A RADIOIMMUNOASSAY FOR PLASMA ARGININE-VASOPRESSIN IN MAN AND DOG: APPLICATION TO PHYSIOLOGICAL AND PATHOLOGICAL STATES J. J. MORTON, P. L. PADFIELD

MARY L. FORSLING *Medical Research Council, Blood Pressure Unit, Western Infirmary, Glasgow, Gil 6NT and f Department of Chemical Pathology, St Bartholomew's Hospital, London, EC1A 7BE AND

(Received 23 September 1974) SUMMARY

A radioimmunoassay has been developed for plasma arginine-vasopressin in man and dog. The mean recovery of added arginine-vasopressin (AVP) was 60 \m=+-\6\m=.\9(s.d.) % and the lower threshold of detection 2\m=.\0pmol/l. A close correlation was found between concurrent radioimmunoassay and bioassay values. The mean concentration found in peripheral venous blood in healthy men after overnight fasting was 5\m=.\3pmol/l (range 4\m=.\6\p=n-\6\m=.\2pmol/l). In man, significant increases in plasma AVP occurred after dehydration (5\m=.\9\p=n-\9\m=.\5pmol/l) and significant decreases after oral water-loading (5\m=.\2\p=n-\3\m=.\7pmol/l). During i.v. infusion of graded doses of synthetic AVP in normal men, plasma levels were closely correlated with infusion rate. On stopping the infusion, plasma vasopressin fell exponentially with a half-life of between 7 and 8 min. In man, plasma AVP was unaffected by tilting head-up for 2 h, or by a non-hypotensive bleeding of 500 ml in 10 min. In the dog, haemorrhage of 5 ml/kg and over caused proportionate increases in AVP in the circulation. In normal men, plasma vasopressin was significantly correlated with concurrent urinary osmolality. Five patients with oat-cell bronchial carcinoma and hyponatraemia showed a marked increase of plasma vasopressin. Five patients with diabetes insipidus had significantly reduced, but detectable, levels of plasma AVP. The plasma concentration in these patients did not increase after water restriction.

INTRODUCTION

investigation of the physiological and pathological role of the antidiuretic hormone (ADH) in man has been limited for lack of methods sufficiently sensitive and simple to measure the hormone in plasma. Bioassay techniques can be very sensitive but are ham¬ pered by the limited number of samples which can be routinely processed. Also, early radioimmunoassays for arginine-vasopressin (AVP) were limited in their application due to lack of sensitivity. Recently more sensitive radioimmunoassay methods have been developed (Beardwell, 1971; Oyama, Kagan & Glick, 1971; Edwards, Chard, Kitau, Forsling & Landon, 1972; Johnston, 1972; Husain, Fernando, Shapiro, Kagan & Glick, 1973; Robertson, Mahr & Athar, 1973; Skowsky, Rosenbloom & Fisher, 1974). This report describes the development and validation of a radioimmunoassay for AVP sufficiently sen¬ sitive to measure plasma levels at physiological concentrations and its application in man and the dog, with particular reference in man to the role which volume, concentration and distribution of the blood play in the control of ADH secretion. Preliminary reports of part of this work have been published (Morton & Waite, 1972; Morton, 1973). Ì Present address : Department of Physiology, Windeyer Building, Middlesex Hospital Medical School, Cleveland Street, London, W. 1. The

MATERIALS AND METHODS

AVP

(A. B. Ferring, Malmö, Sweden), 150i.u./ml (1 pu. 2-5 pg 2-3 Synthetic IO-3 pmol) was used for immunizing rabbits, for the preparation of iodinated hormone and as a standard. After high-voltage electrophoresis of this preparation and staining with a peptide reagent (1 % ninhydrin in acetone, 1 % cadmium acetate in 50% acetic acid, 85:15, v/v) only a single spot could be detected. To minimize the rise of dimer formation, AVP was kept in solution at pH 3-4-5 (0-05 M-acetic acid) and stored at —20 °C. Lysinevasopressin and oxytocin (Sandoz A.G., Basel) were used in cross-reaction studies. Florisil, 100-200 mesh (Koch-Light Ltd, Bucks) was used for the extraction of AVP from plasma. Nanograde acetone was supplied by Mallinckrodt Chemical Works, St Louis, U.S.A. All other reagents were of analytical grade. Dilutions and incubations were made with 0-05 mTris-HCl buffer (pH 7-5), prepared using boiled deionized water containing 0-2 % neomycin sulphate (Glaxo, Greenford, England) and 0-3 % human serum albumin (A. B. Kabi, Stockholm, Sweden). Arginine-vasopressin was conjugated to rabbit serum albumin (Pentex =

=

Inc., Kenkakee, Illinois, U.S.A.); Na125I (Code IMS. 30; Radiochemical Centre, Amersham) was

used for

labelling the

AVP.

Antisera to A VP raised in New Zealand White rabbits

AVP were by injection of AVP rabbit serum albumin as previously described (Morton & Waite, 1972). Antisera were stored undiluted in 0-5 ml aliquots at —20 °C. For routine incubations the chosen antiserum was used at a final dilution of 1:20000.

Antisera to

synthetic

conjugated

to

Iodination of A VP A slight modification of the method of Greenwood, Hunter & Glover (1963) was used. All reagents, except AVP, were prepared in 0-25 M-phosphate buffer at pH 7-5. To 2 mCi Na125I (15-20/¿I) in a small round-bottomed glass tube was added 50/d phosphate buffer. The following reagents were then added in rapid succession: 4 pg AVP (10/d), 100pg chloramine (25/d), 100 pg sodium metabisulphite (1-7 ml) and 2 mg potassium iodide reaction was carried out with continual stirring. A 10/d sample was removed (200 µ\). The and the remainder transferred to a DEAE Sephadex G-25 column for purification by the method of Dusterdieck & McElwee (1971). The effectiveness of the purification step in removing degraded material and free isotope was assessed by paper electrophoresis, in 0-07 M-barbitone buffer (pH 8-6), of the reaction mixture before and after the column fractionation. To determine the amount of iodinated AVP present in the purified fraction, duplicate serial dilutions were incubated with antibody (final dilution, 1:20000) and a constant amount of the same iodinated AVP, diluted to give 10000 c.p.m./50 µ\ (approximately 10 pg). This was then compared with a standard curve containing serial dilutions (250-2 pg) of synthetic AVP, antibody (1:20000) and the same constant amount of label (10000 c.p.m.). The purified iodinated AVP was diluted accurately to 5 pg/50/d and stored at -20 °C in 5 ml aliquots. Separation of bound from free labelled A VP of Separation antibody-bound from free labelled AVP was achieved by a modification (Dusterdieck & McElwee, 1971) of the method of Herbert, Lau, Gottlieb & Bleicher (1965) using plasma-coated charcoal.

Blood sampling and extraction of AVP from plasma Identical procedures were used for human and dog blood. Whenever possible, samples of plasma and urine for concurrent osmolality estimation were also obtained. Blood, usually 10-20 ml, was taken from a peripheral artery or vein into a sterilized plastic bottle containing 50-100 i.u. heparin and was immediately centrifuged at 3500 g for 15 min at 5 °C. The plasma was removed and either extracted immediately or deepfrozen at —20 °C in plastic containers. Frozen samples were extracted within 2 weeks of freezing; there being no change in AVP concentration found in two separate plasma pools stored frozen over this period of time (see later). Arginine-vasopressin was extracted from plasma by a modified Florisil technique (Beardwell, 1971). The Florisil was pretreated by washing several times with distilled water to remove fines, and then washed in sequence with M-HCI, distilled water, M-NaOH, distilled water, and finally nanograde acetone. The acetone was removed by washing with distilled water and the Florisil activated by heating at 120 °C for 18 h. Each plasma sample (5-10 ml) was extracted by mixing for exactly 20 min with 100 ± 10 mg Florisil in a glass test-tube at room temperature. The extract was centrifuged and the plasma supernatant sucked off. The Florisil was washed with boiled distilled water (5 ml), washed again for 30 min with acetone (5 ml), and the acetone was discarded. The Florisil was again washed with water and finally acidified by washing with 0-2 M-HCI (5 ml). The AVP was eluted from the Florisil by mixing for 20 min with 2 ml aqueous acetone (1:9, v/v). After centrifugation the eluate was transferred to a small siliconized glass test-tube and dried in a water-bath at 37 °C under a stream of air. The Florisil was eluted a second time and the eluates were combined and dried. The dried plasma extracts were dissolved in 300 µ\ Tris buffer by vigorous mixing. In¬ soluble material was removed by centrifugation and 200 µ were taken from the supernatant for incubation in serial dilution. Standard curves Known amounts of AVP were serially diluted from 250 pg to 2 pg in 200 µ\ Tris buffer. To each tube was added antiserum (200 µ\) at a dilution of 1:10000, and 5 pg iodinated AVP (50 p\). Two control tubes, one containing antiserum and iodinated AVP only and the other containing iodinated AVP only, were also included to provide information on the maximum binding achieved and the effectiveness of the separation procedure used. Standard curves were constructed by plotting the percentage of labelled AVP remaining bound to the antiserum against amount of standard AVP added on linear/log paper.

Incubations Incubations were carried out in Tris buffer, using Luckham's LP3 polystyrene tubes, at 5 CC for 5 days. Each incubation batch comprised unknown extracts, dilutions for two standard curves and quality control tubes (see below).

Quality control To obtain standard curves which were as reproducible as possible, large batches of labelled AVP, antibody (appropriately diluted) and standard were prepared. These were frozen in suitably sized aliquots and, when required, were thawed, used and then discarded. When one of these components was exhausted, newly prepared material was first checked using the previous preparations of the other two. Not more than one component was changed at one time. Control of the incubation procedure was imposed by including in each incubation an extract of tumour tissue obtained from a patient with excessive secretion of ADH

any

associated with oat-celled bronchial carcinoma (Schwartz, Bennet, Curelop & Bartter, 1957). This extract was split into suitably sized aliquots and a between-incubation mean value obtained from 15 separate incubations (Table 1). The extraction procedure was checked by including in each batch of extractions a sample from a plasma pool where the mean value was similarly estimated. Both incubations and extractions were considered acceptable when the values for both the tumour extract and the plasma pool samples fell within two standard deviations of their mean values, giving 95 % confidence limits. Differences in the values obtained for various samples in the same extraction batch were considered to be genuine when they were greater than twice the coefficient of variation for within-batch replicates. Units

Plasma AVP concentrations are quoted (1-Opmol AVP/1 1-08 pg AVP/ml).

as

pmol/1 according

to the

Système

International

=

RESULTS

Antisera to A VP

The formation of antibodies in the rabbits was indicated by the onset of polyuria of varying severity. Antiserum 102/5 used in the assay was obtained from the rabbit with the most severe polyuria. This antiserum, as well as having the highest titre [a dilution of 1:26000 binding 50 % of added 125I-labelled AVP (10 pg)], was chosen as the most suitable because it also showed the greatest avidity as indicated by its equilibrium constant of 3-5 10101/mol (Nisonoff & Pressman, 1958), and also produced the most sensitive standard curves with respect to the limits of detection of the assay. Oxytocin cross-reacted very slightly ( < 0-03 %) and lysine-vasopressin 8 % with antiserum

102/5.

lodination of A VP The specific activity of 125I-labelled AVP varied between 560 and 916 Ci/mmol in six separate iodinations. Electrophoresis of the reaction mixture before purification showed three radio¬ active fractions, one containing 83-91 % of the total radioactivity applied to the paper and two smaller fractions containing respectively 2-10 % and 5-7 % of the radioactivity applied. After purification of the reaction mixture on DEAE Sephadex, electrophoresis of the peak radioactive fraction obtained from the column showed the presence of the large radioactive fraction only, containing more than 97 % of the total radioactivity applied. As well as monitoring the fractions from the DEAE column for radioactivity, each frac¬ tion was assayed for pressor activity using the rat blood pressure preparation (Peart, 1955). This showed the presence of a biologically active but non-radioactive peak which emerged from the column before the radioactive fraction which did not contain biological activity. The lack of biological activity in the radioactive peak may be due to one or other or both of the following: its destruction following labelling with 125I or the fact that only a relatively small proportion of the AVP is isotopically labelled resulting in insufficient being present for detection by the pressor assay used. However, that the radioactive peak from the column was in fact 12SI-labelled AVP was shown after incubation with excess antiserum which bound more than 95 % of the radioactivity. A further check on its immunological purity was carried out by comparing the ability of known and increasing amounts of the 125I-labelled AVP to displace a constant amount (5 pg) of label from the antiserum with that of known increasing amounts of standard unlabelled hormone. Only 125I-labelled AVP preparations showing similar binding characteristics to the standard hormone were used in the assay. The use of a DEAE Sephadex column for purification of the reaction mixture produces both

chemically pure specific activity.

a

label and

a

label free from unlabelled AVP with

a

resultant increase in

There was no detectable decrease in the binding of the labelled peptide with storage for up to 6 weeks from the date of preparation. Similarly, paper electrophoresis, after 4 weeks, showed no detectable increase in either 'degraded label' or free isotope.

60-

40-

20

-131

16 AVP

Fig.

1. The mean±2

s.d.

62

125

250

(fmol)

for 12 separate standard

curves.

Incubation condition: 5 pg 125I-labelled

arginine-vasopressin (AVP) antiserum 102/5 (1:20000), 5 days at 5 °C.

Standard curves and unknown incubations 1 Fig. shows the mean, with two standard deviations for each point, of twelve separate standard curves set up on consecutive occasions using the same label, the same antiserum dilution and the same standard preparation. If the sensitivity of the method is defined as the smallest quantity of AVP which can be distinguished from the zero standard, given the error for both determinations, then 4-0 fmol/tube is the smallest amount of AVP which can be determined with 95 % confidence. Given 10 ml plasma, a recovery of 60 % and an initial dilution of 1:3, this means that 2 pmol AVP/1 plasma is the lowest concentration which can be determined. The effect of incubation time and temperature on the standard curve was tested by incu¬ bating at 5 °C for 18-24 h, for 45-48 h and for 5 days, and at room temperature for 3 h. The last of these conditions produced a maximum binding of only 40 %. The standard curves at 18-24 h and 45-48 h were very similar while that at 5 days showed an increase in apparent sensitivity, the fall in percentage bound from the zero standard increasing from between 4 and 6 % at 4 fmol. The longest period of incubation was therefore chosen for

optimal sensitivity. Plasma extracts

were

incubated in serial dilution.

Fig.

2 shows the

relationship between

the volume of extract incubated and the quantity of AVP found for 17 separate plasma extracts. The result indicates a close immunological similarity between the human plasma extracts and the synthetic AVP standard. Similar serial dilutions of dog plasma extracts also

closely paralleled the AVP standard

curve.

a,

>




>