European Journal of Pharmacology, 35 (1976) 369--378 © North-Holland Publishing Company, Amsterdam -- Printed in The Netherlands
THE DEVELOPMENT OF A RELIABLE AND SENSITIVE BIOASSAY FOR GASTRIN IN BODY FLUIDS F R E D HALTER, B E R N H A R D and LAJOS V A R G A
KOHLER, J O H A N N A L E U E N B E R G E R , G O R D O N
M. SMITH*
Gastrointestinal Unit, University Hospital, Inselspital, Berne, Switzerland Received 24 March 1975, revised MS received 28 August 1975, accepted 9 September 1975
F. HALTER, B. KOHLER, J. LEUENBERGER, G.M. SMITH and L. VARGA, The development of a reliable and sensitive bioassay for gastrin in body fluids, European J. Pharmacol. 35 (1976) 369---378. The sensitivity and reliability of the Ghosh and Schild rat stomach preparation was improved by implantation of bulky intragastric cannulae, recirculation of the perfusate and measurement of conductivity instead of pH. With these procedures, the threshold values for gastrin I were in the range of 1--2 ng. In attempts to increase the sensitivity further it was shown that neither vagotomy nor antrectomy influenced the sensitivity of the method. The threshold values were not lowered by attempts to reduce gastrin metabolism through nephroligature or small bowel resection. Although rats fitted with portacaval shunts did ,raise the threshold, the limited increase in sensitivity was also achieved by selecting rats that gave an initially high response to a test dose of gastrin thus avoiding the complicated shunt operation. The assay procedure enabled statistically valid measurements of gastrin in plasma and from tumour extracts from patients suffering from the Zollinger--Ellison Syndrome to be assayed using a 2 + 2 block design. Perfused rat stomach preparation
Zollinger--Ellison syndrome
1. Introduction Early studies on the identification and extraction of gastrin were performed using the stomach preparations of anaesthetized and nonanaesthetized cats (Edkins, 1905; Komaroy, 1942; Uvn~ and Em~s, 1961; Blair et al., 1962). Subsequently, the anaesthetized cat preparation was used to estimate circulating gastrin in the ng range (Blair et al., 1968; Blair and Wood, 1968; Herring and Blair, 1969; Blair et al., 1970; Ingram et al., 1971). This bioassay has certain limitations. Since the anaesthetized cat preparation does not give a reliable secretion of acid, acid production has had to be estimated during continuous changes in sensi* School of Pharmacy, Robert Gordon's Institute of Technology, Schoolhill, Aberdeen, Scotland.
Gastrin
Bioassay
tivity. As the true threshold limit for synthetic human gastrin is 50 ng (Blair et al., 1970) volumes of 50 ml of plasma have to be injected per animal to estimate gastrin in the 1 ng range per animal which is the level of circulating gastrin in patients with the Zollinger--Ellison syndrome (Blair and Wood, 1968). Furthermore the use of the cat renders such a bioassay exceedingly expensive. As an alternative animal the rat has been widely used since Komarov et al. (1944) has shown that crude extracts of gastrin stimulate gastric acid production in the anaesthetized rat. Lai (1964) used the rat stomach preparation (Ghosh and Schild, 1958) to assay gastrin with a statistically valid experimental design. Although the assay was not sensitive enough to give reliable quantitative results for circulating gastrin, it was used by many workers in the attempt to diagnose the Zollinger--Ellison
370
syndrome pre-operatively by qualitative or semi-quantitative assays (Moore et al., 1967; Wilson et al., 1968; Murray-Lyon et al., 1968; Thompson et al., 1970). The reperfusion assay developed as a modification of the Ghosh and Schild procedure by Smith et al. (1970) enabled Colin-Jones et al. (1969) to assay the gastrin content of plasma from a patient with the ZoUinger--Ellison syndrome and to estimate the gastrin subsequently found in the tumour. This was achieved with fiducial limits of about 20% which is acceptable for this type of assay. The sensitivity threshold in their assay was however limited and reached 10 ng, and quantitative assays were only possible in the range of 50--200 ng (Colin-Jones and Lennard-Jones, 1972). The use of the more sensitive gastrin immunoassay to diagnose the Zollinger--Ellison syndrome (McGuigan and Trudeau, 1968) indicated that the biological assay would n o t detect many cases of this disease. The immunoassay will however not replace the biological assay since the former does not indicate if the immunoreactive material actually causes gastric secretion. Immunoassay and bioassay techniques should therefore be used alongside one another. Ideally b o t h methods should be equally sensitive and precise. In recent years a number of technical modifications have improved the sensitivity and reliability of the Ghosh and Schild preparation. Continuous reperfusion of the gastric effluent integrated the acid secreted thus increasing the sensitivity of the preparation (Smith et al., 1970). Measuring the conductivity of the gastric effluent instead of pH has greatly improved the linearity and reliability of the measurements of the acid secreted (Lawrence and Smith, 1969, 1974). A large cannula was inserted into the stomach to reduce the intragastric dead space (Parsons, 1969; Black et al., 1972). This enabled the secreted acid to be washed out more efficiently which significantly shortened the duration of the acid response. The work reported here describes an a t t e m p t to increase the sensitivity of the Ghosh and Schild rat preparation to gastrin by combining the most
F. HALTER ET AL. POTENTIOMETR!C RECORDER 1 I CONDUCTOSCOPE ]
STIRRER f HEATING COIL
o
, ,t
J OESOPHAGUS I I '~ I l
3CALEI Fig. 1. Diagrammatic representation of the reperfusion conductivity system for measuring gastric acid secretion.
relevant recent technical improvements described above. This work was performed to facilitate the preoperative diagnosis of the Zollinger--Ellison syndrome and to correlate the gastrin bio- and immunoassays on a large scale.
2. Material and methods 2.1. General
Male Wistar rats (140--240 g) were housed singly in special cages to prevent coprophagy. 18 hr before each experiment the food was replaced by sugar cubes and water was available ad libitum. The rats were anaesthetized by an intramuscular injection of urethane (1.5 g/kg b o d y weight) and small additional doses being given when necessary. Both the trachea and the external jugular vein were cannulated. The a b d o m e n was opened through a midline incision 3--4 cm long. Damage to the parietal cell-containing part of the stomach (corpus) was meticulously avoided throughout the operation. The oesophagus was ligated 5 mm above the cardia taking care to avoid the vagal
BIOASSAY OF GASTRIN IN BODY FLUIDS
371
nerves. An incision was made into the rumen (fundus) and the stomach was gently washed with warm saline. A perspex cannula with a ball-shaped head was inserted through the incision in the rumen and the tip passed o u t through a second opening near the pylorus. A large asymmetric collecting funnel was then introduced into the rumen. The eccentric part of the cannula was placed near the cardia to reduce the dead space in that part of the stomach {fig. 1).
the funnel-shaped cannula and delivered into a beaker containing the conductivity cell (type Siemens 9510). The perfusate in the beaker was mixed b y an electromagnetic stirrer. The temperature of the perfusate, operating table and beaker was held constant at 37°C b y an external heating system. The conductivity cell was connected to a conductoscope (Methrom 365 B) and the conductance of the perfusate recorded by a potentiometric recorder.
2. 2. Vagotomy
2. 5. Calibration o f the measuring system
Vagal fibres were exposed in the cervical region and care taken to avoid the abdominal vagi during gastric surgery. When the vagi were cut there was a change in the respiration rate. Successful cervical vagotomy was confirmed by testing each rat with 0.1 units of regular insulin at the end of each experiment. All animals failed to respond to insulin while a near maximal acid secretion of long duration was produced in control animals.
Aliquots of 0.1 ml N/100 HC1 were added to 20 ml of 5% glucose solution at 37°C, continuously stirred and the conductivity of the solution measured. The relationship between I-I* concentration and conductance was linear between 0--300/~Mhos (fig. 2). Human gastric juice was collected after stimulation with pentagastrin 6 pg/kg. The juice was diluted 1 : 10 and the acid concentration was 1.1 #Eq/0.1 ml. Aliquots of 0.1 ml of the diluted juice were added to 20 ml of 5% glucose and conductivity correlated to the acidity. The slopes of both curves are almost identical (fig. 2).
2. 3. Antrectomy The distal antral vessels were severed by electrocoagulation, carefully avoiding damage to the gastro-epiploic arteries. The antrum was cut at the pylorus and blood vessels of the upper part of the lesser curvature situated 1--2 mm distal to the cardia were cut by electrocoagulation before the antrum was removed. The ball-shaped cannula was then introduced and ligated at the antral-corporal junction. At the end of the experiment, the stomach was opened and inspected. Since the border between the antrum and corpus is clearly visible in the rat incomplete antrectomy was easily detected. These rats were eliminated from the study.
IJMhos
16oj
2. 4. Perfusion o f the gastric lumen The stomach was continuously perfused at a rate of 6 ml/min with 28 ml of 5% glucose solution and the perfusate collected through
8
.1'2 'pEq HCl
Fig. 2. Calibration curve for HCI and diluted gastric juice (GJ).
372
F. H A L T E R E T AL. Conductivity (~Mhos)
2. 6. Preparation o f standard and test
200
2. 6.1. Standard Synthetic human gastrin I (HG I) (Imperial Chemical Industries, Mereside, Macclesfield, England) was used as a standard throughout the experiments. Ampoules containing 1 mg were diluted with 0.05 M ammonium bicarbonate and divided in aliquots containing 50 pg. Each sample was lyophilised and stored at --20 ° C. Fresh samples were made up for each experiment and HG I dissolved in 0.5 M ammonium bicarbonate in concentrations of 1 0 - 6 4 0 ng per ml so that each standard dose had a volume of 0.1 ml in the range of 1--64 ng used in this study. 2.6.2. Test In the experiments where circulating gastrin was assayed, blood was withdrawn after an overnight fast. A drop of heparin was added to the syringe and 20--40 ml of blood was collected. The plasma was separated by high speed centrifugation (15,000 rpm) and stored in aliquots a t - - 2 0 to --30 ° C. Before the test, the material was thawed to n o t more than 37°C and recentrifuged before injection. The plasma was given as an intravenous infusion over 5 min; the maximum volumes injected were 1.0 ml per dose. 2. 6. 3. Statistical analysis 2 + 2 assays were performed using Latin square designs with 4 rats per assay.
Conductivity (/~Mhos) 2,, /..(V~
,,, . / ~ '
8n, / / / ~ "
0 ~~n_g . . ~ ~ 132 6n9
e4.~i
Fig. 3. 6 r e s p o n s e s o b t a i n e d in 1 rat a f t e r t h e i.v. i n j e c t i o n o f 6 d i f f e r e n t doses of HG I.
3. Results
3.1. Qualitative and quantitative evaluation o f acid secretory response to HG I After an i.v. injection of 16 ng of HG I.an acid secretory response was achieved in each animal. An injection of gastrin produced an increase in the steepness of the slope of the conductivity response (fig. 3). After peak acid secretion the slope gradually returned to the basal level. The response was measured b y recording the difference in conductivity between the extrapolated baseline and the postsecretory level (Smith et al., 1970; Lawrence and Smith, 1974). The reproducibility of repeated responses
TABLE 1 Analysis o f variance. 8 rats were given 6 c o n s e c u t i v e doses o f 16 ng HGI. S o u r c e of v a r i a t i o n
D.F.
Sum of square
M e a n square
F ratio
p 0.2
B I O A S S A Y O F G A S T R I N IN B O D Y F L U I D S
was found b y giving each of 8 rats 7 successive doses of 16 ng HG I. Total secretion, latency of onset and total duration of each response were measured. Following previously published studies (Lai, 1964; Smith et al., 1970) the first response to gastrin was n o t evaluated. The 'within rats' variation of the successive responses was very small, without tachyphylaxis. There was a fourfold difference in the 'between rat' variation (table 1). There was little fluctuation in latency and total duration of response between and within rats. The mean latency was 1.06 min + 0.244 (S.D.). 3. 2. Dose--response curve 6 doses (2--64 ng) of HG I were injected into 6 rats using a Latin square design and the responses obtained from one rat are shown in fig. 3. A dose--response curve was constructed from the mean results. Threshold values were a b o u t 2 ng and the steep part of the curve started at a b o u t 8--10 ng (fig. 4).
373
pMhos 90~
6O
30L
i
1'6 3'2 g4
LOG DOSE GASTRINT..inng Fig. 4. Dose---response curves for gastrin obtained in normal and selected rats. The vertical lines represent the standard error of the mean.
As shown in table 2 there was no difference in the sensitivity of the rats to gastrin before and after vagotomy.
3. 3. The effects o f antrectomy and vagotomy on stimulated acid secretion
3.4. A t t e m p t s to increase the sensitivity o f the assay
Since the complicated operative procedure may have a negative influence on gastric acid secretion, the effects of antrectomy and vagotomy were studied.
3. 4.1. General Theoretically it should be possible to increase the sensitivity of this system b y inhibiting the degradation of gastrin. Attempts were therefore made to increase the sensitivity of the assay s y s t e m b y avoiding the renal circulation (nephroligature), b y small bowel resection and by portacaval shunting.
3. 3.1. A n t r e c t o m y As it was technically not feasible to study gastric acid secretion before and after antrect o m y in the same rat, these experiments were performed on two batches of 6 animals matched for age and b o d y weight. 3 doses of HG I (8, 16, 32 ng) were given in random order to rats in b o t h groups and dose--response curves constructed. The slopes of both curves were similar and there was no significant difference at any dose level (table 2). 3. 3. 2. Vagotomy 8 and 16 ng HG I were given to 4 animals before and after cervical section of both vagi.
3. 4.2. Nephroligature 20 rat~ Were prepared for gastric acid secretory studms according to the standard procedure. g and 16 ng of HG I were administered before and 30 min after bilateral ligature of the renal vessles. There was no difference in the secretory response at b o t h dose levels studied (table 2). 3. 4. 3. Subtotal small bowel resection Using a similar procedure as for nephroliga-
F. HALTER ET AL.
374 TABLE 2 Attempts to increase the sensitivity of the assay. Technique
No. of observations
Dose of gastrin (rig)
Control
Test
Mean responses (pMhos)
-+ S.E.
Mean responses (pMhos)
t-test ~ S.E.
Antrectomy
6 6 6
8 16 32
9.5 20 50.2
0.83 0.8 2.72
5.8 17.3 46.8
0.8 2.35 6.63
NS NS NS
Vagotomy
4 4
8 16
18.8 31.3
3.95 4.92
15.5 26.1
5.38 7.98
NS NS
Nephroligature
20 20
8 16
19.0 39.9
2.35 3.15
22.3 51.9
2.31 5.64
NS NS
Small bowel resection
20 20
8 16
13.4 34.3
1.45 4.18
13.7 33.7
1.71 2.71
NS NS
Portocaval shunt*
10 10 10
8 16 32
7.3 20.2 54.7
1.59 2.88 6.89
18.0 41.4 90.2
4.07 3.85 11.68
p < 0.05 p < 0.01 p < 0.01
1 2 4 8 16
1.6 3.1 9.5 20.7
0.26 0.29 0.81 0.78
1.6 3.6 9.25 15.5
0.63 0.80 0.63 2.26
p < 0.025 p < 0.001 p < 0.025
Selected rats**
4 6(4) 6(4) 6(4) 6
* Control = control group I in text (animals same weight as test animals). Control group II (animals same age as test animals) also produced significantly less gastric acid secretion. ** 6 controls used and 4 test animals.
ture, s u b t o t a l small b o w e l r e s e c t i o n was perf o r m e d in 20 rats. T h e small intestine was removed by electrosurgery from the ligament o f Treitz t o t h e ileocoecal valve. Again n o significant d i f f e r e n c e in s e c r e t i o n was achieved (table 2).
3.4.4. Portacaval shunting A g r o u p o f 10 animals was fitted w i t h e n d to side p o r t a c a v a l s h u n t s a c c o r d i n g t o a techn i q u e previously described (Herz et al., 1 9 7 2 ) . T h e gastric assay was p e r f o r m e d 14 d a y s a f t e r t h e o p e r a t i o n w h e n h e p a t i c a t r o p h y is maximal. 2 c o n t r o l g r o u p s were used o f 10 rats each, one m a t c h e d with age ( c o n t r o l g r o u p I) and a s e c o n d for w e i g h t ( c o n t r o l g r o u p II). This was necessary as rats with p o r t a c a v a l s h u n t s lose a c o n s i d e r a b l e a m o u n t o f weight.
8, 16, 32 ng o f H G I were given t o the rats in b o t h groups a n d t h e d o s e - - r e s p o n s e curves c o m p a r e d . Rats with p o r t a c a v a l s h u n t s prod u c e d a b o u t d o u b l e the a m o u n t o f acid res p o n s e t o each dose o f gastrin t h a n did t h o s e in each c o n t r o l g r o u p (table 2).
3. 4. 5. Selected rats As t h e increase in sensitivity achieved b y p o r t a c a v a l s h u n t i n g was m o d e s t in relation t o t h e difficult and t i m e c o n s u m i n g operative p r o c e d u r e it was a i m e d t o p r o d u c e t h e similar result b y a less c u m b e r s o m e p r o c e d u r e . Rats were selected in this e x p e r i m e n t b y rejecting t h o s e w h i c h did n o t give a r e s p o n s e o f 30 # M h o s or m o r e t o an initial dose o f 16 ng o f H G I. 1, 2, 4 and 8 ng H G I were given t o 4 rats in a 4 X 4 L a t i n square design and
B I O A S S A Y O F G A S T R I N IN B O D Y F L U I D S
the resulting dose--response curve compared with the normal dose--response curve (table 2 and fig. 4). Each rat produced a clear response at the 1 ng level. The secretory response for each dose was significantly higher in comparison to the results obtained in unselected animals.
3.5. Quantitative and qualitative assays with plasma and tumour extracts from patients suffering from the Zollinger--Ellison syndrome 1 ml of native plasma from a patient with histologically proven benign gastrinoma was injected i.v. to a rat over 1rain. This induced an increase in acid secretion similar to that produced by a test dose of 8 ng HG I b y the same animal. The t u m o u r was sent to Prof. R.A. Gregory, who kindly prepared an extract. 0.5 ml of this extract produced gastric acid secretion approximately equivalent to 8 ng of HG I (fig. 5). In order to perform quantitative assays of gastrin in plasma from patients suffering from the Zollinger-Ellison syndrome it was necessary to establish that gastrin can be reliably measured when added to plasmas of healthy individuals. Some injections of plasma over 0.5 ml were badly tolerated by rats under 200 g. This problem was largely overcome by using heavier rats (220--240 g). Human plasma was tested for secretory activity in a preliminary experiment and HG I added to give 16 ng/mi. A 2 + 2 assay was performed, comparing 0.5 and 1 ml of the plasma to 8 and 16 ng of HG I standard. This pMhos
0
,'.
~
Fig. 5. 3 r e s p o n s e s p r o d u c e d b y 1 r a t a f t e r t h e inject i o n o f 0.5, 1.0 a n d 2.0 ml o f a n e x t r a c t f r o m a gollinger--Ellison tumour.
375
pMhos 80ii
S T
iI
60-
/'
JlIll I.,0-
Log Dose
8 16 ng GASTRINI 0.125 0.25 ml Plasma
Fig. 6. M e a n results o f a 2 + 2 L a t i n S q u a r e assay. Plasma (T) o b t a i n e d f r o m a patient with t h e Zollinger--Ellison s y n d r o m e was c o m p a r e d w i t h H G I(S). T h e vertical lines i n d i c a t e t h e s t a n d a r d e r r o r o f t h e mean.
resulted in almost identical secretory responses and the differences between standard and test were not significant at either dose level. A 2 + 2 assay was also performed with a plasma from a patient suffering from a metastasing non-~-cell islet carcinoma of the pancreas. 0.125 ml and 0.25 ml of the plasma resulted in a similar acid secretory response to that of 8 and 16 ng of the standard in a preliminary experiment. In the 2 + 2 assay dose--regression lines of standard and test were parallel (fig. 6). The gastrin content per ml of plasma was 60.2 ng (95% confidence limits 28.5--89.7 ng). A 2 + 2 assay was also performed on a t u m o u r extract from another patient, which resulted in a gastrin content of 38.5 ng (18.7--39.9 ng, 95% confidence limits). The data of this latter assay was submitted to an analysis of variance (table 3). The analysis showed that there was a significant regression and a difference between treatments. On the other hand there was no significant difference between standard and test, between the rats or between the order of
376
F. H A L T E R ET AL.
TABLE 3 Analysis o f variance o f a 2 + 2 assay for HG I a n d t u m o u r - e x t r a c t o f a p a t i e n t suffering f r o m t h e Zollinger--Ellison syndrome. Source o f variation
Sum o f square
D.F.
Mean square
F ratio
p
B e t w e e n s t a n d a r d and t e s t Regression Deviation for parallelism Between treatments Between columns B e t w e e n g r o u p (rats) Error Total
33.06 1072.56 0.06 1105.68 113.19 343.69 144.88 1707.44
1 1 1 3 3 3 6 15
33.06 1072.56 0.06 368.56 37.73 114.56 24.14
1.4 44.4 0.002 15.3 1.6 4.7
>0.05 0.05 < 0.01 :> 0.05 >0.05
responses (columns). There was no significant deviation from parallelism.
4. Discussion A reliable bioassay must be sufficiently sensitive to give dose-dependent responses to the available dose of the test material. The design should be simple enough for routine use. The standard preparations must be tested in the same experiment as the test material and in all respects under strictly comparable conditions. The test preparation should demonstrate similar biological activity as the clinical use of the material (Gaddum, 1959). The described assay more closely satisfies these criteria than any previous biological assay for gastrin. It is relatively simple and can be performed on a readily available animal. Since the implantation of the bulky intragastric cannula into the stomach reduces the dead space, all the acid secreted is quickly washed o u t and passes into the reservoir containing the measuring electrode. The responses t o gastrin are thus considerably shorter than those obtained in the early reperfusion system (Smith et al., 1970) as well as in chronic experiments using conscious rats with gastric fistulae (Halter et al., 1966) and the anaesthetized cat preparation (Blair et al., 1968). The reperfusion conductivity system with
the reduced dead space improves the definition of the stimulated response over the basal secretion. This is due to the shorter latency and the ease of registering the end of each response. The repeated changes in sensitivity of the cat preparation necessitated estimates of the acid secretion to be taken (Blair et al., 1968). Since the anaesthetized rat responds reproducibly to gastrin, the acid secreted could be accurately measured. Six successive doses of the standard can be administered to each animal without tachyphylaxis. Thus with this system all the treatments of a 2 + 2 assay can be given to each rat avoiding the use of incomplete blocks (Lai, 1964). This reduced the error due to the 'between animal' variation. Measuring the acid secreted b y conductivity has several advantages over the use of pH. The conductivity of the effluent is linearly related to the concentration of ions present. Due to the high proportion of H ÷ ions present compared with other contributing ions and the high equivalent conductance of H ÷ ions, the conductivity of the effluent is virtually that of H ÷ ions (Lawrence and Smith, 1974). The use of buffer solutions for the perfusate (Smith et al., 1970) or of an antilog generator (Parsons, 1969) is unnecessary. The overall cost of the equipment for registrating conductivity is cheaper and conductivity electrodes are more stable and easier to calibrate than pH electrodes. Threshold responses can be obtained with 1--2 ng of gastrin, if selected
BIOASSAY OF GASTRIN IN BODY FLUIDS
rats are used in this assay system, and quantitative assays are possible above 8 ng. In comparison to the early reperfusion systems as used by Colin-Jones et al. (1969) this method is not only simpler and more precise, but considerably more sensitive. In the attempts to increase the sensitivity further it was shown that neither vagotomy nor antrectomy influenced the sensitivity of the method. Limited success was achieved with attempts to inhibit the degradation of gastrin. The liver (Thompson et ai., 1969) the renal cortex (Jaffe and Newton, 1969) and the mucosa of the small bowel (Temperley and Stagg, 1971) have been implicated in the metabolism of gastrin. It is likely that nephroligature and subtotal small bowel resection were too traumatic for the animals to secrete more gastric acid. The increase in sensitivity to gastrin that was achieved by portacaval shunting is unlikely to be due to the bypass of the hepatic circulation, as the rate constant for elimination of gastrin was not different in shunted and control animals (Kutz et al., 1974). It is more likely that the parietal cells of the rats with portacavai shunts are more sensitive to gastrin, possibly due to an increased background stimulation by a substance released from the small intestine as postulated by Orloft et al. (1969). The net gain in sensitivity achieved by the complicated shunt operation is however limited and similar sensitivity thresholds can be achieved by using selected rats with a high initial sensitivity to a test dose of gastrin. Portacaval shunting was the only procedure to dramatically alter the rate of acid secretion. The large 'between rat' variation prevented the detection of subtle changes in the response of test animals to gastrin when the animals could not be used as their own controis. By rejecting rats that did not give an initially high response to gastrin it is possible to lower the detection threshold for the qualitative measurement o f gastrin levels in cases of suspected Zollinger-Eliison syndrome. Even though the sensitivity threshold of the gastrin immunoassay lies within the lower
377
pg range (McGuigan and Trudeau, 1968; Hansky and Cain, 1969; Yallow and Berson, 1970; Ganguli and Hunter, 1972) a correlation study with the bioassay would yield important information. The amount of gastrin measured by the immunoactive peptides is not necessarily identical to the amount of biologically active peptides present in the plasma. Considerably higher values are obtained using pentagastrin to produce the antibody than when the heptadecapeptide sequence of gastrin is used (Byrnes et al., 1970). Since the sole presence of gastrin cannot be shown by using a specific gastrin antagonist, the bioassay can only indicate net stimulant activity. Other gastric acid stimulants or inhibitors may be present in low concentrations. It is therefore not surprising that the initial comparisons of the gastrin bio- and immunoassays showed a poor correlation between the two methods. The bioassay results were consistently higher by a factor of 2--3 (Temperley and Stagg, 1971). A correlation of both assay techniques on a larger scale is thus necessary for the evaluation of the use of both techniques for the diagnosis of the Zollinger--Ellison syndrome. The results of such a study using this assay will be reported elsewhere.
Acknowledgements We thank Dr. Ph. Heitz, Basel and Dr. P. Fasel, Lausanne for kindly sending us plasma and tumour samples o f their patients, and Prof. R.A. Gregory, Liverpool and Dr. P.C. Ganguli for preparing the tumour extracts. We are greatly indebted to Mr. J. Birrer, ICI, Switzerland for the kind gift of synthetic human gastrin I.
References Black, J.W., W.A.M. Duncan, G.J. Durant, C.R. GaneUin and E.M. Parsons, 1972, Definition and antagonism o f histamine H2-receptors, Nature (London) 236, 385.
378 Blair, E.L., Y. Farra, D.D. Richardson and P. Steinbock, 1970, The half-life of exogenous gastrin in the circulation, J. Physiol. (London) 208, 299. Blair, E.L., A.A. Harper and J.D. Reed, 1962, An assay technique for gastrin, J. Physiol. (London) 163, 47P. Blair, E.L., R.M. Keenlyside, D.J. Newell, J.D. Reed and D.D. Richardson, 1968, Assay of gastrin by means of its gastric acid stimulating activity, J. Physiol. (London) 198, 613. Blair, E.L. and D.D. Wood, 1968, The estimation of gastrin activity in blood, J. Physiol. (London) 194, 44P. Byrnes, D.L., J.D. Young, D.J. Chishold and L. Lazarus, 1970, Serum gastrin in patients with peptic ulceration, Brit. Med. J. 2, 626. Colin-Jones, D.G., R.M.L. Copping, D.D. Gibbs and M.M. Sharr, 1969, Malignant Zollinger--Ellison syndrome with gastrin containing skin metastases, Lancet i, 492. Colin-Jones, D.G. and J.E. Lennard-Jones, 1972, The detection and measurement of circulating gastrinlike activity by bioassay, Gut 13, 88. Edkins, J.S., 1905, On the chemical mechanism of gastric secretion, Proc. Roy. Soc. London, Ser. B 76, 376. Gaddum, J.H., 1959, Pharmacology, 5th edn. (Oxford University Press, London) p. 508. Ganguli, P.C. and W.M. Hunter, 1972, Radio-immunoassay of gastrin in human plasma, J. Physiol. (London} 220, 499. Ghosh, M.N. and H.O. Schild, 1958, Continuous recording of acid gastric secretion in the rat, Brit. J. Pharmacol. Chemother. 13, 54. Halter, F., P.C. Ganguli, M. Marshall and G.P. Crean, 1966, Preliminary communication of the effect of gastrin II on acid secretion in conscious rats with chronic gastric fistula, Gastroenterologia 106, 194. Hansky, J. and M.D. Cain, 1969, Radio-immunoassay of gastrin in human serum, Lancet ii, 1388~ Herring, D.W. and E.L. Blair, 1969, Gastrin activity in the plasma of normal subjects and patients with duodenal ulceration, Brit. J. Surg. 56,707. Herz, R., V. Sautter, F. Robert and J. Bircher, 1972, The Eckfistula rat: definition of an experimental model, European J. Clin. Invest. 2, 390. Ingrain, G., D.W. Herring, C.W. Venables and E.L. Blair, 1971, Plasma-gastrin activity in patients with recurrent peptic Ulceration, Brit. J. Surg. 58, 298. Jaffe, B.M. and W.T. Newton, 1969, Distribution and localization of radio-iodinated gastrin, Surg. Forum. 20, 312. Komarov, S.A., 1942, Studies on gastrin. I. Methods of isolation of a specific gastrin secretagogue from the pyloric mucous membrane and its chemical properties, Rev. Can. Biol. 1 , 1 9 1 . Komarov, S.A., H. Shay, M. Rayport and S.S. Fels, 1944, Some observations on gastric secretion in normal rats, Gastroenterology 3, 406.
F. HALTER ET AL. Kutz, K., R. Herz, F. Halter, J. Leuenberger and J. Bircher, 1974, Increased responsiveness of gastric mucosa to gastrin stimulation in the Eck-fistula rat, Gastroenterology 66, 73. Lai, K.S., 1964, Studies on gastrin. Part I. A method of biological assay of gastrin, Gut 5, 327. Lawrence, A.J. and G.M. Smith, 1969, Measurement of gastric acid secretion by conductivity, Brit. J. Pharmacol. Chemotherap. 35, 371P. Lawrence, A.J. and G.M. Smith, 1974, Measurement of gastric acid secretion by conductivity, European J. Pharmacol. 25, 383. McGuigan, J.E. and W.L. Trudeau, 1968, Immunochemical measurement of elevated levels of gastrin in the serum of patients with pancreatic tumor of the Zollinger--Ellison variety, New Engl. J. Med. 278, 1308. Moore, F.T., J.E. Murat, G.L. Endahl, J.L. Baker and R.M. Zollinger, 1967, Diagnosis of ulcerogenic tumor of the pancreas by bioassay, Amer. J. Surg. 113, 735. Murray-Lyon, I.M., A.L.W.F. Eddleston, R. William, M. Brown, B.M. Hogbin, A. Bennett, J.C. Edwards and K.W. Taylor, 1968, Treatment of multiplehormone-producting malignant islet-cell tumour with streptozotocin, Lancet ii, 895. Orloff, M.J., H. Villar-Valdes, H. Rosen, A.G. Thompson and J.G. Chandler, 1969, Humoral medication of the intestinal phase of gastric secretion and of acid hypersecretion associated with portacaval shunts, Surgery 66, 118. Parsons, M.E., 1969, Quantitative studies of drug-induced acid gastric secretion, Ph.D. Thesis, University of London. Smith, G.M., A.J. Lawrence, D.G. Colin-Jones and H.O. Schild, 1970, The assay of gastrin using the perfused rat stomach, Brit. J. Pharmacol. Chemotherap. 38, 206. Temperley, J.M. and B.H. Stagg, 1971, Bioassay and radio-immunoassay of plasma gastrin in a case of Zollinger--Ellison syndrome, Seand. J. Gastroenterol. 6, 735. Thompson, C.G., I.G.M. Cleator and W. Sircus, 1970, Experiences with a rat bioassay in the diagnosis of the Zollinger--Ellison syndrome, Gut 11,409. Thomspon, J.C., D.D. Reeder, W.D. Davidson, A.C. Charters, W.L. Bruckner, C.A.E. Lemmi and J.H. Miller, 1969, Effect of hepatic transit of gastrin, pentagastrin and histamine measured by gastric secretion and by assay of hepatic vein blood, Ann. Surg. 170, 493. Uvn~s, B. and S. E m ~ , 1961, A method for biologic assay of gastrin, Gastroenterology 40, 644. Wilson, S.D., J.A. Mathison, W.J. Schulte and E.H. Ellison, 1968, The role of bioassay in the diagnosis of ulcerogenic tumor of the pancreas by bioassay, Archs. Surg. 97, 437. Yallow, R.S. and S.A. Berson, 1970, Radio-immunoassay of gastrin, GastroenteroLogy 58, 1.
THE DEVELOPMENT OF A RELIABLE AND SENSITIVE BIOASSAY FOR GASTRIN IN BODY FLUIDS F R E D HALTER, B E R N H A R D and LAJOS V A R G A
KOHLER, J O H A N N A L E U E N B E R G E R , G O R D O N
M. SMITH*
Gastrointestinal Unit, University Hospital, Inselspital, Berne, Switzerland Received 24 March 1975, revised MS received 28 August 1975, accepted 9 September 1975
F. HALTER, B. KOHLER, J. LEUENBERGER, G.M. SMITH and L. VARGA, The development of a reliable and sensitive bioassay for gastrin in body fluids, European J. Pharmacol. 35 (1976) 369---378. The sensitivity and reliability of the Ghosh and Schild rat stomach preparation was improved by implantation of bulky intragastric cannulae, recirculation of the perfusate and measurement of conductivity instead of pH. With these procedures, the threshold values for gastrin I were in the range of 1--2 ng. In attempts to increase the sensitivity further it was shown that neither vagotomy nor antrectomy influenced the sensitivity of the method. The threshold values were not lowered by attempts to reduce gastrin metabolism through nephroligature or small bowel resection. Although rats fitted with portacaval shunts did ,raise the threshold, the limited increase in sensitivity was also achieved by selecting rats that gave an initially high response to a test dose of gastrin thus avoiding the complicated shunt operation. The assay procedure enabled statistically valid measurements of gastrin in plasma and from tumour extracts from patients suffering from the Zollinger--Ellison Syndrome to be assayed using a 2 + 2 block design. Perfused rat stomach preparation
Zollinger--Ellison syndrome
1. Introduction Early studies on the identification and extraction of gastrin were performed using the stomach preparations of anaesthetized and nonanaesthetized cats (Edkins, 1905; Komaroy, 1942; Uvn~ and Em~s, 1961; Blair et al., 1962). Subsequently, the anaesthetized cat preparation was used to estimate circulating gastrin in the ng range (Blair et al., 1968; Blair and Wood, 1968; Herring and Blair, 1969; Blair et al., 1970; Ingram et al., 1971). This bioassay has certain limitations. Since the anaesthetized cat preparation does not give a reliable secretion of acid, acid production has had to be estimated during continuous changes in sensi* School of Pharmacy, Robert Gordon's Institute of Technology, Schoolhill, Aberdeen, Scotland.
Gastrin
Bioassay
tivity. As the true threshold limit for synthetic human gastrin is 50 ng (Blair et al., 1970) volumes of 50 ml of plasma have to be injected per animal to estimate gastrin in the 1 ng range per animal which is the level of circulating gastrin in patients with the Zollinger--Ellison syndrome (Blair and Wood, 1968). Furthermore the use of the cat renders such a bioassay exceedingly expensive. As an alternative animal the rat has been widely used since Komarov et al. (1944) has shown that crude extracts of gastrin stimulate gastric acid production in the anaesthetized rat. Lai (1964) used the rat stomach preparation (Ghosh and Schild, 1958) to assay gastrin with a statistically valid experimental design. Although the assay was not sensitive enough to give reliable quantitative results for circulating gastrin, it was used by many workers in the attempt to diagnose the Zollinger--Ellison
370
syndrome pre-operatively by qualitative or semi-quantitative assays (Moore et al., 1967; Wilson et al., 1968; Murray-Lyon et al., 1968; Thompson et al., 1970). The reperfusion assay developed as a modification of the Ghosh and Schild procedure by Smith et al. (1970) enabled Colin-Jones et al. (1969) to assay the gastrin content of plasma from a patient with the ZoUinger--Ellison syndrome and to estimate the gastrin subsequently found in the tumour. This was achieved with fiducial limits of about 20% which is acceptable for this type of assay. The sensitivity threshold in their assay was however limited and reached 10 ng, and quantitative assays were only possible in the range of 50--200 ng (Colin-Jones and Lennard-Jones, 1972). The use of the more sensitive gastrin immunoassay to diagnose the Zollinger--Ellison syndrome (McGuigan and Trudeau, 1968) indicated that the biological assay would n o t detect many cases of this disease. The immunoassay will however not replace the biological assay since the former does not indicate if the immunoreactive material actually causes gastric secretion. Immunoassay and bioassay techniques should therefore be used alongside one another. Ideally b o t h methods should be equally sensitive and precise. In recent years a number of technical modifications have improved the sensitivity and reliability of the Ghosh and Schild preparation. Continuous reperfusion of the gastric effluent integrated the acid secreted thus increasing the sensitivity of the preparation (Smith et al., 1970). Measuring the conductivity of the gastric effluent instead of pH has greatly improved the linearity and reliability of the measurements of the acid secreted (Lawrence and Smith, 1969, 1974). A large cannula was inserted into the stomach to reduce the intragastric dead space (Parsons, 1969; Black et al., 1972). This enabled the secreted acid to be washed out more efficiently which significantly shortened the duration of the acid response. The work reported here describes an a t t e m p t to increase the sensitivity of the Ghosh and Schild rat preparation to gastrin by combining the most
F. HALTER ET AL. POTENTIOMETR!C RECORDER 1 I CONDUCTOSCOPE ]
STIRRER f HEATING COIL
o
, ,t
J OESOPHAGUS I I '~ I l
3CALEI Fig. 1. Diagrammatic representation of the reperfusion conductivity system for measuring gastric acid secretion.
relevant recent technical improvements described above. This work was performed to facilitate the preoperative diagnosis of the Zollinger--Ellison syndrome and to correlate the gastrin bio- and immunoassays on a large scale.
2. Material and methods 2.1. General
Male Wistar rats (140--240 g) were housed singly in special cages to prevent coprophagy. 18 hr before each experiment the food was replaced by sugar cubes and water was available ad libitum. The rats were anaesthetized by an intramuscular injection of urethane (1.5 g/kg b o d y weight) and small additional doses being given when necessary. Both the trachea and the external jugular vein were cannulated. The a b d o m e n was opened through a midline incision 3--4 cm long. Damage to the parietal cell-containing part of the stomach (corpus) was meticulously avoided throughout the operation. The oesophagus was ligated 5 mm above the cardia taking care to avoid the vagal
BIOASSAY OF GASTRIN IN BODY FLUIDS
371
nerves. An incision was made into the rumen (fundus) and the stomach was gently washed with warm saline. A perspex cannula with a ball-shaped head was inserted through the incision in the rumen and the tip passed o u t through a second opening near the pylorus. A large asymmetric collecting funnel was then introduced into the rumen. The eccentric part of the cannula was placed near the cardia to reduce the dead space in that part of the stomach {fig. 1).
the funnel-shaped cannula and delivered into a beaker containing the conductivity cell (type Siemens 9510). The perfusate in the beaker was mixed b y an electromagnetic stirrer. The temperature of the perfusate, operating table and beaker was held constant at 37°C b y an external heating system. The conductivity cell was connected to a conductoscope (Methrom 365 B) and the conductance of the perfusate recorded by a potentiometric recorder.
2. 2. Vagotomy
2. 5. Calibration o f the measuring system
Vagal fibres were exposed in the cervical region and care taken to avoid the abdominal vagi during gastric surgery. When the vagi were cut there was a change in the respiration rate. Successful cervical vagotomy was confirmed by testing each rat with 0.1 units of regular insulin at the end of each experiment. All animals failed to respond to insulin while a near maximal acid secretion of long duration was produced in control animals.
Aliquots of 0.1 ml N/100 HC1 were added to 20 ml of 5% glucose solution at 37°C, continuously stirred and the conductivity of the solution measured. The relationship between I-I* concentration and conductance was linear between 0--300/~Mhos (fig. 2). Human gastric juice was collected after stimulation with pentagastrin 6 pg/kg. The juice was diluted 1 : 10 and the acid concentration was 1.1 #Eq/0.1 ml. Aliquots of 0.1 ml of the diluted juice were added to 20 ml of 5% glucose and conductivity correlated to the acidity. The slopes of both curves are almost identical (fig. 2).
2. 3. Antrectomy The distal antral vessels were severed by electrocoagulation, carefully avoiding damage to the gastro-epiploic arteries. The antrum was cut at the pylorus and blood vessels of the upper part of the lesser curvature situated 1--2 mm distal to the cardia were cut by electrocoagulation before the antrum was removed. The ball-shaped cannula was then introduced and ligated at the antral-corporal junction. At the end of the experiment, the stomach was opened and inspected. Since the border between the antrum and corpus is clearly visible in the rat incomplete antrectomy was easily detected. These rats were eliminated from the study.
IJMhos
16oj
2. 4. Perfusion o f the gastric lumen The stomach was continuously perfused at a rate of 6 ml/min with 28 ml of 5% glucose solution and the perfusate collected through
8
.1'2 'pEq HCl
Fig. 2. Calibration curve for HCI and diluted gastric juice (GJ).
372
F. H A L T E R E T AL. Conductivity (~Mhos)
2. 6. Preparation o f standard and test
200
2. 6.1. Standard Synthetic human gastrin I (HG I) (Imperial Chemical Industries, Mereside, Macclesfield, England) was used as a standard throughout the experiments. Ampoules containing 1 mg were diluted with 0.05 M ammonium bicarbonate and divided in aliquots containing 50 pg. Each sample was lyophilised and stored at --20 ° C. Fresh samples were made up for each experiment and HG I dissolved in 0.5 M ammonium bicarbonate in concentrations of 1 0 - 6 4 0 ng per ml so that each standard dose had a volume of 0.1 ml in the range of 1--64 ng used in this study. 2.6.2. Test In the experiments where circulating gastrin was assayed, blood was withdrawn after an overnight fast. A drop of heparin was added to the syringe and 20--40 ml of blood was collected. The plasma was separated by high speed centrifugation (15,000 rpm) and stored in aliquots a t - - 2 0 to --30 ° C. Before the test, the material was thawed to n o t more than 37°C and recentrifuged before injection. The plasma was given as an intravenous infusion over 5 min; the maximum volumes injected were 1.0 ml per dose. 2. 6. 3. Statistical analysis 2 + 2 assays were performed using Latin square designs with 4 rats per assay.
Conductivity (/~Mhos) 2,, /..(V~
,,, . / ~ '
8n, / / / ~ "
0 ~~n_g . . ~ ~ 132 6n9
e4.~i
Fig. 3. 6 r e s p o n s e s o b t a i n e d in 1 rat a f t e r t h e i.v. i n j e c t i o n o f 6 d i f f e r e n t doses of HG I.
3. Results
3.1. Qualitative and quantitative evaluation o f acid secretory response to HG I After an i.v. injection of 16 ng of HG I.an acid secretory response was achieved in each animal. An injection of gastrin produced an increase in the steepness of the slope of the conductivity response (fig. 3). After peak acid secretion the slope gradually returned to the basal level. The response was measured b y recording the difference in conductivity between the extrapolated baseline and the postsecretory level (Smith et al., 1970; Lawrence and Smith, 1974). The reproducibility of repeated responses
TABLE 1 Analysis o f variance. 8 rats were given 6 c o n s e c u t i v e doses o f 16 ng HGI. S o u r c e of v a r i a t i o n
D.F.
Sum of square
M e a n square
F ratio
p 0.2
B I O A S S A Y O F G A S T R I N IN B O D Y F L U I D S
was found b y giving each of 8 rats 7 successive doses of 16 ng HG I. Total secretion, latency of onset and total duration of each response were measured. Following previously published studies (Lai, 1964; Smith et al., 1970) the first response to gastrin was n o t evaluated. The 'within rats' variation of the successive responses was very small, without tachyphylaxis. There was a fourfold difference in the 'between rat' variation (table 1). There was little fluctuation in latency and total duration of response between and within rats. The mean latency was 1.06 min + 0.244 (S.D.). 3. 2. Dose--response curve 6 doses (2--64 ng) of HG I were injected into 6 rats using a Latin square design and the responses obtained from one rat are shown in fig. 3. A dose--response curve was constructed from the mean results. Threshold values were a b o u t 2 ng and the steep part of the curve started at a b o u t 8--10 ng (fig. 4).
373
pMhos 90~
6O
30L
i
1'6 3'2 g4
LOG DOSE GASTRINT..inng Fig. 4. Dose---response curves for gastrin obtained in normal and selected rats. The vertical lines represent the standard error of the mean.
As shown in table 2 there was no difference in the sensitivity of the rats to gastrin before and after vagotomy.
3. 3. The effects o f antrectomy and vagotomy on stimulated acid secretion
3.4. A t t e m p t s to increase the sensitivity o f the assay
Since the complicated operative procedure may have a negative influence on gastric acid secretion, the effects of antrectomy and vagotomy were studied.
3. 4.1. General Theoretically it should be possible to increase the sensitivity of this system b y inhibiting the degradation of gastrin. Attempts were therefore made to increase the sensitivity of the assay s y s t e m b y avoiding the renal circulation (nephroligature), b y small bowel resection and by portacaval shunting.
3. 3.1. A n t r e c t o m y As it was technically not feasible to study gastric acid secretion before and after antrect o m y in the same rat, these experiments were performed on two batches of 6 animals matched for age and b o d y weight. 3 doses of HG I (8, 16, 32 ng) were given in random order to rats in b o t h groups and dose--response curves constructed. The slopes of both curves were similar and there was no significant difference at any dose level (table 2). 3. 3. 2. Vagotomy 8 and 16 ng HG I were given to 4 animals before and after cervical section of both vagi.
3. 4.2. Nephroligature 20 rat~ Were prepared for gastric acid secretory studms according to the standard procedure. g and 16 ng of HG I were administered before and 30 min after bilateral ligature of the renal vessles. There was no difference in the secretory response at b o t h dose levels studied (table 2). 3. 4. 3. Subtotal small bowel resection Using a similar procedure as for nephroliga-
F. HALTER ET AL.
374 TABLE 2 Attempts to increase the sensitivity of the assay. Technique
No. of observations
Dose of gastrin (rig)
Control
Test
Mean responses (pMhos)
-+ S.E.
Mean responses (pMhos)
t-test ~ S.E.
Antrectomy
6 6 6
8 16 32
9.5 20 50.2
0.83 0.8 2.72
5.8 17.3 46.8
0.8 2.35 6.63
NS NS NS
Vagotomy
4 4
8 16
18.8 31.3
3.95 4.92
15.5 26.1
5.38 7.98
NS NS
Nephroligature
20 20
8 16
19.0 39.9
2.35 3.15
22.3 51.9
2.31 5.64
NS NS
Small bowel resection
20 20
8 16
13.4 34.3
1.45 4.18
13.7 33.7
1.71 2.71
NS NS
Portocaval shunt*
10 10 10
8 16 32
7.3 20.2 54.7
1.59 2.88 6.89
18.0 41.4 90.2
4.07 3.85 11.68
p < 0.05 p < 0.01 p < 0.01
1 2 4 8 16
1.6 3.1 9.5 20.7
0.26 0.29 0.81 0.78
1.6 3.6 9.25 15.5
0.63 0.80 0.63 2.26
p < 0.025 p < 0.001 p < 0.025
Selected rats**
4 6(4) 6(4) 6(4) 6
* Control = control group I in text (animals same weight as test animals). Control group II (animals same age as test animals) also produced significantly less gastric acid secretion. ** 6 controls used and 4 test animals.
ture, s u b t o t a l small b o w e l r e s e c t i o n was perf o r m e d in 20 rats. T h e small intestine was removed by electrosurgery from the ligament o f Treitz t o t h e ileocoecal valve. Again n o significant d i f f e r e n c e in s e c r e t i o n was achieved (table 2).
3.4.4. Portacaval shunting A g r o u p o f 10 animals was fitted w i t h e n d to side p o r t a c a v a l s h u n t s a c c o r d i n g t o a techn i q u e previously described (Herz et al., 1 9 7 2 ) . T h e gastric assay was p e r f o r m e d 14 d a y s a f t e r t h e o p e r a t i o n w h e n h e p a t i c a t r o p h y is maximal. 2 c o n t r o l g r o u p s were used o f 10 rats each, one m a t c h e d with age ( c o n t r o l g r o u p I) and a s e c o n d for w e i g h t ( c o n t r o l g r o u p II). This was necessary as rats with p o r t a c a v a l s h u n t s lose a c o n s i d e r a b l e a m o u n t o f weight.
8, 16, 32 ng o f H G I were given t o the rats in b o t h groups a n d t h e d o s e - - r e s p o n s e curves c o m p a r e d . Rats with p o r t a c a v a l s h u n t s prod u c e d a b o u t d o u b l e the a m o u n t o f acid res p o n s e t o each dose o f gastrin t h a n did t h o s e in each c o n t r o l g r o u p (table 2).
3. 4. 5. Selected rats As t h e increase in sensitivity achieved b y p o r t a c a v a l s h u n t i n g was m o d e s t in relation t o t h e difficult and t i m e c o n s u m i n g operative p r o c e d u r e it was a i m e d t o p r o d u c e t h e similar result b y a less c u m b e r s o m e p r o c e d u r e . Rats were selected in this e x p e r i m e n t b y rejecting t h o s e w h i c h did n o t give a r e s p o n s e o f 30 # M h o s or m o r e t o an initial dose o f 16 ng o f H G I. 1, 2, 4 and 8 ng H G I were given t o 4 rats in a 4 X 4 L a t i n square design and
B I O A S S A Y O F G A S T R I N IN B O D Y F L U I D S
the resulting dose--response curve compared with the normal dose--response curve (table 2 and fig. 4). Each rat produced a clear response at the 1 ng level. The secretory response for each dose was significantly higher in comparison to the results obtained in unselected animals.
3.5. Quantitative and qualitative assays with plasma and tumour extracts from patients suffering from the Zollinger--Ellison syndrome 1 ml of native plasma from a patient with histologically proven benign gastrinoma was injected i.v. to a rat over 1rain. This induced an increase in acid secretion similar to that produced by a test dose of 8 ng HG I b y the same animal. The t u m o u r was sent to Prof. R.A. Gregory, who kindly prepared an extract. 0.5 ml of this extract produced gastric acid secretion approximately equivalent to 8 ng of HG I (fig. 5). In order to perform quantitative assays of gastrin in plasma from patients suffering from the Zollinger-Ellison syndrome it was necessary to establish that gastrin can be reliably measured when added to plasmas of healthy individuals. Some injections of plasma over 0.5 ml were badly tolerated by rats under 200 g. This problem was largely overcome by using heavier rats (220--240 g). Human plasma was tested for secretory activity in a preliminary experiment and HG I added to give 16 ng/mi. A 2 + 2 assay was performed, comparing 0.5 and 1 ml of the plasma to 8 and 16 ng of HG I standard. This pMhos
0
,'.
~
Fig. 5. 3 r e s p o n s e s p r o d u c e d b y 1 r a t a f t e r t h e inject i o n o f 0.5, 1.0 a n d 2.0 ml o f a n e x t r a c t f r o m a gollinger--Ellison tumour.
375
pMhos 80ii
S T
iI
60-
/'
JlIll I.,0-
Log Dose
8 16 ng GASTRINI 0.125 0.25 ml Plasma
Fig. 6. M e a n results o f a 2 + 2 L a t i n S q u a r e assay. Plasma (T) o b t a i n e d f r o m a patient with t h e Zollinger--Ellison s y n d r o m e was c o m p a r e d w i t h H G I(S). T h e vertical lines i n d i c a t e t h e s t a n d a r d e r r o r o f t h e mean.
resulted in almost identical secretory responses and the differences between standard and test were not significant at either dose level. A 2 + 2 assay was also performed with a plasma from a patient suffering from a metastasing non-~-cell islet carcinoma of the pancreas. 0.125 ml and 0.25 ml of the plasma resulted in a similar acid secretory response to that of 8 and 16 ng of the standard in a preliminary experiment. In the 2 + 2 assay dose--regression lines of standard and test were parallel (fig. 6). The gastrin content per ml of plasma was 60.2 ng (95% confidence limits 28.5--89.7 ng). A 2 + 2 assay was also performed on a t u m o u r extract from another patient, which resulted in a gastrin content of 38.5 ng (18.7--39.9 ng, 95% confidence limits). The data of this latter assay was submitted to an analysis of variance (table 3). The analysis showed that there was a significant regression and a difference between treatments. On the other hand there was no significant difference between standard and test, between the rats or between the order of
376
F. H A L T E R ET AL.
TABLE 3 Analysis o f variance o f a 2 + 2 assay for HG I a n d t u m o u r - e x t r a c t o f a p a t i e n t suffering f r o m t h e Zollinger--Ellison syndrome. Source o f variation
Sum o f square
D.F.
Mean square
F ratio
p
B e t w e e n s t a n d a r d and t e s t Regression Deviation for parallelism Between treatments Between columns B e t w e e n g r o u p (rats) Error Total
33.06 1072.56 0.06 1105.68 113.19 343.69 144.88 1707.44
1 1 1 3 3 3 6 15
33.06 1072.56 0.06 368.56 37.73 114.56 24.14
1.4 44.4 0.002 15.3 1.6 4.7
>0.05 0.05 < 0.01 :> 0.05 >0.05
responses (columns). There was no significant deviation from parallelism.
4. Discussion A reliable bioassay must be sufficiently sensitive to give dose-dependent responses to the available dose of the test material. The design should be simple enough for routine use. The standard preparations must be tested in the same experiment as the test material and in all respects under strictly comparable conditions. The test preparation should demonstrate similar biological activity as the clinical use of the material (Gaddum, 1959). The described assay more closely satisfies these criteria than any previous biological assay for gastrin. It is relatively simple and can be performed on a readily available animal. Since the implantation of the bulky intragastric cannula into the stomach reduces the dead space, all the acid secreted is quickly washed o u t and passes into the reservoir containing the measuring electrode. The responses t o gastrin are thus considerably shorter than those obtained in the early reperfusion system (Smith et al., 1970) as well as in chronic experiments using conscious rats with gastric fistulae (Halter et al., 1966) and the anaesthetized cat preparation (Blair et al., 1968). The reperfusion conductivity system with
the reduced dead space improves the definition of the stimulated response over the basal secretion. This is due to the shorter latency and the ease of registering the end of each response. The repeated changes in sensitivity of the cat preparation necessitated estimates of the acid secretion to be taken (Blair et al., 1968). Since the anaesthetized rat responds reproducibly to gastrin, the acid secreted could be accurately measured. Six successive doses of the standard can be administered to each animal without tachyphylaxis. Thus with this system all the treatments of a 2 + 2 assay can be given to each rat avoiding the use of incomplete blocks (Lai, 1964). This reduced the error due to the 'between animal' variation. Measuring the acid secreted b y conductivity has several advantages over the use of pH. The conductivity of the effluent is linearly related to the concentration of ions present. Due to the high proportion of H ÷ ions present compared with other contributing ions and the high equivalent conductance of H ÷ ions, the conductivity of the effluent is virtually that of H ÷ ions (Lawrence and Smith, 1974). The use of buffer solutions for the perfusate (Smith et al., 1970) or of an antilog generator (Parsons, 1969) is unnecessary. The overall cost of the equipment for registrating conductivity is cheaper and conductivity electrodes are more stable and easier to calibrate than pH electrodes. Threshold responses can be obtained with 1--2 ng of gastrin, if selected
BIOASSAY OF GASTRIN IN BODY FLUIDS
rats are used in this assay system, and quantitative assays are possible above 8 ng. In comparison to the early reperfusion systems as used by Colin-Jones et al. (1969) this method is not only simpler and more precise, but considerably more sensitive. In the attempts to increase the sensitivity further it was shown that neither vagotomy nor antrectomy influenced the sensitivity of the method. Limited success was achieved with attempts to inhibit the degradation of gastrin. The liver (Thompson et ai., 1969) the renal cortex (Jaffe and Newton, 1969) and the mucosa of the small bowel (Temperley and Stagg, 1971) have been implicated in the metabolism of gastrin. It is likely that nephroligature and subtotal small bowel resection were too traumatic for the animals to secrete more gastric acid. The increase in sensitivity to gastrin that was achieved by portacaval shunting is unlikely to be due to the bypass of the hepatic circulation, as the rate constant for elimination of gastrin was not different in shunted and control animals (Kutz et al., 1974). It is more likely that the parietal cells of the rats with portacavai shunts are more sensitive to gastrin, possibly due to an increased background stimulation by a substance released from the small intestine as postulated by Orloft et al. (1969). The net gain in sensitivity achieved by the complicated shunt operation is however limited and similar sensitivity thresholds can be achieved by using selected rats with a high initial sensitivity to a test dose of gastrin. Portacaval shunting was the only procedure to dramatically alter the rate of acid secretion. The large 'between rat' variation prevented the detection of subtle changes in the response of test animals to gastrin when the animals could not be used as their own controis. By rejecting rats that did not give an initially high response to gastrin it is possible to lower the detection threshold for the qualitative measurement o f gastrin levels in cases of suspected Zollinger-Eliison syndrome. Even though the sensitivity threshold of the gastrin immunoassay lies within the lower
377
pg range (McGuigan and Trudeau, 1968; Hansky and Cain, 1969; Yallow and Berson, 1970; Ganguli and Hunter, 1972) a correlation study with the bioassay would yield important information. The amount of gastrin measured by the immunoactive peptides is not necessarily identical to the amount of biologically active peptides present in the plasma. Considerably higher values are obtained using pentagastrin to produce the antibody than when the heptadecapeptide sequence of gastrin is used (Byrnes et al., 1970). Since the sole presence of gastrin cannot be shown by using a specific gastrin antagonist, the bioassay can only indicate net stimulant activity. Other gastric acid stimulants or inhibitors may be present in low concentrations. It is therefore not surprising that the initial comparisons of the gastrin bio- and immunoassays showed a poor correlation between the two methods. The bioassay results were consistently higher by a factor of 2--3 (Temperley and Stagg, 1971). A correlation of both assay techniques on a larger scale is thus necessary for the evaluation of the use of both techniques for the diagnosis of the Zollinger--Ellison syndrome. The results of such a study using this assay will be reported elsewhere.
Acknowledgements We thank Dr. Ph. Heitz, Basel and Dr. P. Fasel, Lausanne for kindly sending us plasma and tumour samples o f their patients, and Prof. R.A. Gregory, Liverpool and Dr. P.C. Ganguli for preparing the tumour extracts. We are greatly indebted to Mr. J. Birrer, ICI, Switzerland for the kind gift of synthetic human gastrin I.
References Black, J.W., W.A.M. Duncan, G.J. Durant, C.R. GaneUin and E.M. Parsons, 1972, Definition and antagonism o f histamine H2-receptors, Nature (London) 236, 385.
378 Blair, E.L., Y. Farra, D.D. Richardson and P. Steinbock, 1970, The half-life of exogenous gastrin in the circulation, J. Physiol. (London) 208, 299. Blair, E.L., A.A. Harper and J.D. Reed, 1962, An assay technique for gastrin, J. Physiol. (London) 163, 47P. Blair, E.L., R.M. Keenlyside, D.J. Newell, J.D. Reed and D.D. Richardson, 1968, Assay of gastrin by means of its gastric acid stimulating activity, J. Physiol. (London) 198, 613. Blair, E.L. and D.D. Wood, 1968, The estimation of gastrin activity in blood, J. Physiol. (London) 194, 44P. Byrnes, D.L., J.D. Young, D.J. Chishold and L. Lazarus, 1970, Serum gastrin in patients with peptic ulceration, Brit. Med. J. 2, 626. Colin-Jones, D.G., R.M.L. Copping, D.D. Gibbs and M.M. Sharr, 1969, Malignant Zollinger--Ellison syndrome with gastrin containing skin metastases, Lancet i, 492. Colin-Jones, D.G. and J.E. Lennard-Jones, 1972, The detection and measurement of circulating gastrinlike activity by bioassay, Gut 13, 88. Edkins, J.S., 1905, On the chemical mechanism of gastric secretion, Proc. Roy. Soc. London, Ser. B 76, 376. Gaddum, J.H., 1959, Pharmacology, 5th edn. (Oxford University Press, London) p. 508. Ganguli, P.C. and W.M. Hunter, 1972, Radio-immunoassay of gastrin in human plasma, J. Physiol. (London} 220, 499. Ghosh, M.N. and H.O. Schild, 1958, Continuous recording of acid gastric secretion in the rat, Brit. J. Pharmacol. Chemother. 13, 54. Halter, F., P.C. Ganguli, M. Marshall and G.P. Crean, 1966, Preliminary communication of the effect of gastrin II on acid secretion in conscious rats with chronic gastric fistula, Gastroenterologia 106, 194. Hansky, J. and M.D. Cain, 1969, Radio-immunoassay of gastrin in human serum, Lancet ii, 1388~ Herring, D.W. and E.L. Blair, 1969, Gastrin activity in the plasma of normal subjects and patients with duodenal ulceration, Brit. J. Surg. 56,707. Herz, R., V. Sautter, F. Robert and J. Bircher, 1972, The Eckfistula rat: definition of an experimental model, European J. Clin. Invest. 2, 390. Ingrain, G., D.W. Herring, C.W. Venables and E.L. Blair, 1971, Plasma-gastrin activity in patients with recurrent peptic Ulceration, Brit. J. Surg. 58, 298. Jaffe, B.M. and W.T. Newton, 1969, Distribution and localization of radio-iodinated gastrin, Surg. Forum. 20, 312. Komarov, S.A., 1942, Studies on gastrin. I. Methods of isolation of a specific gastrin secretagogue from the pyloric mucous membrane and its chemical properties, Rev. Can. Biol. 1 , 1 9 1 . Komarov, S.A., H. Shay, M. Rayport and S.S. Fels, 1944, Some observations on gastric secretion in normal rats, Gastroenterology 3, 406.
F. HALTER ET AL. Kutz, K., R. Herz, F. Halter, J. Leuenberger and J. Bircher, 1974, Increased responsiveness of gastric mucosa to gastrin stimulation in the Eck-fistula rat, Gastroenterology 66, 73. Lai, K.S., 1964, Studies on gastrin. Part I. A method of biological assay of gastrin, Gut 5, 327. Lawrence, A.J. and G.M. Smith, 1969, Measurement of gastric acid secretion by conductivity, Brit. J. Pharmacol. Chemotherap. 35, 371P. Lawrence, A.J. and G.M. Smith, 1974, Measurement of gastric acid secretion by conductivity, European J. Pharmacol. 25, 383. McGuigan, J.E. and W.L. Trudeau, 1968, Immunochemical measurement of elevated levels of gastrin in the serum of patients with pancreatic tumor of the Zollinger--Ellison variety, New Engl. J. Med. 278, 1308. Moore, F.T., J.E. Murat, G.L. Endahl, J.L. Baker and R.M. Zollinger, 1967, Diagnosis of ulcerogenic tumor of the pancreas by bioassay, Amer. J. Surg. 113, 735. Murray-Lyon, I.M., A.L.W.F. Eddleston, R. William, M. Brown, B.M. Hogbin, A. Bennett, J.C. Edwards and K.W. Taylor, 1968, Treatment of multiplehormone-producting malignant islet-cell tumour with streptozotocin, Lancet ii, 895. Orloff, M.J., H. Villar-Valdes, H. Rosen, A.G. Thompson and J.G. Chandler, 1969, Humoral medication of the intestinal phase of gastric secretion and of acid hypersecretion associated with portacaval shunts, Surgery 66, 118. Parsons, M.E., 1969, Quantitative studies of drug-induced acid gastric secretion, Ph.D. Thesis, University of London. Smith, G.M., A.J. Lawrence, D.G. Colin-Jones and H.O. Schild, 1970, The assay of gastrin using the perfused rat stomach, Brit. J. Pharmacol. Chemotherap. 38, 206. Temperley, J.M. and B.H. Stagg, 1971, Bioassay and radio-immunoassay of plasma gastrin in a case of Zollinger--Ellison syndrome, Seand. J. Gastroenterol. 6, 735. Thompson, C.G., I.G.M. Cleator and W. Sircus, 1970, Experiences with a rat bioassay in the diagnosis of the Zollinger--Ellison syndrome, Gut 11,409. Thomspon, J.C., D.D. Reeder, W.D. Davidson, A.C. Charters, W.L. Bruckner, C.A.E. Lemmi and J.H. Miller, 1969, Effect of hepatic transit of gastrin, pentagastrin and histamine measured by gastric secretion and by assay of hepatic vein blood, Ann. Surg. 170, 493. Uvn~s, B. and S. E m ~ , 1961, A method for biologic assay of gastrin, Gastroenterology 40, 644. Wilson, S.D., J.A. Mathison, W.J. Schulte and E.H. Ellison, 1968, The role of bioassay in the diagnosis of ulcerogenic tumor of the pancreas by bioassay, Archs. Surg. 97, 437. Yallow, R.S. and S.A. Berson, 1970, Radio-immunoassay of gastrin, GastroenteroLogy 58, 1.
