GASTROENTEROLOGY
1990;99:1576-1590
Human Pepsinogen A Isozymogen Patterns in Serum and Gastric Mucosa ANTONIE ZWIERS, BART CRUSIUS, GERARD PALS, AB J. M. DONKER, STEPHAN G. M. MEUWISSEN, and REINIER W. TEN KATE Department of Internal Medicine, Department of Gastroenterology, Genetics, Free University Hospital, Amsterdam, The Netherlands
The pepsinogen A isozymogen pattern in gastric mucosa is genetically determined and can be visualized in nondenaturating polyacrylamide gel electrophoresis of supernatants of sonified gastric mucosal biopsies by demonstrating ‘proteolytic activity after converting pepsinogen into pepsin by acid. Pepsinogen isozymogens are present in very low concentrations in the blood but can now be demonstrated in serum by a newly developed immunoblotting procedure. This study investigated whether the serum pepsinogen A isozymogen pattern adequately reflects the pepsinogen A phenotype. Serum and gastric mucosal pepsinogen A isozymogen patterns were compared in 72 subjects from the routine endoscopy program. A close correlation was found between the relative intensities of the pepsinogen A isozymogens in the serum and the gastric mucosal patterns. Increasing the pepsinogen A release into the circulation by oral omeprazole did not affect the pepsinogen A patterns in the blood. It is concluded that the serum pepsinogen A pattern reflects the pepsinogen A phenotype in humans. In addition, no preferential release of a pepsinogen A isozymogen into the circulation was observed. Thus, immunoblotting of serum provides a new and reliable tool to study pepsinogen genetics in humans. Because a relationship was previously shown between specific pepsinogen A phenotypes and gastric malignancies in humans, serum pepsinogen A patterns may provide a tool to detect subjects who are at risk of gastric cancers. epsinogens are proteins synthesized by the gastric mucosa. They are the precursors of the gastric pepsins (~2). By polyacrylamide gel electrophoresis (PAGE) of sonified human gastric mucosa biopsies, it has been shown that seven pepsinogen isozymogens are present in humans. They are numbered Pgl-Pg7
P
and Institute of Human
in order of decreasing anodal mobility (3). Biochemically they can be divided into pepsinogen A (PGA), isozymogens Pgl-5, and pepsinogen C (PGC), isozymogens Pg6-7 (4). Pepsinogen A isozymogens are the precursors of pepsin A (EC 3.4.23.1); PGC isozymogens are the precursors of pepsin C (EC 3.4.23.3) (1). Pepsinogen A isozymogen patterns have been shown to be genetically determined (5,6), and different PGA phenotypes are determined according to the relative intensities of the PGA isozymogens in PAGE. Pepsinogen patterns are visualized after PAGE by demonstrating the presence of proteolytic activity in the gels after incubation with a hydrochloric acidcontaining hemoglobin solution (activity staining) (6). High PGA concentrations in urine also allow visualization by activity staining of PGA isozymogen patterns, but PGA concentrations in serum are too low for this staining method. It has been shown that the PGA patterns in urine often do not match those in gastric mucosa (7). This difference has been ascribed to a different reabsorption of the various PGA isozymogens by the renal tubules (8). Recently, we developed an immunoblot technique sensitive enough to visualize PGA isozymogen patterns in serum (9). In this study we have investigated whether serum PGA patterns reflect the gastric PGA isozymogen patterns. Materials and Methods Patients A blood sample and a gastric mucosa specimen from the greater curvature were obtained from each of 72 patients (42 male, 30 female;
mean
age, 52 years:
range,
23-82 years)
Abbreviations used in this paper: BSA, bovine serum albumin; PAGE, polyacrylamide gel electrophoresis; PGA, pepsinogen A. o 1999 by the American Gastroenterological Association 001s-5095/90/$3.00
December 1990
PEPSJNOGEN A PATTERNS IN SERUM AND GASTRIC MUCOSA
during routine endoscopy at the Department of Gastroenterology. Endoscopy was performed with an Olympus endoscope (Gif (2, Olympus Corporation of America, New Hyde Park, NY), and a standard-grasp forceps was used. In 35 subjects, endoscopy showed no gross abnormalities. Eight subjects had gastritis, and 11 had active duodenal ulcers. Severe erosive esophagitis was found in 13 subjects, and Barrett’s esophagus was diagnosed in four of them. Two subjects proved to have benign and three to have malignant gastric ulcers. The release of pepsinogen from the gastric mucosa into the systemic circulation can be increased by oral administration of omeprazole (10). From one additional subject a total of nine serum samples were obtained within 48 hours before and during oral administration of omeprazole, 40 mg once daily, for 2 days. Informed consent was obtained from each subject. The protocol had been approved by the Ethical Committee of the Free University Hospital. Sample
Preparation
Samples were kept frozen at -20°C until analysis. Serum PGA concentrations were measured by enzymelinked immunosorbent assay (ELISA) as previously described (11). Both native serum and supernatant of sonified gastric mucosal biopsies were treated according to the routine protocol for pepsinogen electrophoresis in our laboratory (6). Samples were diluted in electrophoresis buffer up to a volume of 50 ILL.Subsequently, 25 PL of stacking gel buffer containing 40% (wt/vol] sucrose and 0.1 g/L bromophenol blue was added. Bromophenol blue was used to monitor both the albumin fraction and the electrophoresis front. Polyacrylamide
Gel Electrophoresis
Gels were prepared according to a slight modification of the method routinely used in our laboratory (6). Modifications consisted of an increase in buffer concentrations to reduce the effect of high protein content of the samples: stacking gel buffer, 51 mmol/L Tris, 49 mmol/L H,PO, (pH 5.5); separating gel buffer, 71 mmol/L Tris, 60 mmol/L HCl (pH 7.5); and electrode buffer, 8 mmol/L Tris, 30 mmol/L diethylbarbituric acid (pH 7.0). The gels were 1.5 mm thick. The length of the separating gel was 128 mm, and the length of the stacking gel was 7 mm. Electrophoresis was performed in the Protean vertical slab-gel system (Bio-Rad, Richmond, CA], with a LKB 2197 power supply (LKB, Bromma, Sweden) at 25 mA constant current per gel until the tracking dye was completely separated from the albuminbound bromophenol blue (approximately 1 hour]. Subsequently, electrophoresis was continued for 2 hours at 60 mA constant current per gel. During electrophoresis, the system was kept at a temperature of 4’C. Staining
for Pepsin
Activity
After electrophoresis of the gastric mucosal biopsies, the isozymogens were converted into pepsin by soaking the
1577
gel for 20 minutes at 37’C in a solution of 0.1 mol/L HCl and 2% bovine hemoglobin, followed by incubation in 0.1mol/L HCl for 40 minutes at 37OC. After incubation, the gels were stained for protein during 1 hour in a solution containing 0.2% Coomassie brilliant blue R250 in 10% (vol/vol) acetic acid and 25% (vol/vol] isopropanol. Destaining was done in 10% acetic acid.
Immunoblotting After PAGE of the serum samples, protein blotting was performed with the Trans-blot system and power supply (model 250/2.5; Bio-Rad, Richmond, CA] (12). Protein transfer was performed at 70 V constant voltage for 1 hour. Proteins were blotted onto polyvinylidene difluoride membrane (Immobilon; Millipore Corp., Bedford, MA) (13). After the blotting was completed, the membrane was immediately incubated for 1 hour at 37°C in “blocking” solution [5% (wt/vol] bovine serum albumin (BSA), 0.9% (wt/vol) NaCl in 20 mmol/L Tris-HCI, pH 7.41 with gentle agitation. The membrane was washed in 0.1% BSA, 0.9% NaCl, 20 mmol/L Tris-HCI, pH 7.4, three times for 5 minutes each. The blot was incubated overnight at room temperature with goat anti-pepsinogen antibodies at 10 &mL in 25 mL incubation buffer [l% BSA, 0.05% [vol/vol) Tween-20,0.9% NaCl, 20 mmol/L Tris-HCI, pH 7.41. Incubation was performed with PGA antibodies. Affinity-purified antibodies against human pepsinogens were obtained from PGAimmunized goats as described elsewhere (14). To eliminate possible cross-reactivity, the buffer was supplemented with 10% [vol/vol) native human plasma (pepsinogens in this plasma were destroyed first by heating the plasma for 45 minutes at 57’C). The blot was then washed as described above and incubated for 2 hours at room temperature with biotin-labeled rabbit anti-goat immunoglobulin G (IgG) (l/ 1000 dilution of the stock solution in incubation buffer: Sigma Chemical Co., St. Louis, MO). After another wash cycle, the blot was incubated with extravidin alkaline phosphatase conjugate (l/l000 dilution of the stock solution; Sigma Chemical Co.] for 2 hours at room temperature in incubation buffer. After washing as described above, the blot was placed in enzyme-substrate buffer to develop color. The enzyme-substrate buffer consisted of 100 mmol/L TrisHCI (pH 9.55), 100 mmol/L NaCl, 2 mmol/L MgCl,, 0.03 mmol/L phenazine methosulfate. 0.3 mmol/L 5-bromo-4chloro-3-indolyl phosphate p-toluidine salt, and 0.2 mmol/L nitroblue tetrazolium in high-performance liquid chromatography-grade water. When the incubation was completed (lo-15 minutes), the blot was rinsed in water and dried at 37°C.
Quantitation
of the lsozymogen
Patterns
The gels and blots were photographed with a Polaroid MP-4 Land Camera with Polaroid type 665 black and white instant pack film (Polaroid Corp., Cambridge, MA). For the activity-stained gels, a yellow filter was used. Semiquantitative analysis of the PGA isozymogens was performed by densitometric scanning of the Polaroid nega-
GASTROENTEROLOGY
1578 ZWIERS ET AL.
tives with the LKB 2202 Ultrascan laser densitometer (LKB, Bromma, Sweden) (12). The area under the curve of each PGA isozymogen was calculated with the LKB 2190-001 gelscan qrogram and expressed as percentage of the total area unuer the curve. Results A wide range of serum levels was found: 8-341 pg/L (mean, 76.1; SD, 53.0). In all subjects, adequate visualization of PGA isozymogen patterns in native serum was obtained, and patterns in gastric mucosa were identical to those in serum but not to those in urine (Figure 1). The relative intensity of each individual PGA isozymogen in serum was plotted against the corresponding value in gastric mucosa. An excellent linear correlation was obtained between the relative intensities of each individual PGA isozymogen in gastric mucosa and serum with correlation coefficients for Pg2, Pg3, Pg4, and Pg5 of 0.8068.0.8636.0.8598, and 0.9414, respectively (P -C0.001) (Figure 21.This correlation was present irrespective of the endoscopic diagnosis. There was no correlation between the intensity of an isozymogen in gastric mucosa or in serum and the total serum PGA concentration. Administration of omeprazole in one subject resulted in an increase in serum level from 51 to 101 pg/L. Serum PGA patterns remained unaffected [Figure 3).
A
Pg 7: 5: 4: 3: 2:
Pg 5: 4: 3: 2:
29 50 21 0
26 53 21 0
10 56 34 0
Figure 1. Gastric mucosal PGA patterns (A) obtained with activation staining (lanes 1 and 4) and the corresponding serum PGA patterns obtained by immunoblot (lanes 2 and 5) in two subjects. Corresponding urine PGA patterns ace shown in lane 3 and 6. Densitometric scans of the PGA patterns (B) are shown, as well as the calculated relative intensities of the PGA isozymogens (C).
Vol. 99, No. 6
Discussion In the past, PGA phenotype was assessed by visualization of PGA isozymogen patterns in gastric mucosal biopsy specimens. This was achieved by PAGE of supernatants of sonified gastric mucosal biopsies and subsequent demonstration of proteolytic activity in the gels after conversion of the pepsinogen isozymogens into pepsin by acidification. This technique requires the presence of a considerable amount of PGA. Serum levels of PGA are too low to be visualized in native serum by activity staining, and the use of larger volumes of serum is prohibited by the presence of large amounts of albumin because albumin migrates along with the pepsinogens due to its strong negative charge. In a previous study (9), we have shown that the immunoblotting technique as used in the present study is at least 50 times more sensitive than the activity staining and enabled us to visualize PGA patterns in serum. Moreover, we have shown that PGA isozymogen patterns obtained with the immunoblot technique after PAGE from sonified gastric mucosal biopsies or urine are identical to those obtained with activity staining. The present study shows that PGA patterns obtained by immunoblot in serum are identical to those in the gastric mucosa over the whole range of serum concentrations. Pepsinogen A phenotypes are determined by the relative intensities of the major PGA isozymogens Pg2-Pg5. We did not find any correlation between intensity of one of the PGA isozymogens in either gastric mucosa or serum and the serum concentration of total PGA. This confirms previous observations that no correlation exists between a PGA phenotype and the serum PGA level (15). In addition, oral omeprazole did not affect serum PGA patterns. Therefore, our results suggest that there is no preferential release of a PGA isozymogen into the circulation by the gastric mucosa. Moreover, serum PGA patterns are not affected by an increased release of PGA into the circulation by the gastric mucosa. Pepsinogen A phenotypes have attracted the attention of clinicians for several reasons. The familial occurrence of peptic ulcers, to which pepsin contributes, led to the investigation of the genetics of the PGA isozymogens (16,171. Our results suggest that determination of serum PGA patterns is a reliable, noninvasive method to establish human PGA phenotype that provides a new tool to investigate the genetics of the pepsinogens in humans. An association between PGA isozymogen patterns in gastric mucosa with an intense Pg5 band and malignant or premalignant lesions of the gastric mucosa has been reported (18-21). For obvious reasons, gastric mucosal biopsy specimens cannot be obtained on a large scale for genetic studies. Urinary patterns are
December
PEPSINOGEN
1990
A PATTERNS IN SERUM AND GASTRIC MUCOSA
1579
pg4 FO.8598
pg2
r-0.8068
bp
E 2 I
01 0
20
40
80
60
gastric
mucosa
1
0
100
20
40 gastric
%
60
80
100
80
100
mucoaa %
pg3
r=0.8636 100
80
80
#
.
60
5 !fl al
.
..’
#
.
.
._-“’
60
5
. .--. *: .
2
40
-
*
i
40
*
.
:
20
c
20
0
40
gastric
60
80
100
mucoaa %
0
20
40
60
gastric
mucoaa
%
Figure 2. The relations between the dative intensities (expressed as percent of the total intensity) of the PGA isozymogens Pg2, Pg2, Pg4, and Pg5 in gastric mucosa (x-axis) and in serum (y-axis) from 72 subjects.
not adequate for that purpose either (7). The determination of PGA phenotype by immunoblot technique of native serum is readily applicable in large population studies and might therefore contribute to early detection of subjects at risk for gastric malignancies.
A
Pg
5:. 4: 3:
12
3
4
5
6
7
6
gr
B
C
% Pg
%
%
%
%
%
%
%
%
5:
56
62
62
56
63
56
60
60
62
4:
25
21
23
25
25
27
26
27
25
3:
17
17
15
17
12
15
14
13
14
Figure 2. Immunoblots of serum PGA patterns (A) of one subject after a gradual increase of the serum PGA concentration by oral adminfstration of omeprazole from 51 fig/L (lane 1) to 101 fig/L (hue 9). Densitometric scans of the PGA patterns (lt) are shown, as well as the calculated relative intensities of the PGA isozymogens (c).
References Foltmann B. Purification, structure and activation of pepsinogens. Prog Clin Biol Res 1985;173:1-13. Samloff IM. Pepsinogens I and II: purification from gastric mucosa and radioimmunoassay in serum. Gastroenterology 1982;82:26-33. Samloff IM. Slow moving protease and the seven pepsinogens. Electrophoretic demonstration of the existence of eight proteolytic fractions in human gastric mucosa. Gastroenterology 1969;57:659-669. 4. Eriksson AW, Foltmann B, Frants RR, Gedde-Dahl T Jr, Samloff IM, Taggart RTh. Nomenclature Committee report. Prog Clin Biol Res 1985:173:23-30. 5. Samloff IM, Liebman WM, Glober GA, More JO, Indra D. Population studies of pepsinogen polymorphism. Am J Hum Genet 1973;25:178-180. 6. Frants RR, Pronk JC, Pals G. Dbfize J. Westerveld BD, Meuwissen SGM, Kreuning J, Eriksson AW. Genetics of urinary pepsinogen: a new hypothesis. Hum Gen 1984;65:385-390.
1580 ZWIERS ET AL.
7. Pals G, Westerveld BD, DBfize J. Prank JC, Brand H, Flipse M, Verwey C, Meuwissen SCM, Eriksson AW. Discrepancies between gastric mucosal and urinary pepsinogen A patterns and in vitro synthesis and secretion of human pepsinogen. Dig Dis Sci 1988;33:135-143. 8. ten Kate RW. Pals G, Pronk lC, Bank RA. Eriksson AW. Donker AJM. Meuwissen SGM. Renal handling of pepsinogens A and C in man. Clin Sci 1988;75:649-654. 9. Zwiers A, Toonstra C, Pals G, Donker AJM, Meuwissen SGM, ten Kate RW. Immunoblot technique to visualize serum pepsinogen A isozymogen patterns. J Clin Path01 (in press]. 10. ten Kate RW, Tuynman HARE, Festen HPM. Pals G, Meuwissen SGM. The effect of high dose omeprazole on gastric pepsin secretion and serum pepsinogen levels in man. Eur J Clin Pharmacol1988;35:173-170. 11.Pals G, RBslnen V, Frants RR, Kostense PJ, Meuwissen SGM, Eriksson AW. Enzyme linked immunosorbent assay and radioimmunoassay of serum pepsinogen A. Stand J Clin Lab Invest 1987;47:29-33. 1.2. Towbin H, Staehelin Th. Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Nat1 Acad Sci USA 1979;76:4350-4354. 13.Pluskal MG, Przekop MB, Kavonian MR, Vecoli C, Hicks DA. ImmobilonQ PVDF transfer membrane: a new membrane substrate for western blotting of proteins. Biotechniques 1988;4:272282. 14.Pals G, Westerveld BD, Pronk JC. Frants RR, Defize J, Kostense PJ. Biemond I, Meuwissen SGM, Eriksson AW. Enzyme linked immunosorbent assay of serum pepsinogens A and C in normal controls and patients with superficial or atrophic gastritis. In: Pals G, Westerveld BD. Human pepsinogens: genetic and clinical aspects. Ph.D. Thesis, Free University, Amsterdam, 1986. pp 21-30. 15.Pals G, Westerveld BD, Defize J. Pronk JC, Biemond I. Meuwis-
GASTROENTEROLOGY
Vol. 99, No. 6
sen SGM, Eriksson AW. Relations between serum pepsinogen levels, pepsinogen phenotypes, ABO blood groups, age and sex in blood donors. Ann Hum Biol1985;12:403-411, 16.Samloff IM, Townes PL. Pepsinogens: genetic polymorphism in man. Science 1970;168:144-145. 17.Ellis A, McConnell RB. Duodenal ulcer and urinary pepsinogen phenotypes. Gastroenterology 1982;83:1261-1263. 18.Whitecross DP, Armstrong C, Clarke AD, Piper DW. The pepsinogens of human gastric mucosa. Gut 1973;14:850-855. 19.Ellis A, Hughes S. McConnell RB. Gastric neoplasms and pepsinogen phenotypes. Br J Cancer 1982;46:289-290. 20.DBfize J, Derodra JK, Riddell RH, Hunt RH. Changes in rat and human pepsinogen phenotypes induced by N’-methyl-N’-nitroN-nitrosoguanidine. Cancer 1988;62:1958-1961. 21.Westerveld BD, Pals G, Defize J, Pronk JC, Frants RR, Ooms EMC, Kreuning J. Eriksson AW. Meuwlssen SGM. Pepsinogen A polymorphism in gastric mucosa and urine, with special reference to patients withgastric cancer. Clin Genet 1986:30:108118.
Received April 12.1990. Accepted May 24,199O. Address requests for reprints to: Prof. Stephan G. M. Meuwissen, Department of Gastroenterology, Free University Hospital, P.O. Box 7057,1007 MB Amsterdam, The Netherlands. This study was supported by grants C87 642 and C87 719 of the Dutch Kidney Foundation (Nier Stichtmg Nederland). Presented as a poster at the American Gastroenterological Association, San Antonio, Texas, May 13-16, 1990, and at the World Congress of Gastroenterology, Sydney, Australia, August 1990. The authors acknowledge the help of the nursing staff of the endoscopy unit of the department of Gastroenterology of the Free University Hospital and the technical assistance of A. M. Tolk and T. A. Abrahami de Melverda.
1990;99:1576-1590
Human Pepsinogen A Isozymogen Patterns in Serum and Gastric Mucosa ANTONIE ZWIERS, BART CRUSIUS, GERARD PALS, AB J. M. DONKER, STEPHAN G. M. MEUWISSEN, and REINIER W. TEN KATE Department of Internal Medicine, Department of Gastroenterology, Genetics, Free University Hospital, Amsterdam, The Netherlands
The pepsinogen A isozymogen pattern in gastric mucosa is genetically determined and can be visualized in nondenaturating polyacrylamide gel electrophoresis of supernatants of sonified gastric mucosal biopsies by demonstrating ‘proteolytic activity after converting pepsinogen into pepsin by acid. Pepsinogen isozymogens are present in very low concentrations in the blood but can now be demonstrated in serum by a newly developed immunoblotting procedure. This study investigated whether the serum pepsinogen A isozymogen pattern adequately reflects the pepsinogen A phenotype. Serum and gastric mucosal pepsinogen A isozymogen patterns were compared in 72 subjects from the routine endoscopy program. A close correlation was found between the relative intensities of the pepsinogen A isozymogens in the serum and the gastric mucosal patterns. Increasing the pepsinogen A release into the circulation by oral omeprazole did not affect the pepsinogen A patterns in the blood. It is concluded that the serum pepsinogen A pattern reflects the pepsinogen A phenotype in humans. In addition, no preferential release of a pepsinogen A isozymogen into the circulation was observed. Thus, immunoblotting of serum provides a new and reliable tool to study pepsinogen genetics in humans. Because a relationship was previously shown between specific pepsinogen A phenotypes and gastric malignancies in humans, serum pepsinogen A patterns may provide a tool to detect subjects who are at risk of gastric cancers. epsinogens are proteins synthesized by the gastric mucosa. They are the precursors of the gastric pepsins (~2). By polyacrylamide gel electrophoresis (PAGE) of sonified human gastric mucosa biopsies, it has been shown that seven pepsinogen isozymogens are present in humans. They are numbered Pgl-Pg7
P
and Institute of Human
in order of decreasing anodal mobility (3). Biochemically they can be divided into pepsinogen A (PGA), isozymogens Pgl-5, and pepsinogen C (PGC), isozymogens Pg6-7 (4). Pepsinogen A isozymogens are the precursors of pepsin A (EC 3.4.23.1); PGC isozymogens are the precursors of pepsin C (EC 3.4.23.3) (1). Pepsinogen A isozymogen patterns have been shown to be genetically determined (5,6), and different PGA phenotypes are determined according to the relative intensities of the PGA isozymogens in PAGE. Pepsinogen patterns are visualized after PAGE by demonstrating the presence of proteolytic activity in the gels after incubation with a hydrochloric acidcontaining hemoglobin solution (activity staining) (6). High PGA concentrations in urine also allow visualization by activity staining of PGA isozymogen patterns, but PGA concentrations in serum are too low for this staining method. It has been shown that the PGA patterns in urine often do not match those in gastric mucosa (7). This difference has been ascribed to a different reabsorption of the various PGA isozymogens by the renal tubules (8). Recently, we developed an immunoblot technique sensitive enough to visualize PGA isozymogen patterns in serum (9). In this study we have investigated whether serum PGA patterns reflect the gastric PGA isozymogen patterns. Materials and Methods Patients A blood sample and a gastric mucosa specimen from the greater curvature were obtained from each of 72 patients (42 male, 30 female;
mean
age, 52 years:
range,
23-82 years)
Abbreviations used in this paper: BSA, bovine serum albumin; PAGE, polyacrylamide gel electrophoresis; PGA, pepsinogen A. o 1999 by the American Gastroenterological Association 001s-5095/90/$3.00
December 1990
PEPSJNOGEN A PATTERNS IN SERUM AND GASTRIC MUCOSA
during routine endoscopy at the Department of Gastroenterology. Endoscopy was performed with an Olympus endoscope (Gif (2, Olympus Corporation of America, New Hyde Park, NY), and a standard-grasp forceps was used. In 35 subjects, endoscopy showed no gross abnormalities. Eight subjects had gastritis, and 11 had active duodenal ulcers. Severe erosive esophagitis was found in 13 subjects, and Barrett’s esophagus was diagnosed in four of them. Two subjects proved to have benign and three to have malignant gastric ulcers. The release of pepsinogen from the gastric mucosa into the systemic circulation can be increased by oral administration of omeprazole (10). From one additional subject a total of nine serum samples were obtained within 48 hours before and during oral administration of omeprazole, 40 mg once daily, for 2 days. Informed consent was obtained from each subject. The protocol had been approved by the Ethical Committee of the Free University Hospital. Sample
Preparation
Samples were kept frozen at -20°C until analysis. Serum PGA concentrations were measured by enzymelinked immunosorbent assay (ELISA) as previously described (11). Both native serum and supernatant of sonified gastric mucosal biopsies were treated according to the routine protocol for pepsinogen electrophoresis in our laboratory (6). Samples were diluted in electrophoresis buffer up to a volume of 50 ILL.Subsequently, 25 PL of stacking gel buffer containing 40% (wt/vol] sucrose and 0.1 g/L bromophenol blue was added. Bromophenol blue was used to monitor both the albumin fraction and the electrophoresis front. Polyacrylamide
Gel Electrophoresis
Gels were prepared according to a slight modification of the method routinely used in our laboratory (6). Modifications consisted of an increase in buffer concentrations to reduce the effect of high protein content of the samples: stacking gel buffer, 51 mmol/L Tris, 49 mmol/L H,PO, (pH 5.5); separating gel buffer, 71 mmol/L Tris, 60 mmol/L HCl (pH 7.5); and electrode buffer, 8 mmol/L Tris, 30 mmol/L diethylbarbituric acid (pH 7.0). The gels were 1.5 mm thick. The length of the separating gel was 128 mm, and the length of the stacking gel was 7 mm. Electrophoresis was performed in the Protean vertical slab-gel system (Bio-Rad, Richmond, CA], with a LKB 2197 power supply (LKB, Bromma, Sweden) at 25 mA constant current per gel until the tracking dye was completely separated from the albuminbound bromophenol blue (approximately 1 hour]. Subsequently, electrophoresis was continued for 2 hours at 60 mA constant current per gel. During electrophoresis, the system was kept at a temperature of 4’C. Staining
for Pepsin
Activity
After electrophoresis of the gastric mucosal biopsies, the isozymogens were converted into pepsin by soaking the
1577
gel for 20 minutes at 37’C in a solution of 0.1 mol/L HCl and 2% bovine hemoglobin, followed by incubation in 0.1mol/L HCl for 40 minutes at 37OC. After incubation, the gels were stained for protein during 1 hour in a solution containing 0.2% Coomassie brilliant blue R250 in 10% (vol/vol) acetic acid and 25% (vol/vol] isopropanol. Destaining was done in 10% acetic acid.
Immunoblotting After PAGE of the serum samples, protein blotting was performed with the Trans-blot system and power supply (model 250/2.5; Bio-Rad, Richmond, CA] (12). Protein transfer was performed at 70 V constant voltage for 1 hour. Proteins were blotted onto polyvinylidene difluoride membrane (Immobilon; Millipore Corp., Bedford, MA) (13). After the blotting was completed, the membrane was immediately incubated for 1 hour at 37°C in “blocking” solution [5% (wt/vol] bovine serum albumin (BSA), 0.9% (wt/vol) NaCl in 20 mmol/L Tris-HCI, pH 7.41 with gentle agitation. The membrane was washed in 0.1% BSA, 0.9% NaCl, 20 mmol/L Tris-HCI, pH 7.4, three times for 5 minutes each. The blot was incubated overnight at room temperature with goat anti-pepsinogen antibodies at 10 &mL in 25 mL incubation buffer [l% BSA, 0.05% [vol/vol) Tween-20,0.9% NaCl, 20 mmol/L Tris-HCI, pH 7.41. Incubation was performed with PGA antibodies. Affinity-purified antibodies against human pepsinogens were obtained from PGAimmunized goats as described elsewhere (14). To eliminate possible cross-reactivity, the buffer was supplemented with 10% [vol/vol) native human plasma (pepsinogens in this plasma were destroyed first by heating the plasma for 45 minutes at 57’C). The blot was then washed as described above and incubated for 2 hours at room temperature with biotin-labeled rabbit anti-goat immunoglobulin G (IgG) (l/ 1000 dilution of the stock solution in incubation buffer: Sigma Chemical Co., St. Louis, MO). After another wash cycle, the blot was incubated with extravidin alkaline phosphatase conjugate (l/l000 dilution of the stock solution; Sigma Chemical Co.] for 2 hours at room temperature in incubation buffer. After washing as described above, the blot was placed in enzyme-substrate buffer to develop color. The enzyme-substrate buffer consisted of 100 mmol/L TrisHCI (pH 9.55), 100 mmol/L NaCl, 2 mmol/L MgCl,, 0.03 mmol/L phenazine methosulfate. 0.3 mmol/L 5-bromo-4chloro-3-indolyl phosphate p-toluidine salt, and 0.2 mmol/L nitroblue tetrazolium in high-performance liquid chromatography-grade water. When the incubation was completed (lo-15 minutes), the blot was rinsed in water and dried at 37°C.
Quantitation
of the lsozymogen
Patterns
The gels and blots were photographed with a Polaroid MP-4 Land Camera with Polaroid type 665 black and white instant pack film (Polaroid Corp., Cambridge, MA). For the activity-stained gels, a yellow filter was used. Semiquantitative analysis of the PGA isozymogens was performed by densitometric scanning of the Polaroid nega-
GASTROENTEROLOGY
1578 ZWIERS ET AL.
tives with the LKB 2202 Ultrascan laser densitometer (LKB, Bromma, Sweden) (12). The area under the curve of each PGA isozymogen was calculated with the LKB 2190-001 gelscan qrogram and expressed as percentage of the total area unuer the curve. Results A wide range of serum levels was found: 8-341 pg/L (mean, 76.1; SD, 53.0). In all subjects, adequate visualization of PGA isozymogen patterns in native serum was obtained, and patterns in gastric mucosa were identical to those in serum but not to those in urine (Figure 1). The relative intensity of each individual PGA isozymogen in serum was plotted against the corresponding value in gastric mucosa. An excellent linear correlation was obtained between the relative intensities of each individual PGA isozymogen in gastric mucosa and serum with correlation coefficients for Pg2, Pg3, Pg4, and Pg5 of 0.8068.0.8636.0.8598, and 0.9414, respectively (P -C0.001) (Figure 21.This correlation was present irrespective of the endoscopic diagnosis. There was no correlation between the intensity of an isozymogen in gastric mucosa or in serum and the total serum PGA concentration. Administration of omeprazole in one subject resulted in an increase in serum level from 51 to 101 pg/L. Serum PGA patterns remained unaffected [Figure 3).
A
Pg 7: 5: 4: 3: 2:
Pg 5: 4: 3: 2:
29 50 21 0
26 53 21 0
10 56 34 0
Figure 1. Gastric mucosal PGA patterns (A) obtained with activation staining (lanes 1 and 4) and the corresponding serum PGA patterns obtained by immunoblot (lanes 2 and 5) in two subjects. Corresponding urine PGA patterns ace shown in lane 3 and 6. Densitometric scans of the PGA patterns (B) are shown, as well as the calculated relative intensities of the PGA isozymogens (C).
Vol. 99, No. 6
Discussion In the past, PGA phenotype was assessed by visualization of PGA isozymogen patterns in gastric mucosal biopsy specimens. This was achieved by PAGE of supernatants of sonified gastric mucosal biopsies and subsequent demonstration of proteolytic activity in the gels after conversion of the pepsinogen isozymogens into pepsin by acidification. This technique requires the presence of a considerable amount of PGA. Serum levels of PGA are too low to be visualized in native serum by activity staining, and the use of larger volumes of serum is prohibited by the presence of large amounts of albumin because albumin migrates along with the pepsinogens due to its strong negative charge. In a previous study (9), we have shown that the immunoblotting technique as used in the present study is at least 50 times more sensitive than the activity staining and enabled us to visualize PGA patterns in serum. Moreover, we have shown that PGA isozymogen patterns obtained with the immunoblot technique after PAGE from sonified gastric mucosal biopsies or urine are identical to those obtained with activity staining. The present study shows that PGA patterns obtained by immunoblot in serum are identical to those in the gastric mucosa over the whole range of serum concentrations. Pepsinogen A phenotypes are determined by the relative intensities of the major PGA isozymogens Pg2-Pg5. We did not find any correlation between intensity of one of the PGA isozymogens in either gastric mucosa or serum and the serum concentration of total PGA. This confirms previous observations that no correlation exists between a PGA phenotype and the serum PGA level (15). In addition, oral omeprazole did not affect serum PGA patterns. Therefore, our results suggest that there is no preferential release of a PGA isozymogen into the circulation by the gastric mucosa. Moreover, serum PGA patterns are not affected by an increased release of PGA into the circulation by the gastric mucosa. Pepsinogen A phenotypes have attracted the attention of clinicians for several reasons. The familial occurrence of peptic ulcers, to which pepsin contributes, led to the investigation of the genetics of the PGA isozymogens (16,171. Our results suggest that determination of serum PGA patterns is a reliable, noninvasive method to establish human PGA phenotype that provides a new tool to investigate the genetics of the pepsinogens in humans. An association between PGA isozymogen patterns in gastric mucosa with an intense Pg5 band and malignant or premalignant lesions of the gastric mucosa has been reported (18-21). For obvious reasons, gastric mucosal biopsy specimens cannot be obtained on a large scale for genetic studies. Urinary patterns are
December
PEPSINOGEN
1990
A PATTERNS IN SERUM AND GASTRIC MUCOSA
1579
pg4 FO.8598
pg2
r-0.8068
bp
E 2 I
01 0
20
40
80
60
gastric
mucosa
1
0
100
20
40 gastric
%
60
80
100
80
100
mucoaa %
pg3
r=0.8636 100
80
80
#
.
60
5 !fl al
.
..’
#
.
.
._-“’
60
5
. .--. *: .
2
40
-
*
i
40
*
.
:
20
c
20
0
40
gastric
60
80
100
mucoaa %
0
20
40
60
gastric
mucoaa
%
Figure 2. The relations between the dative intensities (expressed as percent of the total intensity) of the PGA isozymogens Pg2, Pg2, Pg4, and Pg5 in gastric mucosa (x-axis) and in serum (y-axis) from 72 subjects.
not adequate for that purpose either (7). The determination of PGA phenotype by immunoblot technique of native serum is readily applicable in large population studies and might therefore contribute to early detection of subjects at risk for gastric malignancies.
A
Pg
5:. 4: 3:
12
3
4
5
6
7
6
gr
B
C
% Pg
%
%
%
%
%
%
%
%
5:
56
62
62
56
63
56
60
60
62
4:
25
21
23
25
25
27
26
27
25
3:
17
17
15
17
12
15
14
13
14
Figure 2. Immunoblots of serum PGA patterns (A) of one subject after a gradual increase of the serum PGA concentration by oral adminfstration of omeprazole from 51 fig/L (lane 1) to 101 fig/L (hue 9). Densitometric scans of the PGA patterns (lt) are shown, as well as the calculated relative intensities of the PGA isozymogens (c).
References Foltmann B. Purification, structure and activation of pepsinogens. Prog Clin Biol Res 1985;173:1-13. Samloff IM. Pepsinogens I and II: purification from gastric mucosa and radioimmunoassay in serum. Gastroenterology 1982;82:26-33. Samloff IM. Slow moving protease and the seven pepsinogens. Electrophoretic demonstration of the existence of eight proteolytic fractions in human gastric mucosa. Gastroenterology 1969;57:659-669. 4. Eriksson AW, Foltmann B, Frants RR, Gedde-Dahl T Jr, Samloff IM, Taggart RTh. Nomenclature Committee report. Prog Clin Biol Res 1985:173:23-30. 5. Samloff IM, Liebman WM, Glober GA, More JO, Indra D. Population studies of pepsinogen polymorphism. Am J Hum Genet 1973;25:178-180. 6. Frants RR, Pronk JC, Pals G. Dbfize J. Westerveld BD, Meuwissen SGM, Kreuning J, Eriksson AW. Genetics of urinary pepsinogen: a new hypothesis. Hum Gen 1984;65:385-390.
1580 ZWIERS ET AL.
7. Pals G, Westerveld BD, DBfize J. Prank JC, Brand H, Flipse M, Verwey C, Meuwissen SCM, Eriksson AW. Discrepancies between gastric mucosal and urinary pepsinogen A patterns and in vitro synthesis and secretion of human pepsinogen. Dig Dis Sci 1988;33:135-143. 8. ten Kate RW. Pals G, Pronk lC, Bank RA. Eriksson AW. Donker AJM. Meuwissen SGM. Renal handling of pepsinogens A and C in man. Clin Sci 1988;75:649-654. 9. Zwiers A, Toonstra C, Pals G, Donker AJM, Meuwissen SGM, ten Kate RW. Immunoblot technique to visualize serum pepsinogen A isozymogen patterns. J Clin Path01 (in press]. 10. ten Kate RW, Tuynman HARE, Festen HPM. Pals G, Meuwissen SGM. The effect of high dose omeprazole on gastric pepsin secretion and serum pepsinogen levels in man. Eur J Clin Pharmacol1988;35:173-170. 11.Pals G, RBslnen V, Frants RR, Kostense PJ, Meuwissen SGM, Eriksson AW. Enzyme linked immunosorbent assay and radioimmunoassay of serum pepsinogen A. Stand J Clin Lab Invest 1987;47:29-33. 1.2. Towbin H, Staehelin Th. Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Nat1 Acad Sci USA 1979;76:4350-4354. 13.Pluskal MG, Przekop MB, Kavonian MR, Vecoli C, Hicks DA. ImmobilonQ PVDF transfer membrane: a new membrane substrate for western blotting of proteins. Biotechniques 1988;4:272282. 14.Pals G, Westerveld BD, Pronk JC. Frants RR, Defize J, Kostense PJ. Biemond I, Meuwissen SGM, Eriksson AW. Enzyme linked immunosorbent assay of serum pepsinogens A and C in normal controls and patients with superficial or atrophic gastritis. In: Pals G, Westerveld BD. Human pepsinogens: genetic and clinical aspects. Ph.D. Thesis, Free University, Amsterdam, 1986. pp 21-30. 15.Pals G, Westerveld BD, Defize J. Pronk JC, Biemond I. Meuwis-
GASTROENTEROLOGY
Vol. 99, No. 6
sen SGM, Eriksson AW. Relations between serum pepsinogen levels, pepsinogen phenotypes, ABO blood groups, age and sex in blood donors. Ann Hum Biol1985;12:403-411, 16.Samloff IM, Townes PL. Pepsinogens: genetic polymorphism in man. Science 1970;168:144-145. 17.Ellis A, McConnell RB. Duodenal ulcer and urinary pepsinogen phenotypes. Gastroenterology 1982;83:1261-1263. 18.Whitecross DP, Armstrong C, Clarke AD, Piper DW. The pepsinogens of human gastric mucosa. Gut 1973;14:850-855. 19.Ellis A, Hughes S. McConnell RB. Gastric neoplasms and pepsinogen phenotypes. Br J Cancer 1982;46:289-290. 20.DBfize J, Derodra JK, Riddell RH, Hunt RH. Changes in rat and human pepsinogen phenotypes induced by N’-methyl-N’-nitroN-nitrosoguanidine. Cancer 1988;62:1958-1961. 21.Westerveld BD, Pals G, Defize J, Pronk JC, Frants RR, Ooms EMC, Kreuning J. Eriksson AW. Meuwlssen SGM. Pepsinogen A polymorphism in gastric mucosa and urine, with special reference to patients withgastric cancer. Clin Genet 1986:30:108118.
Received April 12.1990. Accepted May 24,199O. Address requests for reprints to: Prof. Stephan G. M. Meuwissen, Department of Gastroenterology, Free University Hospital, P.O. Box 7057,1007 MB Amsterdam, The Netherlands. This study was supported by grants C87 642 and C87 719 of the Dutch Kidney Foundation (Nier Stichtmg Nederland). Presented as a poster at the American Gastroenterological Association, San Antonio, Texas, May 13-16, 1990, and at the World Congress of Gastroenterology, Sydney, Australia, August 1990. The authors acknowledge the help of the nursing staff of the endoscopy unit of the department of Gastroenterology of the Free University Hospital and the technical assistance of A. M. Tolk and T. A. Abrahami de Melverda.
