Higher Proximal Duodenal Mucosal Bicarbonate Secretion Is Independent of Brunner's Glands in Rats and Rabbits

Transcription

1 GASTROENTEROLOGY 1995;19: Higher Proximal Duodenal Mucosal Bicarbonate Secretion Is Independent of Brunner's Glands in Rats and Rabbits MARK A. AINSWORTH, MICHAEL A. KOSS, DANIEL L. HOGAN, and JON I. ISENBERG Division of Gastroenterology, University of California, San Diego, California Back~[round & Aims: Duodenal bicarbonate secretion is impaired in patients with duodenal ulcer. Before characterization of any cellular transport defect is possible, the origin of duodenal bicarbonate (epithelial cells and/ or Brunner's glands) must be determined. The aim of this study was to determine the role of Brunner's glands in duodenal bicarbonate secretion. Methods: Rats, which have Brunner's glands only in the proximal duodenum, and rabbits, which have Brunner's glands throughout the duodenum, were anesthetized. Basal and stimulated (with HCI, prostaglandin E2, and vasoactive intestinal polypeptide [VIP]) bicarbonate secretion was measured in three isolated intestinal segments: proximal duodenum, distal duodenum, and proximal jejunum. Mucosal surface area and Brunner's gland thickness was quantitated in each segment. Resuits: Secretion rates in proximal and distal duodenum and proximal jejunum were significantly different. Normalized proximal-to-distal duodenal gradients in bicarbonate secretion were similar in the two species despite significantly different gradients of Brunner's gland thickness. In rabbits, gradients of bicarbonate secretion and Brunner's gland thickness were not correlated. In both species, HCI, prostaglandin E2, and VIP stimulated secretion in all three segments. If the agonists specifically stimulated Brunner's gland bicarbonate secretion, relationships between gradients of bicarbonate secretion and Brunner's gland thickness would have been anticipated. This was not observed. Conclusions: The higher rates of bicarbonate secretion in the proximal duodenum than in the distal duodenum and proximal jejunum are independent of Brunner's glands. D uodenal bicarbonate, in concert with mucus, is an important first line of defense against acid injuryj -4 Proximal duodenal mucosal bicarbonate secretion is impaired in patients with duodenal ulcer, 5 suggesting that duodenal bicarbonate is a factor in the pathogenesis of ulcer. The lower bicarbonate secretory rates in patients with duodenal ulcer do not seem to be related to inflammation or microscopically detectable changes but more likely represent the result of a cellular transport defect(s). 6 Before characterization of any transport defect is possible, it is important to identify the tissue from which duodenal bicarbonate originates. Proximal duodenum, which in mammalian species contains Brunner's glands, has higher bicarbonate secretory rates than the more distal duodenum, which has no Brunner's glands] Bullfrogs (Rana catesbeiana) have no Brunner's glands in the proximal duodenum, yet have a proximalto-distal gradient, s This suggests that Brunner's glands do not make a substantial contribution to total duodenal bicarbonate secretion. Because separate collection of Brunner's gland secretion and epithelial secretion has not been possible, the role of Brunner's glands in duodenal bicarbonate secretion remains undefined. The purpose of the present study was to determine the role of Brunner's glands in duodenal bicarbonate secretion by comparing proximal-to-distal gradients of bicarbonate secretion in mammalian species with (rabbits) and without (rats) Brunner's glands in the distal duodenum. Materials Animals and Methods Male New Zealand white rabbits ( kg) and male Sprague-Dawley rats (25-35 g) were kept under standardized conditions of light (12 hours light and 12 hours dark) and temperature (21 C). Animals were deprived of food for hours (in mesh-bottom cages to prevent coprophagia) but had free access to water. As suggested and approved by the University of California, San Diego, Animal Subjects Committee, rabbits were anesthetized with pentobarbital (3 mg/kg as a bolus intravenously [IV], followed by 1-2 mg" kg -I h -I as a continuous IV infusion) and rats were anesthetized with thiobutabarbital (Inactin; Research Biochemicals Inc., Natick, MA; 12 mg/kg intraperitoneally). Surgical Procedures In both species, a tracheostomy was performed to secure a free airway. The femoral artery and vein were cannulated for continuous monitoring of blood pressure and infusion of Ringer's solution and drugs, respectively. The abdomen was opened by a midline incision, and the pancreatic and choledochal ducts (in rabbits) or the common pancreatic and bile 1995 by the American Gastroenterological Association /95/$3.

2 October 1995 BRUNNER'S GLANDS AND BICARBONATE SECRETION 1161 duct (in rats) were catheterized with polyethylene tubing (PE 2 and 1, respectively). In rabbits, three separate intestinal segments, each 5-cm long, were cannulated in situ between PE-24 tubing: the proximal duodenum starting.5 cm distal to the pylorus, the distal duodenum starting 2 cm distal to pylorus, and the proximal jejunum starting immediately caudal to the ligament of Treitz. In rats, three separate segments, each 1.6-cm long, were also cannulated in situ between PE-9 tubing: the proximal duodenum starting at the pylorus, the distal duodenum starting immediately distal to the entrance of the common pancreatic and bile duct, and the proximal jejunum starting immediately caudal to the ligament of Treitz. Thus, in each animal, proximal small intestinal secretions were measured simultaneously from three separate isolated segments. Because the pancreatic duct joins the common bile duct close to the duodenum in rats, care was taken to ligate the duct immediately at the entrance into the duodenum and to perform the cannulation proximal to this point. After completion of surgical procedures, the abdomen was closed by sutures and the wound kept moist with isotonic saline (37 C). Rectal temperature was maintained at 37 C using a heating blanket, and the animals were left undisturbed for 45 minutes before starting the experiments. To prevent dehydration, a balanced electrolyte solution was infused continuously IV (rats, 2 ml/h; rabbits, 6.5 ml/h). An arterial catheter was connected to a pressure transducer for continuous monitoring of blood pressure (Propaq 14/EL; Protocol Systems Inc., Beaverton, OR). Arterial plasma bicarbonate concentration was determined at the start and end of each experiment (Micro 13 ph/blood Gas Analyzer; Instrumentation Laboratories, Lexington, MA). As predefined by protocol, to avoid any potentially confounding variables caused by hypotension and systemic acidosis, animals that became hypotensive (mean arterial blood pressure of <6 mm Hg) or acidotic (arterial [HCO 3 ] below 15 mmol/l) were excluded from analyses. Fourteen percent of the rat and 25% of the rabbit experiments were excluded from analyses. The higher exclusion in rabbits was caused in large part by hypotension and acidosis during administration of vasoactive intestinal polypeptide (VIP). Experimental Protocol Each intestinal segment was perfused continuously with 154 mmol/l NaC1 (rats,.5 ml/min; rabbits, 1 ml/min at 37 C). The effluents were collected separately using gravity drainage in 15-minute periods, the volumes were measured to the nearest.1 ml, and samples were taken for immediate determination of bicarbonate concentration. Proximal and distal duodenal as well as jejunal bicarbonate secretion were measured in 15-minute intervals for 2.5 hours under the following conditions with 6 rabbits and 6 rats in each experiment. Basal secretion: control experiment. NaC1 (54 mmol/l) was perfused into each segment for 15 minutes. Effect of prostaglandin E2. After measuring basal secretion for 3 minutes, stepwise increasing concentrations of natural prostaglandin E 2 (PGE2) (1 7, 1-6, 1-5, and 1-4 mol/l; Upjohn, Kalamazoo, MI) were simultaneously infused intraluminally into each isolated segment. Each dose was infused for 3 minutes. Stock solutions of PGE2 (1 5, 1-4, 1-3, and 1-2 moi/l) in ethanol were prepared and stored at -7 C until use and diluted (1:1) in 154 mmol/l NaCI immediately before use. The PGE2 dose range was selected to include doses with a minimal effect and doses with near-maximal effect. 7'9 Effect of VlP. After measuring basal secretion for 3 minutes, stepwise increasing doses of VIP were infused IV (rabbits, 2, 1, 5, and 25, pmol" kg -1" h -1 as a continuous infusion; rats, 2, 1, 5, and 25, pmol/ kg as a bolus; Salk Institute, San Diego, CA). Each dose was prepared immediately before use and infused for 3 minutes. Bolus injection was preferred in rats because of the low infusion rate (2 ml/h) and the potential risk of adsorption of VIP to the infusion lines. To minimize VIP adsorption to the infusion lines, VIP was dissolved in 154 mmol/l NaC1 containing 1 mmol/l ascorbic acid and.1% bovine serum albumin. Similar to PGE2, the VIP dose range included doses with a minimal effect and doses with a near-maximal effect on proximal duodenal bicarbonate secretion as reported in rats. 1 Effect of luminal acidification. After measuring basal secretion for 3 minutes, duodenal and jejunal segments were acidified simultaneously by infusing 3 mmol/l (rabbits) or 1 mmol/l (rats) HC1 (made isotonic by addition of sodium chloride) for 5 minutes, followed by two 5-minute washout periods (saline infusion). During these periods, duodenal effluent was collected anaerobically. A lower HCI concentration was used in rats because previous studies have shown that concentrations above 1 mmol/l induce mucosal damage. H Immediately after the animals were killed, the duodenal and jejunal segments were excised. The tissue was placed on a millimeter grid, and the surface area was measured. The segments were fixed in buffered formalin for at least 24 hours. Three evenly spaced cross sections were obtained from each segment. The sections were carefully oriented and embedded in paraffin to allow perpendicular cutting to the serosal surface. Staining was performed with H&E. The tissues were then coded and randomized. In each section, the thickness of the Brunner's glands was determined as the average of three random measurements in each cross section (using a micrometer at 1 magnification). Brunner's gland thickness was determined independently by three observers unaware of the species and segmental origin of the tissue. Analyses Bicarbonate concentration in nonacidified samples was determined in triplicate using a validated back-titration method as described in detail previously. ~2 In acidified samples, bicarbonate concentration was determined anaerobically by measuring ph and Pc2 (Micro 13; Instrumentation Laboratories) and calculating the bicarbonate concentration using Henderson-Hasselbalch equation. Previous comparisons of back-titration and the ph and PCO 2 techniques have shown excellent agreement. 13

3 1162 AINSWORTH ET AL. GASTROENTEROLOGY Vol. 19, No. 4 Calculations and Statistical Analysis For each segment, bicarbonate output, expressed as Bmol" h 1. cm-2, was calculated as bicarbonate concentration multiplied by effluent volume per hour divided by area of the respective segments. Results are expressed as mean _+ SEM. To relate bicarbonate output and Brunner's gland thickness, they were corrected for the surface area of the individual segments as measured on a millimeter grid. However, this does not take into account the contribution of the villi or the microvilli. To compare the results between species, gradients of bicarbonate output and Brunner's gland thickness between segments were normalized individually as a percentage of the entire gradient from proximal duodenum to proximal jejunum.* Statistical analysis of differences between segments or changes "within experiments" was performed using analysis of variance and Newman-Keuls multiple comparisons test. Analysis of differences between the two species was performed using Student's unpaired t test. P values of <.5 were considered significant. A E 2 u} u~ c 15 a ca Im J~ u) "O C _= 1 1 u) "1. r. C 5 II1 P<O.5 rn.s. ],--] T N P<.5 f Results Arterial Blood Pressure Except for VIP experiments, the mean arterial blood pressure remained constant throughout the experiments in both species. In rabbits, the mean arterial pressures at the start and the end of the experiments were 84 _+ 1 and mm Hg, respectively. In rats, the corresponding values were and , respectively. VIP dose dependently decreased mean arterial pressure in rabbits (from to mm Hg) and rats (from to mm Hg). Thickness of Brunner's Gland Layer In rabbits, Brunner's glands constituted a continuous submucosal layer extending beneath the entire duodenal surface from pylorus and close to the ligament of Treitz. The Brunner's gland layer was about threefold thicker in the proximal duodenum than the distal duodenum (P <.5; Figure 1). Close to the ligament of Treitz, the Brunner's gland layer was dispersed into isolated small groups of glands. Brunner's glands were absent from the proximal jejunum. In rabbits, Brunner's glands were of two histologically different types: serous and mucous, as described by Carlton. 14 The mucous type was by far the most abundant, constituting more than 9% of the glands in the proximal duodenum. In rats, Brunner's glands were only of the mucous type, thickest at the pylorus, and disappeared approximately 5 mm *The difference between proximal duodenum (PD) and distal duodenum (DD) or between distal duodenum and proximal jejunum (PJ) was divided by the total difference between proximal duodenum and proximal jejunum, i.e., (PD-DD)/(PD-PJ) or (DD-PJ)/(PD-PJ). Proximal Distal Proximal Duodenum Jejunum Figure 1. Thickness of the Brunner's gland layer in the proximal duodenum, distal duodenum, and proximal jejunum in rabbits (E~) and rats ([]). Results are shown as mean ± SEM of 22 experiments in rabbits and 24 experiments in rats. distally. Thus, in rats, Brunner's glands were only present in the proximal duodenum. The mean thicknesses of the Brunner's gland layer in the proximal duodenum in rabbits and rats were similar (Figure 1). Proximal-to-Distal Gradients of Basal Bicarbonate Secretion In both species, proximal duodenal basal bicarbonate output was significantly higher than that by the distal duodenum, which in turn was significantly greater than that by the proximal jejunum (Figure 2). The gradient in basal bicarbonate output from proximal-to-distal duodenum, i.e., the differences in bicarbonate output between proximal and distal duodenum (as a percent of the entire difference between proximal duodenum and proximal jejunum), was comparable in rats and rabbits (7% + 8% and 8% + 6%; P >.3; Figure 3), although the gradient in Brunner's gland thickness was steeper in rats than in rabbits (1% vs. 62% _+ 8%; P <.5). In rabbits, gradients in bicarbonate output from proximal-to-distal duodenum were on average of the same magnitude as the gradient in thickness of the

4 October 1995 BRUNNER'S GLANDS AND BICARBONATE SECRETION 1163 Brunner's gland layer (7% + 8% and 62% + 8%, respectively). However, the individual data (Figure 4) showed no correlation between gradients of bicarbonate output and Brunner's gland thickness (r 2 =.18; P >.4; n = 22). In rats, basal bicarbonate output in the distal duodenum was significantly higher than in the proximal jejunum, although Brunner's glands were absent from both segments (Figure 3). This indicates that gradients in bicarbonate output exist in the absence of gradients in Brunner's glands. Furthermore, the decrement in HCO 3- output from distal duodenum to proximal jejunum was comparable in both species (rabbits, 3% + 8%; rats, 2% + 6%; P >.3; Figure 3), although a gradient of Brunner's glands from distal duodenum to proximal jejunum was only present in rabbits. Proximal-to-Distal or VIP-Stimulated Gradients of HCl-, PGE~-, Bicarbonate Secretion In both species, luminal acidification significantly stimulated bicarbonate output in each three intestinal segments (Table 1). Although duodenal bicarbonate secretion in response to luminal acidification was higher 25.~" 2 4 E u m o E 15 o. o 1 ' O O O -r" ' T l I P<O.5 Proximal Distal Duodenum I P<O.5 r I P<O.5 P<O.5 I 1 r Proximal Jejunum Figure 2, Basal bicarbonate secretion in the proximal duodenum, distal duodenum, and proximal jejunum in rabbits ([3) and rats (~). Results are shown as mean ± SEM of the 2.5-hour basal secretion in 6 experiments in each species. in the proximal than the distal duodenum in rabbits, the net increases above basal were similar in the proximal and distal duodenum (P, NS). In both species, PGE2 dose dependently stimulated bicarbonate output. At the highest dose (1-4 mol/l), the changes reached significance in both duodenal segments but not in the jejunal segment (Table 1). VIP also significantly increased bicarbonate output in all three segments in rats but had no significant effect in rabbit proximal duodenum (n = 8) (Table 1). In rabbits, VIP doses above 1 pmol" kg -1 h -1 caused acidosis and a decrease in intestinal bicarbonate output; therefore, only results from infusion of 1 pmol" kg -1" h -1 are shown in Table 1. If an agonist specifically stimulated bicarbonate secretion from Brunner's glands, this would steepen the bicarbonate output gradient from areas with Brunner's glands to those with few or absent glands. In both species, none of the agonists tested significantly steepened the proximal-to-distal duodenal gradient. In fact, in rabbits, HC1, PGE2, and VIP decreased the proximal-to-distal gradients by about 12% (range, 1%- 15%). In rats, the O ~3 = '~, "$' O "~,9 = ~1 Q " 6 "6 g 4O 2 Proximal Distal Proximal Duodenum Duodenum Jejunum A ~ ~ 2 =~ ~ D. := ~ B l ==- U~.~' E 8,-'~. eo? E 4 g $ Figure 3. Gradients of basal bicarbonate output () and Brunner's gland thickness ([]) in (A) rabbits (n = 22) and (B) rats (n = 24). Gradients are normalized as a percentage of the entire gradient from proximal duodenum to proximal jejunum. Note that bicarbonate output gradients are similar in both species, although the distal duodenum is free of Brunner's glands in the rats. *P <.5 vs. gradients of Brunner's gland thickness in rabbits, 2

5 1164 AINSWORTH ET AL. GASTROENTEROLOGY Vol. 19, No ~ 8o c 7O ~ 6o -'-- 5 ~ 4 'm 3 2 -r- 1 y = -.27.x r =.18, n = 22, P >.1 i i i i i i i i i i Brunner glands gradient (%) Figure 4. Correlation between proximal-to-distal duodenum gradients of basal bicarbonate output and Brunner's gland layer thickness in rabbits. Gradients are expressed as a percentage of the entire gradient from proximal duodenum to proximal jejunum. Each point represents a single animal. reductions caused by HC1, PGE2, and VIP were similar (range, 4%- 19%). Furthermore, if an agonist specifically stimulated bicarbonate secretion from Brunner's glands, the agonisr would strengthen the correlation between gradients of bicarbonate secretion and Brunner's gland thickness. However, significant correlations between proximal-todistal Brunner's gland gradients and stimulated bicarbonate output were not observed (range of correlation coefficients,.6-.38; P >.1). Discussion Most mammalian species have submucosal glands (Brunner's glands) of the mucous type in the proximal part of the duodenum. The thickness and extent of the Brunner's glands have great species variation In humans, the Brunner's gland area extends approximately to the papilla of Vater. In rats, the area extends onehalfway down to the entry of the bile duct; in rabbits, it extends almost to the beginning of the jejunum. 14 The composition of the pure Brunner's gland secretion is unknown. Histological and immunohistochemical studies suggest that the glands secrete mucus and epidermal growth factor Collection of fluid from acute or chronic duodenal pouches containing Brunner's glands has yielded a colorless, viscous fluid with a high content of protein and bicarbonate Because it was not possible to separate the Brunner's gland secretion from the fluid secreted by the surface epithelial cells, the fluid collected was not pure Brunner's gland secretion but rather a mixture of Brunner's gland secretion and epithelial secretion. Thus, these and other similar studies claiming to measure Brunner's gland secretion (for reviews, see Lang and Tansy 23 and Kirkegaard 24) do not prove that the bicarbonate originates directly from the Brunner's glands. On the contrary, it is known that at least part of the bicarbonate secreted by the proximal duodenum originates from the mucosa because duodenal segments free of Brunner's glands also secrete bicarbonate. 25 The fact that the proximal part of the duodenum containing Brunner's glands has a higher bicarbonate secretion rate than the more distal parts without Brunner's glands 7'12 indirectly suggests that Brunner's glands contribute to bicarbonate secretion. However, this proximal-to-distal gradient of bicarbonate output is also present in amphibians (bullfrog, Rana catesbeiana) without Brunner's glands, 8 suggesting that the bicarbonate gradient is independent of Brunner's glands. The present study shows that the proximal-to-distal gradient of duodenal bicarbonate output in two mammalian species is equally steep irrespectively of whether Brunner's glands are present or absent in the distal duodenum. Furthermore, in rabbits, the proximal-to-distal gradient of duodenal bicarbonate output was independent of the thickness of the Brunner's gland layer. Finally, in both species, with and without Brunner's glands in the distal duodenum, the distal duodenum secretes significantly more bicarbonate than the adjacent part of the proximal jejunum. These findings suggest that the higher rates of proximal duodenal bicarbonate secretion are probably independent of Brunner's glands and instead Table 1. Effect of HCl, PGE2, and VIP on Net Increases in Bicarbonate Output Above Basal &HC3- secretion (#mol. cm -2. h -~) Proximal Distal Proximal duodenum duodenum jejunum Rabbits HCl a a a PGE a a VIP ~ Rats HCl a 5.9 _+ 1.4 a 4.4 _+.8 a PGE a " VIP a a " NOTE. Results are net increases (mean + SEM; n - 6) above basal (see Figure 1). HCI was infused intraduodenally during a period of 5 minutes. The results for PGE2 and VIP are the maximal responses. The maximal response to PGE2 was obtained at 1,4 mol/l and to VIP at 25, pmol/kg IV (rats) and 1 pmol. kg -1o h -1 IV (rabbits). ap <.5 vs. basal.

6 October 1995 BRUNNER'S GLANDS AND BICARBONATE SECRETION 1165 related to inherent qualities of the proximal duodenal surface epithelia (e.g., intrinsic cellular secretory differences and/or differences caused by the larger total surface area [villi and microvilli] in the proximal duodenum than the distal segments). A definitive answer to this issue must await future direct collection of pure Brunner's gland juice. This does not detract from the finding that Brunner's glands contribute (e.g., by mucus or epidermal growth factor) to the observed greater acid and peptic resistance of the proximal duodenum compared with areas free of Brunner's glands] 6 Epidermal growth factor is an important trophic factor for the intestinal mucosa27'28; it stimulates bicarbonate and mucus secretion 29 and can prevent cysteamine-induced duodenal ulcers in rats. 3 Intraluminal HC1 significantly stimulated bicarbonate output in both duodenal and jejunal segments in both species. However, previous studies indicate that acidinduced stimulation of bicarbonate output probably occurs by different mechanisms in the duodenum and the jejunum. 31 Whereas the increase in duodenal bicarbonate output is caused by an increase in active bicarbonate transport, the increase in jejunal bicarbonate output appears to be caused by increased leakiness of the mucosa due to HCl-induced injury. 3t The functional difference between duodenum and jejunum is also shown by the observation that intraluminal PGE2 significantly stimulated bicarbonate output in both proximal and distal duodenum but had no effect on the proximal jejunum in both species. Chikhlssa et al. also noted that the sensitivity to VIP was highest in the duodenum and that VIP-stimulated bicarbonate secretion was greater in the duodenum than in the jejunum and ileum. 32 In the present study, VIP caused a small, yet significant, increase in jejunal bicarbonate secretion. Most agents that modulate bicarbonate output in duodenal segments containing Brunner's glands produce a corresponding effect in areas devoid of Brunner's glands.* This suggests either that Brunner's glands do not contribute to bicarbonate secretion or, less likely, that Brunner's glands and epithelial ceils share identical regulatory mechanisms with equivalent responses to luminal acidification, PGE2, and VIP. If any of the agonists tested specifically stimulated Brunner's gland bicarbonate secretion, a correlation between gradients of duodenal bicarbonate secretion and Brunner's gland thickness would have been anticipated. No such relationship was observed. In conclusion, the findings suggest that the higher rates of proximal duodenal bicarbonate secretion are independent of Brunner's glands but instead are related to inherent qualities of the proximal duodenal surface epithelia. References 1. Flemstrom G. Gastric and duodenal mucosal bicarbonate secretion. In: Johnson LR, Christensen J, Jackson M J, Jacobson ED, Walsh JH, eds. Physiology of the gastrointestinal tract. 2nd ed. New York: Raven, 1987: Wenzl E, Fell W, Starlinger M, Schiessel R. Alkaline secretion. A protective mechanism against acid injury in rabbit duodenum. Gastroenterology 1987;92: Leung FW, Miller JC, ReedyTJ, Guth PH. Exogenous prostaglandin protects against acid-induced deep mucosal injury by stimulating alkaline secretion in rat duodenum. Dig Dis Sci 1989;34: Ainsworth MA, Fenger C, Svendsen P, Schaffalitzky de Muckadell OB. Effect of stimulation of mucosal HC3- secretion on acidinduced injury to porcine duodenal mucosa. Scand J Gastroenterol 1993; 28: Isenberg JI, Selling JA, Hogan DL, Koss MA. Impaired proximal duodenal mucosal bicarbonate secretion in patients with duodenal ulcer. N Engl J Med 1987;316: Basuk PM, Hogan DL, Marin M, Ballesteros MA, Weinstein WM, Isenberg Jl. Diminished duodenal mucosal bicarbonate secretion in duodenal ulcer is independent of mucosal histologic abnormalities (abstr). Gastroenterology 1989;96:A Isenberg JI, Flemstrom G, Johansson C. Mucosal bicarbonate secretion is significantly greater in the proximal versus distal duodenum in the in vivo rat. In: Allen A, Flemstrom G, Garner A, Silen W, Turnberg LA, eds. Mechanisms of mucosal protection in the upper gastrointestinal tract. New York: Raven, 1984: Simson JN, Merhav A, Silen W. Alkaline secretion by amphibian duodenum. I. General characteristics. Am J Physiol 1981; 24:G41-G Granstam SO, Flemstrom G, Nylander O. Bicarbonate secretion by the rabbit duodenum in vivo: effects of prostaglandins, vagal stimulation and some drugs. Acta Physiol Scand 1987; 131: Flemstrom G, Kivilaakso E, Briden S, Nylander O, Jedstedt G. Gastroduodenal bicarbonate secretion in mucosal protection. Possible role of vasoactive intestinal peptide and opiates. Dig Dis Sci 1985;3:63S-68S. 11. Livingston EH, Passaro EPJ, Miller J, Guth PH. Spectrum of injury produced in the duodenum by perfusion with luminal acid in the rat. Gastroenterology 1992; 13: Isenberg JI, Hogan DL, Koss MA, Selling JA. Human duodenal mucosal bicarbonate secretion. Evidence for basal secretion and stimulation by hydrochloric acid and a synthetic prostaglandin E1 analogue. Gastroenterology 1986;91: Isenberg JI, Hogan DL, Thomas FJ. Duodenal mucosal bicarbonate secretion in humans: a brief review. Scand J Gastroenterol Suppl 1986; 125: Carlton A. The distribution of Brunner's glands in the duodenum of mammals. Proc Zool Soc London 1935; 1: Landboe-Christensen E, Bohn CL. Topographical aspects of the Brunner glands in swine. Acta Pathol Microbiol Scand 1944; 54(Suppl): Treasure T. The ducts of Brunner's glands. J Anat 1978; 127: Heitz PU, Kasper M, van-noorden S, Polak JM, Gregory H, Pearse AG. Immunohistochemical Iocalisation of urogastrone to human duodenal and submandibular glands. Gut 1978; 19: Kirkegaard P, Olsen PS, Nexo E, Hoist J J, Poulsen SS. Effect of vasoactive intestinal polypeptide and somatostatin on secretion of epidermal growth factor and bicarbonate from Brunner's glands. Gut 1984;25: Poulsen SS, Nexo E, Olsen PS, Hess J, Kirkegaard P. Immunohis-

7 1166 AINSWORTH ET AL. GASTROENTEROLOGY Vol. 19, No. 4 tochemical localization of epidermal growth factor in rat and man. Histochemistry 1986; 85: Florey HW, Harding HE. Further observations on the secretion of Brunner's glands. J Pathol Bacteriol 1934;39: Kirkegaard P, Lundberg JM, Poulsen SS, Olsen PS, Fahrenkrug J, H6kfelt T, Christiansen J. Vasoactive intestinal polypeptidergic nerves and Brunner's gland secretion in the rat. Gastroenterology 1981; 81: Kirkegaard P, Poulsen SS, Haise C, Loud FB, Skov-Olsen P, Christiansen J. The effect of cysteamine on the Brunner gland secretion in the rat. Scand J Gastroenterol 1981;16: Lang IM, Tansy MF. Brunner's glands. Int Rev Physiol 1983; 28: Kirkegaard P. Brunner's gland secretion and duodenal ulcer. An experimental study in the rat. Copenhagen: FADL's Forlag, 1985: Flemstrom G, Garner A, Nylander O, Hurst BC, Heylings JR. Surface epithelial HC3- transport by mammalian duodenum in vivo. Am J Physiol 1982;243:G348-G Florey HW, Jennings MA, Jennings DA, O'Connor RC. The reactions of the intestine of the pig to gastric juice. J Pathol Bacteriol 1939;49: Ulshen MH, Lyn-Cook LE, Raasch RH. Effects of intraluminal epidermal growth factor on mucosal proliferation in the small intestine of adult rats. Gastroenterology 1986;91: Goodlad RA, Wilson T J, Lenton W, Gregory H, McCullagh KG, Wright NA. Proliferative effects of urogastrone-egf on the intestinal epithelium. Gut 1987;28(Suppl): Marotta F, Chui DH, Zhong GG, Safran P. Effect of graded intravenous doses of urogastrone on duodenal bicarbonate secretion in conscious rats: evidence of a dose-response pattern. Digestion 199; 47: Kirkegaard P, Olsen PS, Poulsen SS, Nexo E. Epidermal growth factor inhibits cysteamine-induced duodenal ulcers. Gastroenterology 1983; 85: Nylander O, Kvietys P, Granger DN. Effects of hydrochloric acid on duodenal and jejunal mucosal permeability in the rat. Am J Physiol 1989;257:G653-G Chikhlssa AR, Gharzouli A, Charpin G, Descroix-Vagne M, Pansu D. Comparison of VIP-induced electrolyte secretion at three levels in rat small intestine. Reprod Nutr Dev 1992;32: Received December 23, Accepted June 12, Address requests for reprints to: Jon I. Isenberg, M.D., Division of Gastroenterology, UCSD Medical Center, 2 West Arbor Street, San Diego, California Fax: (619) Supported by grants from the National Institutes of Health (AM ) and from the Danish Research Council ( KG). Dr. Ainsworth was a visiting scientist from Denmark and a recipient of a Fogarty International Fellowship from the National Institutes of Health (1 FO5TW4799-2). These data have been published in abstract form (Gastroenterology 1994; 16:A38).