The pancreatic duct epithelium in vitro: Bile acid injury and the effect of epidermal growth factor Carlos Alvarez, MD, Cynthia Nelms, MD, Victor D’Addio, MD, and Barbara L. Bass, MD, Baltimure, Md.

Background. Pancreatic duct epithelial cells form a barrier against parenchymal injury. The capacity of these cells to respond to injury has not been investigated. We hypothesized that qbidermal growth fac- tor (EGF), normally found in pancreatic juice, could protect the duct @ithelium from damage. Methods. An explant system of duct cell culture developed in our lab with the bovine main pancreatic duct was used. Explants were exposed to bile acid (taurodeoxycholic acid [TDGA] 0, 0.05, 0.5, and 1 mmol/L) in the presence or absence of EGF (0, 1, 10, and 100 nmol/L) for 48 hours. Epithelial prolif- eration, damage, and growth out from the explant edge were assessed histologically. Expression of ductal markers and the extent of cell proliferation were determined by immunohistochemistry using specific anti- bodies. Results. Explant duct cells poliferated and demonstrated continued expression of key duct antigens in culture. TDCA produced dose-dependent mucosal damage and reduced epithelial density and growth

from the edge. EGF increased cellular density in the native epithelium, but did not significantly alter growth porn the edge. Mucosal damage created by TDGA exposure was significantly decreased with EGF and both growth from the edge and cell density were preserved. Conclusions. Explants created from the bovine main pancreatic duct serve as an excellent model for the study of duct epithelial cells in vitro. These cells proliferate in response to EGF and are damaged by TDGA at concentrations below those normally associated with detergent-like activity and below levels observed in bile and duodenal secretions. The ability of EGF to protect from this injury suggests a poten- tial physiologic role in the maintenance of the pancreatic duct mucosal barrier (Surgery 1997;122:4 76 84.)

From the Department of Surgery, University of Maryland, and the Surgical Service, Baltimore VAMC, Baltimore,

Md.

THE PANCREATIC DUCT MUCOSAL BARRIER protects the parenchyma from potentially damaging luminal substances such as pancreatic enzymes and the highly alkaline fluid the cells themselves pro- duce.l” Pancreatic injury is readily established in a number of animal models by means that effective- ly break down this mucosal barrier.3-5 The duct mucosa appears particularly sensitive to agents such as bile and ethanol that are associated with clinical pancreatitis. The importance of this “pan- creatic duct mucosal barrier” in pancreatic disease has long been hypothesized, but investigation in this area has been hampered by inadequate in vitro systems.

Presented at the Fifty-eighth Annual Meeting of the Society of University Surgeons, Tampa, Fla., Feb. 13-15, 1997.

Reprint requests: Carlos Alvarez, MD, Surgical Service (112), Baltimore VAh4C, 10 N. Greene Street, Baltimore, MD 21201.

Copyright 0 1997 by Mosby-Year Book, Inc.

0039.6060/97/$5.00+0 11/6/82480

Recent studies have demonstrated increased expression of a variety of peptide growth factors in and around the ductal system in both acute and chronic pancreatitis. 6’ In particular, the levels of both epidermal growth factor (EGF), its relative peptide, transforming growth factor-a (TGF-a), and the cell membrane receptor they share (EGFr), appear to be markedly increased in in- flammatory diseases of the pancreas. Although this upregulation probably represents a positive re- sponse to injury, the effects of these growth factors on pancreatic ducts are not well defined. In other gastrointestinal tissues where its actions have been more extensively studied, EGF has been shown to protect from injury and improve healing.s-10 We hypothesized that it may have a similar cytoprotec- tive effect on the duct mucosa. To isolate ductal responses from acinar, humoral, and neural influ- ences we used a new in vitro model of pancreatic duct cell isolation developed in our laboratory.

476 SURGERY

477 Alvarez et al. surgery August 1997

/ TOTAL LENGM 1

GROWTH-NORMAL -BLEBBED*FLAT-SLOUGHED

Fig. 1. Schema of pancreatic duct explant epithelial morphology and definitions used for the morpheme- tric determination of damage and growth.

METHODS

Explant harvest and culture. Fresh whole bovine pancreas glands were obtained from a local beef- processing plant and transported on ice in 2% antibiotic-antimycotic solution in phosphate- buffered saline (PBS-AA). With a sterile technique, the main pancreatic duct and its major branches were dissected free from the surrounding parenchyma. The duct was then cut open length- wise and divided into 10 to 20 mm by 5 to 10 mm segments, which were extensively washed in cold PBS-AA. The duct segments were then placed on plastic tissue culture dishes on gelatin (Gelfoam, Upjohn) “rafts.” Basal culture medium based on Waymouth’s medium and containing 5% fetal bovine serum, 2% antibiotic-antimycotic, and 0.1 mg/ml soybean trypsin inhibitor was added, and the explants were placed in humidified 95% 0,/5% CO, air and incubated at 37” C for 48 hours. The medium was exchanged after the first 24 hours. This 48-hour recovery period was prompted by preliminary evidence of significant duct epithelial abnormalities and occasional cell sloughing immediately after the tissue handling and dissection. At 48 hours the normal epithelial appearance had returned in most explants.

Experimental conditions. Basal medium was replaced with medium containing the same com- ponents, except that the fetal bovine serum was reduced to 1%. This “test” medium was supple- mented with graded quantities of EGF (0, 10, and 100 nmol/L), taurodeoxycholic acid (TDCA; 0, 0.05, 0.5, and 1 mmol/L), or both. In one set of explants we studied the effect of adding 100 umol/L tyrphostin A47 (LC Labs), an inhibitor of EGFr tyrosine kinase activity, to medium contain- ing 1 mmol/L TDCA in the presence or absence of 100 nmol/L EGF. The explants were returned to the incubator for 48 hours and then harvested. They were then fixed in 4% paraformaldehyde overnight, pan&n-embedded, and cut in 7 !..trn sec- tions. Sections were stained with hematoxylin-eosin for morphometric assessment of the epithelial injury or processed for immunohistochemistry.

Assessment of damage and growth. An optical micrometer was used to determine the length of the epithelial surface of each explant, as measured between the cut edges of the duct (Fig. 1). Mucosal damage was defined as the aggregate length of sloughed, blebbed, or flattened epithelium. These subjective measurements were made by two blind- ed observers. When measurements did not agree between observers, the average of the two scores was used. The degree of mucosal damage is report- ed as a percentage of the total explant length to jus- tify for different explant sizes. As a measure of cel- lular proliferation or loss in the native epithelium, the number of epithelial cell nuclei per high-power (x200) field was counted in three separate seg- ments of a representative section of each explant, averaged, and reported as nuclei per high-power field or epithelial density. We observed that in the absence of damaging conditions, the epithelium consistently extended from the explant edge (where the duct had been cut open) onto the bare connective tissue. The extent of this proliferation (growth from the edge) was measured with the optical micrometer. Statistical comparisons between treatment groups was made using the unpaired Student’s t test, with statistical signifi- cance accepted at p less than 0.05.

Immunohistochemistry. Explant sections were deparaffinized, rehydrated, and washed with PBS. Inherent peroxidase activity was quenched with dilute hydrogen peroxide in methanol for 5 min- utes. The sections were then blocked with nonim- mune sera for 1 hour at room temperature. This was followed by incubation with primary antibodies raised against carbonic anhydrase II (CA II, l:lOO, Biodesign In&), the cystic fibrosis transmembrane conductance regulator (CFTR; 1: 100, Genzyme) , and the EGF receptor (EGFr; l:lOO, Fitzgerald), or nonimmune sera (control) for 1 hour at room tem- perature. To assess cellular proliferation, sections of explants were incubated with antibody raised against the proliferative cell nuclear antibody (PCNA; l:lOO, Neomarker). The sections were then washed and incubated with horseradish per- oxidase (HRP)-linked secondary antibody (1:200,

surgery Volume 122, Number 2

Alvarez et al. 478

Fig. 2. Immunoreactivity (da& staining) against (A) CA II (original magnification x 80)) (B) CFTR (original magnification x 200), and(C) EGFr (original magnification x 200) in the pancreatic duct explant is limit- ed to the epithelial cells. Staining was absent when specific antibodies were replaced with nonimmune sera in adjacent sections.

Vector Labs) for 1 hour, which was followed by exposure to HRP substrate (DAB Reagent, Vector Labs), counterstaining with eosin, and examina- tion by microscopy.

RESULTS

Our preliminary studies revealed that on the basis of histologic appearance and absent infec- tion, explants remained viable in the basal medium for as long as 3 weeks in culture. Infection of the medium was seen in approximately 10% of cases,

increasing in incidence with longer culture peri- ods. Expression of key ductal antigens, GA II, CFTR, and EGFr, was preserved in culture (Fig. 2) and the epithelium remained responsive to secretin stimulation (data not shown).

Incubation of the explants in TDGA produced mucosal damage (Table I; Fig. 3,c and 6). Although the effect was dose dependent, even at very low concentrations of TDGA there was a statistically sig- nificant increase in epithelial damage to more than 20% of the explant surface. Similarly low concen-

479 Alvarez et al. surgery August 1997

Fig. 3. Hematoxylineosin-stained sections of duct explants. A, Epithelial appearance after 48 hours’ stabi- lization, followed by 48 hours’ incubation in nonsupplemented medium (x200). B, Explant epithelium after stabilization and then 48 hours’ incubation with 100 nmol/L EGE Note extension of epithelium out from the cut edge of the explant (closed arrov, x80). C and D, Explants after exposure to 0.5 mmol/L and 1.0 mmol/L TDCA, respectively, showing blebbing (@en arrowhead), sloughing (closed arrouhead) , and flat- tening (open arrow) of the epithelium (x80). Explants exposed to 1 mmol/L TDCA concurrently with (E) 100 nmol/L EGF or (F) both 100 nmol/L EGF and 100 p mol/L tyrphostin A47. The native epithelium and the extension at the edge are preserved (E) with EGF, but damage is extensive (F) with tyrosine kinase inhi- bition (x80).

Table I. Effect of adding either TDGA or EGF alone to pancreatic duct explants for 48 hours in culture

TDCA 0 0.05 0.5 1.0

EGF ~0 1 10 100

mmol/L mmol/L mmoL/L mmol/L nmol/L nmol/L nmol/L nmol/L

n 5-7 5-7 5-7 5-7 5-7 5-7 5-7 5-7 Epithelial growth 55* 13 28 + 9 22 k 14* 17 i: 13* 55* 13 58+21 25+19 51+15

from edge (mm) Percentage of l+O 22 + 16* 24? ll* 59*12* - - - -

damaged mucosa (%) Epithelial cell density 90 I! 5 109 f 12 75 + 6 57 * 10” 90 * 5 73 k 8 92*16 117+11*

(nuc/HPF)

nuc/HPF, Nuclei per high-power field.

*p < 0.05 by unpaired Student’s t test.

trations of TDGA also reduced the extent of epithe- imately 37% of the basal density with maximal TDGA lial growth from the cut edge of the duct (Table I). doses, a reflection of cell sloughing noted on gross Epithelial cell density, on the other hand, increased histology. Expression of duct antigens persisted on slightly but not significantly (p = 0.09) with exposure the remaining cells in :he epithelium, although to 0.05 mmol/L TDCA before decreasing to approx- immunohistochemistry did not allow quantification.

surgery Volume 12.2, Number 2

Alvarez et al. 480

Fig. 4. Immunoreactivity (dark staining) against PCNA in the explant epithelium. A, Control using nonim- mune serum. B-D, Explants after 48 hours’ stabilization and 48 hours’ incubation with 0, 10, and 100 nmol/L EGF, respectively. Note presence of PCNA staining in epithelial cell nuclei in both the native epithelium and the new growth at the cut edge in all sections (x80).

Exposure of the explants to EGF did not pro- duce any significant alteration in the epithelium, including antigen expression as determined by immunohistochemistry. At higher concentrations, however, epithelial hyperplasia became histologi- cally evident (Fig. 3,b). This correlated with an increase in cellular proliferation in the native mucosa, as measured by cell density, which showed a statistically significant 30% increase with high concentrations of EGF (100 nmol/L). EGF did not significantly alter the extension of the epithelium, or growth, from the edge of the explants at any concentration. PCNA immunohistochemistry con- firmed the proliferative response of duct epithelial cells to EGF, with cellular proliferation noted in both the native explant mucosa and the area of new growth at the cut edges (Fig. 4).

At 100 nmol/L, EGF significantly improved all parameters of epithelial damage caused by bile exposure (Figs. 3,e, and 5). Specifically, the extent of mucosal damage was reduced nearly by half at each of the concentrations of TDCA studied. EGF at this concentration also maintained the epithelial cell density and preserved epithelial migration out from the cut edge. Results obtained with 10 nmol/L EGF showed a trend toward a similar response but did not achieve statistical significance (data not shown). Lower concentrations of EGF were not studied. Tyrphostin A47 did not signift- cantly alter the mucosal appearance by itself, nor did it affect the damage induced by TDCA. When

added together with EGF, however, tyrphostin A47 negated EGF’s protective effect (Fig. 3,~).

DISCUSSION

The properties of the pancreatic duct epitheli- urn, particularly its ability to respond to injury, have not been extensively investigated.“sl* This is due in large part to three factors: the inaccessible location of the pancreas, the intimate relationship between acinar and ductal components, and the relative paucity of duct cells, which make up only 5% to 10% of the pancreatic cell mass. Separation of duct cells for in vitro study and isolation of duct cell function in vivo are exceedingly difficult. Most efforts to date have relied on laborious whole- organ digestion, followed by isolation of short duc- tal segments and either immediate study or selec- tive growth in culture of the few ductules obtained.‘* A simpler alternative, use of tumor cell lines bearing antigenic similarities to duct cells, is complicated by the uncertainties of the neoplastic phenotype.

In an attempt to overcome these difficulties, we modified methods previously used for the culture of rabbit esophagus13 to develop an explant model of bovine pancreatic duct culture. Others have used a similar process in the pancreas.‘* The bovine pancreas is an excellent source of tissue because of the large diameter of its central ductal system and the ready availability of specimens from local beef-processing plants. Although study of pancreatic function in this species has not been

481 Alvarez et al. surgery August 1997

;,,

H ij 75 % t 4 SO

OmM O.OSmM 0.5mM 1mM --

r OmM z

O.OSmM 1

O.StllM a

1mM

TDCA

Fig. 5. Morphometric appearance of the duct explant epithelium with bile acid exposure in the presence or absence of EGF. *p < 0.05 vs 100 nmol/L EGF.

extensive, it appears to resemble that of other ruminants, such as the rabbit and the rat. The explanted epithelium preserves the critical rela- tionship between neighboring cells and between cells and the native basement membrane, which may influence tissue repair.141r5 The cells demon- strate normal morphology, including expression of key ductal secretory proteins such as CA II and CFTR, for as long as 3 weeks in culture.

We successfully adapted an established method of in vivo ductal damage, bile-induced injury, to the in vitro setting3,16 and documented that duct epithelial cells are exquisitely sensitive to bile acids. TDCA was used in these experiments because of its well-established ability to damage epithelia in the gastrointestinal tract and pancreas. In preliminary studies we noted that concentrations of TDCA

higher than 1 mmol/L, commonly used to induce pancreatitis with ductal injection or perfusion, resulted in complete epithelial sloughing with incubation times as short as 4 hours. A more grad- ed injury followed the use of bile acid concentra- tions one and two orders of magnitude below that observed in bile. It seems likely that bile acid-relat- ed cytotoxity involves mechanisms other than sim- ple detergent activity observed at higher concen- trations than used here. l7 These findings are consistent with in vitro results in other cell lines.‘” The mechanism for bile-associated cytotoxicity, whether other bile acids incur damage in a similar manner, and the optimal concentration of bile acids necessary to induce damage with shorter exposure times will be the subject of future studies.

In experimental animals disruption of the duct

surgery Volume 122, Number 2

Alvara et al. 482

epithelial barrier by bile acids results in pancreatic inflammation.2-4,16 Similar effects on ductal per- meability have been noted with other factors asso- ciated with clinical pancreatitis including ethanol, hypercalcemia, duct obstruction, and supramaxi- ma1 stimulation.3,5,1Q,20 This has led to the hypoth- esis that the epithelium acts as a barrier in a man- ner analogous to the gastroduodenal mucosa, protecting the pancreatic parenchyma against back diffusion of and damage by digestive enzymes and alkaline juice.’ Although the mechanisms underly- ing the pathophysiology of acute biliary pancreati- tis remain undetermined, the finding that chole- docholithiasis does not induce pancreatitis at every opportunity suggests that other factors may play a role in both establishing an individual’s risk for developing pancreatitis in this setting and deter- mining the severity of a particular episode. As with ulcer disease in the stomach, the balance of forces that enhance or weaken the ductal epithelial barri- er could define an individual’s risk of developing pancreatitis and the course of the disease.

Growth factors like EGF and TGF-a could exert such influence. These peptides are part of a grow- ing family of growth factors that play an important role in the regulation of growth and other func- tions in the gastrointestinal tract.*,tO Although characterized primarily by their mitogenic effects, EGF and TGF-a have also been shown to protect gastrointestinal epithelia from injury.‘,1° The sali- vary glands are the richest source of EGF in most species, but pancreatic juice contains significant amounts (2 to 10 nmol/L) and levels increase with secretory stimulation. 21 Duct cells possess EGFr, which is found on the luminal cell membrane, in contrast to the basal position found in other gas- trointestinal tissues (Fig. 2,~).~~ Thus EGF or TGF- cx or both are probably involved in normal duct cell physiology.

A number of recent reports have noted in- creased expression of EGF, TGF-a, and EGFr in the course of pancreatic inflammation6,’ Although it has been suggested that such overexpression may contribute to pancreatic damage, little is known about the effects of these growth factors on the exocrine pancreas. The available data are often contradictory. In the intact animal EGF has been shown to not affect, to promote, and to inhibit pro- liferation and secretion.21,23 Genetically engi- neered overproduction of TGF-o produces overall pancreatic atrophy but ductal hyperplasia.* In agreement with our results, EGF has been shown to be a mitogen for both primary duct cell cultures and pancreatic tumor cell lines expressing a duct phenotype. 8,15,24 Less consistent is the effect on

cultured acinar cells, for which results on growth and secretory responses to EGF have been vari- ,bleq21,25,26

EGF did not significantly affect epithelial exten- sion from the cut edge of our explants. However, it did promote proliferation in this region of the explants as demonstrated on PCNA staining, a reli- able measure of cellular proliferation.27 Thus we hypothesize that this phenomenon may involve both proliferation, responsive to EGF stimulation, and EGF-independent epithelial cell migration.** The degree of “growth” from the cut edge of the explant correlated well with epithelial cell density in the setting of bile-induced injury, suggesting that cell damage and loss account for the changes noted. When EGF was added alongside TDCA, basal levels of epithelial density and extension from the edge were maintained, supporting the idea that the main effect of EGF in the setting of acute dam- age is cellular protection, at least at this relatively early time. This response was effectively reversed with the addition of tyrphostin A47, an inhibitor of the EGFr tyrosine kinase.2Q The ability of EGF to protect from this injury therefore appears specific and suggests a potential physiologic role for this growth factor in the maintenance of the pancreat- ic duct mucosal barrier.

We wish to express our gratitude for the technical assistance provided by Kathleen Lally and Zhoi Li.

REFERENCES

1. Reber HA, Roberts C, Way LW. The pancreatic duct mucos-

al barrier. Am J Surg 1979;137:128-34.

2. Konok GP, Thompson AG. Pancreatic ductal mucosa as a protective barrier in the pathogenesis of pancreatitis. Am J

Surg 1969;117:18-23.

3. Reber HA, Adler G, Karanjia N, Widdison A. Permeability characteristics of the main pancreatic duct in cats. In: Go

VLW, Gardner JD, Brooks FP, Lebenthal E, DiMagno EP,

Scheele GA, editors. The exocrine pancreas: biology,

pathobiology, and diseases. 2nd ed. New York: Raven Press,

1993. p 527-50.

4. Widdison AL, Alvarez C, Reber HA. The low-pressure duct perfusion model of acute pancreatitis. Eur Surg Res

1992;24:55-61.

5. Gates MC, Singh SM, Peick AL, Harvey MH, Reber HA.

Acute hypercalcemia, pancreatic duct permeability, and pancreatitis in cats. Surgery 1988;104:137-41.

6. Korc M, Friess H, Yamanaka Y, Kobrin MS, Buchler M, Beger HG. Chronic pancreatitis is associated with increased

concentrations of epidermal growth factor receptor, trans-

forming growth factor alpha, and phospholipase C gamma. Gut 1994;35:1468-73.

7. Ebert M, Friess H, Buchler MW, Korc M. Differential distri- bution of human epidermal growth factor receptor family

in acute pancreatitis. Dig Dis Sci 1995;40:2134-42. 8. Barnard JA, Beauchamp RD, Russell WE, Dubois RN,

Coffey RJ. Epidermal growth factor-related peptides and

483 Alvarez et al.

9.

10.

11.

12.

their relevance to gastrointestinal pathophysiology.

Gastroenterology 1995;108:56480.

Riegler M, Sedivy R, Sogukoglu T, et al. Epidermal growth factor promotes rapid response to epithelial injury in rab-

bit duodenum in vitro. Gastroenterology 1996;111:28-36. Uribe JM, Barrett KE. Nonmitogenic actions of growth fac-

tors: an integrated view of their role in intestinal physiolc- gy and pathophysiology. Gastroenterology 1997;112:255-68.

Hootman SR, de Ondarza J. Overview of pancreatic duct physiology and pathophysiology. Digestion 1993;54:32330.

Githens S. Pancreatic duct cell cultures. Annu Rev Physiol

1994;56:419-43.

13.

14.

15.

16.

17.

D’Addio VJ, Sayles JM, Alvarez C, Lally K Bass BL. Bile: a

trophic factor for rabbit esophageal epithelium in vitro.

Surg Forum 1996;47:125-7.

Raghow R. The role of extracellular matrix in postinflam- matory wound healing and fibrosis. FASEB J 1994;8:823-31.

Bhattacharyya E, Panchal A, Wilkins TJ, de Ondarza J, Hootman SR. Insulin, transforming growth factors, and

substrates modulate growth of guinea pig pancreatic duct

cells in vitro. Gastroenterology 1995;109:94452. Luthen RE, Niederau C, Grendell JH. Effects of bile and pancreatic digestive enzymes on permeability of the pan-

creatic duct system in rabbits. Pancreas 1993;8:671-81.

Devor DC, Sekar MC, Frizzell RA, Duffey ME. Taurodeoxycholate activates potassium and chloride con-

ductances via an IP3-mediated release of calcium from

intracellular stores in a colonic cell line (T84).J Clin Invest

1993;92:2173-81. 18. Shekels LL, Beste JE, Ho SB. Tauroursodeoxycholic acid

protects in vitro models of human colonic cancer cells from cytotoxic effects of hydrophobic bile acids. J Lab Clin Med

1996;127:57-66.

19.

20.

21

Harvey MH, Wedgwood KR, Austin JA, Reber HA.

Pancreatic duct pressure, duct permeability and acute pan-

creatitis. Br J Surg 1989;76:859-62. Fallon MB, Gorelick FS, Anderson JM, Mennone A, Saluja

A, Steer ML. Effect of cerulein hyperstimulation on the paracellular barrier of rat exocrine pancreas.

Gastroenterology 1995;108:1863-72.

Jaworek J, Konturek SJ. Distribution, release, and secretory activity of epidermal growth factor in the pancreas. Int J

Pancreatol 1990;6:189-205.

22 Korc M, Chandrasekar B, YamanakaY, Friess H, Buchier M,

Beger HG. Overexpression of the epidermal growth factor

receptor in human pancreatic cancer is associated with concomitant increases in the levels of epidermal growth

factor and transforming growth factor alpha. J Clin Invest

1992;90:1352-60.

23

24

Loser C, Folsch UR. Epidermal growth factor (EGF) fails to

stimulate pancreatic growth and pancreatic polyamine metabolism in rats. Z Gastroenterol 1995;32:21620. Mangino MM, Hubchak S, Scarpelli DG. Stimulation of

DNA synthesis in pancreatic duct cells by gastrointestinal

hormones: interaction with other growth factors. Pancreas 1992;7:271-9.

96 w.r. Logsdon CD. Stimulation of pancreatic acinar cell growth

by CCK, epidermal growth factor, and insulin in vitro. Am J Physiol 1986;251:G487-94.

26. Stryjek-Kaminska D, Piiper A, Caspary WF, Zeuzem S. Epidermal growth factor inhibits amylase secretion and

activation of phospholipase C in response to calcium-mobi- lizing secretagogues in rat pancreatic acini. Z Gastroenterol 1994;32:232-5.

27. Velazquez OM, Zhou D, Seto RW, et al. In vivo crypt surface

hyperproliferation is decreased by butyrate and increased

surgery August 1997

by deoxycholate in normal rat colon: associated in vivo

effects on c-Fos and c-Jun expression. JPEN 1996;20:243-50. 28. Basson MD, Beidler DR, Turowski G, et al. Effect of tyro-

sine kinase inhibition on basal and epidermal growth fac- tor-stimulated human Caco-2 enterocyte sheet migration

and proliferation. J Cell Physiol 1994;160:491-501. 29. Lyall RM, Zilberstein A, Gazit A, Gilon C, Levitzki A,

Schlessinger J. Tyrphostins inhibit epidermal growth factor (EGF)-receptor tyrosine kinase activity in living cells and EGF-stimulated cell proliferation. J Biol Chem

1989;264:14503-9.

DISCUSSION

Dr. Howard A. Reber (Sepulveda, Calif.). So you have really done two things; you have designed a new model and you have used it, you are beginning to use it, to ask and answer some questions.

I have some questions about both the model and some of the conclusions that you have reached. First of all, you have indicated that you really needed to go up to concentrations of 100 nmol/L EGF to see a statistically significant response. Also you have said that the highest concentration in pancreatic juice that is being normally secreted is about 10 nmol/L. You said that EGF concen- trations do increase with secretory stimulation. How high do EGF concentrations in pancreatic juice go? Do they get up to the levels where you have seen the protective effects that you demonstrated in your experiments? If they do get up that high, then a corollary to that ques- tion is how long can they be maintained at that level? Presumably if EGF is going to act in a protective way, it would need to be at that level for a reasonable amount of time.

Another issue is that it seems to me that the only time you have given EGF is along with the taurocholate. Have you done experiments in which you have injured with the taurocholate and then after that injury occurred you have removed the taurocholate solution and put in EGF and then followed the morphologic appearance of the cells to see whether EGF can hasten repair after that injury?

Finally, you have indicated here that the duct cells contain EGF receptors on their luminal surface. That is a very interesting observation. Do you know whether there is any difference in receptor density on the duct cells in different parts of the ductal system? If your hypothesis that EGF may really serve a protective role physiologically is true, you might expect to see more receptors, or have those cells more responsive in some way, in those cells that are in the main duct as opposed to the smaller ducts, and maybe in the main duct closer to the duodenum rather than in the main duct out in the tail, because that is where the highest concentrations of bile are ever likely to be seen. Have you done any studies in which you have seen whether exposure to tauro- cholate or other injurious agents has the ability to up- regulate the receptors? Can you alter receptor density by chronic exposure to the taurocholate?

Dr. Alvarez. There have been only a few reports describing either EGF secretion in pancreatic juice or level of expression in the tissues, and none that correlate the two. The immunohistochemical evidence from both

surgery Volume 122, Number 2

Alvarez et al. 484

clinical and experimental pancreatitis suggests up-regu- lation of EGF expression, whereas examination of pan- creatic juice from both animals and human beings has shown a fivefold to tenfold increase in levels with stimu- lation. The duration of this increase and how much of an increase from basal is seen in juice with pancreatitis is unknown. I would also point out that although we did use a high dose of EGF to achieve statistically significant differences, it is likely that some of the EGF gets adsorbed onto the absorbable gelating sponge rafts used to suspend the explants out of the medium. There may also be a diffusion barrier across the connective tissue underlying the epithelium, so I think the concentration at the ductal cell is considerably less than what we added to the medium.

With regard to your other questions, the short answer is that you have outlined very nicely the work that we are currently undertaking, looking at whether EGF can be effective after the damage has occurred and whether receptor levels change with injury. I do not have any information of receptor distribution along the pancreat- ic duct, although you propose an interesting hypothesis.

Dr. David I. Soybel (West Roxbury, Mass.). The ques- tions that I have are similar to the types of questions that used to be asked about cytoprotection in the stomach.

Is it taurocholate alone that does this, or is there a dif- ferent effect of bile acids? As you know, John Harmon showed many years ago that there are sort of different effects of different bile acids, different levels of injury. To follow that up, what do you think is the mechanism of injury by bile acids in this particular model? The doses of your bile acids and the kind of injuries you are seeing suggest that this might be more than just a detergent-like effect and that might provide you some insight into why EGF is actually protective in this setting.

Do you think that bile regularly bathes the pancreatic duct? Or what sources of injury are you postulating? Here obviously the bile acid is being used as a model of injury. Do you think this is an ongoing physiologic insult?

Dr. Alvarez. Regarding your question on whether other bile acids produce similar injury, I believe that Dr. Reber carried out a series of experiments perfusing the main duct of cats with a variety of bile acids and noted no significant difference with other gastrointestinal epithe- lia in vitro. The mechanism of action generally proposed for bile acid injury has been their detergent action on the cell membrane. However, at low concentrations like we used, below the micellar concentration, the detergent activity is not seen. A number of studies looking at gas- trointestinal epithelia have suggested that there are ion fluxes, particularly calcium, created by bile acid expo- sure. Whether this is receptor mediated or due to non- specific permeability changes is unclear at this time.

Regarding your second question, it has been nearly a century since Opie proposed the common channel the- ory, and I do not think anyone has yet come up with a good explanation of how biliary pancreatitis develops. I am certainly in no position to enter that debate.

Dr. Soybel. That is not the thrust of the question. Whether you believe that bile refluxes when there is a stone obstructing the duct, do you think that under phys- iologic circumstances with changes in sphincter of Oddi motility that bile might regularly be bathing at least the proximal portion of the duct; is there any information about that?

Dr. Alvarez. We know that patients with pancreatitis have reflux more frequently than those being studied for other reasons. I think it has been assumed that a one- time event caused the pancreatitis but yet the reflux has usually been demonstrated at a later time during cholan- giography. In fact, we do not know whether bile normal- ly refluxes into the pancreatic duct. I suspect it does, and as with peptic ulcer disease, that it is the balance of dam- aging and protective factors that dictates whether pan- creatitis develops. But that is a very weak hypothesis at this point.