Prostaglandins and Medicine 2: 167-175, 1979

IMMUNOASSAY OF INDOMETHACIN: THE USE OF [125 IIPROTEIN A TO DETECT SPECIFIC SEROLOGIC BINDING.

Iftekhar Alam, John J. Langone* and Lawrence Levine. Department of Biochemis- try, Brandeis University, Waltham, Massachusetts 02154 andyaboratory of Imu- nobiology, National Cancer Institute, N.I.H., Bethesda, Maryland 20014. (send reprint requests to LL)

ABSTRACT

The sera of rabbits that were injected with indomethacin covalently linked to human albumin bound increasing amounts of [14C]indomethacin during the course of immunization. The serum-binding components chromatographed with the y-globulin fraction and were precipitated with goat-anti-rabbit-y- globulin. When measured by radioimmunoassay with [14C]indomethacin under o P4 timal conditions, 34 ng (95 pmoles) of unlabeled indomethacin inhibited

[ Clindomethacin binding 50%. The antibodies reacted most effectively with indomethacin and the desmethylated analogue, l-(p-chlorobenzoyl)-2-methyl-5- hydroxyindole-3-acetic acid, but not 2-methyl-5-methoxy-indole-3-acetic acid.

An immunoassay based on the use of [125 I]Protein A as tracer for IgG antibody was developed for quantitative determination of indomethacin at the picogram level. Indomethacin, immobilized by covalent linkage to a solid sup- port, bound the rabbit anti-indomethacin. Protein A, labeled with [1251], measured the levels of the bound IgG antibody. Fluid phase indomethacin com- peted with solid-phase indomethacin for the anti-indomethacin which resulted in decreased anti-indomethacin and consequently decreased [ 125 I]Protein A on the immobilized indomethacin-anti-indomethacin complex. The serologic speci- ficity withthe immobilized ligand immunoassay was the same as that found with the [14C]indomethacin radioimmunoassay, but the sensitivity for detection of indomethacin was increased over 300-fold.

INTRODUCTION

Indomethacin is an effective anti-inflammatory, antipyretic and analgesic drug (l), possibly because it is an especially potent inhibitor of prostaglan- din, thromboxane and prostacyclin formation; it blocks the formation of the unstable substrate (the endoperoxide) necessary for their biosynthesis (2). Probably, some of indomethacin's toxicity results from this blockage of the arachidonic acid metabolic cascade. Determination of indomethacin in biolog- ical fluids may be helpful in evaluating its effectiveness and toxicity. Some

analytical procedures that have been used for measurement of indomethacin in blood are: mass fragmentation (3), radioisotope dilution (4), fluorometry (5), and radioimmunoassay (6).

Radioimmunoassay procedures have been developed for estimation of many drugs (7). These assays are specific,and sensitive if the specific radioac- tivity of the labeled drug is high. Six years ago, we prepared in rabbits antibodies directed toward indomethacin. A radioimmunoassay (RIA) was devel- oped with the use of [14C]indomethacin as the radiolabeled ligand. The sen- sitivity of this RIA was in the order of lo-50 ng of unlabeled indomethacin. At that time, in order to increase the sensitivity of the RIA, we linked tyr- amine to indomethacin and labeled the conjugates with [1251]. While these [1251] conjugates did bind to the anti-indomethacin, the binding and inhibi- tion of that binding ere notreproducible, although the inhibition was more sensitive than when Y4 [ Cl-labeled tracer was used.

For the measurement of indomethacin in biolo ical fluid at picogram quan- tities, an immunoassay was developed in which [ 125 I]Protein A was used to quantitate anti-indomethacin. Such assays have been developed for human chor- ionic gonadotropin, IgM, methotrexate (8), leucovorin (9), IgE (10) and in the accompanying paper, thromboxane B2, prostaglandin D2, 13,14-dihydro-prosta- glandin E2, 5,6-dihydro-prostaglandin 12, 6-keto-prostaglandin Flu, 15-hydroxy- 9u,llu(epoxymethano)prosta-5,13-dienoic acid and 15-hydroxy-lla,9a(epoxymeth- ano)prosta-5,13-dienoic acid (11).

MATERIALS AND METHODS

Reagents. The indomethacin used for synthesis of the immunogen, inhibi- tor and immobilized ligand; the indomethacin metabolites and the [14C]indo- methacin (specific activity = 19 mCi/mmole) were gifts of Merck, Sharp and Dohme Research Laboratories, Rahway, N.J. Affi-gel 701 (polyacrylamide beads derived with free amino groups) was purchased from Bio-Rad Laboratories, Rockville Center, N.Y. Protein A was obtained from Pharmacia Fine Chemicals, Piscataway, N.J.

Conjugates for immunization. To a solution of indomethacin (25 mg) in 500 ~1 of pyridine was added dropwise 500 ~1 of distilled water and human al- bumin (25 mg) dissolved in 500 ul of distilled water. 1-Ethyl-3-(3-dimethyl- aminopropyl)-carbodiimide hydrochloride (CDI) (25 mg) was added and the solu- tion was allowed to sit at room temperature overnight. The solution then was applied to a Sephadex G-50 column (1.9 X 34 cm) and eluted with 0.005 M sodium phosphate-O.15 M NaCl buffer, pH 7.2. Successful coupling was confirmed by qualitative analysis of the ultraviolet spectrum at wavelengths greater than 300 nm, where 5-methoxyindole (but not albumin) absorb. We would like to thank Ms. Hilda Gjika for her help in these syntheses.

Immunization. Each rabbit was immunized subcutaneously at multiple sites with 4 mg of the human albumin conjugate in complete Freund's adjuvant. Five weeks later, the rabbits were boosted with 2 mg of the conjugate, this time intermuscularly in the form of an emulsion with complete Freund's adjuvant; about 8 weeks later they were again boosted by the same route and with the same dose. The rabbits were bled each week after the booster injection. In this study, the antiserum used was that obtained after the fourth boost.

168

Immobilized ligands. The solid phase absorbent was prepared by a modifi- cation of a procedure reported previously (8). Ligands (20-50 pg) were added to 5 ml of affi-gel 701 (500 mg) which had been centrifuged and resuspended in 0.003 M sodium phosphate buffer, pH 6.35, containing 1-ethyl-3-(3-dimethyl- aminopropyl) carbodiimide (50 mg). The reaction mixture was stirred at room temperature for 48 hours. The beads were washed at 4" successively with water (3 ml), ethanol (3 ml) and isogel tris buffer (pH 7.2-7.6) (3 X 3 ml). Finally the beads were resuspended in 25 ml isogel tris buffer containing 0.02% NaN3. In the immunoassays, 50 ~1 (1 mg) of this suspension were used.

Iodination of Protein A. This preparation is based on the procedure re- ported in (12). Five hundred UCi of Bolton-Hunter reagent (New England Nuclear, 400 Ci/mmol) in 50 ul of benzene was evaporated in an acid washed tube under N2 at room temperature. To this dry tube were added 50 ~1 of 0.1 M sodium phos- phate buffer, pH 8, followed by the addition of 2 I_rg of PA (in 40 ~1). The reaction mixture was allowed to stand at room temperature for one hour. At the end of the hour, 0.5 ml of 0.1 M sodium phosphate buffer, pH 8, containing 0.2 M glycine was added. Ten minutes after addition, 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 pre- packed column (Pharmacia Fine Chemicals, Piscataway, N.J.). One ml fractions were collected. The iodinated Protein A came out in tubes 3 and 4. PA incor- porated 55% of the Bolton-Hunter reagent. Specific activities ranging from 6,000 to 10,000 Ci/mmol have been obtained.

RESULTS

The binding of [14 Clindomethacin by the antiserum increased during the course of immunization (Fig. 1). Maximal binding was found in the antiserum after the fourth injection; 1 ~1 of the antiserum from this fourth bleeding bound 39% of the added [14C]indomethacin. Subsequent boosts did not increase binding. The indomethacin binding serum proteins are immunoglobulins (the RIA was performed by separating bound from free indomethacin with goat-anti-rabbit immunoglobulin). As can be seen in Fig. 2, the specific immunoglobulins have different sedimentation properties.

As expected, the antibodies that bound [ 14C]indomethacin were inhibited most effectively by indomethacin (Fig. 3). The 5-methoxy moiety was not mark- edly immunodominant as the 1-(p-chlorobenzoyl)-2-methyl-5-hydroxyindole-3- acetic acid (desmethylindomethacin) reacted 50% as effectively. However the p-chlorobenzoyl-moiety is recognized by the anti-indomethacin; it reacted less than 0.1% as effectively as indomethacin. The limit of sensitivity of this RIA was around 10 ng.

In order to develop an RIA with greater sensitivity, indomethacin was coupled to tyramine after activation of indomethacin with oxalyl chloride. The indomethacin-t ramine conjugate was isolated by thin layer chromate raphy and labeled with [ir25 I] by the lactoperoxidase procedure (13). This [ 1$5I]- indomethacin-derivative bound to anti-indomethacin and its sensitivity was less (50% inhibition was found with 4 ng of indomethacin instead of 34 ng found with [14C]indomethacin); however, its dose response curve was shallow. Most importantly, the [1251] preparation was unstable and within weeks lost its capacity to bind reproducibly to the antibodies. The serologic specific- ity was the same as that found with [14C]indomethacin; only indomethacin and

169

3000

.g

,” t E 2 2000 I ._

‘c a %

2 z 1000

a m E t

Normal I 2 3 4 5 Rabbit Strum

Course of immunization

[‘4C] lndomethocin Anti-lndomethocin 0.6

l-

Figure 1. Binding of [14C] by 1 ~1 of anti-indomethacin. The antisera were obtained from rabbits 1 week after injection of human albumin- indomethacin conjugates

0.4 - ;; 1600 7

Figure 2. Sephadex G-200 chromatography of indomethacin anti- serum. Absorbance, 280 nm (0). Binding of [1251]indomethacin by 10 ~1 of fraction (0).

Fraction

170

6 ; 60- z 'E = 40- $

20 -

Figure 3. Serologi specificity of the LY4C]- indomethacin anti-indo- methacin reaction. Indo- methacin (0); desmethyl- indomethacin (0) and des- chlorobenzoylindomethacin (A).

I 1 -,. I

I IO 100 1000 l0,000 l00,000 Nonogroms

desmethylindomethacin inhibited; 2-methyl-5-methoxy-indole-3-acetic acid did not react when tested at the 10 ug level.

Because of the instability of the [125~] derivative, as well as the need to obtain a more sensitive assay for indomethacin, the immunoassay that mea- sures inhibition of the reaction between immobilized ligand and specific anti- body using [1251]Protein A as the marker for bound antibody was developed.

Binding of anti-indomethacin to the immobilized ligands increased during the course of immunization. Within each bleeding of the immunized rabbits, binding of the antibodies to the immobilized homologous ligands was dependent on the concentration of the antiserum (Fig. 4). An antiserum dilution which bound 2,000 to 3,000 cpm of PA out of 30,000 to 40,000 cpm added, was used in the immunoassay. In this solid phase immunoassay, appropriate dilutions of rabbit antiserum (100 ul), immobilized-ligand-beads (50 ul), and test solution (100 ~1) (or buffer to determine maximum binding) were incubated at 37O for 60 min. The beads were washed with two 3 ml protions of buffer by centrifu- gation for 10 min at 1,200 xg at 4' incubation with [125

C; then carried through a second similar I]PA (100 1~1, 30,000 to 40,000 cpm) for one hour at 37O.

After this second incubation, the beads were washed twice with buffer, and the radioactivity bound was determined by counting in a Packard Model 578 Auto- gamma spectrometer. The counter efficiency was 70%. Control tubes which con- tained beads (immobilized ligand) plus appropri dilutions of normal rabbit plasma were included to estimate non-specific [ 853 I]PA binding. Control beads bound 15% of the cpm bound by the immune serum. A typical protocol and re- sults of an immunoassay for indomethacin are shown in Table 1.

The serologic specificity of the immobilized indomethacin-anti-indometh- acin reaction remained the same (Fig. 5). The homologous ligand, indomethacin, was the more effective inhibitor; 0.3 picomoles (107 picograms) in the fluid phase inhibited 50% of the binding. Desmethylindomethacin was 50% less ef- fective; 2-methyl-5-methoxyindole-3-acetic acid at the 400 picomole level (100 ng) did not inhibit. Again the p-chlorobenzoyl-moiety is strongly im- munodominant whereas the 5-methoxy group, while being recognized by the anti- bodies, offers much less to the binding energy of the indomethacin-anti-indo- methacin reaction.

171

0.1 I pl Anti-lndomethocin

DISCUSSION

Binding of mAtein A to a con- stant amount of immobi- lized indomethacin that had been incubated with increments of anti-indo- methacin.

We have described a solid phase immunoassay method for measurement of in- domethacin; less than 100 picograms can be detected. Indomethacin's primary metabolite, 5-OH indomethacin, reacted 50% as effectively while the metabolite lacking the p-chorobenzoyl group was inactive as an inhibitor.

With this sensitivity, indomethacin at sub-nanog am levels, can be meas- ured in plasma. An RIA for indomethacin which uses 5 [ Hlindomethacin also can detect sub-nanogram levels (6). Mass fragmentation is a sensitive analytical procedure for measurement of indomethacin, it allows specific determination of indomethacin in plasma at levels of 25 ng/ml (3). The use of the immobilized indomethacin assay would permit analyses of plasma levels for indomethacin over a longer period of time. With less sensitive procedures (4), the half life of indomethacin was reported to be between 1.8 and 2.3 hours. With the use of mass fragmentation for analyses, the time of sampling was extended and the half life was found to range from 4.0 to 11.2 hours (3). The sensitivity of mass fragmentation was sufficient to obtain data indicative of the real metabolic life of indomethacin and pharmacokinetics of indomethacin as deter- mined by the immobilized indomethacin immunoassay should confirm these find- ings.

ACKNOWLEDGEMENTS

This work was supported by Grant HD-07966 from the National Institute of Child Health and Human Development and Grant CA-17309 from the National Cancer Institute. This is publication number 1232 from the Department of Biochemis- try, Brandeis University, Waltham, Massachusetts 02154.

IA was supported by Training Grant No. AI-07069 from the National Insti- tute of Allergy and Infectious Diseases.

172

0 .

m

c

:: 0

6.

0

a, u 0

4-l

4

173

IOO-

60 -

.! 60-

z c

=40- af

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

I IO Picomoler Inhibitor

I

100 1000

Figure 5. Serologic specificity of the immo- bilized-indomethacin anti-indomethacin reac- tion by fluid phase in- domethacin (0); desmethyl indomethacin (A) and deschlorobenzoylindo- methacin (0).

REFERENCES

Robinson HJ, Vane JR. Prostaglandin synthetase inhibitors. Raven Press, New York, 1974.

Vane JR. Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs. Nature New Biol. 231: 232, 1971.

Palmer L, Bertilsson L, Alvan G, Orme M, Sjoqvist F, Holmstedt B. Indo- methacin: Quantitative determination in plasma by mass fragmentography including pilot pharmacokinetics in man. p. 91 in Prostaglandin Synthe- tase Inhibitors (Robinson JH, Vane JR, eds) Raven Press, New York, 1974.

Duggan PE, Hogans AF, Kwan KC, McMahon FG. The metabolism of indomethacin in man. J. Pharmacol. Exp. Ther. 181: 563, 1972.

Hvidberg E, Lausen HH, Jansen JA. Indomethacin: Plasma concentrations and protein binding in man. Eur. J. Clin. Pharmacol. 4: 110, 1972.

Nelson JC, Berk LS, Lewis JE, Emori HW. Indomethacin: Measurement by radioimmunoassay. Clin. Res. 24: 151A, 1976.

Butler VP, Jr. The immunological assay of drugs. Pharmacological Reviews 29: 103, 1977.

Langone JJ. [1251]Protein A: A tracer for general use in immunoassay. J. Immunol. Meth., in press.

Langone JJ, Levine L. Immunoassay of leucovorin: Use of [1251]Protein A to detect immunological binding. Anal. Biochem., in press.

Langone JJ, Boyle MDP, Borsos T. A solid phase immunoassay for human im- munoglobulin E: Use of [125 I]Protein A as the tracer. Anal. Biochem., in press.

Levine L, Alam I, Langone JJ. The use of immobilized ligands and [1251]- Protein A for immunoassays of thromboxane B2, prostaglandin D2, 13,14-di- hydro-prostaglandin E2, 5,6-dihydro-prostaglandin 12, 6-keto-prostaglandin

174

12.

13.

Flu, 15-hydroxy-9a,lla(epoxymethano)prosta-5,13-dienoic acid and 15-hy- droxy-llu,9a(epoxymethano)prosta-5,13-dienoic acid. Submitted for publi- cation, Prostaglandins and Medicine.

Langone JJ, Boyle MDP, Borsos T. [125 I]Protein A: Applications to the quantitative determination of fluid phase and cell-bound IgG. J. Immunol Meth. 18: 281, 1977.

Morrison M, Bayse GS, Webster RG. Use of lactoperoxidase catalyzed io- dination in immunochemical studies. Immunochemistry 8: 289, 1971.