Radioimmunochemical quantitation of sulfated and non-sulfated gastrins in serum

Clinica Chimrcn Elsevier Biomedical

127 (1983)

29

CCA

Radioimmunochemical non-sulfated Bent Nyboe Andersen,

quantitation of sulfated and gastrins in serum

Laura De Magistris and Jens F. Rehfeld

University Department of Clinical Chemistr~~, Rigshospitalet, DK -2100 Copenhagen, and Medical Depurtment F, Copenhagen County Hospital. DK - 2600 Glostrup (Denmark) (Received

April 29th; revision August 31st, 1982)

Summary A radioimmunochemical procedure which distinguishes sulfated from non-sulfated gastrins has been developed. Two antisera raised against synthetic non-sulfated human hexadecapeptide gastrin were used. No. 2604 binds sulfated and non-sulfated gastrins with equimolar potency, whereas No. 2605 reacts poorly with sulfated gastrin (ID,, for non-sulfated gastrin : ID,, for sulfated gastrin = 0.06). Both antisera bind gastrins of different molecular length with equimolar potency using monoiodinated human gastrin-17 as tracer. The method was validated by fractionating gastrins in serum and in tissue extracts, and by recovery experiments. We found that Component I of gastrin - like the smaller gastrin components - was present in both, sulfated and non-sulfated form. In serum from normal fasting subjects the concentration of non-sulfated gastrin was 12.5 + 0.8 pmol/l (mean f SEM) with a total range of O-44 pmol/l and the corresponding values for sulfated gastrin were 7.5 k 0.5 pmol/l (range O-20 pmol/l). Sulfated gastrin accounted for more than half of the gastrins in only 21% of normal subjects. There was a parallel rise and fall in sulfated and non-sulfated gastrins after a meal and after stimulation with adrenaline.

Introduction The antral hormone gastrin occurs in both, sulfated and non-sulfated forms, in that tyrosine in position 6 from the C-terminus may be sulfated in gastrin-14, -17 and -34 isolated from antral or gastrinoma tissue [l-4]. Moreover, gel and ion-exchange chromatography of serum have shown that the gastrin also circulates in both, sulfated and non-sulfated form [5,6], which suggests that desulfation does not occur during boiling and extraction of the tissue.

Correspondence to Bent Nyboe Andersen, University Blegdamsvej 9, DK-2 100 Copenhagen, Denmark. 0009-8981/83/0000-0000/$03.00

Department

0 1983 Elsevier Biomedical

of Clinical

Press

Chemistry,

Rigshospitalet,

30

Since suifation appears to have only limited effect on the parietal cell in larger mammals 171, quantitation of sulfated versus non-sulfated gastrins in biological fluids has so far received only limited attention. However, recent studies suggest that sulfation of gastrin has considerable pancreozyminic and cholecystokinetic effects in higher mammals [7-91. In order to study the significance of sulfation of gastrin, we have now developed radioimmunoassays which distinguish sulfated from non-sulfated gastrins. Materials Peptides Human heptadecapeptide amide gastrin in sulfated (hG-17s) and non-sulfated (hG-17ns) forms was obtained from J. Walsh, Center for Ulcer Research and Education, Los Angeles, CA, USA. HG-34s hG-3&s, hG-14s and hG-14ns were obtained from R.A. Gregory, University of Liverpool, UK. The amount of hG-14 was quantitated in freshly prepared solutions using the absorbance at 280 nm [lo]. HG-34s and hG-34ns were not available in amounts sufficient for quantitation by light absorbance. Synthetic hG-17ns for isotope labelling and standard were purchased from I.C.I., Alderley Park, UK. Synthetic sulfated octapeptide of CCK was obtained from Squibb, Princeton, NJ, USA, and synthetic Met-Enkephalin from Sandoz, Basel, Switzerland. Antral or gastrinoma tissues were extracted in boiling water (10 ml/g tissue) for 20 min. After homogenization for 5 min the sample was centrifuged at 10000 X g and the supernatant decanted. Antisera Antisera Nos. 2604 and 2605 were raised in rabbits against non-sulfated synthetic human gastrin 2-17. These antisera have previously been characterized in detail [ 11,12]. Briefly, the titers (defined as the final dilution at which the antisera can bind 50% of 2 fmol tracer at equilibrium) were 100000 for Ab. 2604 and 150 000 for Ab. 2605. Data for index of heterogeneity and effective equilibrium constants are reported elsewhere [ 1 I]. The ratio ID,, G- 17ns/ID,, CCK-8s was 0.0 17 for Ab. 2605 and 0.001 for Ab. 2604. Even in concentrations of IO-’ moi/l Met-Enkephalin did not displace the tracer from antisera 2604 and 2605. Gastrin tracer Synthetic hG- 17ns was monoiodinated

at Tyr- 12 as previously

described

[ 131.

Methods Gel chromatography A Sephadex G-50 superfine column 20 x 2000 mm (Pharmacia, Sweden) was eluted with ammonium bicarbonate, pH 8.2, at 4°C and a flow rate of 15 ml/h. The column was calibrated with ‘251-albumin and *‘NaCl for indication of void and

31

total volume, and with hG-1’7s hG-17ns, were collected in all experiments.

hG-34s

and hG-34ns.

Samples

of 2.5 ml

Ion-exchange chromatography Tissue extracts or samples of gastrin peptides were applied to aminoethylcellulose columns (10 x 150 mm) (AE-41, Whatman, Maidstone, UK), eluted with ammonium bicarbonate, pH 8.2, in a continuous gradient (300 ml) from 0.05 to 0.2 mol/l at 20°C with a flow rate of 34 ml/h. Samples of 2.25 ml were collected. Desalting A pool of 50 ml of Component I of gastrin eluted on gel chromatography desalted using an Amicon@ (MA, USA) pressure cell and a U.M. 05 membrane a volume of 3 ml was obtained.

was until

Potency of hG-34 versus hG-I 7 Incubation of hG-34 with trypsin will result in a total conversion to hG-17 [ 141, and estimation of immunoreactivity in a sample of hG-34 before and after trypsin can be used to estimate the potency of hG-34 relative to hG-17 [14]. Thus, samples of hG-34s and hG-34ns were incubated with 1 mg trypsin per ml (Trypsin TPCK, Worthington Biochemical Corporation, NJ, USA) at 37’C. The trypsination was terminated by boiling. The immunoreactivity was measured with Ab. 2604 and 2605 in different dilutions of the samples before and after incubation with trypsin. The mean concentration before and after tryptic cleavage was used to calculate the potency of hG-34 relative to hG-17. Radioimmunoassay The assay was performed as previously described [ 12,151. When serum samples were assayed, the standard curve was made in hormone-free serum [16], which was prepared using 10 g charcoal to 100 ml serum. Charcoal efficiently removed gastrin from the serum of two patients with hypergastrinaemia (from 400 to 0 pmol/l in one patient with pernicious anaemia and from 20800 to 25 pmol/l in a gastrinoma patient). Non-specific interference by protein was studied after adding different concentrations of human albumin (Kabi, Finland) to the assay buffer. Results

Effect of molecular size on the binding Before tryptic cleavage of hG-34ns after trypsin when measured with Ab. the potency before trypsin was 100% trypsin.

to Ah. 2604 and 2605 the potency was 94% and 91% of the potency 2604 and Ab. 2605, respectively. For hG-34s (Ab. 2604) and 93% (Ab. 2605) of that after

Effect of suljation on binding to Ab. 2604 and 2605 The inhibitory dose,, (ID,,) of hG-14, hG-17

and

hG-34

in sulfated

and

I

1.00 1.00

hG- 17s I .on 0.07

hG-17ns I .os 1.06

0.82 0.78

hG-34ns

* 0.88 0.05

HUMAN

1.14 1.20

**

PEPTIDES

with Ab. 2604

0.96 0.20

hG-14s

GASTRIN

were measured

hG- 14ns

in stock solutions

*

OF NATURAL

hG-34s

hG-17ns)

* Exact amounts of G-34 peptides were not available, and the concentration of G-34s and G-34ns ** Ion-exchange chromatography of this peptide suggested that it was impure.

Ab. 2604 Ab. 2605

Synthetic hG- 171-1s

POTENCY (EXPRESSED AS ID,, COMPARED TO THE ID,,, OF SYNTHETIC DISPLACING ANTIBODIES 2604 AND 2605 FROM I’2S-GASTRIN-I-17

TABLE

IN

33

0

50

100

140

130

170

EFFLUENT

samples were (and shown in the Fig.) in dilutions ranging from 1: 2 to 1: 5 with antibodies 2604 and 2605. The recovery of added gastrin was 101% (hG-l7ns), 114 (hG-17sf, 81% (hG-14ns). 69% (hG-14s). 30% (hG-34ns) and 27% (hG-34s) when measured with antibody 2604. -, Ab. 2604; ---_--, Ab. 2605.

non-sulfated form was compared to that of synthetic hG-17ns (Table I). As seen from the table, antiserum 2605 reacted poorly with sulfated gastrin. The reactivity with sulfated and non-sulfated gastrin was further investigated by ion-exchange chromatography of hG-14, hG-17 and hG-34 in sulfated and non-sulfated form. As seen from Fig. 1, the non-sulfated peptides reacted equally with antiserum 2604 and 2605, whereas the peak size of sulfated peptides, as measured by Ab. 2605, was only 5.3% (hG-34s) and 6.3% (hG-17s) of that measured by antibody 2604. HG-14s appeared to contain significant amounts of impurities, because part of the immunoreactivity eluted before and after the main peak, which, when measured with Ab. 2605, was only 7.2% of that measured with Ab. 2604. Binding of ‘Component I to Ab. 2604 and 2605 When 0.5 ml of serum from a gastrinoma patient was subjected to gel chromatography a peak corresponding to Component I with a K,, of 0.23 occurred. This component was also demonstrated with Ab. 2605, but the total immunoreactivity with a K,, from 0.16 to 0.30 was only 54% of the immunoreactivity measured with Ab. 2604. When a desalted trypsinated pool of Component I was subjected to ion-exchange chromato~aphy the i~unoreacti~ty now eluted in two peaks corresponding to hG- 17ns and hG-17s (Fig. 2).

G-17s

G47ns

OJ, 50

1

I

100

150

EFFLUENT VOLUME (ml)

Fig. 2. A desalted trypsinated pool of Component I from serum of a gastrinoma patient has been subjected to ion-exchange chromatography. The elution pattern was monitored with antisera 2604 ) and 2605 f------). Similar columns had previously been calibrated with samples of hG-17ns and (hG- 17s.

A0

260~

AB 2605

i \ _ \ \ I \ \ \ \

x

\

\ \

-\ 0

1

-iGin

‘.dY

CONCI~NT~TION Lpmot/l)

6

-!?I0 (x)CONCI ENTFWION

II,rn”l/ll

n O~~ASTRIN

t 0:ALB”MlN

\

\

\

‘x. 8

?=a &-”

*

500

‘P”‘“‘!”

Fig. 3. Standard curves for antisera 2604 and 2605 in buffer f) and in hormone-free serum (-- - - - -) are shown. The Fig. also shows how different concentrations of human albumin in the incubation medium affects the binding of the antisera to ‘*‘I-gastrin at zero hormone concentration.

35

Interference by protein For both antisera standard curves made in hormone-free serum were parallel, but not completely superimposable on standard curves made in buffer. Approximately 2% more radio-ligand was bound when using hormone-free serum, despite correction for incubation damage (Fig. 3). As also indicated on Fig. 3, different concentrations of human serum albumin influenced the binding of both antisera to the radio-ligand at zero hormone concentration. Detection limit The detection limit for non-sulfated gastrin (hG-17ns) - defined as the lowest concentration significantly different from zero hormone concentration was 2 pmol/l in hormone-free serum (assay in quadruplicate [lo]). For sulfated gastrin (hG-17s) the detection limit was 3.6 pmol/l as this was the lowest concentration where the difference between the binding measured by Ab. 2604 and 2605 was significantly higher than the difference at zero hormone concentration. Precision The intra-assay coefficient of variation for measurement of non-sulfated gastrin (Ab. 2605) in serum was 8.9% at a mean concentration of 13.4 pmol/l, and the inter-assay coefficient of variation (during 6 months) was 21.4%. Corresponding values for sulfated gastrin (Ab. 2604-2605) at a mean concentration of 8.6 pmol/l were 23.3% and 30.6%, respectively. Accuracy Sulfated and non-sulfated gastrins were separated in extracts of human antrum, porcine antrum and gastrinoma tissue using ion-exchange chromatography. The observed and measured ratio between sulfated and non-sulfated gastrin correlated well (Table II). When known amounts (10 different doses) of hG-17ns were added to serum from a normal subject the mean increase in immunoreactivity measured with Ab. 2605 was 103%. When hG-17s was added to serum the mean increase in immunoreactivity (Ab. 2604-2605) was 94% of the calculated increase. TABLE

11

THE OBSERVED DISTRIBUTION OF SULFATED AND NON-SULFATED GASTRIN FOUND AFTER SEPARATION ON ION-EXCHANGE CHROMATOGRAPHY, COMPARED TO DISTRIBUTION MEASURED BY DIRECT ASSAY OF EXTRACTS Extract

Measured

distribution

Observed

distribution

sulfated

non-sulfated

sulfated

non-sulfated

Human

antrum

30.9%

69.1%

35.5%

64.5%

Porcine

antrum

50.5%

49.5%

48.2%

51.8%

Gastrinoma

(1)

57.9%

42.1%

58.4%

41.6%

Gastrinoma

(2)

59.8%

40.2%

57.1%

42.9%

36

TABLE

III

REPEATED MEASUREMENT NORMAL SUBJECTS WITH

1 (2)

2

3

(3) (1) (2) (3) (1) (2) (3)

OF SULFATED EITHER LOW

Non-sulfated

Sulfated

(pmol/l)

(pmol/l)

36 31 32 0 4 5 10 I5 I3

3 5 I I4 20 I6 8 5 7

AND

NON-SULFATED

GASTRIN

IN THREE

Serum concentrations of sulfated and non-sulfated gustrin The concentration of sulfated and non-sulfated gastrin in the fasting state was measured in seven different assays in 100 healthy subjects (61 women, 39 men) with an age of 40.0 + 1.2 years (mean f SEM). The concentration of both sulfated and non-sulfated gastrin followed the normal distribution. The concentration of nonsulfated gastrin was 12.5 f 0.8 pmol/l (mean f SEM) with a total range of O-44 pmol/l; that of sulfated gastrin was 7.5 rt 0.5 pmol/l (range O-20 pmol/l). There was no difference in the degree of sulfation between men and women and it was unchanged with increasing age.

1 30-

MEAL

z 2 J Y= =>

ADRENALINE k -1.h~1 2Sng. 4

I 20-

@,,.

5 2

‘fi

345

OI 0

60

120 180 MINUTES

0’

60 MINUTES

Fig. 4. The concentration of sulfated and non-sulfated gastrin after a protein rich meal (n = 6) and after infusion of adrenaline (n = 3) (25 ng’kg ‘.h-’ for 10 min). Non-sulfated gastrin was measured with antiserum 2605 and sulfated gastrin as the difference between the concentrations measured with antisera 2604 and 2605. Values are given as median. 0 -0, non-sulfated gastrin; n ---n, sulfated gastrin.

31

The ratio between sulfated and non-sulfated gastrin varied between the subjects. In 21% of the subjects sulfated gastrin accounted for more than half of the gastrins. Repeated estimations of sulfated and non-sulfated gastrins in individual sera confirmed that the ratio between sulfated and non-sulfated gastrin was highly variable (Table III). With a protein rich meal sulfated and non-sulfated gastrin increased and declined in parallel (Fig. 4). Intravenous infusion of adrenaline 25 ng . kg- ’ . h- ’ for 10 min slightly increased the concentration of both sulfated and non-sulfated gastrin for 15 min (Fig. 4). Discussion The present study shows that it is possible in a simple manner, i.e. without chromatography, to distinguish sulfated from non-sulfated gastrins. The data indicate that Component I of gastrin circulates in both, sulfated and non-sulfated forms. Moreover, the study showed that the ratio between sulfated and non-sulfated gastrins varied from one subject to another, although on average only 37.5% of the gastrins in circulation are sulfated. The difference between the concentrations of gastrin measured with two antisera was taken as a measure of sulfated gastrin. It is obvious that an antiserum entirely specific for sulfated gas&in is desirable. However, such an antiserum would be difficult to raise, because sulfated gastrin is available only in limited amounts, and because it is uncertain if immunization with sulfated gastrin necessarily would give antibodies without reactivity towards non-sulfated gastrin. Nevertheless, an antiserum entirely specific for the sulfated form of the homologous hormone, cholecystokinin, has been produced [ 171. Hansky et al [ 181 first suggested that sulfated and non-sulfated gastrins in serum could be differentiated using two antisera with different affinity towards sulfated and non-sulfated gastrins. In order to do so, the antisera should fulfil the following criteria: (I) one antiserum should be specific for either non-sulfated or sulfated gastrin; (2) the other should measure sulfated and non-sulfated gastrin equally well; (3) the antibodies should show equal and preferably full binding to the gastrins of different chain lengths; and (4) protein and other serum factors should give either no or equal interference with both antisera. Our antisera almost fulfilled these criteria: the cross-reactivity of antiserum 2605 for sulfated gastrins is very low (= 6%). The cross-reactivity was not dependent on the length of the gastrin peptide. This was also substantiated by the fact that binding of hG-34s to Ab. 2605 was similar before and after tryptic cleavage of hG-34s. Assuming a cross-reactivity of 6% and a concentration of sulfated gastrin of 7.5 pmol/l, it means that the amount of non-sulfated gastrin will be over-estimated with less than 0.5 pmol/l. As previously shown for porcine G-17 [ 121, we found that Ab. 2604 showed equal binding to sulfated and non-sulfated human gastrins from G-14 to G-34. We minimized non-specific interference of serum by using hormone-free serum for the standards. Subsequently, accuracy experiments confirmed that we measured sulfated and non-sulfated gastrin correctly in serum samples.

38

The purpose of the present study was to develop a simple and reliable method for distinction of sulfated from non-sulfated gastrins in serum. In addition, some interesting side-observations were made. We found that gastrin Component I [6] existed in both sulfated and non-sulfated form. This suggests that sulfation occurs early in the post-translational processing of gastrin. It has been assumed that sulfated and non-sulfated gastrins are found in equal amounts in serum [19]. Our study shows that in normal subjects, sulfated gastrin only accounts for 37.5% of the gastrins in serum. This is in harmony with the finding that 36% of gastrin in human antrum is sulfated [4]. The parallel increase and decline of sulfated and non-sulfated gastrin with a meal also suggests that the two gastrin peptides are secreted in parallel from the antrum, and that they have a similar biological half life as in dogs [20]. As previously indicated by Hansky et al [ 1X] and by McCuigan and Herbst [2l] different gastrin antisera have different affinity towards sulfated and non-sulfated gastrin. In a recent study of 17 different antisera, we found [l l] that the affinity towards sulfated gastrin varied from 0.04 to 3.73 (compared to non-sulfated gastrin). But the observation that suIfated gastrin only accounts for 37.5% of gastrin in serum will limit the variation in gastrin concentration measured with antibodies with different reactivity towards sulfated gastrin. References 1 Gregory

RA, Tracy HJ. The constitution and properties of two gastritis extracted from hog antral 1964; 5: 103-114. Gregory RA, Tracy HJ. Isolation of two ‘big gastrins’ from Zollinger-Ellison tumour tissue. Lancet 1972; i.i.: 797-799. Gregory RA. Tracy HJ. isolation of two minigastrins from Zollinger-Ellison tumor tissue. Gut 1974; 15: 683-685. Gregory RA, Tracy HJ, Grossman MI. Isolation of two gastrins from human antral mucosa. Nature 1966; 209: 583. Rehfeld JF, Stadil F. ‘Big gastrins’ in the Zollinger-Ellison syndrome. Lancet 1972; i.i.: 1200. Rehfeld JF, Stadil F, Vikelsoe J. Immunoreactive gastrin components in human serum. Gut 1974; 15: 102-111. Kaminski DL, Despande YG. Effect of gastrin I and gastrin II on canine bile flow. Am J Physiol 1979; 236: E-584-588. Chowdhury JR, Berkowitz JM, Praissman M, Fara JW. Effect of sulfated and non-sulfated gastrin and octapeptide-cholecystokinin on cat gall bladder in vitro. Experientia 1976; 9: 1173- 1175. Jensen SL, Rehfeld JF, Holst JJ, Fahrenkrug J, Nielsen OV, ~haffalit~y de Muckadeli OB. Secretory effects of the gastrins on the isolated perfused porcine pancreas. Am J Physiol 1980; 238: E-186- 194. Rosenquist GL, Walsh JH. Radioimmunoassay of gastrin. In: Glass GBJ, ed. Gastrointestinal hormones. New York: Raven Press, 1980: 769-795. Rehfeld JF, Magistris LD, Andersen BN. Sulfation of gastrin: effect on immunoreactivity. Regul Pept 1981; 2: 333-342. Rehfeld JF, Stadil F, Rubin 8. Production and evaluation of antibodies for the radioimmunoassay of gastrin. Stand J Clin Lab Invest 1972; 30: 221-232. Stadil F, Rehfeld JF. Preparation and evaluation of ‘2sI-synthetic human gas&in I for radioimmunoanalysis. Stand J Clin Lab Invest 1972; 30: 361-368. Magistris LD, Rehfeld JF. A simple enzymatic procedure for radioimmunochemical quantitation of the large molecular forms of gastrin and cholecystokinin. Analyt Biochem 1980; 102: 126.. 133. mucosa. Gut

2 3 4 5 6 7 8 9 10 11 12 13 14

39

15 Stadil F, Rehfeld JF. Determination of gastrin in serum: An evaluation of the reliability of a radioimmunoassay. Stand J Gastroenterol 1973; 8: lOI-- 112. 16 Ratcliffe WA, Ratcliffe JG, McBride AD, Harland WA, Randall TW. The radioimmunoassay of thyroxine in unextracted human serum. Clin Endocrinol 1974; 3: 481-488. 17 Rehfeld JF. Immunochemical studies on cholecystokinin. I. Development of sequence-specific radioimmunoassays for porcine triacontatr~apeptide cholecystokinin. J Biol Chem 1978: 253: 4016-4021. 18 Hansky J, Soveny C, Korman MC. Studies with two gastrin antisera of different specificity for gastrins I and 11. Digestion 1974; IO: 97-107. 19 Dockray GJ. Multiple molecular forms of hormones: Significance and methods of study. In: Bloom SR, Polak JM, eds. Gut hormones, 2nd ed. Edinburgh, London: Churchill-Livingstone, 1981: 43-48. 20 Walsh JH, Debas HT. Grossman MI. Pure human big gastrin. Immunochemical properties. disappearance half time and acid-stimulating action in dogs. J Clin Invest 1974; 54: 477-485. 21 McGuigan JE, Herbst CA. Binding and measurement of different gastrin forms by region-specific antibodies to gastrin. In: Thompson JC, ed. Gastrointestinal hormones. Austin, London: University of Texas Press, 1975: 85-98.