Gastric lipase in alcoholic pancreatitis

LIVER,

PANCREAS,

AND

BILIARY

TRACT

Gastric Lipase in Alcoholic Pancreatitis Comparison of Secretive Profiles Following Pentagastrin Stimulation in Normal Adults and Patients With Pancreatic Insufficiency JACQUES MOREAU, MICHELE BOUISSON, DANIEL BALAS, ALAIN RAVAUD, SYLVIO STUPNIK, LOUIS BUSCAIL, NICOLE VAYSSE, and ANDRE RIBET Service des maladies de l’appareil digestif and INSERM U 151, CHU Rangueil, Toulouse, France

The aims of this study were to evaluate the amount of gastric lipase secreted by the stomach in normal adults and to elucidate a possible adaptative secretion of this enzyme in response to pancreatic insufficiency secondary to alcoholic chronic pancreatitis. Forty-one subjects underwent a gastric intubation. Pentagastrin (6 rg . kg-’ - h-’ IV) significantly increased gastric lipase concentration and output. Stimulated gastric lipase output in seven normal subjects was 12,596 + 2036 U/h (by using tributyrin as substrate). Outputs where higher (P < 0.02) in 17 patients with pancreatic insufficiency who were not drinking alcohol, but were not significantly different in nine patients who continued to drink (20,413 + 1776 U/h and 21,953 + 4973 U/h, respectively). On the other hand, high gastric lipase outputs were found in eight patients with duodenal ulcers and no evidence of pancreatic dysfunction (23,160 f 262 U/h). The time required to reach maximal lipase output (peak output) following pentagastrin stimulation was the same in all groups (~38 minutes) except for the group of patients with pancreatic insufficiency who did not drink alcohol, in whom it was significantly reduced (a26.5 minutes). Secretory patterns of gastric lipase and pepsin were closely comparable. Gastric lipase secretion could be increased in several clinical conditions and particularly in patients with pancreatic insufficiency caused by alcoholic chronic pancreatitis who have been abstinent for a long time.

s suggested by previous studies (l-4], gastric lipase plays a major role in the digestion of dietary fat, particularly when pancreatic lipase is decreased (i.e., prematurity, cystic fibrosis, and chronic pancreatitis). Recent advances have been achieved

A

with regard to the biochemical properties and the tissue origin of this lipolytic enzyme (l-7). Its localization at the cellular level has recently been elucidated (8). Nevertheless, to our knowledge, the amount of lipase secreted by the stomach in adults with pancreatic insufficiency (PI) has never been evaluated in a large number of patients. Thus, the main aim of this study was to compare the total acid lipolytic activity recovered from the gastric juice under basal and pentagastrin-stimulated conditions in normal control subjects and in patients with chronic pancreatitis submitted to this investigation after a long period of abstinence (A + PI]. To further clarify the possibility of a compensatory secretion in response to severe PI, comparative data were obtain from a third group of persistent drinkers with pancreatic dysfunction (NA + PI) and from a fourth group of patients with duodenal ulcer (DU) and high stimulated acid output. These two last groups were chosen to evaluate the effects of alcohol drinking and acid secretion on gastric lipase output.

Materials and Methods Subjects After informed consent was obtained, and in accordance with the recommendations of the human experimental committee, 41 subjects took part in the study. Control subjects. Seven control subjects [aged 35 k S years) were investigated for functional complaints. All had

Abbreviations used in this paper: DU, duodenal ulcer; PI, pancreatic iusuffkiency. 0 1990 by the American Gash-oenterological Association 0016-5085/90/$3.00

176

MOREAU

ET AL.

normal gastric acid secretion and no evidence of pancreatic dysfunction. Abstainers with pancreatic insufficiency. Seventeen patients (aged 33 + 12 years] with severe exocrine PI caused by alcoholic chronic pancreatitis who had been abstinent for at least 6 months were included. Nonabstainers with pancreatic insufficiency. This group consisted of 9 patients with chronic pancreatitis (aged 49 * 9 years] who had a long history of heavy ethanol intake (>150 g/day] and were still drinking at the time of the study. Patients with duodenal ulcer. Eight patients (aged 38 + 7 years] had a recent history of endoscopically proven duodenal ulcer and age- and sex-adjusted values for peak acid output more than 2 SD above normal limits (91,but were not alcoholics and had no pancreatic symptoms. The diagnosis of PI in the second and the third groups was based on a major decrease of all exocrine pancreatic secretions, including pancreatic lipase output tlOO0 U/min after hormonal stimulation by secretin (1 CU . kg-’ - h-‘] and cerulein (75 ng . kg-’ - h-l]. The diagnosis of chronic pancreatitis in these same two groups was based on the presence of two or more of the following criteria: (a] visualization of pancreatic calcifications on a plain x-ray film of the abdomen: (b] presence of pancreatic duct abnormalities strongly suggestive of chronic pancreatitis at endoscopic retrograde cholangiopancreatography; (c] presence of steatorrhea (>7.0 g on a daily dietary fat intake of 100 g]; and (d] abnormal oral glucose tolerance test or diabetes mellitus. Three out of the 26 patients with chronic pancreatitis met only criteria c and d. All three had severe PI as measured by duodenal intubation tests. All chronic pancreatitis patients were asked to stop taking their oral pancreatic supplementation at least 10 days before beginning the study.

Methods Collection of gastric juice. After an overnight fast, subjects were asked to swallow a single-lumen polyvinyl tube (5 mm ID]. Under fluoroscopic control, the tube was placed so that the multiple openings for aspiration were in the antrum. Gastric samples were recovered in ice-chilled tubes using a continuous aspiration device (at -100 cm H,O]. This method allowed recovery of 90% of gastric content. At the end of four 15-minute basal collections, gastric secretion was stimulated by IV infusion of pentagastrin (6 fig - kg-’ . h-‘] (Peptavlon; ICI, Macclesfield, England] and by four 15-minute collection periods. At the end of each collection, total volume and pH were measured. Hydrogen ion concentration and lipolytic and pepsin activities were evaluated on each experimental day. Assay of gastric acidity. Hydrogen ion concentration was determinated by automatic titration to pH 7 with 0.1 mol/L NaOH using end-point titrator (Radiometer, Copenhagen, Denmark]. Gastric acid outputs were calculated from the product of concentration and volume and expressed as millimoles of hydrogen ions per hour. Assay of gastric lipolytic activity. Lipolytic activity was measured using tributyrin (Fluka, Buchs, Switzerland), a short-chain triacylglycerol. as substrate according to the method of Gargouri et al. (3). Schematically, the incubation

GASTROENTEROLOGY Vol. 99, No. 1

medium contained tributyrin (0.5 mL, 110 mmol/L], taurodeoxycholate (2.5 mmol/L] (Sigma, St. Louis, MO] in acetate buffer (2 mmol/L] with CaCI, (3 mmol/L], pH 5.4, at 25°C. Gastric lipase outputs were expressed as micromoles of free fatty acid released per milliliter per minute (U/min]. Assay of pepsin activity. Pepsin activity was mea-

sured at pH 1.6 using bovine albumin, fraction V [Sigma], as substrate (10,ll). Tyrosin residues were estimated by spectrophotometry. One unit of pepsin corresponded to the release of 32.2 Mg/mL of tyrosin. Pepsin outputs were expressed as units per milliliter per minute (U/min].

Statistical Methods All of the variables except one (lipase output] were normally distributed as arithmetic data. Thus lipase output data were subjected to natural log transformation. Analysis of variances (F test] shows that the range of interindividual stimulated lipase output values in drinking patients (NA + PI] was too wide to allow further statistical evaluation. Multiple comparisons between the three other groups were performed using Scheff e’s contrasts method (12). When significant interactions were found, linear regression coefficients were estimated. Precise analysis of secretory profiles were achieved by calculation of the mean time to the peak output and area under the curve. Results Basal and Stimulated Gastric Lipase and Gastric Acid Concentrations and Outputs In the whole population studied, basal and stimulated gastric lipase concentrations (mean + SEM) were 40.40 + 3.45 U/mL and 59.74 + 3.12 U/mL, respectively (P < 0.001). Comparative output data are shown in Table 1. Higher baseline levels of gastric lipase output were found in NA+PI patients than in CS for all 15-minute intervals. Nevertheless, because of variations in the results obtained in alcoholic patients, this difference was not statistically significant (see Materials and Methods). Conversely, stimulated gastric lipase outputs were significantly higher in A + PI patients (20,413 + 1778 U/h) than in controls (12,598 * 2036 U/h) (P < 0.02). The highest mean stimulated gastric lipase output value was found in DU patients (23,180 + 2621 U/h] and was not statistically different from the data obtained in A + PI patients. Peak output following pentagastrin stimulation was also higher in A+PI group than in controls (Table 2). Comparison of the secretory profile among the patient groups [Figure 1) indicates that the time required to observe the maximal response was shorter in the A+PI group than in controls (P < 0.05). This phenomenon was not observed in NA+PI or DU patients. Stimulated gastric acid outputs were significantly

HUMAN GASTRICLIPASEAND PENTAGASTRIN 177

July 1990

Table 1. Basal and Pentagastrin-Stimulated Hourly Acid, Gastric, Lipase, and Pepsin Outputs [Mean if:SEM] Acid (mmoVh) Control A + PI NA+PI DU

n

Age (yrl

7 17 9 8

39 + 3 46 f 3 51 f 3 40 f 3

Basal 2.57' + 1.09 5.52* 1.69 3.88+ 1.63 6.30” zt 1.53

Pepsin (U/h]

Lipase (U/h]

Stimulated 19.96' _+4.16d 36.05' + 3.26 31.15O + 5.57 43.40d f 4.50”

Stimulated

Basal 3020+ 496' 5188+ 942 7556+1508' 5445 f 1002

12.59eb f 2036d 20,413b f 1778 21,953 * 4973 23.180d + 2621

Stimulated

Basal 43.9" + 16.0 65.0+ 13.4 72.2f 22.4 104.6” f 18.2

160.6d A 24.3 230.2f 21.9 222.2+ 63.5 295.3d f 21.6

NOTE. All differences between basal and stimulated outputs in the same group were highly significant. The significant differences same parameter [column) and between the different groups are shown as paired letters. "P < 0.05. bP < 0.02. "P< 0.01. dP < 0.01.

higher in A+PI patients than in controls (Table 1). However, gastric acid secretory profiles appeared parallel in these two groups of subjects (Figure 2). Finally, high stimulated gastric acid outputs were also noticed in NA+ PI and DU patients (Table 1).

Correlations Between Gastric Lipase and Gastric Acid Outputs As shown in Table 3, a positive correlation between gastric lipase and acid outputs was found in control subjects during the first 30 minutes after pentagastrin infusion. Gastric and enzymatic secretions were well correlated over the whole stimulated period for the NA+PI patients, whereas no correlation was observed for A+ PI and DU patients.

Correlation Between Gastric Lipase and Pepsin Basal and stimulated gastric lipase and pepsin outputs are strongly correlated in all groups except in the case of A+PI patients, in whom linear regression coefficients decreased during the stimulation period and became non-statistically significant at the end of the pentagastrin infusion (Table 3).

for the

Correlation Between Gastric Acid and Pepsin Outputs Correlations were similar to those obtained between gastric acid and gastric lipase outputs (data not shown]. Discussion The physiological importance of human gastric lipase in the digestion of dietary fat is still debated (1,4,6j. Several studies suggest that intragastric lipolysis could play a major role in conditions associated with low levels of pancreatic lipase such as prematurity, congenital absence of pancreatic lipase, and cystic fibrosis. Abrams et al. (4) reported that 90% of the total lipolytic activity measured at the ligament of Treitz in patients with severe PI was of gastric origin, compared with 7% in control subjects. Nevertheless, the postprandial duodenal nonpancreatic lipolytic activity in patients with PI (whether caused by alcoholism or cystic fibrosis) was not quantitatively different from activity in controls. The authors suggested that the low duodenal pH and the low level of tryptic activity in patients with PI allowed gastric lipase to be active in the duodenum. Thus gastric lipase secretion may be a compensatory mechanism for fat digestion in

Table 2. Peak Lipase Output, Mean Time to Peak Lipase Output, and Area Under the Curve

Controls A + PI NA+PI DU

n

PLO

7 17 9 8

3738” + 535b.c 7388b f 845 7351Cf 1450 863gazt991

Time to PLO 40.7c * 7.1 26.5” + 3.2 35.0 + 6.1 37.5f 5.7

ALJC 2762’ zt 454’ 4528’ f 427 +64% 5369’ f 1036 +94% 5034'* 530 i-82%

NOTE. Peak lipase output is expressed as units per 15-minute collection period corresponding to the mean * SEM of the highest value obtained during the stimulation period in each group. Time to the PLO is expressed in minutes corresponding to the time required from the beginning of the stimulation to reach the PLO. Significant differences are shown by paired letters. PLO, peak lipase output; AUC, area under the curve. “P < 0.01. bP < 0.01. “P < 0.05.

178 MOREALJ ET AL.

LIPASE

1

6

Figure 1. Mean basal and pentagastrin (PG~stimulated gastric lipase outputs in control subjects [k), abstainers with PI (a), nonabstainers with PI (o), and patients with duodenal ulcer (0).

patients with PI. In these previous studies, however, the amount of gastric lipase was not quantified because only concentrations were measured. Moreover, the choice of chronic alcoholics as control subjects introduced another possible bias. Indeed, compared with control values, we found higher values of basal gastric lipase output in nonabstinent patients. In our study, however, because of scattered results, this difference did not reach statistical significance for stimulated output values. Conversely, preliminary data from our own group have shown a significant decrease in the tissue concentrations of gastric lipase (measured on homogenized gastric biopsies) in chronic alcoholics when compared with abstinent subjects with or without PI (13). These findings were correlated with the degree of mucosal dystrophy measured by means of a precise histological score. Finally, high basal output and low tissue concentration are both consistent with a chronic impairement of gastric mucosa secondary to persistent ethanol intake. One of the main aims of the present study was to evaluate the total amount of gastric lipase secreted in the stomach under basal and stimulated conditions in

GASTROENTEROLOGY

Vol. 99. No. I

normal subjects and in patients with PI. We have found a significant increase in stimulated gastric lipase in abstinent patients with PI when compared with normal controls. Nevertheless, when either the role of alcoholism or the role of gastric acid secretory status is taken into account, the only difference remaining is an earlier response to the stimulation in A+PI patients than in the three other groups. Quantitatively, the amount of gastric lipase activity measured during the l-hour stimulated period ranged from 13-23 kU/h. Because the mean value of the stimulated pancreatic lipase output in normal humans after hormonal stimulation by cholecystokinin is about 130 kU/h (range, 77-322 kU/h) (14), the ratio of gastric lipase to pancreatic lipase is probably around 0.10 or 0.20. This is only a rough estimation because in vitro data suggest that lipolytic activity of gastric lipase measurement by the tributyrin assay (a medium chain triglyceride) is two to four times more efficient than measurement by the triolein assay (a long-chain triglyceride that is more representative of in vivo substrates] (3,4,7). One could argue that in patients with severe PI, this ratio is significantly higher so that preliminary digestion of dietary fat by gastric lipase becomes essential and may occur in the duodenum, where optimal conditions [including low pH and low proteolytic activity) may be present (4). Another quantitative approach to gastric lipase secretion has recently been reported by Moreau et al. (6). In this study, the total gastric lipase content of fundic mucosa from biopsies and organ donor specimens was compared with the total lipase content of the human

ACID

Figure 2. Comparison of gastric acid secretory profiles (mean + SEM) in control subjects (*)I abstainers with PI (a), nonabstainers with PI (o), and patients with duodenal ulcer (0 j

HUMAN GASTRIC LIPASE AND PENTAGASTRIN

July 1990

Table 3.

179

Values of Linear Regression Coefficient Obtained for the Relationships Between Gastric Acid and Gastric Lipase and Between Gastric Lipase and Pepsin Output After Pentagastrin Stimulation for the Different Groups of Patients

n Controls(r) A + PI(r) NA + PI [r) DU(pl

7 17 9 8

H+/L 0.7097" 0.1380 0.9174b 0.6356

L/P 0.9045b 0.7815b 0.9750b 0.8292'

H+/L 0.7503O 0.4688 0.9411b 0.2830

60 Minute

45 Minute

30 Minute

15 Minute

H+/L

L/P 0.6255' 0.6526b 0.8670b 0.8237'

0.5820 0.4028 0.9547b -0.038

L/P 0.8374" 0.5512" 0.941Y 0.7188'

H+/L 0.7470 0.6798b 0.8165b -0.3419

L/P 0.8250" 0.0760 0.9492b 0.8451b

H+, acid; L, lipase; P, pepsin. “P c 0.02or P < 0.05. bP < 0.01.

pancreas. The authors suggest that the total amount of gastric lipase in the fundic mucosa in normal subjects is about five times lower than the amount of pancreatic lipase present in the whole pancreas; the results of the present study agree. In the same report, gastric lipase secretion was shown to decrease in patients aged >50 years. Although the age of drinking patients was slightly higher than the age of subjects in other groups, this difference was not significant. Moreover, most of the patients from this series were t5O years old. Thus it is unlikely that our results have been influenced by age. As already described in the literature (15,16), we found that pentagastrin-stimulated acid output was significantly higher in abstinent patients with chronic alcoholic pancreatitis than in healthy subjects. However, this could not be demonstrated in NA+PI patients. Older publications showed that hypochlorhydria secondary to mucosal atrophy was proportional to the duration and the severity of ethanol abuse (17,18). Our results on pentagastrin-stimulated gastric lipase secretion are not consistent with a simple wash-out phenomenon. Indeed, high gastric lipase output levels were maintained during the l-hour stimulation period and were associated with a significant increase in gastric lipase concentration values. Moreover, there was poor correlation between gastric acid and gastric lipase secretions, especially in DU patients. Similar observations on gastric secretions of enzymes have been made by Szafran et al. (19) in DU patients before and after vagotomy. Conversely, our results are not in agreement with a recent report showing that pentagastrin increased the output but not the concentration of gastric lipase in the gastric juice (20). However, the population studied and the experimental conditions were not comparable to the one used in this study. Surprisingly, in the same report, the authors found that secretin and/or cerulein had no effect on gastric lipase secretion. These findings are once again in disagreement with in vitro data showing that gastric lipase secretion from gastric biopsies was increased after addition of carbamylcholine or cerulein to the incubation medium (211.

The existence of a strong correlation between secretion and content of pepsin and gastric lipase (13,191 suggests that these two enzymes may be secreted by cells under similar regulatory mechanisms. This hypothesis agrees with the immunocytolocalization of human lipase in chief cells of the fundic mucosa as described recently by Moreau et al. (8). However, in A+PI patients the correlation between gastric lipase and pepsin declined during the stimulation period and was not significant during the last 15-minute period. In summary, the secretory profile of gastric lipase in response to pentagastrin stimulation is consistent with an increased sensitivity of chief cells responsible for the secretion of this enzyme in patients with PI. This phenomenon is observed only after a long period of abstinence, when gastric mucosa becomes normal histologically. Stimulated gastric lipase is higher in PI patients than in healthy subjects, but it may increase in circumstances other than PI (i.e., DU). However, the total amount of gastric lipase secreted in the stomach is relatively low compared with the total lipolytic activity usually secreted by the exocrine pancreas following hormonal stimulation. Nevertheless, the role of gastric lipase in the digestion of dietary fat must be considered, particularly when the ratio of gastric lipase to pancreatic lipase is high such as in patients with severe PI. The therapeutic relevance of these observations is currently being investigated.

References 1. Hamosh M, Scalon JW, Ganot D, Like1 M, Scanlon KB, Hamosh P. Fat digestion in the newborn. J CIin Invest 1981;67:838-846, Bodmer MW, Angal S, Yarranton GT, Harris TJR, Lyons A, King DJ, Pieroni G, Riviere C, Verger R. Lowe PA. Molecular cloning of a human gastric lipase and expression of the enzyme in yeast. Biochim Biophys Acta 1987;909:237-244. Gargouri Y, Pieroni G. Riviere C, Sauniere JF, Lowe PA, Sarda L, Verger R. Kinetic assay of human gastric lipase on short and long chain triacyiglycerol emulsions. Gastroenterology 1986;91: 919-925. Abrams CK, Hamosh M, Dutta SK, Hubbard VS. Hamosh P.

180 MOREAU ET AL.

5.

6.

7.

8.

9.

10.

11. 12. 13.

14.

15.

Role of nonpancreatic lipolytic activity in exocrine pancreatic insufficiency. Gastroenterology 1987;92:125-129. Abrams CK, Hamosh M, Lee TC, Ansher AF, Collen MJ, Lewis JH, Benjamin SB, Hamosh P. Gastric lipase: localization in the human stomach. Gastroenterology 1988;95:1460-1464. Moreau H, Laugier R, Gargoury Y, Ferrato F, Verger R. Human preduodenal lipase is entirely of gastric fundic origin. GastroenteroIogy 1988;95:1221-1226. Levy E, Goldstein R, Freier S, Shafrir E. Characterization of gastric Iipolytic activity. Biochim Biophys Acta 1981;664:316326. Moreau H, Bernadac A, Gargouri Y, Benkouka F. Laugier R, Verger R. Immunocytolocalization of human gastric lipase in chief cells of the fundic mucosa. Histochemistry 1989;91:419423. Mignon M, Vallot T. Exploration de la secretion gastrique. In: JJ Bernier, ed. Gastroenterologie. Paris: Flammarion, 1984:239251. Vagne M, Martin F, Lambert R. Dosage de la pepsine dans la secretion gastrique humaine. Influence de differents facteurs sur la proteolyse peptique. Rev Franc Etud CIin Biol1967;12:559565. Samloff M. Pepsinogens, pepsins, and pepsin inhibitors. Gastroenterology 1971;60:586-804. Godfrey K. Statistics in practice: comparing the mean of several groups. N Engl J Med 1985;313:1450-1456. Moreau J. Stupnic S, Bouisson M, Ravaud A, Bertrand C, Balas D, Ribet A. Adaptative secretion of gastric lipase in alcoholic patients with pancreatic insufficiency: a reality? (abstr). Digestion 1987;38:41. Di Magno EP. Human exocrine pancreatic enzyme secretion. In: VLW Go, ed. The exocrine pancreas. New York: Raven, 1988:193-211. Saunders JHB, Cargill JM, Wormsley. Gastric secretion of acid in patients with pancreatic disease. Digestion 1978;17:365-369.

GASTROENTEROLOGY

Vol. 99. No. 1

16. Piubello W, Vantini I, Scuro LA, Novelli P, Benini L, Brocco G, Cavallini G. Gastric secretion, gastroduodenal histological changes and serum gastrin in chronic alcoholic pancreatitis. Am J Gastroenterol1982;77:105-110. 17. Chey WY, Kusakcioglu 0, Dinoso Y. Gastric secretion in patients with chronic pancreatitis and in chronic alcoholics. Arch Intern Med 1968;122:399-402. 18. Dinoso YP, Chey WY, Braverman SP. Gastric secretion and gastric mucosal morphology in chronic alcoholics. Ann Intern Med 1972;130:715-719. 19. Szafran Z, Szafran H, Popiela T, Trompeter G. Coupled secretion of gastric lipase and pepsin in man following pentagastrin stimulation. Digestion 1978;18:310-318. 20. Moreau H, Sauniere JF, Gargourl Y, Pieroni G, Verger R, Sarles H. Human gastric lipase: variations induced by gastrointestinal hormones and by pathology. Stand J Gastroenterol1988:23:10441048. 21. De Negris SJ, Hamosh M, Kasbekar DK, Fink CS, Lee TC, Hamosh P. Secretion of human gastric lipase from dispersed gastric glands. Biochim Biophys Acta 1985;838:67-72.

Received February 16.1989. Accepted January 24.1990. Address requests for reprints to Dr. Jacques Moreau, Service des Maladies de 1’Appareil Digestif, CHU Rangueil, 31054 Toulouse Cedex, France. This study was supported in part by a grant from Conseil Regional Midi-Pryrenees (8800622). The authors thank all the members of the “laboratoire d’explorations fonctionnelles digestives” CHU Rangueil for their generous contribution to this work. We are grateful to Dr. J. Charlet, Professor Y. Plusquellec. and Dr. A. Taberly (Toulouse, France) for statistical analysis and to P. Sarfati for English editing assistance.