Radioimmunoassay of Group I Pepsinogens in Serum

Vol. 66, No.4

GASTROENTEROLOGY 66: 494-R02, 1974 Copyright © 1974 by The Williams & Wilkins Co.

Printed in U.S.A.

RADIOIMMUNOASSAY OF GROUP I PEPSINOGENS IN SERUM L

MICHAEL SAMLOFF, M.D., AND WILLIAM M. LIEBMAN, M.D.

Division of Gastroenterolo!iY, Department of Medicine, Harbor General Hospital, Torrance and the University of California School of Medicine, Los Angeles, California

A radioimmunoassay for the group I pepsinogens (PG 1) in human serum has been developed. The standard PG I was isolated from human urine, and, when labeled with 1251, was shown to be immunochemically homogeneous by radioimmunoelectrophoresis and radioimmunoprecipitation. The antiserum was obtained from rabbits immunized with partially purified PG 1. Standard PG I and serum give parallel inhibition curves against the labeled antigen. The sensitivity of the assay is in the order of 2 ng per ml of serum. The reproducibility of the determinations was demonstrated. PG I in serum is stable when stored at -20 C for at least 7 months. The mean (±sn) serum PG I levels in 69 healthy control subjects and in 71 hospitalized control patients were 102.7 ± 32.5 and 96.2 ± 41.9 ng per ml, respectively. In 14 patients with pernicious anemia and in 12 patients with total gastrectomy, mean serum PG I levels were 10.2 ± 6.4 and 3.8 ± 2.9 ng per ml, respectively. In 5 patients with Zollinger-Ellison syndrome, the mean level was 456.1 ± 187.3 ng per ml. Studies during the past decade have revealed that the gastric fundus, gastric antrum, and proximal duodenum contain multiple pepsinogens,1-4 that these are separable into two immunochemically distinct groups, 5, 6 namely group I (PG I) and group II (PG II), and that the two groups differ in their mucosal distribution and cellular origins. Both PG I and PG II are present in fundic mucosa, and both groups have been localized to the chief and muReceived September 19, 1973. Accepted October 30, 1973. Address requests for reprints to: Dr. 1. Michael Samloff, Harbor General Hospital, 1000 West Carson Street, Torrance, California 90509. This work was supported by Research Grant AM 13233 and Training Grant AM 05654 from the National Institutes of Health, Bethesda, Maryland. The authors thank Madeline Kursar and Violetta Dadufalza for their technical assistance, Victoria Adkins for her secretarial assistance, Dr. Jon Isenberg and Dr. Ed Passaro for many of the sera from patients with Zollinger-Ellison syndrome and total gastrectomy, and Dr. Leonard Goldberg for many of the sera from patients with pernicious anemia. 494

cous neck cells. 4, 7, 8 In contrast, only PG II is found in the gastric antrum and proximal duodenum, and only this group has been identified in the pyloric glands and in Brunner's glands. 4, 7. 8 Studies of the extragastric distribution of these groups have shown that both PG I and PG II are present in serum 9 but that only PG I is found in normal urine. 9 The estimation of the pepsinogens and the pepsins is conventionally performed by proteolytic assays. Numerous methods have been described, but the technique which has gained the widest acceptance is based on the estimation of peptides released during incubation of the sample with a protein substrate at acid pH.10-14 However, these assays are not specific for pepsinogen in that they detect other acidactive proteases which may be present in the sample. In addition, they are not specific for a particular group of pepsi no gens. The finding of multiple pepsinogens has stimulated interest in the development of assays for their separate determination. Methods which have been described in-

April 1974

495

RADIOIMMUNOASSA Y OF PC I IN SERUM

elude the use of specific synthetIc substrates, 15. 16 the estimation of proteolytic activity before and after alkaline inactivation of one group, 17. 18 and immunoassay using electrophoretic immunoprecipitation. 19 These assays have been applied to the estimation of the pepsins and pepsinogens in gastric juice and urine but not to PG I or PG II in serum. The mechanism by which the pepsinogens enter the circulation is not known, but their level in serum or plasma, as estimated by proteolytic assays, has been reported to reflect the secretory potential and morphologic status of the gastric mucosa, 20. 21 to be helpful in the diagnosis of pernicious anemia and duodenal ulcer, 22-25 and to have predictive value for the development of duodenal ulcer 26 -28 and gastric cancer.29 Based on the known limitation of PG I to the oxyntic gland area of the stomach, one would anticipate that the level of PG I in serum might better reflect the functional and morphologic status of the fundic glands than the total acid protease activity of serum. In this report, we describe the development of a competitive binding, double antibody radioimmunoassay for PG I and present initial results for serum PG I levels in control subjects and in patients with total gastrectomy, pernicious anemia, and Zollinger-Ellison syndrome.

Materials and Methods

ing the peak of proteolytic activity from six such filtrations were pooled, concentrated by pressure dialysis to 25 ml, reapplied to the column of Bio-Gel P-l00, and eluted in a similar manner. The fractions containing PG I were pooled, concentrated to 46 ml, dialyzed against 0.1 M phosphate buffer, pH 7.1, and applied to the top of a 2- by 20-cm column of diethylaminoethylcellulose which had been equilibrated with the same buffer. The column was then washed with 500 ml of the equilibration buffer, and PG I was eluted with 0.3 M NaCI in 0.25 M phosphate buffer, pH 7.1 (fig. 2). The fractions containing proteolytic activity were pooled, concentrated to 25 ml, dialyzed against 0.1 M phosphate buffer, pH 7.1, and rechromatographed on another 2- by 20-cm column of diethylaminoethylcellulose using 0.3 M NaCI in 0.1 M phosphate buffer, pH 7.1, to elute the PG I. The fractions containing PG I were concentrated to 6.2 ml,

I o.

B

40

1)

c:

30

£0.4

Q1

Z

0

(Ii

Q

~0.3

202

~0.2

'"~"-

z

~

a::

en0

m

10

0.1

«

10

20

30

40

50

FRACTION NO.

FIG. 1. Gel filtration of concentrated urine on Bio-Gel P-IOO.

Preparation of PC I A 30.6-liter pool of urine was obtained from two volunteers who were known to be pepsinogen phenotype A 30 and who, upon repeated testing, excreted only PG 1. 9 The following steps were carried out at 4 to 6 C in a cold room or in a refrigerated centrifuge. The urine was dialyzed against tap water, centrifuged at 13,200 X g for 30 min and concentrated 200-fold by pressure dialysis. After repeat centrifugation, the supernatant was dialyzed against 0.1 M phosphate buffer, pH 7.3, and applied in approximately 25-ml portions to the top of a 5- by 90-cm column of Bio-Gel P-lOO. The column was equilibrated and eluted with the same buffer (fig. 1). Eluate fractions (25 m!) were collected. The absorbance of these and subsequent eluate fractions was determined at 280 nm, and proteolytic activity was estimated by the radial diffusion assay. 13 The eluate fractions contain-

1)

f11

eO. 5

..

0.2

E

"

0

III N

40

POOL

i " " '\" J "" " ",

11

f11 30 ~

.,

Z

0

:,,''.,',

~ 01

z «

Gl

20

~

'"~... ~

III II:

0

10

UJ

III

«

~ "'OoOO()O
20

30

40

FRACTION NO.

FIG. 2. Chromatography of group I pepsinogens on diethylaminoethylcellulose. The arrow indicates a change in the elution buffer to 0.3 M NaC! in 0.25 M phosphate buffer.

496

Vol. 66. No . 4

SAMLOFF AND LIEBMAN

dialyzed against 0.005 M phosphate buffer. pH 6.8. and applied to a 1.2- by 21.5-cm column of hydroxylapatite which had been equilibrated with the same buffer. The column was then washed with 80 ml of the equilibration buffer followed by 200 ml of 0.025 M phosphate buffer, pH 6.8 (fig. 3). Three peaks of proteolytic activity were obtained, but this was not investigated further. The fractions comprising the peaks of proteolytic activity were pooled, concentrated. dialyzed against deionized water, and lyophilized.

•n ,, .'

: : j ~ ~ !: ~

1

02

~

Purification of !2'I-PG I On the day of radioiodination, from 80 to 90% of the radioactively labeled antigen could be precipitated by an excess of antibody to PG I. This decreased progressively with time, presumably because of radiation damage. Consequently, a portion of the I2SI_PG I was purified before each assay by chromatography on a 0.5by 5-cm column of diethylaminoethylcellulose. The column was equilibrated with 0.12 M phosphate buffer, pH 7.1. Elution with the same buffer resulted in a peak of radioactivity; approximately 5% of this peak could be precipitated by an excess of antibody to PG I. The second peak, eluted by adding 0. 3 M NaCI to the buffer, was used in the assay. From 75 to 80% of this peak was precipitable by excess antibody.

Antisera Antiserum to PG I (anti-PG 1). New Zealand white rabbits were immunized with partially purified PG 1 which had been obtained from urine, as previously described. 6 The antiserum selected for use in this study was obtained from 1 rabbit which had received 17 injections of antigen over a period of 18 months. Seven of the

1)

301'1

1) (j)

:1' \I '\11 :: II ~ V' I::

o

II)

I

W

I:

N

,,I

o Z


3

i

IJ)

II)

.,

I

I

I

t

A

1/ 10

z

o

~

II

II

20 Z

~

,,


Radioiodination of PG I The lyophilized PG 1 was dissolved in 0.15 M phosphate buffer, pH 7.5, to a concentration of 500 Jlg per ml, and labeled with 1251 by a modification of the technique of Greenwood and Hunter.31 The reaction mixture consisted of 50 Jlliters of 0.5 M phosphate buffer, pH 7.5. 5 Jlg of PG I. 250 JlC of 1251, 10 Jlg of chloramine-T, and 25 Jlg of sodium metabisulfite. The organic and inorganic peaks were separated by gel filtration on an 0.8- by 8-cm column of Sephadex G-75 which was equilibrated and eluted with 0.15 M phosphate buffer, pH 7.5. Specific activity was in the range of 15 to 25 JlC per Jlg. The stock solution of 1251_PG I was stored at - 20 C and was suitable for immunoassay for 6 to 7 weeks when purified as described below.

40

20

~

'" ~ ?

b

30

10

40

l

50

FRACTION NO.

FIG. 3. Chromatography of group I pepsinogens on hydroxylapatite. The arrow on the left indicates a change in the elution buffer to 0.025 M phosphate buffer. The arrow on the right indicates a change to 0.8 M phosphate buffer.

injections were given in Freund's complete adjuvant, and the remaining injections were given intravenously and subcutaneously without adjuvant. Antisera to PG II (anti-PC II) and PG I-PG II (anti-PG I-PG 11). These antisera were prepared in rabbits and were used to study the immunochemical homogeneity of the PG I. The isolation of the antigens for immunization and the preparation of the antisera have previously been described. 6

Characterization of PG I Radioelectrophoretic analysis. The 12'1_PG I was electrophoresed in agar gel under the conditions previously described for the electrophoretic analysis of pepsinogens. 4 The plate was then immersed in 50% ethanol for 18 hr, dried, and exposed on X-ray film. The radioautograph revealed only four radioactive bands with the mobilities of Pg 2, Pg 3, Pg 4, and Pg 5 (fig. 4). These fractions of PG 1 have previously been shown to be immunochemically indistinguishable.' Radioimmunoelectrophoresis. The 1251_PG 1 was mixed with an equal volume of a gastric mucosal extract or normal human serum and electrophoresed in agar gel on glass microscope slides as previously described. 6 Rabbit anti-Pg I-PG II or anti-whole human serum was then placed in a lateral trough, and immunodiffusion was allowed to proceed for 24 hr. The slides were then washed, dried, and exposed on X-ray film.

April 1974

RADIOIMMUNOASSA Y OF PG I IN SERUM

497

FIG. 5. Radioimmunoelectrophoresis. A, stained slide. The well contained a mixture of l2'I-labeled group I pepsinogens (PG I) and gastric mucosal extract. The trough contained rabbit anti-PG I-PG II. The precipitin arc with the faster anodal mobility (anode on left) is the PG I precipitin arc. The precipitin arc which formed closest to the well is the PG II precipitin arc. B, radioautograph. The arc on the radioautograph corresponds to the PG I precipitin arc.

FIG. 4. Radioautograph obtained by electrophoresis of "'I-labeled group I pepsinogens in agar gel.

The stained precipitin arcs were then compared with the radioautographic arcs. During immunoelectrophoresis, two precipitin arcs formed between the mixture of gastric mucosal extract and 125I_PG I and the rabbit anti-PG I-PG II (fig. 5, A). The radioautograph revealed radioactivity only in the PG I precipitin arc (fig. 5, B). When 125I_PG I was added to human serum and radioimmunoelectrophoresis was conducted against rabbit anti-whole human serum, none of the precipitin arcs were radioactive. There was, however, a faint radioactive zone on the radioautograph which was not associated with a visible precipitin arc. The radioactive zone was anodal to the precipitin arc for albumin (fig. 6). PG I has faster anodal mobility than albumin and the radioactive zone is attributed to the presence of antibodies to PG I in the rabbit anti-whole human serum. This might be expected since the immunizing antigen, human serum, contains PG I, and human PG I is antigenic in rabbits.

FIG. 6. Radioimmunoelectrophoresis. A, stained slide. The well contained a mixture of 125I-labeled group I pepsinogens (PG 1) and normal human serum. The trough contained rabbit anti-whole human serum. The anode is to the left. B, radioautograph. None of 'the precipitin arcs produce a radioautographic arc. The radioactive zone (arrow), which is located anodal to the precipitin arc for albumin, is attributed to the interaction of 125I-PG I with anti-PG I in the anti-whole human serum.

Radioimmunoprecipitation. The 12 5I_PG I

was diluted to provide approximately 100,000 counts per min per ml, and 100-~liters portions were incubated with an equal volume of a 1: 10 dilution of normal rabbit serum (NRS), rabbit anti-whole human serum, rabbit anti-PG I, or rabbit anti-PG II. The total volume of the incubation mixture was made up to 1.0 ml with 0.01 M phosphate-buffered saline, pH 7.5, containing 1% NRS. After standing for 72 hr at 4 C, 100 ,uliters of goat anti-rabbit 'YG were added to each tube, and the incubation was continued for an additional 24 hr. The tubes were then centrifuged at 4 C, and the radioactivity in the precipitates was determined. In table 1 it is shown that 4.0% of the radioactivity was non-

498

SAMLOFF AND LIEBMAN TABLE

Vol. 66,No . 4

1. Immunoprecipitation of ' 25I-group

I pepsinogens Rabbit serum

Radioactivity in precipitate %

Normal Anti-human serum Anti-PG I . Anti-PG II . . . .. . . . . . . . . .

4.0 7.9 74.0 5.3

specifically precipitated by NRS. This was not appreciably different from that precipitated with anti-PG II. The finding of 7.9% of the radioactivity in the precipitate formed with rabbit anti-whole human serum is compatible with the previous evidence that this anti-serum contains anti-PG I in low titer. Identification of PC I and PC II in serum and preparation of depepsinogeniz ed human serum. Twenty-five milliliters of normal human serum were dialyzed against 0.1 M phosphate buffer, pH 7.3, and applied to a 5- by go-cm column of Bio-Gel P-100 which had been calibrated with PG I and PG II in an extract of gastric mucosa. The eluate fractions corresponding to the calibrated elution volume of PG I and PG II were pooled, concentrated to 10 ml by pressure dialysis, and examined by electrophoretic analysis. • The zymogram revealed that this pool contained Pg 2 through Pg 7 (fig. 7). Pg 2 through Pg 5 are components of PG I, and Pg 6 and Pg 7 are the components of PG II. Pg 1, the fastest migrating fraction of PG I, was not seen. This fraction is probably present in serum in low concentration , as it has been identified in urine .' The fractions of serum which eluted ahead of PG I and PG II were pooled and concentrated to 25 m!. This concentrate did not contain any detectable PG I by radioimmunoassay. Consequently, gel filtration was routinely used to prepare the depepsinogenized normal human serum (DNHS) which was used in the radioimmunoassay to maintain an equivalent concentration of serum protein in each tube.

Conditions of Radioimmunoassay The general scheme for the radioimmunoassay was that of a competitive binding, double antibody system. The conditions were determined by examining the influence of different buffers, the duration of incubation, and the concentration and type of protein on the proportion of 125I_PG I bound to anti-PG I during incubation with standard PG lor human serum. Based on the results of these experiments, the following conditions were selected.

FIG. 7. Zymogram of pepsinogens in serum . The stained band indicated by the arrow is due to albumin. The proteolytic band corresponding to pepsinogen 6 was faint and is not well seen in the photograph.

The diluent buffer was 0.01 M phosphate-buffered saline, pH 7.5, containing 0.01 % sodium azide. The reagents were added to and the incubations were carried out in 10- by 75-mm disposable glass test tubes. The final volume of the initial incubation mixture was 1.0 ml and contained the following in the order given. (a) 0.1 M disodium ethylenediaminetetraacetate, 1% human serum albumin, and 5% NRS in diluent buffer (100 ~liters) . (b) PG I standard or test serum (10 to 100 ~liters) . The PG I standard was made up in 20% DNHS in diluent buffer. (c) Diluent buffer (500 to 590 ~liters). (d) DNHS or diluent buffer (100 ~li­ ters). The DNHS was appropriately diluted to maintain an equivalent concentration of human serum protein in each tube. (e) 12SI_PG I to provide approximately 10,000 counts per min in diluent buffer containing 2.0 mg per ml of human serum albumin and 0.75% NRS (100

~liters ) .

(f) Anti-PG I in a 1 :5000 dilution (100

The stock anti-PG I was stored at -20 C as a 1: 100 dilution in diluent buffer containing 2% NRS. The working solution was made up in diluent buffer containing 2.0 mg per ml of human serum albumin and 0.75% NRS. After incubation for 72 hr at 4 C, goat antirabbit 'YG was added in excess of that required to precipitate all of the rabbit 'Y-G, and incubation was continued for an additional 24 hr at 4 C .The tubes were then centrifuged at 2900 X g at 4 C for 30 min , and the radioactivity in the precipitate was counted in an automatic 'Y counter having a 4-sodium iodide crystal. All standards and test sera were set up in triplicate. Nonspecific binding in the precipitate was determined by setting up blank tubes containing diluent buffer instead of anti-PG I. The specific radioactivity precipitated was calculated by subtracting the mean radioactivity in the blank tubes from that in the standards and unknowns. ~liters) .

Characterization of Radioimmunoassay Calibration curve. A standard curve in which the concentration of PG I ranged from 0.02 to 100 ng per ml of incubation mixture is shown in figure 8. The sensitivity of t he assay is in the older of 0.2 ng per ml of incubation mixture and the curve is essentially linear when the concentration of PG I is between 1.0 and 10 ng per mI. In routine assays, unknown sera were assayed at a 1: 25 dilution (40 ~liters) . Sera which gave results not on the linear portion of the curve were l'eassayed at a higher or lower dilution. Comparison of standard curve and serum dilution curve. From 20 to 100 ~liters of serum were run simultaneously with a standard curve. The two curves were not distinguishable (fig. 9). 0

I

50

"'....0:: cr:

U 40 OJ

a: Q.

r

....

s:

~
u

30

i5

;;:

Reproducibility and effect of storage. Fifteen sera were assayed on two occasions, approximately 2 months apart. The sera were kept frozen in the interim. Of the 15 paired determinations, 13 differed by less than 5% of the mean value (fig. 10). These results suggested that PG I in serum was stable at - 20 C for at least 2 months. Eight sera with PG I levels between 70.0 and 165.2 ng per ml were assayed again after 7 months of storage at - 20 C. The mean (± SE) serum PG I 0 w 60

f-

'"

ij;

u

w

50

~

a:

a.

>-

f-

~

40

f-

U

<1 0

is

30

<1

a: u

.:;;; u

20

w

a. V>

.... Z

10

W U

a:

w

a.

.5

2

5

10

20

50

GROUP I PEPSINOGENS ng /ml

FIG. 9. Comparison of standard curve and serum dilution curve. The concentration of group I pepsi nogens in whole serum was calculated from the average of the five determinations. The concentration at each dilution of serum was calculated from the respective dilution factor and plotted against the observed percentage of radioactivity precipitated .

~

(/) (/)

« E "-

0>

c:

(/)

I

z

w

C>

o z

iii


a: u

499

RADlOIMMUNOASSA Y OF PC I IN SERUM

Aprii1974

Q.

20

W

Q.

U

"'

Q.

Vl

....

Q.

10

::>

o

Z

OJ

a:

U

a:

"'

C>

Q.

.02

5 10 20 GROUP I PEPSINOGENS ng / ml

.05 0.1 0.2

0.5

I

2

50 100

FIG. 8. Standard curve for radioimmunoassay of group I pepsinogens.

GROUP I PEPSINOGENS ng/ml ASSAY

n

FIG. 10. Reproducibility of determinations. The interval between assay I and assay II was 2 months .

500

Vol. 66,No. 4

SAMLOFF AND LIEBMAN

level at the initial assay was 118.6 ± 12.9, and at 7 months it was 119.8 ± 12.0 ng per ml. Repeatability. Four sera known to contain from low to very high levels of PG I were each assayed 15 times in a single assay. The coefficients of variation ranged from 4.1 to 6.1% (table 2).

Subjects and Sera Healthy control subjects. Sera were obtained from 69 healthy subjects ranging in age from 6 to 67 years (mean 30.2 years). Most of the subjects were hospital personnel and none gave a history of chronic dyspeptic symptoms. Hospital control patients. Sera were obtained from the clinical chemistry lab of the hospital from 71 patients without history or evidence of a gastroduodenal disorder, chronic alcoholism, or renal insufficiency. The patients ranged in age from 6 to 86 years (mean 44.6 years). Patients. Sera were obtained from 14 subjects with pernicious anemia, from 13 patients who had had a total gastrectomy for Zollinger-Ellison syndrome, and from 5 unoperated subjects with Zollinger-Ellison syndrome. Of the latter group, 3 subjects have been verified at surgery to have the Zollinger-Ellison syndrome, and 2 patients are suspected of having the ZollingerEllison syndrome because of elevated levels of serum gastrin and gastric hypersecretion.

Results Serum PG I levels. The mean (±sn) level of serum PG I in the 69 normal control subjects with 102.7 ± 32.5 ng per ml with a range of 49.5 to 212.0 ng per ml. This did not differ significantly from the mean value of 96.2 ± 41.9 ng per ml for the hospital control subjects, but the range of values, 17.6 to 225.0 ng per ml, was greater in the latter group (fig. 11). The 14 patients with pernicious anemia had a mean serum PG I level of 10.2 ± 6.4 ng per ml with a range of 3.6 to 40 ng per ml. Only 1 patient had a value over 25 ng per ml. The mean serum PG I level in 12 patients with total gastrectomy was 3.8 ± 2.9 ng per ml with a range of <2.0 to 8.5 ng per m!. (Sera with PG I levels below the detectable level, 2 ng per ml, were arbitrarily assigned a value of 1.0 to calculate the mean). One additional patient with an alleged total gastrectomy had a serum PG I level of 36 ng per ml and was not included in the mean value. The mean level of serum PG I in the 5 patients

TABLE 2.

Serum

1 2 3 4

Repeatability of determinations Mean

Standard deviation

ng/ml

ng/ml

%

26.8 80.5 247.6 463.4

1.6 3.3 11.9 19.2

6.1 4.2 4.8 4.1

Coefficient of variation

500-

400


Z

~ 300

o Z

-

~

'" Q.

Q.

200-

ffi

-

=> o

100 - -

ar ::t::'

-it

.~ -Iii

--:a"- -.:!.-

- ~'= - ':----

o

Healthy Control

I

.I. r-.-

Hospital Pernicious Totol Z- E Control Anemia Gastreef. Syndrome

FIG. 11. Serum group I pepsinogen levels in 69 healthy control subjects, 71 hospital control subjects, 14 patients with pernicious anemia, 13 patients with total gastrectomy, and 5 patients with Zollinger-Ellison syndrome. The horizontal lines indicate the mean values.

with Zollinger-Ellison syndrome was 456.1 ± 187.3 ng per ml with a range of 315 to 77 4 ng per ml. The finding of detectable levels of PG I in the sera of some patients with total gastrectomy was further studied by examining the urine from 2 of these subjects by electrophoretic analysis. 4 PG I was detected in the 100-fold concentrated urine of each patient.

Discussion Our current information about pepsinogen levels in serum has been limited to results obtained with proteolytic assays. Such assays detect both PG I and PG II, as

April 1974

RADIOIMMUNOASSA Y OF PG I IN SERUM

well as other acid proteases which may be present in serum. In several previous studies, the mean level of serum proteolytic activity in patients with total gastrectomy or pernicious anemia has been in the range of one-fourth to over one-third of the mean activity found in control subjects.21, 22. 24. 2 5. 3 2 This high background activity has usually been attributed to acid proteases other than pepsinogen, but, in light of recent findings, a proportion of this activity may have been due to PG II in serum which originated from the pyloric glands or Brunner's glands. In this study, the mean levels of PG I in patients with pernicious anemia and total gastrectomy were, respectively, about 10 and 4% of the mean values found for the control subjects. The finding of low levels of PG I in some patients with total gastrectomy appears to be supported by the finding of PG I in the urine of such patients by electrophoretic analysis. The source of this PG I is uncertain, but it may arise from heterotopic gastric mucosa in the .duodenum or other portions of the alimentary tract. 33 - 36 The occurrence of heterotopic gastric mucosa in the duodenum is well recognized. In a systematic study, Hoedemaeker33 found fully developed islets oigastric mucosa in the duodenal cuff in 33% of 158 gastric resection specimens. Chief · cells and mucous neck cells, the cellular origins of PG 1,7 were identified in many of the heterotopic glands. In some patients, the heterotopic mucosa may assume the appearance of macroscopic nodules or polyps.35. 36 An alternative explanation for the low levels of serum PG I after total gastrectomy is unresected homotopic gastric mucosa. Two patients with an estimated 90% gastrectomy for diffuse hemorrhagic gastritis have had serum PG I levels of 12.6 and 31.8 ng per ml. These values are higher than in 12 of 13 patients with a total gastrectomy. One patient with an alleged total gastrectomy had a serum PG I concentration of 36.0 ng per ml. This finding implies that the gastrectomy may not have been complete and suggests that the determination of serum PG I by radioimmunoassay may be of value in evaluating the completeness of gastrectomy.

501

The number of control patients examined thus far is not sufficient to establish the normal range of values, but these initial results suggest that the lower limit of normal is approximately 50 ng per ml. There was no overlap in values between the healthy control group and the patients with pernicious anemia, but such was observed between a few of the hospital control subjects and some patients with pernicious anemia. We have no information about the morphologic or functional status of the gastric mucosa in the control patients, but 7 of the 8 patients with levels of less than 50 ng per ml were in their sixth to ninth decades, ages associated with an increased incidence of atrophic gastritis. 37 This association has been suggested as the explanation for the finding that serum gastrin levels increase with advancing age. 38 Patients with Zollinger-Ellison syndrome have been reported to have an increased chief cell mass . 39 This increase appears to be reflected by elevated levels of PG I in serum. The mean value of serum PG I in the patients with Zollinger-Ellison syndrome was approximately 4-fold greater than in the control subjrcts, and there was no overlap in values between this group and either of the control groups. These initial data suggest that the extremes of gastric secretory function, pernicious anemia, and Zollinger-Ellison syndrome are characterized, respectively, by low and high levels of serum PG I, and that control subjects have levels between these extremes. Studies are currently in progress to fully characterize serum PG I levels in additional control subjects and in other groups of patients. REFERENCES 1. Seijffers MJ, Segal HL, Miller LL : Separation of pepsinogen I, pepsinogen II, and pepsinogen III from gastric mucosa. Am J Physiol 205: 1099-1105, 1963 2. Kushner I, Rapp W, Burtin P: Electrophoretic and immunochemical demonstration of the existence of four human pepsinogens. J Clin Invest 43:1983-1993, 1964 3. Hirsch-Marie H: Mise en evidence et separation de pepsinogenes et hydrolys'~s acides extragastriques. Bioi Gastroenterol (Paris) 2: 109- 122, 1968

502

SAMLOFF AND LIEBMAN

4. Samloff 1M: Slow moving protease and the seven pepsinogens. Electrophoretic demonstration of the existence of eight proteolytic fractions in human gastric mucosa. Gastroenterology 57:659-669. 1969 5. SamloffIM: Immunologic studies of human group I pepsinogens. J Immunol 106:962-968, 1971 6. Samloff 1M, Liebman WM: Purification and immunochemical characterization of group II pepsinogens in human seminal f1uid. Clin Exp Immunol 11:405-414, 1972 7. Samloff 1M: Cellular localization of group I pepsinogens in human gastric mucosa by immunofluorescence. Gastroenterology 61: 185-188, 1971 8. Samloff 1M, Liebman WM: Cellular localization of the group II pepsinogens in human stomach and duodenum. Gastroenterology 65:36-42, 1973 9. Samloff 1M, Townes PL: Electrophoretic heterogeneity and relationships of pepsinogens III human urine, serum, and gastric mucosa. Gastroenterology 58:462-469, 1970 10. Anson ML, Mirsky AE: The estimation of pepsin with hemoglobin. J Gen Physiol 16:59-63, 1932 11. Ilic V, Spray GH: Estimation of serum pepsinogen. Gut 7:415-419, 1966 12. Uete T, Was a M, Shimogami A: A simplified method for the determination of pepsinogen in blood and urine. Clin Chern 15:42-55, 1969 13. Samloff 1M, Kleinman MS: A radial diffusion assay for pepsinogen and pepsin. Gastroenterology 56:30-34, 1969 14. Berstad A: A modified hemoglobin substrate method for the estimation of pepsin in gastric juice. Scand J Gastroenterol 5:343-348, 1970 15. Chiang L, Sanchez-Chiang L, Wolf S. et al: The separate determination of human pepsin and gastricsin. Proc Soc Exp Bioi Med 122:700-704, 1966 16. Huang WY, Tang J: On the specificity of human gastric sin and pepsin. J Bioi Chern 244:1085-1091, 1969 17. Turner MD, Tuxill JL, Miller LL, et al: Measurement of pepsin I (gatricsin) in human gastric juice. Gastroenterology 53:905-911, 1967 18. Seijffers MJ, Tkatch R: Assay of two pepsin fractions in human gastric juice by alkali inactivation. Gastroenterology 59:528-533, 1970 19. Hirsch-Marie H: Dosage immunochimique du pepsinogene et de la pepsine gastrique et urinaire par la methode de Laurell. Clin Chim Acta 24:411-416, 1969 20. Chinn AB: Studies on a blood serum proteolytic enzyme with particular reference to gastric secretory function. Gastroenterology 25:14-23, 1953 21. Bock OAA, Arapakis G, Witts LJ, et al: The serum pepsinogen level with special reference to the histology of the gastric mucosa. Gut 4:106-111, 1963

Vol. 66,No. 4

22. Mirsky lA, Futterman P, Kaplan S: Blood plasma pepsinogen. II. The activity of the plasma from "normal" subjects, patients with duodenal ulcer and patients with pernicious anemia. J Lab Clin Med 40:188-199, 1952 23. Spiro HM, Ryan AE, Jones CM: The utility of the blood pepsin assay in clinical medicine. N Engl J Med 253:261-266, 1955 24. Hoar CS, Browning JR: Plasma pepsinogen in peptic ulcer disease and other gastric disorders. A clinical and laboratory evaluation. N Engl J Med 255:153-158, 1956 25. Singh AK, Shin ton NK: Serum pepsinogen in the differentiation of megaloblastic anaemia. J Clin Pathol 18:349-352, 1965 26. Mirsky IA: Physiologic, psychologic and social determinants in etiology of duodenal ulcer. Am J Dig Dis 3:285-314, 1958. 27. Yessler PG, Reiser MF, Rioch DM: Etiology of duodenal ulcer. 1. Serum pepsinogen and peptic ulcer in inductees. JAMA 169:451-456, 1959 28. Niederman JC, Spiro HM, Sheldon WH: Blood pepsin as marker of susceptibility to duodenal ulcer disease. Arch Environ Health 8:540-546, 1964 29. Pastore JO, Kato H, Belsky JL: Serum pepsin and tubeless gastric analysis as predictors of stomach cancer. A 10-year follow-up study, Hiroshima. N Engl J Med 286:279-284, 1972 30. Samloff 1M, Townes PL: Pepsinogens: Genetic polymorphism in man. Science 168:144-145,1970 31. Greenwood FC, Hunter WM: The preparation of I' II-labelled human growth hormone of high specific activity. Biochem J 89:114-123, 1963 32. Noland KE: Serum pepsinogen levels: relation to gastric secretion and gastric biopsy. Med J Aust 45:831-833, 1958 33. Hoedemaeker PJ: Heterotopic gastric mucosa in the duodenum. Digestion 3: 165-173, 1970 34. Taylor AL: The epithelial heterotopias of the alimentary tract. J Pathol Bacteriol 30:415-449, 1927 35. Belber JP, Musick R: Ectopic gastric mucosa in the duodenum (abstr). Gastroenterology 58:1063, 1970 36. Johansen AA, Hansen OH: Macroscopically demonstrable heterotopic gastric mucosa in the duodenum. Scand J Gastroenterol 8:59-63, 1973 37. Siurala M, Eramma E, Ngberg W: Pernicious anemia and atrophic gastritis. Acta Med Scand 166:213-223, 1960 38. Trudeau WL, McGuigan JE: Serum gastrin levels in patients with peptic ulcer disease. Gastroenterology 59:6-12, 1970 39. Neuberger P, Lewin M, Bonfils S: Parietal and chief cell populations in four cases of the Zollinger-Ellison syndrome. Gastroenterology 63:937-942, 1972