The development of a radioimmunoassay for 12-L-hydroxyeicosatetraenoic acid

PROSTAGLANDINS

THE DEVELOPMENT OF A RADIOIMMUNOASSAYFOR 12-L-HYDROXYEICOSATETRAENOIC ACID Lawrence Levine, Iftekhar Alam, Hilda Gjika, Thomas J. Carty* and Edward J. Goetzl** Department of Biochemistry,Brandeis University, Waltham, MA 02254; *Pfizer Central Research, Groton, CT 06340; **The Howard Hughes Medical Institute Laboratory at Harvard Medical School, the Departments of Medicine, Harvard Medical School, and the Robert B. Brigham Hospital, Boston, MA 02115 ABSTRACT Antibodies directed toward 12-L-hydroxyeicosatetraenoic acid (12-L-HETE)were generated in rabbits by immunizationwith conjugates of 12-L-HETE and hunranserum albumin. The concentration of antibodies was determined by incubating immune plasma with 12-L-HETE that had been covalently linked to a solid support, washing the 12L-HETE support, and measuring the quantity of bound antibodies by reaction with [125I]ProteinA. The addition of 0.5 ng-10 ng of fluid-phase 12-L-HETE to the standard mixture of solid-phase 12-LHETE and anti-12-L-HETEplasma inhibited by 21-80% the binding of antibodies and consequently of [125IlProteinA to the solid support. The 12-OH function positioned between two double bonds was the immunodominant determinant of this antigen-antibodyreaction, but the carboxyl function also was recognized. This radioimmunoassaywas used to detect and quantitate 12-L-HETE resolved by high pressure liquid chromatography. INTRODUCTION Since the first demonstrationby Levine and Van Vunakis (1) that the prostaglandins are antigenic when covalently bound to a carrier molecule, antibodies have been produced to various cyclooxygenase products of arachidonic acid metabolism (2) and utilized in specific radioimmunoassays(2). The antigenicity of these haptenic groups in the conjugates probably reflects the cyclic portion of their structure. Historically, the immunogenic activity of lipids is poor and antibodies have been produced only by immunizationwith "lipids" that contain rigid structure, e.g. lipids that contain sugar moieties such as cytolipin H (3), globoside (4), and sulfatide (5). The sugar moieties were immunodominant,but the antibodies may have recognized some lipid function as well (6,7). It has been difficult to raise antibodies to lipids that have a high degree of flexibility. The lipoxygenaseproducts of arachidonic acid metabolism are unsaturated fatty acids that are structurallymore rigid than saturated fatty acids and possess one or more hydroxyl functions. A conjugate of the platelet-derivedlipoxygenaseproduct, 12-L-hydroxyeicosatetraenoicacid, and human serum albumin has been utilized

NOVEMBER

1980VOL.20NO.S

923

PROSTAGLANDINS

successfully to raise highly specific antibodies in rabbits. The concentrationof the resultant antibodies was assessed by utilizing [125I]ProteinA to quantitate the amount of antibody bound to 12-Lhydroxyeicosatetraenoicacid that had been coupled covalently to a solid support. ?&is procedure had been used to develop radioimmunoassays for methotrexate (8), leucovorin (9), indomethacin (lo), as well as cyclooxygenaseproducts of arachidonic acid metabolism (11). The antibodies to 12-L-hydroxyeicosatetraenoic acid recognize predominantly the 12-hydroxy group between the double bonds at positions 10 and 14, as well as the carboxyl function and possibly the sequence of the cis/trans double bonds. MATERIALS AND METHODS acid (12-L-HETE)was Reagents. 12-L-hydroxyeicosatetraenoic prepared by incubating indomethacin-treatedhuman platelets (Pioneer Blood Service, NY) with [l-14Clarachidonicacid (New England Nuclear), specific radioactivity 106 dpm/mg,accordingto the published method (12). Following ether extraction of an acidified reaction mixture, the methyl ester of 12-L-HETE was synthesized using diazomethane, and isolated by chromatographyon thin layer plates (Silica Gel 60 F254, EM Reagents) developed with isopropyl ether. The ester was hydrolyzed with dilute base. The products, 12-L-HETE and its methyl ester, analyzed by mass spectrometry and NMR spectroscopyrespectively, had properties in agreement with previously published data (13,14). Preparation and characterizationof the ll-HETE, 8-HETE and 5-HETE used in this study have been reported (15). Conjugate for immunization. The 12-L-HETE-humanserum albumin conjugate was prepared according to the procedure of Bauminger (16). Eight mg of 12-L-HETE was dissolved in 100 1.11 of N,N-dimethylformamide. Three mg of dicyclohexylcarbodiimideand 3.5 mg of N-hydroxysuccinimide were added and the mixture was stirred at room temperature for 30 min. The precipitated N,N'-dicyclohexylureawas removed by centrifugationand the supernatant fluid was added to a solution of 12.5 mg of human serum albumin in 500 ~1 of 0.1 N sodium bicarbonate. The reaction mixture was stirred at 4O for 2 hours and then was dialyzed in the cold against 5 changes of 2 liters of phosphate buffer (0.15 M NaCl; 0.005 M sodium phosphate, pH 7.5). After dialysis, the volume was 2.5 ml. The conjugates were stored at -20'. Antisera. Two rabbits were immunized by injecting a total of 2.0 mg of the conjugate in complete Freund's adjuvant intramuscularly in the hind limbs and in the toe pads. Two weeks later, and in each of the following 3 weeks, 40 ml of blood was collected in 4.0,ml of 0.1 M EDTA, 0.14 M NaCl, pH 7.0. The plasma was obtained by centrifugation and stored at -20'. Three months after the primary injection, the rabbits were boosted by the same route with 3.0 mg of the conjugate in complete Freund's adjuvant. Plasma was obtained one week later and for successive weeks after the boost. A third course of immunization similar to the second boost was given 3 months

924

NOVEMBER 1980VOL.20N0.5

PROSTAGLANDINS

later, and again plasma was obtained weekly for several weeks. rabbits produced antibodies to 12-L-HETE.

Both

Immobilized ligands. Solid-phase absorbents were prepared by a modification of the reported procedure (17). Eighty ng of 12-L-HETE in 10 yl of ethanol, was added to 0.3 ml of 0.003 M sodium phosphate buffer (pH 7) at 0' followed by the addition of 50 mgs of l-ethyl3-(3-dimethylaminopropyl) carbodiimide. This clear solution was added immediately to a stirring suspension of affi-gel 701 (BIO-RAD Laboratories, Richmond, CA) which, after centrifugation, had been resuspended (500 mg) in 5 ml of 0.003 M sodium phosphate buffer (pH 7). Stirring was continued at room temperature for 18 hr. Five mgs of the CD1 was then added and stirring was continued for another 4 hours. The affi-gel suspension was centrifuged at 2,000 rpm for 10 minutes. The beads were washed at 4" successively with water (3 ml), ethanol (3 ml) and Tris buffer (pH 7.2). Finally, the beads were resuspended in 25 ml Tris buffer containing 0.1% gelatin and 0.02% NaN3. In the immunoassays, 50 ~1 (1 mg) of this suspension was used. iodination of Protein A. The preparative procedure employed has been described (8). Five hundred pCi of Bolton-Hunter reagent (New England Nuclear, 2000 Ci/mmol) in 50 1.11 of benzene was evaporated with a gentle stream of nitrogen in the same "combi-v-vial" shipped by New England Nuclear, Boston, immediately prior to use. Five ug of Protein A in 25 ~1 of 0.1 M sodium phosphate buffer (pH 8) was added to the vial at O'C. The reaction mixture was allowed to stand at room temperature for two hours. Then, 0.5 ml of 0.1 M sodium phosphate buffer (pH 8) containing 0.2 M glycine was added and the mixture was incubated at room temperature for 10 min. The reaction mixture was chromatographed over G-25 Sephadex with 0.05 M sodium phosphate buffer, pH 7.5, containing 0.15 M NaCl and 0.2% gelatin. The G-25 M Sephadex column was a prepacked column (Pharmacia Fine Chemicals, Piscataway, NJ). One ml fractions were collected. The iodinated Protein A eluted in tubes 3 and 4. Protein A incorporates 20-55% of the Bolton-Hunter reagent. With Bolton-Hunter reagent of higher specific activity and less Protein A, specific activities as high as 10,000 Ci/mmol have been obtained (11). High pressure liquid chromatography. A Waters Associates (Milford, MA) instrument with a pPorasi1 silica gel column was used for the high pressure liquid chromatography. The solvent program system for the resolution of arachidonic acid metabolites consisted of isocratic elution with hexane : glacial acetic acid (125:1, vol:vol; Solvent A) for 11 minutes, followed by a linear gradient--from Solvent A to chloroform: methanol : glacial acetic acid (125:5:1, vol: vol:vol; Solvent B) over 76 minutes, and then isocratic elution with 100% Solvent B for 13 minutes. The flow rate for the entire program was set at 1 mlfmin.

NOVEMBER 1980VOL.20NO.5

925

PROSTAGLANDINS Radioimmunoassay. This radioimmunoassay which measures binding of specific antibodies to immobilfzed ligands by uptake of [125IiProtein A has been described in detail (11). RESULTS The quantity of rabbit immunoglobulins (IgG) that bound to the solid-phase 12-L-HETE increased during the course of immunization as measured by the progressively greater amounts of [125I]Protein A attached to the washed beads (Fig. 1). This binding was specific for 12-L-HETE as it was not observed with succinylnicotine-beads, 6-keto-PGEl-beads or N-dimethylguanosine-beads. The extent of binding was dependent on the volume of immune plasma that had been incubated with the solid-phase 12-L-HETE (Fig. 2). The quantity of anti-12-L-HETE antibodies in 0.0125 pl of immune plasma from a later bleeding was readily quantitated b,y binding to the solid-phase 12-L-HETE as reflected in the adherence of 7,600 cpm of a total of 61,000 cpm of [125IlProtein A, as compared to the adherence of 634 cpm with 0.0125 1.11 of non-immune plasma. The ability of fluid-phase 12-L-HETE to inhibit the binding of antibodies to the solid phase 12-L-HETE is shown in Fig. 3. Increments of 0.5 ng10 ng of fluid-phase 12-L-HETE in the standard mixture of solidphase 12-L-HETE and anti-12-L-HETE plasma inhibited by 21-80% the binding of antibodies and consequently of [125I]Protein A to the solid support. In contrast, the presence of related lipids in the fluid phase failed to inhibit the binding (Table 1). Interference in this immune system by heterologous ligands can not be evaluated until the actual extent of cross reaction with all of the heterologous hydroxyfatt,y acids ;; measured. The 12-L-HETE inhibition curve was used to quantitate [ C]12-L-HETE resolved by high pressure liquid chromatography (Fig. 4). The position and quantity of 12-LHETE in the chromatogram were assessed by the 14C radioactivity and the 12-L-HETE radioimmunoassay (Fig. 4). As estimated from the specific activity of the 12-L-HETE before separation by HPLC, 8,000 ng of 12-L-HETE was recovered, whereas 2,200 ng of 12-L-HETE was measured by radioimmunoassay. The reason for this 3.6-fold difference is not known, but it may reflect heterogeneity in the [14C]12-L-HETE before HPLC. Estimations of 12-L-HETE as measured by radioimmunoassay and radiochemical assay after incubation of [14C]arachidonic acid with platelet lipoxygenase were in agreement (18). Quantitation of the binding of specific IgG to the 12-L-HETEbeads depends on the density of ligand on the beads as well as the specific activity of the [125I]Protein A. Thus, for the studies of the specificity of the 12-L-HETE immune system as well as its use in the measurement of 12-L-HETE resolved by high pressure liquid chromatography, one preparation of 12-L-HETE-beads and one preparation of [125I]Protein A were used. However, the specificities obtained with other preparations of immobilized ligands and [125I]Protein A were similar and the measurement of chromatographic properties of

926

NOVEMBER 1980VOL.20NO.S

PROSTAGLANDINS

Figure 1. Binding of [1251]Protein A by the anti-12-L-HETE antibodies in 0.5 ~1 of immune plasma. The immune plasmas were obtained from a rabbit three weeks after the primary injection of human albumin-12-L-HETEconjugates, and three weeks after each "booster" injection.

5.0 -

“0 4.0 Figure 2. Binding of [12%]‘; Protein A to anti-12-L-HETEantibodies in various volumes of rabbit P g 3.0 immune plasma obtained three weeks 0 after the third course of immuniza2 tion. The increments of immune plasma had been incubated with 1 mg r 2.08 of solid-phase 12-L-HETE. t k l.OU

0

/

I

I

I

0.01

0.1

I

*I

NOVEMBER 1980VOL.20NO.S

Immune Plarmo

927

PROSTAGLANDINS

80

.g 60 ‘Z .c = #

40

20 0 I ng l2-OH

IO Eicosotatraenoic

Acid

Figure 3. Inhibition of the binding of anti-12-LHETE antibodies to solid-phase 12-L-HETE by varying quantities of fluid-phase 12-L-HETE. The immune plasma was obtained 6 weeks after the third course of immunization. This inhibition curve was generated with three preparationsof 12-L-HETE. Duplicate analyses were made for each level. The values agreed within 20% of the mean values shown.

928

NOVEMBER 1980VOL.20NO.S

Froc tion

60

/

?

E2 X

Figure 4. Separation by high pressure liquid chromatography of arachidonic acid (AA), 12-L-HETE, PGB2, PGD2, 6-keto-PGF1, (6-k-Flu), TXB2, PGE2, and PGF20. The standard arachidonic acid metabolites were radiolabeled with 3H (0, o) and 14C (A). [l4C]12-L-HETE was also measured serologically by inhibition of the binding of antibodies to solid-phase 12-L-HETE (Fig. 3).

40

/

//-

lsocratic ---

PROSTAGLANDINS

TABLE 1.

SEROLOGIC SPECIFICITY OF THE 12-L-HETE IMMUNE SYSTEM Serologic Activity (ng for 50% inhibition)

Ligand 12-L-HETE Methylester-12-HETE 12-L-hydroxyheptadecatrienoic acid (HHT) ll-HETE 5-HETE 8-HETE Ricinoleic acid Arachidonic acid PGEl, PGE2, PGFlo, PGF20, 6-keto-PGF1,, TXB2

> > > > >

1.9 9.5 22.0 10. a lo. a 10. b 10. a 10. a

.lOO. c

a 0% inhibition with 10 ng b 23% inhibition with 10 ng c 0% inhibition with 100 ng

12-L-HETE was identical. DISCUSSION The 12-L-HETE hapten, coupled to human albumin, was not strongly immunogenic. The rabbit plasma that had the highest antibody titer when assessed by the solid-phase 12-GHETE assay did bind univalent 13H]12-L-HETE,but 4 ul of the immune plasma only bound 37% of the total of 3000 cpm. In the'solid-phase12-L-HETE assay, the binding of 12-L-HETE by 0.012 ~1 of the IgG antibodies in immune plasma was easily detectable. This 300-fold greater sensitivity in measurement of binding could reflect the multivalent properties of the solid-phase 12-L-HETE or the contribution of the "link" in the ligand. However, a 12-L-HETE[1251]tyrosineligand bound even less effectively than the [3Hil2-LHETE, suggesting that the multivalence of the ligand was responsible for the more efficient binding. As shown by Xarush, a multivalent antigenic particle can have an association constant 104-fold higher than that between the same antibodies and the univalent hapten (19). While the use of the multivalent-immobilized-12-L-HETE to detect anti-12-L-HETE is advantageous, the lower affinity of the univalent 12-L-HETE for the antibody receptor sites decreases its competitive effectiveness. An ideal immobilized-12-L-HETEantigen would be one that increases the association of 12-L-HETE-antibodiesenough so that binding can readily be measured but not enough to make competition by the univalent 12-L-HETE extremely unfavorable. The ability

930

NOVEMBER 198OVOL.20NO.S

PROSTAGLANDINS

to detect different levels of anti-12-L-HETEbound to the multivalent Immobilized-12-L-HETEwill depend also on the specific activity of the radiolabeled Protein A. Calibration curves of inhibition must be obtained with every preparation of solid-phase 12-L-HETE and with each preparation of [125I]ProteinA. Another potential disadvantage of the immobilized-multivalentligand radioimmunoassayis that the immobilized-ligandcomplex may contain a ,uniqueantigenic determinant that is lacking in the free hapten, e.g. the amide bond "link." The 12-L-HETE in biological fluids, lacking this determinant,might have a lower association constant for the antibodies and thus might compete less effectively. Nevertheless, even with this hapten of weak immunogenic properties, we have been able to generate and quantitativelyestimate homologous antibodies. Such antibodies are specific; they react with the homologous 12-L-HETE, but react poorly with related ligands. The immunodominance of the 12-OH function between two double bonds is clear (Table 1). The contribution to binding of the carboxyl function also can be seen from the decreased inhibition with the methyl ester of 12-L-HETE. Most likely, the OH function, or its position, is associated with altered conformations of these unsaturated fatty acids which in turn affects their association with the anti-12-L-HETE. Lipoxygenase-derived12-L-HETE is produced by platelets (ZO), rat mast cells (Zl), human neutrophils (22) and probably human lymphocytes (23). Purified 12-L-HETE has chemotactic and chemokinetic activities for neutrophils and eosinophils_in vitro and stimulates the peritoneal accumulation of neutrophils and eosinophils in guinea Pigs (15). Several other arachidonic acid metabolites of the lipoxygenase pathway have been characterized (24-26). Clearly, other HETEs will also prove to be immunogenic, and of more importance, molecules such as the leucotrieneswith potentially immunodominantcysteinylcontaining functions as well as the OH functions also would generate specific antibodies. Development of radioimmunoassaysfor SRS-like compounds will be helpful in evaluating their roles in anaphylaxis and asthma. ACKNOWLEDGEMENTS This work was supported by Grant GM-27256, Grant CA-17309, and Grant HL-19777 from the National Institutes of Health. This is publication 1311 from the Department of Biochemistry,Brandeis University, Waltham, Massachusetts 02254. LL is a Research Professor of Biochemistry of the American Cancer Society (Award PRP-21). IA was supported by Training Grant No. AM 07251 from the National Institutes of Health.

NOVEMBER 1980VOL.20NO.S

931

PROSTAGLANDKNS

REFERENCES 1.

Levine, L. and Van Vunakis, H. Antigenic Activity of Prostaglandins. Biochem. Biophys. Res. Commun. -4?:888-896. 1972.

2.

Granstroem, E. Radioimmunoassayof Prostaglandins.Prostaglandins -15:3-l?. 1978.

3.

Joffe, S., Rapport, M., and Graf, L. Identificationof an Organ Specific Lipid Hapten in Brain. Nature 19?:60-62. 1963.

4.

Koscielak, J., Hakomori, S., and Jeanloz, R.W. Glycolipid Antigen and its Antibody. Immunochemistry5:441-455. 1968.

5. Hakomori, S. Preparation and Properties of Anti-Sulfatide Serum. J. Immunol. 112:424-426. 1974. 6. Rapport, M.M., and Graf, L. ImmunochemicalReactions of Lipids. Prog. Allergy -13~273-331. 1969. 7. Alving, C.R. Immune Reactions of Lipids and Lipid Model Membranes. In: The Antigens Vol. IV. (M. Sela, ed.) Academic Press, New York, 1977. pp. l-72. 8.

Langone, J. [125I]ProteinA: a Tracer for General Use in Immunoassay. J. Immunol. Methods -24:269-285. 1978.

9.

Langone, J.J., and Levine, L. Immunoassay of Leucovorin: Use of 1251-LabeledProtein A to Detect ImmunologicalBinding. Anal. Biochem. -95:4?2-478. 1979.

10. Alam, I., Langone, J.J., and Levine, L. Immunoassay of Indomethacin. The Use of [125I]ProteinA to Detect Specific Serologic Binding. Prostaglandinsand Medicine 2:16?-175. 1979. 11. Levine, L., Alam, I., and Langone, J.J. The Use of Immobilized Ligands and [125I]ProteinA for Immunoassaysof Thromboxane B2, Prostaglandin D2, 13,14-Dihydro-Prostaglandin E2, 5,6-DihydroProstaglandin 12, 6-Keto-ProstaglandinFlu, 15-Hydroxy-9a,lla(Epoxymethano)Prosta-5,13-Dienoic Acid and 15-Hydroxy-lla,9a(Epoxymethano)Prosta-5,13-Dienoic Acid. Prostaglandinsand Medicine L:l??-189. 1979. 12. Goetzl, E.J., Woods, J.M., and Gorman, R.R. Stimulation of Human Eosinophil and Neutrophil PolymorphonuclearLeukocyte Chemotaxis and Random Migration by 12-L-Hydroxy-5,8,10,14-eicosatetraenoic Acid. J. Clin. Invest. -59:1?9-183. 1976. 13. Hamberg, M., and Samuelsson, B. Prostaglandin Endoperoxides: Novel Transformationsof Arachidonic Acid in Human Platelets. Proc. Natl. Acad. Sci. USA -?1:3400-3404.1974.

932

NOVEMBER 1980VOL.20NO.S

PROSTAGLANDINS

14. McGuire, J.D., Kelly, R.C., Gorman, R.R., and Sun, F.F. Preparation and Spectral Properties of 12-Hydroxyleicosatetraenoic Acid (HETE). Prep. Biochem. 8:147-153. 1978. 15. Goetzl, E.J., Weller, P.F., and Sun, F.F. The Regulation of Human Eosinophil Function by Endogenous Mono-Hydroxy-Eicosatetraenoic Acids (HETEs). J. Immunol. 124:926-933. 1980. 16.

Bauminger, S., Zor, U., and Lindner, H.R. Radioimmunological Assay of Prostaglandin Synthetase Activity. Prostaglandins3: 313-324. 1973.

17. Langone, J.J., Boyle, M.D.P., and Borsos, T. A Solid Phase Immunoassay for Human ImmunoglobulinE: Use of 1251-LabeledProtein A as the Tracer. J. Immunol. Methods -18:281-293. 1977. 18. Eskra, J.D., Levine, L., and Carty, T.J. Preparation of [3H]-12L-Hydroxyeicosatetraenoic Acid and Its Use in Radioimmunoassay. Prostaglandinsand Medicine 5, No. 3. 1980. 19. Hornick, C.L., and Karush, F. The Interaction of Hapten-Coupled Bacteriophage $X174 with Antihapten Antibody. In: Topics in Basic Immunology (M. Sela and M. Prywes, eds.) Academic Press, New York, 1969. pp. 29-36. 20. Nugteren, D.H. Arachidonate Lipoxygenase in Blood Platelets. Biochim. Biophys. Acta 380:299-307. 1975. 21. Roberts, L.J. II, Lewis, R.A., Oates, J.A., and Austen, K.F. Prostaglandin, Thromboxane, and 12-Hydroxy-5,8,10,14-Eicosatetraenoic Acid Production by Ionophore-StimulatedRat Serosal Mast Cells. Biochim. Biophys. Acta 575:185-192. 1979. 22. Goetzl, E.J., and Sun, F.F. Generation of Unique MonohydroxyEicosatetraenoicAcids from Arachidonic Acid by Human Neutrophils. J. Exp. Med. 150:406-411. 1979. 23. Parker, C.W., Stenson, W.F., Huber, M.G., and Kelly, J.P. Formation of Thromboxane B2 and HydroxyarachidonicAcids in Purified Human Lymphocytes in the Presence and Absence of PHA. J. Immunol. 122~1572-1577.1979. 24. Murphy, R.C., Hammarstrom, S., and Samuelsson, B. Leukotriene C: A Slow Reacting Substance from Murine Mastacytoma Cells. Proc. Natl. Acad. Sci. USA -76:4275-4279.1979. 25. Morris, H.R., Taylor, G.W., Piper, P.J., Samhoun, M.N. and Tippins, J.R. Slow Reacting Substances (SRSs): The Structure Identificationof SRSs from Rat Basophil Leukaemia (RBL-1) Cells. Prostaglandins-19:185-201. 1980.

NOVEMBER 1980VOL.20NO.S

933

PROSTAGLANDINS

26. Morris, H.R., Taylor, G.W., Piper, P.J. and Tippins, J.R. Structure of Slow-reacting Substance of Anaphylaxis from Guinea-Pig Lung. Nature 285:104-106.1980.

Editor: John E. Pike Received: 6-9-80 Accepted: 9-9-80

934

NOVEMBER 1980VOL.20NO.5