Gut, 1990, 31, 11-16 11
Bismuth subsalicylate reduces peptic injury of the
oesophagus in rabbits Gut: first published as 10.1136/gut.31.1.11 on 1 January 1990. Downloaded from
H P Tay, R C Chaparala, J W Harmon, J Huesken, N Saini, F Z Hakki, E J Schweitzer
Abstract models ofoesophagitis HCl inflicts little injury to Bismuth subsalicylate was tested in an in vivo the intact oesophageal epithelium, except at very perfused rabbit model of oesophagitis for its high concentrations and after prolonged periods ability to prevent the mucosal injury caused by of exposure."4 16 The bile acids typically increase pepsin. Treatment efficacy was assessed the oesophageal mucosal permeability, but cause under both a treatment-before-injury protocol only a relatively modest degree of accompanying and a treatment-after-injury protocol. morphologic change.' In contrast, pepsin not Oesophageal mucosal barrier function was only causes a significant increase in the perme- evaluated by measuring flux rates of H+, K+, ability of the oesophageal mucosa, but also and glucose. The degree of oesophagitis was causes a frankly haemorrhagic, severe morpho- determined by gross and microscopic exami- logic injury.36 The present study was therefore nation of the mucosa by several independent carried out with pepsin as the injurious agent, as observers. Results showed that under both it appears to be the most damaging of the treatment protocols, bismuth subsalicylate components of an acidic gastroesophageal significantly reduced the pepsin induced dis- refluxate fluid. ruption of the mucosal barrier, as weli as the We performed in vivo experiments to deter- morphologic changes. Bismuth subsalicylate mine whether bismuth subsalicylate had when given after exposure to pepsin was also protective effects against the pepsin induced found to protect against the morphologic oesophageal injury. Bismuth subsalicylate was injury in a dose dependent manner. Experi- tested in a treatment-before-injury and a ments in vitro suggested that bismuth sub- treatment-after-injury protocol. After bismuth salicylate inhibits the proteolytic action of was found to be protective in vivo, the antipepsin pepsin by interacting with pepsin, rather than activity of bismuth subsalicylate was tested in with the pepsin substrate. We conclude that vitro to attempt to clarify its mechanism of bismuth subsalicylate can protect the oeso- action. http://gut.bmj.com/ phageal mucosa against peptic injury, prob- ably through inactivation ofpepsin. Methods
The oesophageal mucosal injury found in IN VIVO EXPERIMENTS patients with reflux oesophagitis occurs as a result ofabnormal exposure ofthe mucosa to the Oesophagealperfusion on September 24, 2021 by guest. Protected copyright. corrosive gastric contents. We have used an in An in vivo perfused rabbit model ofoesophagitis, vivo perfused rabbit model of oesophagitis to adapted from Chung'2 was used as described.' 36 elucidate the mechanisms of oesophageal In brief, New Zealand white rabbits weighing mucosal injury' and to test agents which might 3-5 kg were lightly anaesthetised with an intra- be useful in the prevention or treatment of such muscular injection of ketamine/xylazine. The injury. oesophagus was cannulated in the neck at The present study was designed to test the the pharyngoesophageal junction and in the efficacy of bismuth subsalicylate in preventing abdomen at the gastroesophageal junction with experimental oesophagitis. Bismuth subsali- plastic tubing (id 0-317 mm). The oesophagus cylate has been used for many years as an over- was then perfused with 45 ml of various test Departments of Surgery and Pathology, the-counter drug for the relief of heartburn, solutions at 10 ml/min using a peristalic pump Washington VA Medical which is often associated with oesophagitis. (Harvard Apparatus Co, Millis, MA) and a Center and the Schools of Bismuth compounds have been shown to reduce recirculating system. The perfusate solution was Medicine, George Washington University experimental gastric mucosal injury in the rat,7'8 stirred mechanically (Haake, Berlin, Germany), and Georgetown and they have been used successfully to treat and a thermoregulator (Haake) kept the temp- University, Washington, peptic ulcer disease in man.9 '° Ulcer healing with erature of the perfusate at 37°C. A pH stat/ DC, USA H P Tay bismuth compounds has been shown to be autoburette system (Radiometer, Cophenhagen, R C Chaparala similar to cimetidine," '3 even though it lacks a Denmark) containing 0 4 N HC1 constantly J W Harmon significant acid neutralising capacity. Because maintained the pH of the perfusate at 2. J Huesken the mucosa of the N Saini oesophagus and stomach are Each experiment consisted of a 30 minute F Z Hakki both exposed to the same caustic gastric con- 'exposure period' in which the perfusate solution E J Schweitzer tents, we investigated whether bismuth is contained pepsin at pH 2. This was followed by a Correspondence to: John W efficacious against mucosal injury of the 40 minute 'flux period' in which the perfusate Harmon, Chief, Surgical Service, Washington VA oesophagus. solution contained no pepsin, but was used for Medical Center, 50 Irving Some of the potentially injurious components measuring transmucosal flux rates of H+, K+, Street NW, Washington, DC 20422, USA. of the oesophageal refluxate fluid include hydro- and glucose. In the treatment-before-injury Accepted for publication chloric acid (HCI), proteolytic enzymes such as protocol, for 10 minutes before the initiation of 10 May 1989 pepsin, and the bile acids.3 In experimental the 30 minute exposure period, the oesophagus 12 Tay, Chaparala, Harmon, Huesken, Saini, Hakki, Schweitzer
was perfused with either saline 145 mmol/l Calculation offluxes (control group) or saline plus bismuth subsali- Oesophageal mucosal permeability to ions and cylate (The Procter and Gamble Co, Cincinnati, small molecules was determined by calculating
OH) 30 mg/ml (study group), bopth at pH 2. At H+, K+, and glucose flux rates. Hydrogen ion Gut: first published as 10.1136/gut.31.1.11 on 1 January 1990. Downloaded from the start of the exposure period, porcine pepsin flux out of the lumen was given by the amount of (Sigma Chemical Co, St Louis, Missouri) was HCI delivered by the pH stat/autoburette added to the perfusate of both the control and apparatus to maintain the pH ofthe solution at 2. study groups to obtain a final perfusate concen- Concentrations of K+ were assayed by ion selec- tration of 1 mg/ml. The pH of the perfusate was tive electrode potentials (Beckman System E4A, maintained at 2-0 with HC1. At the end of this Beckman Instruments, Fullerton, CA), and con- exposure period, the perfusate was discarded centrations of glucose were assayed by the and the entire system, including the oesophagus, hexokinase reaction (Centrifichem System 500, was irrigated with 100 ml isotonic saline in Union Carbide Corp, Rye, NY). Fluxes were preparation for the subsequent flux period. calculated using the following formula: The exposure period in the treatment-after- injury protocol was similar to that in the Net flux=(CF-CI)XVI treatment-before-injury protocol, except that where VI is the initial volume of the perfusate at the bismuth subsalicylate was added after an the beginning of the flux period, CI is the initial initial oesophageal exposure to pepsin. The 30 concentration of K+ or glucose in the perfusate minute exposure period of both the control and solution, and CF is the final concentration. study groups began with perfusion of the oeso- 3H-PEG was used as an impermeable volume phagus with a solution containing 1 mg/ml of marker during the initial experiments. It was not porcine pepsin at pH 2. Ten minutes after the used in the later experiments because it was initiation of the exposure period, an amount of found to be unnecessary, as have others who bismuth subsalicylate was added to the perfusate have used this animal model have reported.'4 solution of the study groups to obtain a concen- Polyethylene glycol was deemed unnecessary tration of either 15 mg/ml, 30 mg/ml, or 60 mg/ when it was found in the initial experiments ml of bismuth subsalicylate. Concomitantly, a that the volume and PEG recovery rates were quantity of pepsin and HCI was added to the between 96-99%, and the volume flux was perfusate to maintain the pepsin concentration at relatively low (less than 0 5 ml per experiment). 1 mg/ml and the pH at 2. At the end of the exposure period, the perfusate was discarded and preparation was made for the flux period, as Determination ofoesophagitis index in the treatment-before-injury protocol. Immediately after removal, the oesophagus was
The methods used during the flux period were fixed in 10% formalin. A photograph ofthe fixed http://gut.bmj.com/ identical for both the treatment-before-injury specimen was taken for later grading of the and the treatment-after-injury protocols. The degree of gross morphologic injury. Histologic flux solution contained 1-5 g/l polyethylene slides were prepared from sections taken from glycol (PEG, Fisher Scientific Co, Fair Lawn, the proximal, central, and distal portions of the NJ) 100 mCi/l 3H-PEG (New England Nuclear, esophagus and stained with haematoxylin and Boston, Mass), and 10 mM HCl at pH 2. The eosin. Photographs and slides of all specimens
osmolality was brought to 280 mosmol/l with were scored for the severity of oesophagitis by on September 24, 2021 by guest. Protected copyright. mannitol. A 4 ml aliquot of the flux solution was several observers unaware of the treatments taken from the reservoir at the beginning and given, using a system previously described.36 In end of the 40 minutes flux period for later brief, the gross oesophagitis index was deter- analyses of K+ and glucose. A pH of 2 was mined by the following criteria: 1 =normal constantly maintained in the perfusate solution appearance; 2=erythema or other abnormal with the pH stat/autoburette apparatus. After appearance, but no haemorrhage; 3=non- completion of the flux period, the animal was confluent mucosal haemorrhage; 4=confluent killed with an intracardiac bolus ofpentobarbital intramural haemorrhages (Fig 1). The micro- and the oesophagus was excised. scopic oesophagitis index was scored by these
Figure 1: Grades ofgross oesophagitis. Numbers correspond to the gross oesophagitis index (GOI).
I~~~~~~~~~~~~~~~~~~~~~~*I., :-j~~~~~~~~~~~~~~~~~~~. .. .§ ...... E~~~~~~~~~~~~~~~~~~~~~~~~~ ...... ?3,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~7v ..'~~~~~~~~~~.. ..4 Bismuth subsalicylate reduces oesophagitis 13
Figure 2: Grades of microscopic oesophagitis. Numbers correspond to the microscopic oesophagitis
index (MOI). Gut: first published as 10.1136/gut.31.1.11 on 1 January 1990. Downloaded from
. w~~~. *w *.*:
-J "'' *b
:VW,- *.V
criteria: 1 =normal oesophagus; 2 = submucosal oesophagitis grades) were compared for signifi- oedema or separation of epithelial layers; 3 = cant differences (p<0 05) with the Mann- focal areas of intramural haemorrhage or partial Whitney test. epithelial loss; 4=large areas of haemorrhage or complete epithelial desquamation (Fig 2). Results http://gut.bmj.com/ In vitro experiments IN VIVO EXPERIMENTS A series of in vitro experiments were performed to determine whether bismuth subsalicylate Treatment-before-injury protocol could inhibit the proteolytic action of pepsin on When administered before exposure of the oeso- haemoglobin, and if so, whether this inhibition phageal mucosa to pepsin, bismuth subsalicylate was caused by the interaction of bismuth sub- reduced both the permeability changes and the
salicylate with pepsin or its substrate. The morphologic injury caused by pepsin. As shown on September 24, 2021 by guest. Protected copyright. biochemical principle underlying this experi- in Table I, significantly lower transmucosal flux ment was that if an enzyme inhibitor acts by rates of H+, K+, and glucose occurred in the binding to the substrate molecule, then the group treated with bismuth subsalicylate than in inhibition observed at a low substrate concentra- the untreated group (p<005). Treatment with tion can be reduced by providing more substrate bismuth subsalicylate before pepsin exposure to the reaction. Alternatively, if the inhibition is also significantly reduced the morphologic caused by the inhibitor binding to the enzyme, injury, as reflected by both the gross and micro- then the inhibition will not be reduced by scopic oesophagitis indices. The gross oeso- increasing the concentration of substrate.'7 In phagitis index ofthe untreated control group was the experiment, 50 mg/ml pepsin was dissolved 3-2 (0-2), compared with the group treated with in one ofthe following solutions: (a) HCI at pH 2, bismuth subsalicylate, which was graded as 2-0 or (b) HC1 plus 30 mg/ml bismuth subsalicylate (0-3) (p<001, Fig 3). Similarly, the microscopic at pH 2. The peptic activity of these solutions esophagitis index was reduced from 3-1 (0 2) in was then assayed by spectrophotometrically the untreated group to 2-1 (02) in the group measuring the amount of tyrosine released from treated with bismuth subsalicylate (p<0 001, 2 5 ml of 1, 2, 3, or 4 g% solutions of bovine Fig 3). haemoglobin substrate.'8
Treatment-after-injury protocol Statistical analysis The efficacy ofbismuth subsalicylate in reducing Each experimental group contained between five the oesophageal injury when administered after and nine animals. All results are expressed as the exposure to pepsin was similar to that observed mean±one standard error of the mean (SE). The when it was administered before the pepsin differences between means of groups of con- exposure. Flux rates ofH+, K+, and glucose were tinuous data (H+, K+, and glucose flux rates) significantly lower in the groups exposed to 15 were assessed for statistical significance mg/ml, 30 mg/ml, and 60 mg/ml of bismuth (p<005) using the Student's unpaired t test. subsalicylate than those in the untreated control Non-parametric data (gross and microscopic group (p<005, Table II). Bismuth subsalicylate 14 Tay, Chaparala, Harmon, Huesken, Saini, Hakki, Schweitzer
TABLE I Transmucosalflux rates in the treatment-before- 4. injury experiments []Gross ElMicroscopic Flux ratet -Tif Gut: first published as 10.1136/gut.31.1.11 on 1 January 1990. Downloaded from HI K+ Glucose ,m 3- (pLmol/10 min) (pmol/I0 min) (unol/lO min) 0 T*4* No treatment 53 (6) 7 (1) 5 (1) -E2c.L BSS treatment 28 (3)* 3 (1)* 3 (1)* a0'WA2_ Results are expressed as mean (SE). Bismuth subsalicylate 30 mg/ml. Q *Significantly different from no treatment group, p<005; offlux: H+ of lumen. tDirection -out lumen, K+, glucose=into I I I1- I Control 15 30LrIL60 also reduced both the gross and microscopic Bismuth subsalicylate (mg/ml) esophageal indices in a dose dependent manner Figure 4: Gross and microscopic oesophagitis indicesfrom experiments in which bismuth subsalicylate was administered (Fig 4). The gross oesophagitis index of the 10 minutes after pepsin exposure. *=Significantly different untreated group was 3-4 (0 2), while the treated (p 120 4- 0-0 Pepsin alone l 0-0* Pepsin and bismuth subsalicylate FIGross 100 | Microscopic aCx 1-I 8 3- 80 60 CD1) * * 0. 0 * * 40 e + 0)2- * * 20 1 0 1 2 3 4 Control 30 Haemoglobin substrate concentration (g/100 ml) Bismuth subsalicylate (mg/ml) Figure 5: In vitro pepsin assay. Significantly lower pepsin Figure 3: Gross and microscopic oesophagitis indicesfrom activity was measured when bismuth subsalicylate was present experiments in which bismuth subsalicylate was administered than when it was not (*=p<0.05). Bismuth subsalicylate before pepsin exposure. *=Significantly different (p<001) apparently interacts directly with pepsin, rather than the from the control group, which was exposed to pepsin but with substrate, because increasing the substrate concentration did no bismuth subsalicylate treatment. not increase the peptic activity. Bismuth subsalicylate reduces oesophagitis 15 There are several possible mechanisms by to prevent mucosal injury by different which this protection could occur. Bismuth mechanisms. compounds are known to complex with proteins. In conclusion, these studies indicate that Bismuth subsalicylate could therefore interfere bismuth subsalicylate can prevent the oeso- Gut: first published as 10.1136/gut.31.1.11 on 1 January 1990. Downloaded from with the peptic digestion of the oesophageal phageal mucosal injury caused by pepsin. Its mucosa by complexing with either pepsin or with effect derives, at least in part, from its capacity to the pepsin substrate protein in the mucosa. Our interact with pepsin. in vitro experiment would indicate that the former occurs and, indeed, pepsin inactivation In conducting the research described in this report, the investigat- by bismuth compounds has been described by ors adhered to the Guide for laboratory facilities and care as others.2 19 20 promulgated by the Committee on the Guide for Laboratory Animal Facilities and Care of the Institute of Laboratory Animal Alternatively, we could not rule out the Resources, National Academy of Sciences, National Research possibility of a topical protective effect of Council. This work was funded in part by Proctor and Gamble. The authors acknowledge the assistance of Barbara L Bass, Raj bismuth subsalicylate by binding to the oeso- Lakshman, and his staff, and William Andrews. phageal mucosa in addition to the pepsin inacti- vation. It has been suggested that bismuth 1 Harmon JW, Johnson LF, Maydonovitch CL. Effect of acid subsalicylate and other bismuth salts encourage and bile salts on the rabbit esophageal mucosa. Dig Dis Sci gastric and duodenal ulcers to heal because of 1981; 26: 65-72. 2 Lillemoe KD, Johnson LF, Harmon JW. Alkaline their ability to bind with various glycoproteins esophagitis: a comparison of the ability of components of and mucopolysaccharides at the ulcer base. gastroduodenal contents to injure the rabbit esophagus. Gastroenterology 1983; 85: 621-8. Bismuth compounds are thought to form a 3 Lillemoe KD, Johnson LF, Harmon JW. Role of the com- barrier over the base ofan ulcer which protects it ponents of the gastroduodenal contents in experimental acid esophagitis. Surgery 1982; 92: 276-84. from the noxious luminal contents.225 When 4 Schweitzer EJ, Bass BL, Batzri S, Harmon JW. Bile acid complexed with gastric mucus, bismuth has accumulation by rabbit esophageal mucosa. Dig Dis Sci 1986;31: 1105-13. been shown to drastically retard the migration of 5 Schweitzer EJ, Bass BL, Batzri S, Young PM, Huesken J, hydrogen ions.26 These local effects have not yet Harmon JW. Lipid solubilization during bile salt-induced esophageal mucosal barrier disruption in the rabbit. J Lab been described in the oesophagus, but could Clin Med 1987; 110: 172-9. have contributed to the protective effect 6 Schweitzer EJ, Bass BL, Johnson LF. Sucralfate prevents experimental peptic esophagitis in rabbits. Gastroenterology observed in vitro in the present study. In a 1985; 88:611-9. clinical setting administration of bismuth sub- 7 Goldenberg MM, Honkomp LJ, Burrous SE, et al. Protective effect of Pepto-Bismol Liquid on the gastric mucosa of rats. salicylate in a liquid form (Pepto Bismol, Proctor Gastroenterology 1975; 69: 636-40. and Gamble), which could coat the oesophagus 8 Wilson TR. Effect of tripotassium citrato bismuthate (TDB) on the healing ofexperimental gastric ulcers in rats. Postgrad during swallowing, would be desirable to take MedJ7 1975; 51 (suppl 5): 22-5. advantage ofany topical properties. 9 Eberhardt R, Kasper G, Dettmer A, Hochter W, Hagena D. Effect of oral bismuth subsalicylate on campylobacter http://gut.bmj.com/ Bismuth is reported to exhibit other interest- pyloridis and on duodenal ulcer. Gastroenterology 1987; 92: ing properties which are pertinent to its potential 1379. 10 Lee Fl, Samloff IM, Hardman M. Comparison of tri- applicability to the treatment of esophagitis. potassium dicitrato bismuthate tablets with ranitidine in Colloidal bismuth was recently shown to sup- healing and relapse ofduodenal ulcers. Lancet 1984; i: 1299- 302. press the activity and output ofpepsin in patients 11 Martin DF, May SJ, Tweedle DEF, Hollanders D, with duodenal and gastric ulcers.27 This effect Ravenscroft MM, Miller JP. Difference in relapse rates of was still present 24 hours after the medication duodenal ulcer after healing with cimetidine or tripotassium dicitrato bismuthate. Lancet 1981; i: 7-10. on September 24, 2021 by guest. Protected copyright. had been discontinued, suggesting the inhibition 12 Tytgat GNJ, Hameeteman W, Van Olffen GH. Sucralfate, bismuth compounds, substituted benzimidazoles, was sustained.28 A long lasting agent would be trimipramine and pirenzepine in the short- and long-term desirable in treating reflux oesophagitis, where treatment of duodenal ulcer. Clin Gastroenterol 1984; 13: 543-68. protection is frequently needed during periods of 13 Dekker W, Reisma K. Double-blind controlled trial with nocturnal reflux.2' In addition, its activity colloidal bismuth subcitrate in the treatment ofsymptomatic duodenal ulcers, with special references to blood and urine against campylobacter like organisms may be levels. Ann Clin Res 1979; 11: 94-7. beneficial, as these bacteria may contribute to the 14 Chung RSK, Magri J, DenBesten L. Hydrogen ion transport in the rabbit esophagus. AmJ3 Physiol 1975; 299: 496-500. pathogenesis ofoesophagitis in some cases.30 15 Salo J, Kivilaakso E. Role of luminal H+ in the pathogenesis of Current therapies for oesophagitis have tradi- experimental esophagitis. Surgery 1983; 92: 61-8. 16 Orlando RC, Nabila A, Turjman NA. Mucosal protection by tionally included measures aimed at reducing sucralfate and its components in acid-exposed rabbit reflux, either by increasing lower oesophageal esophagus. Gastroenterology 1987; 93: 352-61. 17 Bergmeyer HU, ed. Methods of enzymatic analysis, vol 2. sphincter tone or decreasing intra-abdominal Deerfield Beach, Florida: Verlag Chemie International, pressure. Furthermore, the mainstay of therapy 1981: 1046-57. 18 Decker LA. Worthington enzyme manual. Freehold, NJ: rests with reduction of gastric acidity. None of Worthington Biochemical Corporation, 1977. the existing therapies, however, are specifically 19 Bateson PR. A comparative in vitro evaluation of a new bismuth salt, bismuth aluminate. J Pharm Pharmacol 1958; directed at diminishing peptic injury of the 10:123-31. mucosa, even though part of the efficacy of 20 Bateson PR. The effect of bismuth carbonate and other antacids on the activity ofpepsin. Medicine 1954; 8: 370-4. antacids may derive from a reduction of the 21 Koo J, Ho J, Lam SK. Selective coating of gastric ulcer by refluxate pepsin activity through a pH rise. We tripotassium dicitratobismuthate in the rat. Gastroenterology 1982;82:864-70. recently found that the mucosal protective agent 22 Soutar RL, Coghill SB. Interaction of tripotassium dicitrato sucralfate was highly effective in preventing bismuthate with macrophages in the rat and in vitro. Gastroenterology 1986; 91: 84-93. experimental peptic oesophagitis in the rabbit.6 23 Brogden RN, Pinder RM, Sawyer PR. Trioptassium dicitrato- This benefit was attributed to a topical protec- bismuthate: a report of its pharmacological properties and therapeutic efficacy in peptic ulcer. Drugs 1976; 12: 401-11. tion when it was found that sucralfate did not 24 Wieriks J, Hespe W, Jaltly D. Pharmacological properties of inactivate pepsin. The present study suggests colloidal bismuth subcitrate. ScandJ3 Gastroenterol 1986; 17 (suppi 80): 11-6. that, through direct inactivation of pepsin, 25 Wilson TR. The pharmacology of tripotassium dicitratobis- bismuth subsalicylate could provide an added muthate (TDB). PostgradMed3' 1975; 51 (suppl 5): 18-21. 26 Lee SP. A potential mechanism of action of colloidal bismuth benefit to a clinical therapeutic regimen which subcitrate: diffusion barrier to hydrochloric acid. Scand J might include a combination of agents acting Gastroenterol 1982; 17 (suppl 80): 17-21. 16 Tay, Chaparala, Harmon, Huesken, Saini, Hakki, Schweitzer 27 Baron JH, Barr J, Batten J, Sidebotham R, Spencer J. Acid, 29 DeMeester TR, Johnson LF, Guy JJ, Toscano MS, Hall AW, pepsin, and mucus secretion in patients with gastric and Skinner DB. Patterns of gastroesophageal reflux in health duodenal ulcer before and after colloidal bismuth subcitrate and disease. Ann Surg 1976; 184: 459-70. (De-Nol). Gut 1986; 27: 486-90. 30 Borkent MV, Beker JA. Treatment of ulcerative reflux 28 Colin-Jones DG. There is more to healing ulcers than sup- oesophagitis with colloidal bismuth subcitrate in combina- pressing acid. Gut 1986; 27: 475-80. tion with cimetidine. Gut 1988; 29: 385-9. Gut: first published as 10.1136/gut.31.1.11 on 1 January 1990. Downloaded from http://gut.bmj.com/ on September 24, 2021 by guest. Protected copyright.