 7. Davenport HW: Protein-losing gastropathy produced by sulfhydryl reagents. Gastroenterology60: 870, 1971. 6. Davenport HW, Cohen BJ, Bree M, et al: Damage to the gastric mucosa: effects of salicylates and stimulation. Gastroenterolcgy 49: 169, 1965. 9. Johnson LR: Histamine liberation by gastric mucosa of pylorus-ligated rats damaged by acetic or salicylic acids. Proc Sot Exp Biol Med 121: 364, 1966. IO. Johnson LR. Overholt BF: Release of histamine into gastric venous blood following injury by acetic or salicylic acid. Gastroenterology 52: 505, 1967. II. Wlodek GK, Leach RK: Effects of histamine, feeding and insulin
21.
384
22.
23.
24.
25. 26.
27.
28. 29.
hypoglycemia on net ionic fluxes in gastric pouches. Arci Surg 93: 175, 1966. Altamirano M: Back diffusion of H+ during gastric secretion Am J Physiol216: I, 1970. Moody FG, Davis WL: Hydrogen and sodium permeation o canine gastric mucosa during histamine and sodium thio cyanate administration. Gasfroentero/ogy 59: 350, 1970. Fromm D, Schwartz JH, Quijano R: Transport of H+ and othe electrolytes across isolated gastric mucoSa of the rabbit. An J Physiol226: 166, 1975. Cheung LY, Moody FG, Reese RS: Effect of aspirin, bile salt an ethanol on gastric mucosal blood flow. Surgery 77: 766 1975. Fromm D, Schwartz JH, Quijano R: The effects of cyclic AMF and related agents on acid secretion by isolated rabbit gastric mucosa. Gastroenferology 69: 453, 1975. Fromm D, Robertson R: Effects of alcohol on ion transport b) isolated gastric and esophageal mucosa. Gastroenterolog; 70: 220, 1976. Fromm D, Schwartz JH, Quijano R: Effects of salicylate and hilt salt on ion transport by isolated gastric mucosa of the rabbi Am J Physioi 230: 319, 1976. Fromm D, Silen M, Robertson R: Histamine effects on H’ permeability by isolated gastric mucosa. Gastroenterolog~ 70: 1076, 1976. Fromm D: Ion selective effects of salicylate on antral mucosa Gastroenterology 71: 743, 1976. Smith P. O’Brien P, Fromm D, et al: Secretory state of gastric mucosa and resistance to injury by exogenous acid. Am, Surg 133: 61, 1977. Fromm D, Schwartz JH, Robertson R, et al: Ion transport across isolated antral mucosa of the rabbit. Am J Pnysio/ 23 1: 1763 1976. Wieth JO: Effect of some monovalent anions on chloride ant sulphate permeability of human red cells. J b%ysio/ (Lo@ 207: 581. 1970. Schwartz JH. Fromm D. Conner T, et al: Effect of salicylatea on ion fluxes across the turtle urinary bladder. fed Proc 32: 217, 1973. Levitan H, Barker JL: Membrane permeability: cation selectivib reversibly altered by salicylate. Science 178: 63, 1972. llani A: Ion discrimination by “millipore” filters saturated with organic solvents. I. Cation selectivity, mobility, and relative permeability of membranes saturated with bromobenzene. Biochim Biophys Acta 94: 405, 1965. llani A: Ion discrimination by “millipore” filters saturated with organic solvents. II. The significance of the hydrophobic medium. Biochim Biophys Acta 94: 415. 1965. llani A, Tzivoni D: Hydrogen ions in hydrophobic membranes. Biochim Biophys Acta 163: 429, 1968. Wieth JO: Paradoxical temperature dependence of sodium and potassium fluxes in human red cells. J Physio/ 207: 563, 1970.
The American Journal of Surgery