JPM Vol. 28, No. 4 December 1992:185-190

A Specific Assay for Leukotriene Bq in Human Whole Blood Jytte Laboratory

Fogh,

Lars K. Poulsen,

and Hans

Bisgaard

of Medical Allergology (J.F., L.K.P.), Allergy Unit; and Department National University Hospital, Copenhagen N, Denmark

of Pediatrics

(H.B.),

Leukotrienes (LTs) are potent mediators of inflammatory and allergic responses, and are present in biological fluids in minute amounts, that is, in the picogram range. The aim of this study was to develop and validate a method for determination of LTB4 synthesized in vitro in human whole blood. Heparinized blood was stimulated with calcium-ionophore A23187 at 37°C. After 30 min ceils were separated by centrifugation. LTB4 was analyzed by radioimmunoassay (RIA). When sample preparation was restricted to protein precipitation with acetone, interference was demonstrated by lack of parallelism between standard and sample dilution curves. Purification was, therefore, extended by combinations of the following steps: 1) protein precipitation, 2) lipid extractions, and 3) high-performance liquid chromatography (HPLC). One of two commercially available LTB4 standards was found to contain multiple components, several of which were immunoreactive in RIA. Even for the standard containing pure LTB4, interference was demonstrated by lack of parallelism between sample and standard dilution curves. Testing eight combinations of varying purification steps, we found that only a three-step purification procedure, including 1) solidphase extraction, 2) protein precipitation at -2o”C, and 3) HPLC, was able to eliminate interference in RIA. Using this procedure, the recovery was 78%. Stimulation of whole blood from normal subjects with calcium-ionophore showed optimal LTB4 production at 10 JLM ionophore, yielding 6.6 ng LTBJmL blood. Keywords:

Leukotriene Radioimmunoassay

Bh; Whole

blood;

Introduction Leukotrienes (LTs) are proinflammatory mediators biosynthesized from cell membrane-derived arachidonic acid via the S-lipoxygenase @-LO) pathway. Elucidation of the metabolism of arachidonic acid by individual blood cells has shown that various cell types, including neutrophils and basophils, have the capacity to generate LTs. Blood cell interactions in the synthesis of LTB4 have been demonstrated: released arachidonate from platelets serves as a precursor for LTB4 production by neutrophils (Marcus et al., 1982), platelet-derived IZHPETE stimulates the LTB4 synthesis in leucocytes (Maclouf et al., 1982), and leucocyte-derived LTA4 can be modified to LTB4 by erythrocytes

Address reprint requests to Dr. J. Fogh, Laboratory of Medical Allergology, Allergy Unit, National University Hospital ‘lTA 7542, Tagensvej 20, DK-2200 Copenhagen N, Denmark. Received March 1992; revised and accepted July 1992.

High-performance

liquid

chromatography;

(Fitzpatrick et al., 1984). These findings suggest that transcellular metabolism may be an important biochemical factor affecting the generation of LTs, and, accordingly, studies of LT synthesis in monocellular systems may be biased. An integrated production model can be obtained by using whole blood. We plan to exploit this for an investigation of microbial and allergenic stimuli inducing LT formation in vitro. Further, putative 5-LO inhibitors may be tested in vitro taking into account the possible plasma protein binding. Several investigators have described methods for quantifying LTB4 in human whole blood using different methods for sample preparation and analysis (Gresele et al., 1986; Sweeney et al., 1987; Patrignani and Canete-Soler, 1987; Zakrzewski et al., 1988; Fradin et al.., 1989). Gresele et al. (1986) compared several extraction procedures before radioimmunoassay (RIA) and adopted a sample preparation with acetone extraction before RIA. Later reports have employed solid-phase

Journal of Pharmacological and Toxicological Methods 28, 185-190(1991) 0 1992 Elsevier Science Publishing Co., Inc., 655 Avenue

of the Americas,

New York, NY 10010

1056-8719/92/$6.00

186

JPM Vol. 28, No. 4 December 1992:185-190

extractions and chromatographic methods for quantization or for purification prior to quantization by immunoassays (Sweeney et al., 1987; Patrignani and Canete-Soler, 1987; Zakrzewski et al., 1988; Fradin et al., 1989). We adopted the method described by Gresele et al. (1986) but were not able to eliminate interference, thus the aim of this study was to develop and validate a specific assay for LTB4 quantization in stimulated whole blood.

Material and Methods Reagents LTB4 and 3H-LTB4 were obtained from Amersham (Buckinghamshire, UK). Rabbit antisera raised against LTB4 was obtained from Advanced Magnetics Inc. (Cambridge, Massachusetts, USA). The cross-reactivities (data provided by the manufacturer) were less than 1% for LTC4, LTD4, LTE4, 5-,12-, and 15-HETE (hydroxyeicosatetraenoic acid), prostaglandin (PG)s A,, AZ, Bi, Bz, Dz, El, and Ez, 6-keto-PGFi,, 6-ketoPGE,, 13,14-dihydro-15keto-PGFzu, 13,lbdihydro15keto-PGEz, tromboxane (TX)Bz, and bicycloPGE*. LTB4 RIA kit was supplied by NEN DuPont (Dreieich, Germany) and Amersham (UK). A Tris-buffer (10 mM Tris-HCl, 0.9% NaCl, 0.1% gelatine, 0.1% sodiumazide, pH 8.6) served as buffer for the LTB4-RIA. Charcoal (0.4%) (Sigma, St. Louis, Missouri, USA) was suspended in (0.9% NaCl, 0.1% gelatine, 0.1% sodiumazide, pH 8.6). Aqualyte Plus liquid scinti~ation cocktail was obtained from Baker Analyzed Reagent (Deventer, Holland), methanol, hexane, and petroleum benzine (bp. 40”-60°C) from Merck (Darmstadt, Federal Republic of Germany), glacial acetic acid from Nordisk Droge (Copenhagen, Denmark) and Ca2+ -ionophore A23 187 from Sigma. To reduce adsorption of LT we exclusively used tubes, HPLC vials, and pipette tips made of polypropylene.

Venous blood was drawn from healthy volunteers with Exetainer vacuum blood sampling system containing lithium heparinate (15 III/5 mL blood). Aliquots (1 mL) of whole blood were immediately transferred to tubes containing varying amounts of l-m&f calciumionophore A23187 (final concentrations: O-40 FM) and incubated at 37°C for 30 min. Plasma was separated by centrifugation (2000 g, 15 min, at 20°C).

Purification

Procedures

Solid-phase extraction. The plasma was acidified by adding two volumes of hydrochloric acid, pH 3.0. Samples were purified by solid-phase extraction, Sep-pak Cra (Waters, Massachusetts, USA). The Sep-pak cartridges were pre-wet with 5 mL of methanol followed by 5 mL of distilled water. Acidilied samples were added and the cartridges were washed with 2 mL of distilled water, 2 mL of 25% methanol, and 2 mL of petroleum benzine. Elution was made with 3 mL of methanol. Until further handling, the eluted fractions were either frozen ( - 20°C) or immediately evaporated to dryness by directing a gentle stream of nitrogen gas onto the surface of the solution. During this procedure the samples were placed in a water bath at 30°C. The residues were reconstituted in 250 pL of high-performance liquid chromatography (HPLC) mobile phase (see below) and stored in HPLC vials. The vials were filled with argon, capped, and frozen ( - 20°C). Protein precipitation. Protein precipitation of plasma samples was carried out using 2 volumes of ethanol, O”C, or 9 volumes of acetone, 0°C. Following the solid-phase extraction, a sediment was formed during freezing, and the samples were, therefore, centrifuged for 3 min at 4700 g (Figure 1). Liquid extraction. Different liquid extraction procedures were investigated. Extraction of acidic and neutral lipids. These were made by adjusting pH to between 3 and 4 using HCl before extracting lipids into ethyl acetate (2 volumes, performed twice). The organic phase was isolated, the solvent was removed under a stream of nitrogen gas, and the residue was dissolved in 1 mL of RIA buffer or 250 p.,L of HPLC mobile phase. Extraction of LT& after protein precipitation and removal of neutral and basic lipids. Attempts to isolate LTB4 were also made by a three-step process involving 1) protein precipitation with 2 volumes of ethanol or acetone at O’C, 2) pH adjustment to between 10 and 11 and removal of basic and neutral lipids by extraction into petroleum benzine or hexane, and 3) isolation of LTB4 from the remaining aqueous/acetone or aqueous/ethanol phase of LTB4 by acidification (pH 3-4) and extraction into ethyl acetate or chloroform. The extract was evaporated to dryness under nitrogen and reconstituted in 1 mL RIA buffer or 250 ~24 of HPLC mobile phase. Reversed-phase high-performance liquid chromatography (RP-HPLC). The equipment used for developing the purification method for LTB4 consisted of a pump model N-60OOA, an automated injector WISP 710B, and an absorbance detector 440, all from Waters. Of the test solution, 200 ILL were applied to a LiChrosorb RP-18 (5 pm, 250 x 4 mm) column (Merck) in a mobile phase of 70: 30: 0.01 methanol/water/acetic acid, pH

187

J. FOGH ET AL. SPECIFIC ASSAY FOR LTBG IN HUMAN WHOLE BLOOD

~~pi~~on of proteins and ~~a~ Seppak (C-18)

I f

Evaporation

I Precipitation of proteins

i j

i ;

(k)

I RP-HPLC

j I EVeponttian

stituted in 20 mL of buffer per vial, giving a working titer of 1:4000. The sample antiserum mixture was incubated at 4°C on a vibrator for 1 hr. 3H-LTB4 was diluted to a total count of about 2500 cpm per 100 pL. The tracer preparation was kept at O”C, and 100 ~.LLof the solution were added to each test tube. The test tubes were then placed at 4°C for 20-28 hr before separating bound and unbound tracer with 200 FL of cold charcoal solution (stirred to homogeneity for 20 min on a magnetic stirrer). The tubes were agitated vigorously for at least 5 set and incubated at 4°C for 15 min before centrifugation (5000 r-pm, 4°C 30 mm). Supernatants were decanted into liquid scintillation counting vials, 4 mL of liquid scintillation counting cocktail were added, and B-irradiation was measured. LTB4 standard curves were prepared by twofold dilutions in the range of 7.8-2000 pg/mL. Results of the RIA analysis are given as % B/Be, B giving the actual count of a sample, and BO, the result with no “cold” LTB4 present. _

I

Results Purity of Standard

Figure 1. List of purification steps tested alone or in combinations indicated by a straight (-) or dotted (. . . .) line. The procedure indicated by a straight line shows the purification sequence eliminating interference, whereas all sequences of procedures including dotted lines showed interference in RIA,

5.1 with NH40H, at a flow rate of 1.0 rnL/min. The retention time for LTB4 was determined from the LTB4 stock solution (Amersham) at a wavelength of 280 nm. Purity of standards from NEN DuPont and Amersham was analyzed by isocratic RP-HPLC identical to the setup for sample preparation. Fractions of 1 mL were collected, and the fractions eluting at retention times identical to those of synthetic standards were either frozen until further handling or the solvent was removed immediately by directing a stream of nitrogen gas onto the surface of the solution at 30°C. The sample was redissolved in 1 mL of RIA buffer and kept under an argon atmosphere at - 20°C until RIA analysis.

Radioimmunoassay

for LTB4

Of the standard or sample aliquots, 100 p,L were placed in polypropylene tubes (triplicate determinations), and 100 PL of antise~m were added. The LTB4 antiserum was lyophilized on delivery and was recon-

The first step in validation of the analysis was to check the purity of the standards. Standards from two companies were investigated. The chromatograms of the NEN-standards (three deliveries, two lot nos.) displayed several peaks indicating impurities. One of these chromatograms is depicted in Figure 2(A). Standards from Amersham (two lot nos.) displayed only one peak, as illustrated in Figure 2(B). Fractions exerting immunoreactivity as determined by RIA are indicated with ” + .” The NEN standard showed immunoreactivity in several fractions, whereas immunoreactivity of the Amersham standard was restricted to fractions with retention time identical to the single peak. The LTB, standard from Amersham was used for all further experiments.

Elimination

of Interference

When the sample preparation was restricted to protein precipitation with acetone interference was demonstrated by a lack of parallelism between standard and sample dilution curves. Subsequently, protein precipitation and extraction procedures alone or in combinations were tested. Eight combinations of purification steps (outlined in Figure l), including three different procedures for liquid extraction were tested; however, none could eliminate interference. A comparison of dilution curves for one of these combinations (protein precipitation followed by solid-phase extraction) is shown in Figure 3(A). Determinations of the LTB4 content in various dilutions of the same sample, produced different conclusions about

188

JPM Vol. 28, No. 4 December 1992:185-190

Table 1. Reproducibility Coefficient of Variation RIA

(NEN Dupont)

Determined as (in %) for LTB4

COW. (pg/mL)

137 16 18

Intraassay C.V. (%) (n = 7) Interassay C.V. (%) (n = 7) Abbreviations: Cone., abbreviations as per text.

J+rf,+,+,+,+,+,+,~,+,+,

,

,

,

,

,

t

)

,

/

+ = immunoreactivjty (Amersham)

67 8 26

~OnceRtration;

35 5 30 other

dard curve would be 6000, 7800, or 15,000 pg/mL, respectively. In other words, using the upper or the lower part of the standard curve would produce a variation of 250%. This is evidenced by the lack of parallelism between sample and standard dilution curves in Figure 3(A). Of the eight combinations of purification steps, only solid-phase extraction followed by protein precipitation at - 20°C before isolation by HPLC led to parallelism between standard and sample dilution curves (Figure 3(B)).

3

Recovery and Variability 3H-LTB4 was added to whole blood, and recovery after sample preparation was determined to be 78%. In&a- and interassay variations were determined in a setup comprising three concentrations+ Each was run in triplicate on seven different days. Intraassay coefficient of variation (c.v.) was calculated as the mean of the daily C.V. Results are shown in Table 1 in which it is seen that intraassay C.V. increased for higher concentrations, whereas the interassay C.V. decreased.

B

+ = immunoreactivity

Figure 2. HPLC-chromatograms of LTB4 standards. Standards obtained from (A) NEN DuPont and (B) Amersham. Coilected fractions were tested in RIA. “ f ” indicates immunoreactivity of the fraction.

the concentration in the undiluted sample. The LTB4 content of a diluted sample was determined using the standard curve to correlate %B/Bo to concentration. Despite the use of only the steepest part of the standard curve, that is, between 20% and 80% B/Be, LTB4 concentrations in a sample seemed to vary depending on the degree of dilution used for determination. Thus in the example of Figure 3(A), if the sample was tested in dilutions giving B/Bos of 20%, 50%, or 70%, then the corresponding concentrations calculated from the stan-

The dose-dependent LTB4 release after calcium ionophore stimulation of heparinized whole blood is shown in Figure 4 (a = 2). It appears that a very low spontaneous production takes place, and that the stimulation peaks at lo-20 FM.

Discussion Different methods for analysis of LTB4 have been reported, including RIA (Gresele et al., 1986; Sweeney et al, 1987; Patrignani and Cante-Soler, 1987; Zakrzewski et al., 1988), enzyme immunoassay (Fradin et al., 1989), HPLC (Sweeney et al., 1987; Zakrzewski et al., 1988), and gas chromatography-mass spectrometry (Fradin et al., 19891, the former two being based on immunoreactivity for recognition, the latter two on physical/chemical behavior only. The immunoassays

J. FOGH ET AL. SPECIFIC ASSAY

189 FOR LTBI IN HUMAN WHOLE BLOOD

% B/B,

%B/B0

100

,.,,.,,., 100

1000 pg

LTB,/ml

L,,, 1000

w

LTB,/ml

Figure 3. Dilution

curves of standard and samples after two different sample purification procedures before RIA analysis. (A). Protein precipitation and solid-phase extraction. (B). Solid-phase extraction, protein precipitation at - 20°C and HPLC (0) Standards, (W and A) Samples. Sample dilution curves are placed so that concentrations of standard and samples are equal at B/B0 = 50%.

offer good sensitivity, but the specificity depends on the antibodies available. The chemical/physical methods offer good specificity, but require expensive equipment and often offer a sensitivity lower than that of the immunoassays (Sweeney et al., 1987). In this study we aimed at establishing the sensitivity of RIA while at the same time ensuring specificity.

pg

LTB4/ml 8000

blood r

7000

c

6000

5000

4000

3000

2000

1000

0 0

25

5.0

10

20

40

pLM A23107

Figure 4. Dose dependence

on net LTB4 synthesis after calcium-ionophore stimulations of whole blood from normal subjects (30 min, 37°C). Mean and range of stimulations performed on two subjects.

In developing an assay for LTB4 our first step was to validate the RIA itself. We found that one of two commercially available standards contained impurities as evidenced by the HPLC chromatogram. Furthermore, these impurities exhibited immunoreactivity in RIA. This would invalidate the RIA, because the standard curve does not represent a dilution of pure LTB4. Although we did not systematically study different antisera, all of the three preparations studied seemed to react with other components than LTB4, as illustrated in Figure 2(A) for NENs antiserum. Also antiserum from Amersham and Advanced Magnetics showed interference as evidenced by a lack of parallelism between standard and sample dilution curves, using other purification procedures than the one finally selected. The latter antiserum was chosen because it could be purchased separately frp an assay-kit. To obtain specificity of an assay, researchers have used various methods for purification of LT-containing samples before analysis, including protein precipitation by organic solvents (Gresele et al., 1986; Sweeney et al., 1987; Zakrzewski et al., 1988; Fradin et al., 1989), lipid extractions (Gresele et al., 1986; Sweeney et al., 1987; Patrignani and Canete-Soler, 1987; Zakrzewski et al., 1988; Fradin et al., 1989), thin-layer chromatography (Patrignani and Canete-Soler, 1987), and HPLC (Sweeney et al., 1987; Zakrzewski et al., 1988; Fradin et al., 1989). In our experiments parallelism between standard and sample dilution curves could not be obtained despite purifications by protein precipitation or lipid extraction on their own or in combinations. Nor did HPLC alone or in combination with the previous steps ensure proper purification. The only feasible

190

method comprised 1) solid-phase extraction followed by 2) protein precipitation at - 20°C and 3) HPLC. The latter was not able to remove interference in the RIA when the protein precipitation was performed at room temperature, although the only fraction showing immunoreactivity was eluted at a retention time identical to that of the standard. The interference was only removed after cooling of the sample to - 20°C and centrifugation prior to HPLC. We speculate that the precipitate formed might be of lipoprotein character, as the solubility of some lipoproteins in organic solvents decreases with decreasing temperature, and because this group of compounds may interfere with antibodies raised against lipids. The methods reported by Sweeney et al. (1987) and by Zackrzewski et al. (1988) using RIA without prior purification or after protein precipitation and solidphase extraction, were applied for the analysis of exogenous LTB4. Stimulation with calcium-ionophore not only induces LTB4 synthesis, but also activates arachidonic acid metabolism, in general. Using the same purification procedure we were not able to establish specificity in the RIA, which may be explained by the biosynthesis of related compounds during stimulation and by differences in antibody specificity. Antibody specificity may also account for the fact that protein precipitation as described by Gresele et al. (1986) was found not to eliminate interference. In the present study we have validated a method for LTB* in whole blood by combining the following criteria: 1) retention time of immunoreactive LTB4 in HPLC identical to synthetic standard, 2) purity of standard and radiolabelled standard verified by HPLC, 3) parallelism between sample and standard dilution curves, and 4) recovery from whole blood. Our data demonstrate the ins~ciency of validating specificity of an RIA exclusively by showing the immunoreactivity of only one HPLC fraction, as impurities may be retained by &he column, or one peak may contain more substances. Validation of an assay should include demonstration of parallelism between sample and standard dilution curves. Although no systematic variability was found, the overall interassay C.V. seems quite high. It should be emphasized that the experiments included for calculation of interassay C.V. were carried out by three different operators. Moreover, it is possible that a higher reproduction could be obtained by further standardization of incubation time and decantation procedure. Using the developed purification method and analvsis, we have investigated the dose-response relationship between calcium-ionophore and LTB4 production. Maximal production of LTB4 was seen at lo-20

JPM Vol. 28, No. 4 December 1992:185-190

p.M ionophore, which is below that found by Gresele et al., 1986. It should be emphasized that the content of the plasma following stimulation is the net LTB4 produced, that is, the difference between produced and consumed LTB4 in the culture. A higher level of LTB4 consumption may very well be induced at higher levels of ionophore, and the decline in LTB4 at 40pM ionophore may be caused by this. Interindividual differences may also account for the discrepancy, and further studies are necessary to elucidate this. Due to the biochemical homology between the LTs, we speculated that a similar approach as described above could be applied to analyze LT&. Using the LTB4 purification method before analyzing LTC, production in whole blood by RIA, we have also found parallelism between sample and standard dilution curves (to be reported elsewhere). Several biological factors including the kinetics of production and degradation vary between LTB4 and LTC*, thus further investigations are required to optimize the analysis for LTC4 following stimulations in whole blood. The authors thank Karsten Fogh for valuable discussions regarding the analysis, Anne Larsen for secretarial assistance, and Helene Jensen for skiiful technical assistance. This study was supported by grants from the Lundbeck Foundation, Thorvald Madsens Foundation, and Danish Medical Research Council.

References Fitzpatrick F, Liggett W, McGee J, Bunting S, Morton D, Samueisson B (1984) Metabolism of leukotriene A4 by human erythrocytes. A novel cellular source of leukotriene Bq. J Biol Chem 259: 11403-l 1407. Fradin A, Zirrolli JA, Maclouf J, Vausbinder L, Henson PM, Murphy RC (1989) Platelet-activating factor and leukotriene biosynthesis in whole blood. A model for the study of transcellular a~chidonate metabolism. J ~~munol 143(11):3680-3685. Gresele P, Arnout J, Coene MC, Deckmyn H, Vermylen J. (1986) Leukotriene Bd production by stimulated whole blood: Comparative studies with isolated polymorphonuclear cells. Biochem Biophys Res Commun 137(1):334-342. MacIouf J, de Laclos BF, Borgeat P (1982) Stimulation of leukotriene biosynthesis in human blood leukocytes by platelet-derived 12 hydroperoxy-icosatetraenoic acid. Proc Natf Acad Sci USA 79: 6042-6046. Marcus AJ, Brockman MJ, Safier LB, Ullman HL, Islam N (1982) Formation of leukotrienes and other hydroxy acids during platelet-neutrophil interactions in vitro. B&hem Biophys Res Commun 109:130-137. Patrignani P, Canete-Soler R (1987) Biosynthesis, characterization and inhibition of leukotriene Bd in human whole blood. Prostag/andjns 33(4):539-551. Sweeney FJ, Eskra JD, Catty TJ (1987) Development of a system for evaluating 5-lipoxygenase inhibitors using human whole blood. Prostaglandins Leukotrienes Medicine 28:73-93. Zakrzewski JT, Sampson AP, Evans$M, Barnes NC, Piper PJ, Costello JF (1988) The metabolism “in vitro” of leukotriene Bq in blood of normal subjects and asthmatic patients. Prostaglundins 35{6):849-883.