Radioimmunoassay of thromboxane B2 in horse plasma

Research ill Veterinary Science 1987. 41. 150-153

Radioimmunoassay of thromboxane B2 in horse plasma H. CAMBRIDGE, J. A. REYNOLDSON*, J. D. DUNSMORE, School of Veterinary Studies, Murdoch University, Murdoch, Western Australia 6150, Australia, B. J. HILBERT, College of Veterinary Medicine, St Paul, Minnesota, USA

A radioimmunoassay for thromboxane ~ (TXB 1) in unextracted horse plasma was evaluated. ~ensitivity of the assay was 14·0 (SD 5' 6) pg ml- I of plasma. Interassay and intra-assay variation were 21· 3 per cent and 4·3 per cent, respectively. The percentage of tracer bound in unextracted plasma in the absence of TXB1 was often higher than that in buffer. Therefore standard curves were obtained using standards diluted in plasma from horses treated with aspirin or plasma. Standard in charcoal treated TX~-free curves determined in plasma and buffer were parallel. This assay was used to determine the half-life of exogenous TXB 1 in horses. The mean value of 20·7 minutes (n =3) is similar to that determined in other species. Storage of plasma samples at - 20 0 e for four months was found to alter the TX~ levels in an unpredictable manner. Mean recovery was 102·4 per cent (SI) 44'1). Effect of variability in sample collection and handling was assessed, but no consistent source of artifactual generation of TX~ was found. POSSIBLE roles of eicosanoids in the pathogenesis of non-reproductive diseases of domestic animals are an area of current research interest (Higgins 1985). In horses acute increases in the eicosanoids thromboxane A 2 (TXA 2) and prostacyclin (PGI 2) have been demonstrated in experimental endotoxic shock, by measurement of their metabolites .thromboxane B2 (TXB 2) and 6-keto-prostaglandin Fl. in plasma (Bottoms et aI1982). The potent vasoactive effects of these substances suggests that they are responsible for the haemodynamic dysfunction associated with this syndrome. Eicosanoids are generated during carrageenin-induced inflammation (Higgins and Lees 1984) and serum levels are reduced after administration of non-steroidal anti-inflammatory drugs (Kopp et al 1985) in the horse. Involvement of TXA 2 in the pathogenesis of abnormal blood flow and colic in horses with arteritis caused by infection with larvae of Strongylus vulgaris has also been postulated (White 1981). The effectiveness of non-steroidal anti-inflammatory drugs in

treating signs of this disease, and the abundant platelet aggregates found on scanning electron micrographs of arteritis lesions support this theory (White et aI1983). However, no reports of local or peripheral TXB 2 levels in affected horses have appeared. The purpose of this study was to evaluate a radioimmunoassay for TXB 2 in plasma of horses as a means of assessing endogenous TXA 2 generation in normal horses and those with active S vulgaris lesions. Radioimmunoassay is the most convenient method for TXB 2 measurement and although measurement of in vitro or local TXB 2 generation is widely accepted, use of radioimmunoassay to assess peripheral plasma levels in humans has produced widely diverging and often contradictory results (Morris et al 1981). This may be caused by alternative paths for TXA2 metabolism (Kindahl and Granstrom 1980) or to variation in sample handling, extraction, radioimmunoassay technique and artifactual production during blood collection. In humans, unextracted plasma can be assayed satisfactorily provided antisera of sufficient specificity are available (McCann et al 1981). These authors found that in humans with no known thrombotic disorders, basal levels of plasma TXB 2 were at or below the limit of detection of their assay, about 30 pg ml "'. Since human platelets possess a greater capacity for TXA 2 generation than equine platelets (Meyers et al 1980), resting levels in normal horses could be expected to be even lower and therefore an iodinated tracer was used to maximise the sensitivity of the assay. Materials and methods

Reagents Non-radioactive TXB 2 was a gift from Upjohn (Kalamazoo). Antiserum to TXB 2 raised in rabbits was obtained from Seragen (Boston, Massachusetts) and the 1251 labelled TXB 2 tyrosyl-methyl ester from New England Nuclear (Boston). The cross-reactivity of this antiserum was reported as less than O' I per cent with prostaglandins and their metabolites. The specific activity of the tracer was 2200 Ci rnmol: I •

150

Radioimmunoassay of TXB1 in horses Tracer and antiserum were diluted with phosphate buffered saline (pH 7' 4) containing I g sodium nitrate and I g of bovine y-globulin per litre. Polyethylene glycol 6000 and Norit activated charcoal was obtained from Sigma Chemical (St Louis). Standards were prepared in buffer, in TXBrfree charcoal treated plasma or in plasma from horses which had received acetylsalicyclic acid (aspirin) 20 mg kg-I by mouth on the day before blood collection. The anticoagulant solution used for sample collection consisted of 20 ~I aspirin (2' 5 mg ml- I ) and 10~1 EDTA (70 mg ml- I) per ml of blood.

151 Aspirin

40

~

~

10

i

i

3

Radioimmunoassay One hundred microlitres of sample or standard, 100 ~I of an appropriate dilution of antiserum, to bind 30 to 40 per cent of tracer in the absence of TXB 2 , and 100 ~I of tracer (10,000 to 14,000 cpm) were added to duplicate or triplicate polypropylene test tubes maintained in an ice bath. Appropriate tubes for determining total counts and non-specific binding were included. The tubes were vortexed and incubated overnight at 4°C. Separation of bound and free tracer was achieved by chemical precipitation of the immunoglobulins. Five hundred ~I of a 22· 5 per cent aqueous solution of ice cold polyethylene glycol 6000 was added to each tube. Tubes were vortexed and centrifuged immediately at 2200 g for 40 minutes at 4°C. The supernatants were removed by vacuum aspiration and the bound fractions in the pellets counted in a gamma counter (Searle 1197).

Sample collection Unless otherwise stated 10 ml of blood was drawn from a jugular vein using a 10 ml polypropylene syringe and an 18 G I' 5 inch (3' 8 ern) needle. Blood was immediately transferred to a polypropylene tube containing the anticoagulant solution, mixed by gentle inversion and placed in ice. Before venepuncture the hub of the needle was filled with anticoagulant. The samples were centrifuged as soon as possible (within 30 minutes of collection) at 1800 g and 4°C for 15 minutes. Plasma was stored in polypropylene tubes at - 20°C and assayed after thawing for the first time.

TXBrfree plasma The method of Chard (1978) was used to produce ligand-free plasma; Activated charcoal was added to plasma (10 g 100 ml : I) and the mixture stirred overnight at 4°C. Charcoal was removed by centrifugation followed by millipore filtration. Aspirin (17 mg kg-I by mouth) has been shown to block cyclooxygenase effectively in horses (Kopp et aI1985). This

7 Time (days)

FIG 1: Effect of two doses of aspirin (20 mg kg - t) on plasma TXB2 levels in two horses. The dotted line represents the limit of sensitivity of the radioimmunoassay and values below this level represent concentrations indistinguishable from zero

dose produced a drop in serum TXB 2 to a level barely detectable by their radioimmunoassay. In humans, Sturm et al (1984) showed that a dose of 300 mg daily for five days produced plasma TXB 2 levels of 0 pg ml I. Aspirin (20 mg kg - I by mouth) was found to suppress TXB 2 levels to below the limit of the assay (this study, Fig I).

TXB1 half-life estimation Three thoroughbred cross mares were used. Each horse received 20 mg kg-I aspirin by mouth on the day before the experiment to block in vivo generation of thromboxane. TXB 2 (250 ~g) dissolved in 500 ~I of ethanol and made up to 5 ml with normal saline was injected as a bolus into a jugular vein. Samples were collected from the contralateral vein via an indwelling 12 G cannula. Results

Assay characterisation The percentage of tracer bound in the absence of unlabelled TXB 2 (per cent Bo) was often higher in unextracted horse plasma than in buffer. Thus plasma samples which contained low levels of TXB 2 bound a higher percentage of tracer than the '0' binding tube, and falsely low levels were indicated. To overcome this problem standards were diluted in plasma collected from aspirin treated horses. Standard curves obtained in buffer and plasma were parallel indicating that, apart from the initial change in specific binding, non-specific interference in the assay by plasma was not a problem. Curves generated in TXBrfree plasma obtained by charcoal treatment were also parallel to buffer curves but often '0' binding was also less than in plasma from aspirin treated horses.

H. Cambridge. J. A. Reynoldson. J. D. Dunsmore. B. J. Hilbert

152

•

120

TABLE 1: TXB2 levels (pg ml-'I in five horses with different treatment of samples followed by centrifugation at 4°C Processing of samples Immediate 1 h on ice 1 h at 20°C

Horse

228 245 247

15·0 <14 14·5

264

31·5

263

28·0

18·0 <14 <14

26·5 36·5

64·0 <14 16·0 30·0

39·5

Sampling with stasis 24·0 <14 <14 32·0

34·5

measurement of samples containing 25,50 or 100 pg ml ' ! of TXB2 • Coefficients of variation for the assayed levels in the three pools were 18· 5 per cent (n=5), 23'4 per cent (n=15), and 21'9 per cent (n = 5). respectively (Fig 2). The intra-assay coefficient of variation was 4· 3 per cent (SD 4· 3 per cent). obtained from 65 triplicate samples. Non-specific binding was not increased by the use of plasma for standard dilution (mean 8· 3 per cent. SD I· 98 per cent, n = 25).

I

20

40

I

I

60 8) TXB2 added (pg mi'!)

I

100

FIG 2: Correlation between added TXB2 and assayed TXB 2 in 25 samples of equine plasma. The correlation coefficient is 0·822 and the least squares linear regression line is plotted

The sensitivity of the assay was arbitrarily defined as the amount of TXB2 which decreased tracer binding to 85 per cent of Bu. The mean value from 10 standard curves was 14· 0 (SD 5' 6) pg ml- I. Interassay variation was assessed by repeated 5000

t%o ; 2·35 min t%fJ ; 20·7 min K. 1 ; 0'lOBmin- 1 V d ; 1·515litres per kg- 1

1000

E

~

f

l-

Half-life estimation The mean elimination half-life of TXB2 was 20' 7 minutes (n = 3). Therefore, after the bolus injection, TXB 2 levels dropped below the detection limit of the assay after four half-lives or 83 minutes (Fig 3).

Effect of variation in sample collection and handling The effect of stasis in the syringe during blood sampling, delayed centrifugation and variable sample temperature after collection were assessed. Results from five horses are summarised in Table I. Prolonged storage of samples has also been shown to affect TXB2leveis in a random manner (McCann et al 1981). Table 2 compares TXB2 levels in aliquots of plasma first assayed a few days after collection and again after four months of storage at - 20°C. Mean recovery was 102'4 per cent (SD 44·1 per cent). Discussion

200

These results indicate that TXB2 can be assayed

I

TABLE 2: Effect on TXB2 levels (pg ml-'I of storage at -20°C for up to sevan days or for four months

e,

50

Storage time Up to 7 days 4 months

o

10 ...---,r-"""T--r--,--or--.....,r---r-..., i , i 8) 30 40 50 60 70 10 20 Time (min)

o

FIG 3: Semilogarithmic plasma disposition curve following intravenous injection of 0·25 mg of TXB2 in three horses (pooled data ± 1 SO). Least squares regression lines show the elimination and distribution phasesof the biexponential curve

145·0 140·0 32·0 123·0 103·3

o

23·8 74·2 117·0 58·9 115·5 105·0

o

Mean per cent recovery 102·4 ± SO44·1

% recovery 51·2

83·6

184·1 93·9 101·6 100·0

Radioimmunoassay of TXB2 in horses with a high degree of sensitivity in unextracted horse plasma, despite the inconstant differences in per cent Bo between plasma and buffer. Increased specific binding in unextracted human plasma ranging from 3 to 13 per cent was reported by Morris et al (1981) and a similar effect in horse plasma was mentioned by Granstrom and Kindahl (1978). The reasons for this are unclear, but as standard curves in buffer and plasma are parallel this problem can be largely overcome by the use of plasma from aspirin treated horses. This approach has also been used for prostanoid measurement in unextracted dog (Prosdocimi et al 1985) and rabbit plasma (Deckmyn et al 1983). A problem arises where absolute rather than relative values are required or where single samples from a number of animals must be compared, as the true '0' for these is unknown and probably different. Where large numbers of serial samples are to be collected, it is desirable to obtain plasma from each animal after aspirin treatment so that individual baselines can be established. The interassay variation encountered in this study is higher than that for most radioimmunoassays. This may be a reflection of the variability in initial binding characteristics between individual plasma samples, as well as variation in levels with prolonged storage. The half-life of exogenous TXB 2 is short and comparable to that established by radio-isotopic means in the rabbit (McCann et al 1981) and the monkey (Kindahl 1977). The short half-life implies that for peripheral levels of TXB 2 to be elevated for long periods, and to be detectable in plasma, endogenous production must be fairly continuous. The large volume of distribution (1'515 litres kg I) is consistent with the high lipid solubility of TXB2 • Morris et al (1981) showed that TXB 2 levels were increased by difficulty in drawing blood samples and other factors such as delay before centrifugation. TXB2 is present in equine serum in nanogram quantities under normal conditions (Kopp et a11985) and is generated as a result of platelet activation or damage. There is clearly great potential for artifactual TXB 2 production during sampling procedures particularly if the blood is not completely and rapidly anticoagulated. In the present study, however, no factor which consistently contributed to artificial production was identified (Table I). This may reflect species or individual variation in development of the thromboxane generating system. In general a strictly standardised system of blood collection should circumvent artifactproblems, It is not known whether the levels of TXB2

153

encountered in horses before aspirin administration represent a true 'circulating' TXB2 or a basal amount of artifactual production. An alternative noninvasive method involves measurement of the 2,3 dinor TXB2 metabolites in urine but blood sampling is technically simpler and more practical in the horse. In addition this study has shown that background levels of either circulating or artifactual TXB2 are low and therefore pathological increases should be detectable. Acknowledgements This work was carried out with the support of the Australian Equine Research Foundation to whom we are most grateful. The authors wish to thank Upjohn for the gift of unlabelled TXB2 • References BOTTOMS, G. D., TEMPLETON, C. B., FESSLER. J. F., JOHNSON, M. A., ROESEL, O. F., EWART, K. M. & ADAMS, S. B. (t982) American Journal of Veterinary Research 43, 999-t002 CHARD, T. (1978) Laboratory Techniques in Biochemisrry and Molecular Biology. Eds T. S. Work and S. Work. Amsterdam, North-Holland. Vo16. I" 341 DECKMYN, H., VAN HOUTTE, E., VERSTRAETE, M. & VERMYLEN, J. (1983) Biochemical Pharmacology 32, 27572762 GRANSTROM, E. & KINDAHL, H. (1978) Advances in Prostaglandin and Thromboxane Research. Vol 5. Ed J. C. Frolich. New York, Raven Press. PI" 119-210. HIGGINS, A. J. (1985) Journal of Veterinary Pharmacology and Therapeutics 8, 1-18 HIGGINS, A. J. & LEES, P. (1984) Journal of Veterinary Pharmacology and Therapeutics 7,65-72 KINDAHL, H. (1977) Prostaglandins 13, 6t9-629 KINDAHL, H. & GRANSTROM, E. (1980) Prostaglandins, Prestacyclin and Thromboxancs Measurement. Eds J. M. Boeynaems and A. G. Herman. The Hague, Martinus Nijhoff', PI" 1-17 KOPP, K. J., MOORE, J. N., BYARS, T. D. & BROOKS, P. (1985) Equinl! Veterinary Journal 17, 322-324 McCANN, D., TOKARSKY, J. & SORKIN, R. P. (I98t) Clinical Chemistry 27, t417-1420 MEYERS, K. M., KATZ, J. B.. CLEMMONS, R. M., SMITH, J. B. & HOLMSEN, H. (1980) Thrombosis Research 20, 13-24 MORRIS, H. G .. SHERMAN, N. A. & SHEPPERDSON, F. T. (1981) Prostaglandins 21, 771- 789 PROSDOCII\II, M., FINESSO, M., GORIO, A., LANGUINO, 1.. R., DEL MASCHIO, A.• CASTAGNOLl, M. N., DE GAETANO, G. & DEJANA, E. (1985) American Journal of Physiology 248, H493-H499 STURM, M., BARDEN, A., BEILIN, L. J. & TAYLOR, R. R. (1984) Clinicul und Experintental Pharmacology and Physiology 11,611-619 WHITE, N. A. (1981) Journal of the American Veterinary Medicat Association 178,259-262 WHITE, N. A.. MOORE, J. N. & DOUGLAS, 1\1. (1983) Equine Veterinary Journal IS, 349-353

Accepted March 3, 1986