Europ. J. clin. Invest. 5, 311-318 (1975)
Radioimmunoassay of Prostaglandins Fa Eland E, in Human Plasma F. Dray, B. Charbonnel, and J. Maclouf Unit; de Radioimmunologie Analytique, Institut Pasteur, 28 rue du Dr. ROUX, Paris, France Received: May 31, 1974, and in revised form: January 6, 1975
Abstract. Antibodies against prostaglandins (PG)F2a,
and E2 were obtained in rabbits immunized with respectively PG F20, PG El and PG E2 conjugated to bovine serum albumin by carbodiimide. A radioimmunoassay capable of measuring 7 pg of PG Fa, 2 pg of PG E2 and 14 pg of PG El in human peripheral plasma is described. Plasma samples (pH 3, citric acid) are extracted with cyclohexane: ethyl acetate, 1 : l and then chromatographed on silicic acid columns to separate the prostaglandins into three fractions: fraction I, PG A, PG B and some unknown immunoreactive compounds; fraction 11, PG E and fraction 111, PG Fa. The recovery is 80 % f 6.2. Mean plasma levels adults of PG Fa and PG E, expressed in pg/ml: -PG Fa 12 f 2.8 (n= 25 men), 8 f 2.3 (n= 18 women, follicular phase), 7 f 1.4 (n= 18 women, luteal phase). -PG E I 40.5 f 7.6 (n= 13 men), 38 f 17.1 (n= 10 women). -PG E2 4.5 f 1 (n= 12 adult subjects). The major characteristics of the method described herein are the following: a large volume of plasma has to be processed (10 ml or more for PG Fa and PG El, 5 ml or more for PG E2). - a chromatographic step is necessary to separate the different prostaglandins wh{ch makes it possible to circumvent problems of immunological cross reactivity and interference with unknown immunoreactive compounds. - great care has been taken in collection of blood samples, especially to insure complete removal of blood cells namely platelets.
-
Key words: Prostaglandins, radioimmunoassay in human plasma.
Introduction
Standard
In the last three years, the possibility of using antiprostaglandin sera with sufficient binding affinities has allowed the development of radioimmunoassay for prostaglandins (PGs) at the picogram level (1-5, Table 7). This mekhod has been used to measure in particular, PG Fa and PG E in human plasma and serum. Meanwhile, the different values published are very scattered - 0.1 to 1.5 ng/ml for PG Fa as well as for PG E whose mean levels are higher - and seem too high from a physiological point of view. This paper concerns the critical study of a radioinrmunoassay of PG Fa, PG El and PG E2 in human plasma. The values presented are much lower than those published in the literature.
Unlabelled prostaglandins were a gift from Dr. J.E. Pike Upjohn Co. Tritiated prostaglandins were purchased from the New England Nuclear Corporation (PG Fla 79 Ci/ mmol, PG F2a 14 Ci/mmol, PG El 87 Ci/mmol and PG E2 125 Ci/mmol);their purity was checked by column chromatography on silicic acid. Standard solutions were prepared in methanol and stored at -2OOC. Prior to use an appropriahe aliquot was removed, dried and the prostaglandin redissolved in PBSG.
Materials and Methods Solvent Ethyl acetate (Merck, for spectroscopy), cyclohexane, benzene and methanol (Merck, for analysis) were used without further purification. Buffer (PBSG): 0.1 M phosphate pH 7.4 with 0.9 %. NaC1, sodium azide and 0.1 X gelatine.
0.1 %
Production of Antibodies 1. Prepration of I m o g e z s PG F2a, El or E2 (5 mg + 2 x lo5 dpm) was separately evaporated to dryness in a 25 ml conical flask. While maintaining a constant pH of 5.5, the following were added successively: -an aqueous solution of sodium carbonate (0.02 M, EtOH 10 %), 10 mg bovine serum albumin and 5 mg l-ethyl-3- (-3-diethylamine-propyl) carbodiimide, HC1. After 24 hours at room temperature, the reaction mixture was dialysed for 4 days against a phosphate buffered saline, 0.1 5 M NaC1, 0.01 M phosphate pH 7.4. The number of hapten molecules fixed per molecule of protein was determined by the measurement of specific activity of the radioactive material
31 2
F. Dray d al. : Radioimmunoassay of Prostaglandins Fa, El and E2
before and after coupling: 18 molecules of PG Fa, 22 molecules of PG El and 20 molecules of PG E2 were fixed per molecule of BSA.
I
Solvent 2 Solvent 1 Solvent L 2. I m i z a t i m . The immunization was carried B/EA/MeOH 60 :LO: 20 out using a modification of the method proposed by Vaitukaitis et aZ. (6). A volume of each dialysate corresponding to 0 . 2 mg of conjugate per animal was emulsified with complete Freund's adjuvant and the emulsion injected intradermally into 5 adult male rabbits at 30 - 40 points: on the same day one ampoule of Bordetella Pertussis was injected subcutaneously. The evolution of the titre of the antisera was followed each week, from the third week. The booster was only carried out on those animals which had responded, when titre had diminished significantly. This booster consisted 6 ml 13 ml L ml of a daily injection of conjugate for three days, intramuscular the first day and intravenous the Fig. 1 . Chromatographic separation of PG E and PG following two days. A study of the specificity Fa on silicic acid column (see the text for legends) and affinity was carried out on the antisera with a sufficient titre, and the best of these, after sterile filtration and-addition of 0.02 X sodium azide were stored at 4OC or at - 20°C after dilution tion 11: PG El and 5 ml of solvent 2 (fraction to 1/2 with glyerol. 111: PG Fa). Each of the last two frgctions 11 and 111, was evaporated in air at 50 C and redisPlasma Extraction solved in the assay buffer:0.8 ml for fraction I1 (0.3 ml for recovery, 2 x 0.1 ml for PG El assay, The blood, taken in the morning from a fasting 2 x 0.1 ml for PG E2 assay) - 0.6 ml for fraction 0 subject, was collected at 4 C under vacuum in a I11 (0.3 ml for recovery 2 x 0.1 ml for P C Fa plastic tube containing an anticoagulant (EDTA, ,assay). disodium salt-dry 14 mg, for 10 ml of plasma). It We have tried to eliminate many fatty sub0 was centrifuged immediately at 4 C for I5 min. at stances by prior plasma extraction with petroleum 2450 g. The plasma was decanted at once and mainether or hexane, or by three extractions, first 0 tained at 4 C if the extraction was carried out in by the method described, then addition of alkali the 60 min. following collection, or stored at to pH 8 and extraction with phosphate buffer then -2OOC. re-acidification and extraction with ethyl acetate. Extraction was effected in glass tubes. To this The recovery by this method was correct (= 8 5 X) plasma (10 ml or more) were added 1800 dpm (3.9pg) but there was no modification in the final results. of 3H-PG Flay 1800 dpm (2.9 pg) of 3H-PG El for the calculation of recovery. The plasma was then acidified to pH 3 with citric acid. The prostaglan- Radioimmunoassay System dins were extracted twice with three times the For each assay (i.e., PG Fa (1+2),PG Ei and plasma volume of a mixture of cyclohexane/ethyl PG E2) 0.1 ml of tritiated PG Flu, PG El or PG E2 acetate ( 1 : l ) with vigorous shaking for 15 min. (7500 dpm) and 0.1 ml of either unlabelled PG (3 After centrifugation at 250 g for 10 min., the organic phase was 0separated, reduced by evaporation to 100 pg for PG F2a, 5 to 300 pg for PG El, 2 to 100 pg for PG E2) or the unknown sample as a under vacuum at 50 C and evaporated to dryness in silicic acid fraction was incubated with 0.1 ml a siliconized glass conical tube. of appropriately diluted antiserum: final dilution 1/36OOO for anti PG Fza, 1/54000 for anti PG El Chromatographic Separation of PG Fa and PG E and 1/5000 for anti PG E2. Under these conditions, approximately 35 Z of tritiated hapten was bound We used microcolumns (0.5 x 6.0 cm) of 0.5 g in the absence of unlabelled hapten. The incunon-activated silicic acid (100 Mesh-Mallinckrodt) bation time was 2 h at 4 O C , although equlibrium equilibrated in 2 ml of solvent 1 (benzene/ethylwas attained after 30 min. The free fraction was acetate, 60:40). Each column was washed with 5.0 separated from the fraction bound to the antibody ml of solvent 2 (benzene/ethyl-acetate/methanol, with dextran-charcoal whose working conditions 60:40:20) and 1.5 ml of solvent 1. Sample residues have been studied ( 7 ) . One ml of dextran-charcoal were vortexed in 0.2 ml of solvent 3 (benzenelethyl- (Norit A, 250 mg, dextran T 70, 25 mg/AQO ml PBSG) acetate/methanol 60:40:10) then in 0.8 ml of was added to the reaction mixture at 4 C. Ten solvent 1 and successively applied to the columns. minutes later, the tubes were centrifuged for PG fractions were obtained (Fig. I ) by developing 5 min. at 2000 g, the supernatant decanted into the columns serially with 5.5 ml of solvent 1 counting vials and 10 ml Instagel (Packard) added. (PG A, PG B and pigments) 13 ml of solvent 4 Counting efficiency in an "Intertechnique" Spectro(benzene/ethyl acetate/methanoL, 60:40:2) (fracmeter was 27 X.
F. Dray et aZ. : Radioimunoassay of Prostaglandins Fa, EI and E2
RasuZts
B/Box 100
Antibody Binding Parameters
90
1. A f f i n i t y and T i t r e . The association constant (Ka) and the antibody concentration (n = sites number) of each selected antiserum were calculated from a curve established according to Scatchard, obtained by equilibrium dialysis. They are as follows : -PG F2a antiserum # A193 708. Ka = 6 x 109 M-1 n = 2 x IO-I@l
80-
-PG El antiserum# A179 585 Ka = 4.8 x 308M-1
n = 5 x lO-l@l
-PG E2 antiserum # A193 709 Ka = 1.8 x IOI@l-l
n = 0.3 x I O - I h
-
Table I . Cross reactions of prostaglandins (X) with anti PG serum calculated from mass displacing 50 % of the corresponding tracer Prostaglandins
PG F2a-AS
PG El-AS 1
PG E2-AS 0.04
-
-
70
6050-
LO
2. S p a i f i c i t y (Table I ) . # 1 PG F2a antiserum: the percentage cross-reactions (CR %) were calculated from the relative amounts of PG F2a and the test compound required to reduce the intial binding of 3H-PG F20 by 50 9.. Table 1 shows that this antiserum cross-reacted completely with PG F2a. This property allowed us to use as tracer 3H-PG Flu, which had a high specific activity. Since PG F2a and PG F l u were not separated on silicic acid chromatography, our results correspond exactly to the sum of PG Flu and PG F2a. The other prostaglandins were very weakly recognized, CR
Radioimmunoassay of Prostaglandins Fa Eland E, in Human Plasma F. Dray, B. Charbonnel, and J. Maclouf Unit; de Radioimmunologie Analytique, Institut Pasteur, 28 rue du Dr. ROUX, Paris, France Received: May 31, 1974, and in revised form: January 6, 1975
Abstract. Antibodies against prostaglandins (PG)F2a,
and E2 were obtained in rabbits immunized with respectively PG F20, PG El and PG E2 conjugated to bovine serum albumin by carbodiimide. A radioimmunoassay capable of measuring 7 pg of PG Fa, 2 pg of PG E2 and 14 pg of PG El in human peripheral plasma is described. Plasma samples (pH 3, citric acid) are extracted with cyclohexane: ethyl acetate, 1 : l and then chromatographed on silicic acid columns to separate the prostaglandins into three fractions: fraction I, PG A, PG B and some unknown immunoreactive compounds; fraction 11, PG E and fraction 111, PG Fa. The recovery is 80 % f 6.2. Mean plasma levels adults of PG Fa and PG E, expressed in pg/ml: -PG Fa 12 f 2.8 (n= 25 men), 8 f 2.3 (n= 18 women, follicular phase), 7 f 1.4 (n= 18 women, luteal phase). -PG E I 40.5 f 7.6 (n= 13 men), 38 f 17.1 (n= 10 women). -PG E2 4.5 f 1 (n= 12 adult subjects). The major characteristics of the method described herein are the following: a large volume of plasma has to be processed (10 ml or more for PG Fa and PG El, 5 ml or more for PG E2). - a chromatographic step is necessary to separate the different prostaglandins wh{ch makes it possible to circumvent problems of immunological cross reactivity and interference with unknown immunoreactive compounds. - great care has been taken in collection of blood samples, especially to insure complete removal of blood cells namely platelets.
-
Key words: Prostaglandins, radioimmunoassay in human plasma.
Introduction
Standard
In the last three years, the possibility of using antiprostaglandin sera with sufficient binding affinities has allowed the development of radioimmunoassay for prostaglandins (PGs) at the picogram level (1-5, Table 7). This mekhod has been used to measure in particular, PG Fa and PG E in human plasma and serum. Meanwhile, the different values published are very scattered - 0.1 to 1.5 ng/ml for PG Fa as well as for PG E whose mean levels are higher - and seem too high from a physiological point of view. This paper concerns the critical study of a radioinrmunoassay of PG Fa, PG El and PG E2 in human plasma. The values presented are much lower than those published in the literature.
Unlabelled prostaglandins were a gift from Dr. J.E. Pike Upjohn Co. Tritiated prostaglandins were purchased from the New England Nuclear Corporation (PG Fla 79 Ci/ mmol, PG F2a 14 Ci/mmol, PG El 87 Ci/mmol and PG E2 125 Ci/mmol);their purity was checked by column chromatography on silicic acid. Standard solutions were prepared in methanol and stored at -2OOC. Prior to use an appropriahe aliquot was removed, dried and the prostaglandin redissolved in PBSG.
Materials and Methods Solvent Ethyl acetate (Merck, for spectroscopy), cyclohexane, benzene and methanol (Merck, for analysis) were used without further purification. Buffer (PBSG): 0.1 M phosphate pH 7.4 with 0.9 %. NaC1, sodium azide and 0.1 X gelatine.
0.1 %
Production of Antibodies 1. Prepration of I m o g e z s PG F2a, El or E2 (5 mg + 2 x lo5 dpm) was separately evaporated to dryness in a 25 ml conical flask. While maintaining a constant pH of 5.5, the following were added successively: -an aqueous solution of sodium carbonate (0.02 M, EtOH 10 %), 10 mg bovine serum albumin and 5 mg l-ethyl-3- (-3-diethylamine-propyl) carbodiimide, HC1. After 24 hours at room temperature, the reaction mixture was dialysed for 4 days against a phosphate buffered saline, 0.1 5 M NaC1, 0.01 M phosphate pH 7.4. The number of hapten molecules fixed per molecule of protein was determined by the measurement of specific activity of the radioactive material
31 2
F. Dray d al. : Radioimmunoassay of Prostaglandins Fa, El and E2
before and after coupling: 18 molecules of PG Fa, 22 molecules of PG El and 20 molecules of PG E2 were fixed per molecule of BSA.
I
Solvent 2 Solvent 1 Solvent L 2. I m i z a t i m . The immunization was carried B/EA/MeOH 60 :LO: 20 out using a modification of the method proposed by Vaitukaitis et aZ. (6). A volume of each dialysate corresponding to 0 . 2 mg of conjugate per animal was emulsified with complete Freund's adjuvant and the emulsion injected intradermally into 5 adult male rabbits at 30 - 40 points: on the same day one ampoule of Bordetella Pertussis was injected subcutaneously. The evolution of the titre of the antisera was followed each week, from the third week. The booster was only carried out on those animals which had responded, when titre had diminished significantly. This booster consisted 6 ml 13 ml L ml of a daily injection of conjugate for three days, intramuscular the first day and intravenous the Fig. 1 . Chromatographic separation of PG E and PG following two days. A study of the specificity Fa on silicic acid column (see the text for legends) and affinity was carried out on the antisera with a sufficient titre, and the best of these, after sterile filtration and-addition of 0.02 X sodium azide were stored at 4OC or at - 20°C after dilution tion 11: PG El and 5 ml of solvent 2 (fraction to 1/2 with glyerol. 111: PG Fa). Each of the last two frgctions 11 and 111, was evaporated in air at 50 C and redisPlasma Extraction solved in the assay buffer:0.8 ml for fraction I1 (0.3 ml for recovery, 2 x 0.1 ml for PG El assay, The blood, taken in the morning from a fasting 2 x 0.1 ml for PG E2 assay) - 0.6 ml for fraction 0 subject, was collected at 4 C under vacuum in a I11 (0.3 ml for recovery 2 x 0.1 ml for P C Fa plastic tube containing an anticoagulant (EDTA, ,assay). disodium salt-dry 14 mg, for 10 ml of plasma). It We have tried to eliminate many fatty sub0 was centrifuged immediately at 4 C for I5 min. at stances by prior plasma extraction with petroleum 2450 g. The plasma was decanted at once and mainether or hexane, or by three extractions, first 0 tained at 4 C if the extraction was carried out in by the method described, then addition of alkali the 60 min. following collection, or stored at to pH 8 and extraction with phosphate buffer then -2OOC. re-acidification and extraction with ethyl acetate. Extraction was effected in glass tubes. To this The recovery by this method was correct (= 8 5 X) plasma (10 ml or more) were added 1800 dpm (3.9pg) but there was no modification in the final results. of 3H-PG Flay 1800 dpm (2.9 pg) of 3H-PG El for the calculation of recovery. The plasma was then acidified to pH 3 with citric acid. The prostaglan- Radioimmunoassay System dins were extracted twice with three times the For each assay (i.e., PG Fa (1+2),PG Ei and plasma volume of a mixture of cyclohexane/ethyl PG E2) 0.1 ml of tritiated PG Flu, PG El or PG E2 acetate ( 1 : l ) with vigorous shaking for 15 min. (7500 dpm) and 0.1 ml of either unlabelled PG (3 After centrifugation at 250 g for 10 min., the organic phase was 0separated, reduced by evaporation to 100 pg for PG F2a, 5 to 300 pg for PG El, 2 to 100 pg for PG E2) or the unknown sample as a under vacuum at 50 C and evaporated to dryness in silicic acid fraction was incubated with 0.1 ml a siliconized glass conical tube. of appropriately diluted antiserum: final dilution 1/36OOO for anti PG Fza, 1/54000 for anti PG El Chromatographic Separation of PG Fa and PG E and 1/5000 for anti PG E2. Under these conditions, approximately 35 Z of tritiated hapten was bound We used microcolumns (0.5 x 6.0 cm) of 0.5 g in the absence of unlabelled hapten. The incunon-activated silicic acid (100 Mesh-Mallinckrodt) bation time was 2 h at 4 O C , although equlibrium equilibrated in 2 ml of solvent 1 (benzene/ethylwas attained after 30 min. The free fraction was acetate, 60:40). Each column was washed with 5.0 separated from the fraction bound to the antibody ml of solvent 2 (benzene/ethyl-acetate/methanol, with dextran-charcoal whose working conditions 60:40:20) and 1.5 ml of solvent 1. Sample residues have been studied ( 7 ) . One ml of dextran-charcoal were vortexed in 0.2 ml of solvent 3 (benzenelethyl- (Norit A, 250 mg, dextran T 70, 25 mg/AQO ml PBSG) acetate/methanol 60:40:10) then in 0.8 ml of was added to the reaction mixture at 4 C. Ten solvent 1 and successively applied to the columns. minutes later, the tubes were centrifuged for PG fractions were obtained (Fig. I ) by developing 5 min. at 2000 g, the supernatant decanted into the columns serially with 5.5 ml of solvent 1 counting vials and 10 ml Instagel (Packard) added. (PG A, PG B and pigments) 13 ml of solvent 4 Counting efficiency in an "Intertechnique" Spectro(benzene/ethyl acetate/methanoL, 60:40:2) (fracmeter was 27 X.
F. Dray et aZ. : Radioimunoassay of Prostaglandins Fa, EI and E2
RasuZts
B/Box 100
Antibody Binding Parameters
90
1. A f f i n i t y and T i t r e . The association constant (Ka) and the antibody concentration (n = sites number) of each selected antiserum were calculated from a curve established according to Scatchard, obtained by equilibrium dialysis. They are as follows : -PG F2a antiserum # A193 708. Ka = 6 x 109 M-1 n = 2 x IO-I@l
80-
-PG El antiserum# A179 585 Ka = 4.8 x 308M-1
n = 5 x lO-l@l
-PG E2 antiserum # A193 709 Ka = 1.8 x IOI@l-l
n = 0.3 x I O - I h
-
Table I . Cross reactions of prostaglandins (X) with anti PG serum calculated from mass displacing 50 % of the corresponding tracer Prostaglandins
PG F2a-AS
PG El-AS 1
PG E2-AS 0.04
-
-
70
6050-
LO
2. S p a i f i c i t y (Table I ) . # 1 PG F2a antiserum: the percentage cross-reactions (CR %) were calculated from the relative amounts of PG F2a and the test compound required to reduce the intial binding of 3H-PG F20 by 50 9.. Table 1 shows that this antiserum cross-reacted completely with PG F2a. This property allowed us to use as tracer 3H-PG Flu, which had a high specific activity. Since PG F2a and PG F l u were not separated on silicic acid chromatography, our results correspond exactly to the sum of PG Flu and PG F2a. The other prostaglandins were very weakly recognized, CR
