ANALYST, FEBRUARY 1993, VOL. 118 137 Development of a Radioimmunoassay for the Determination of Buprenorphine in Biological Samples Lode Debrabandere, M. Van Boven and P. Daenens Laboratory of Toxicology, University of Louvain (KULeuven), E. Van Evenstraat, 4, B 3000 Leuven, Belgium The development of a specific and sensitive radioimmunoassay for the detection of buprenorphine in urine samples is described. With minor adjustments, the assay was also applied to the analysis for buprenorphine in plasma samples. The 2-diazobenzoic acid derivative of buprenorphine has been prepared as a hapten. The immunization of rabbits with the hapten-bovine serum albumin conjugate resulted in the production of antibodies, which cross-reacted with N-dealkylbuprenorphine up to about the 90% level.The antibodies showed very low cross-reactivities with the 3-O-glucuronides and with the structural analogue etorphine. The assay was mainly used to pre-screen for buprenorphine in urine samples of persons suspected of Temgesic misuse and to determine buprenorphine in plasma samples. A linear calibration graph for buprenorphine was obtained after logit-log regression [ Y = 0.383 (s, 0.059) - 0.535 X(s, 0.025); r = 0.997 (s, O.OOl)]. The spiking recovery study showed a linear regression of Y (observed) = 0.94 + 0.84 X (expected); r = 0.997. Intra- and inter-assay relative standard deviations were <4.35 and <6.36%, respectively. A comparison study of the high-performance liquid chromatographic determination ( X ) to the radioimmunoassay (v) resulted in the following regression equation for the urine samples: Y = 1.44 + 1.64X(n = 32; r = 0.910) and Y= 0.007 + 1.58 X ( n = 10; r = 0.930) for plasma specimens. The minimum detectable dose of the immunoassay was calculated to be 10 pg ml-1 (Student’s t-distribution, p = 0.01, degrees of freedom = 8).Keywords: Buprenorphine detection; drug abuse; biological sample; radioimmunoassa y Buprenorphi ne , (2S)-2-[ ( - ) (5R ,6R ,7R, 14S)-9a-cyclopropyl- methyl-4,5-epoxy-3-hydroxy-6-methoxy-6,14-ethanomorphin- an-7-yl]-3,3-dimethyIbutan-2-ol, is a very potent oripavine derivative with mixed agonist and antagonist opiate p receptor activities. 1 4 Buprenorphine can be administered sublingually (Temgesic Sublingualis; buprenorphine 0.2 mg), or by intravenous or intramuscular injection (Temgesic 10 ampoules for injection; buprenorphine 0.3 mg).Relatively soon after the introduction of buprenorphine, some cases of misuse were reported.5 13 An immunological method for the determination of buprenorphine in plasma samples was developed by Bartlett et al. 13 The antisera were obtained by immunizing rabbits with, respectively, 3-O-car- boxymethylbuprenorphine and N-hemisuccinylnorbuprenor- phine. The antibodies raised against the first hapten show an important cross-reaction with buprenorphine 3-O-glucuro- nide, while the antibodies against the second hapten cross- react with N-dealkyl buprenorphine and the structurally re- lated product etorphine. Diagnostic Products (DPC; Los Angeles, CA, USA) recently introduced the buprenorphine 1251 radioimmunoassay (RIA) kit, which was designed for the qualitative and quantitative determination of buprenorphine in equine urine.It was the purposc of this work to prepare a more specific and sensitive immunoassay. This RIA was developed, on the basis of antibodies elicited by 2-diazobupre- norphine-bovine serum albumin (BSA). The spacer between buprenorphine and BSA is established at C-2, and therefore, the C-3, C-6 and C-7 substituent and the nitrogen are free to serve as potential antigenic determinants. A similar spacer and linkage position has already been examined for the development of a highly specific RIA for the determination of morphine in brain tissue.15 The test has been applied to the determination of buprenorphine in human urine samples, and the results have been compared with those obtained using the commercially available kit.Experimental Materials and Methods Iodine-125 in NaOH, IMS 30 (606.8 Mbq pg-1 of iodine), was obtained from Amersham International (Amersham, Buck- inghamshire, UK). Buprenorphine hydrochloride and N-dealkylbuprenorphine were synthesized in the laboratory according to a modified method of Kleeman and Engel.16 Etorphine and diprenorphine hydrochlorides were obtained from C-Vet (Bury St. Edmunds, Suffolk, UK). Other reagents were obtained from the following sources: ammonium sulfate, ethanol, 4-aminobenzoic acid, chloramine-T, sodium metabi- sulfite, sodium nitrite and poly(ethy1ene glycol) (PEG) Type 6000 from Merck (Darmstadt, Germany); and BSA (fraction 5 ) and 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (98%) from Janssen Chimica (Beerse, Bel- gium).Goat antiserum to rabbit y-globulin, normal rabbit carrier serum (NRS) and Freund’s complete adjuvant were obtained from Calbiochem Biochemicals and Immuno- chemicals (San Diego, CA, USA). Norit Supra A , Norit Supra B and Norit Extra C were gifts from Norit (Amersfoort, The Netherlands). The SpectdPor membranes [relative molecular mass ( M , 12 000-14000)] for the dialysis of the BSA conjugate were obtained from Spectrum Medical Industries (Los Angeles, CA, USA). To determine the cross-reactivity of the glucuronides of buprenorphine and N-dealkylbuprenor- phine in the assay, the samples were analysed before and after hydrolysis with (3-glucuronidase (40 U ml-1; 1 U = 16.67 nkat)-arylsulfatase (20 U ml-1) from Helix pornatia (Merck).Prior to analysis, samples (100 pl) from patients were diluted 1 + 1 with 0.2 moll-1 citrate buffer (pH 5.0) and 100 p1 of P-glucuronidase-arylsulfatase, followed by incubation over- night at 37 “C. Purification of the hapten was carried out by high-perfor- mance liquid chromatography (HPLC) on a semi-preparative silica column (15 g of LiChrosorb Si-60, 10 pm, suspended in carbon tetrachloride and compressed at 350 x 102 kPa in a 25 X 1.25 cm stainless-steel column) using a Merck-Hitachi 6002 pump and an ultraviolet (UV) detector (Model 440, Waters Associates, Milford, MA, USA). The structure of thc synthesized buprenorphine and of N-deal kylbuprenorphine, as well as that of the synthesized hapten, was confirmed by nuclear magnetic resonance (NMR) spectrometry (JEOL FX90Q NMR spectrometer, Tokyo, Japan) and mass spec- trometry (MS) (HP 5995A; Hewlett-Packard, Avondale, PA, USA).The radioactivity of the tracer was counted on a y-counter (Berthold BF 5300, Wildbad, Germany).138 ANALYST, FEBRUARY 1993, VOL. 118 Preparation of the Hapten The hapten was synthesized by coupling buprenorphine at the C-2 position with the diazonium salt of 4-aminobcnzoic acid. To a 0.1 mot I-' HCI solution (15 ml) of 4-aminobenzoic acid (89 mg) a solution of NaN02 (45.6 mg) was added drop by drop until the reaction mixture immediately coloured blue on a freshly prepared iodine-starch paper. All solutions were kept at 0 "C throughout the reaction. Buprenorphine (300 mg) was added to 0.1 mol 1-1 NaOH, and the pH of the solution was adjusted to 11 by adding 0.1 mol 1-l HCI.The cold solution of diazobenzoic acid was added to the buprenorphine solution, and the mixture was stirred for 2 h. The pH was kept constant during the whole synthesis. After all the diazoben- zoic acid solution had been added, the mixture was kept at 0 "C for a further 30 min. The pH of the reaction mixture was adjusted to 4.5. The precipitate formed was filtered off, dried and purified by preparative HPLC. The mobile phase for the HPLC system consisted of a mixture of dichloromethane, methanol and ammonia solution, which was pumped at 5 ml min--l, in accordance with the following program. Starting from 100% dichloromethane, the mobile phase was changed in 30 min to 70% dichloromethane and 30% methanol-ammonia solution (99 + I ) .The hapten eluted at about 27 min. After each injection of the reaction mixture the hapten fraction was collected. The compound exhibited UV absorption at 280 and 360 nm. The organic layer was dried with anhydrous sodium sulfate, the solvent was evaporated at 40°C under a stream of nitrogen, and the residue was identified by MS and NMR. The total yield of the synthesis was 19%. Immunization A 750 pl aliquot of the dialysed hapten-BSA conjugate (15 nmol) was emulsified with an equal volume of complete Frcund's adjuvant, and two New Zealand rabbits were immunized by multiple intramuscular injections of the water- oil emulsion. The first two booster injections were given at a 2 week interval, while the remaining injections were given at intervals of approximately 4 weeks.Blood was collected from the central ear artery. The first blood samples were collected 10 d after the second booster injection; the second blood samples 1 month later and afterwards at intervals of 14 d. The sera from the two rabbits were tested for their ability to bind [12sI]iodobuprenorphine and for their cross-reactivity against the major metabolite N-dealkylbuprenorphine and the struc- tural analogue etorphine. The titre of the serum was obtained by treating diluted antiserum (1 + 50-1 + 800) as described in the method of second antibody separation of bound and free ligand. The dilution that results in 50% binding of the tracer was selected to generate calibration curves for buprenorphine. Preparation of the Immunogen The hapten (6.6 mg) was dissolved in 600 pl of dimethyl sulfoxide (DMSO) and 400 pl of phosphate buffer (pH 7.4; 0.05 rnol I-') was slowly added.In another tube, 15 mg of BSA was dissolved in 5 ml of phosphate buffer and a solution of 30 mg of l-ethyl-3-(3-dimethylaminopropyl)carbodiimide in 0.5 ml of phosphate buffer was added. The hapten solution was mixed in portions with the BSA-carbodiimide solution over 20 h, under stirring at room temperature. Previous experiments had shown that in order to keep the hapten in solution, the addition of at least 60% of DMSO was necessary. The compounds of low relative molecular mass were removed from the solution by dialysis with a cellulose membrane having a cut-off value of 12000 to 14000. The mixture was dialysed in phosphate buffer (0.05 mol I-'), changing the buffer three times a day.After the first day a 0.005 mol 1 - 1 buffer was used. The dialysis was stopped after 3 d. Degree of Incorporation The degree of incorporation of the hapten was ascertained by UV spectroscopy. 17 The diazobenzoic acid derivative of buprenorphine is yellow with a maximum absorption at 360 nm; BSA does not exhibit any absorption at this wavelength. Synthesis of the Tracer The iodination of buprenorphine with chloramine-T has already been described in detail.18 To a solution of 100 pl of phosphate buffer (0.066 mol 1 - 1 , pH 7.0) was added 10 p1 (3.7 MBq; 0.5 nmol) of a sodium [1~sI]iodidc solution and 10 pl of a buprenorphine hydrochloride solution (30% v/v methanol, 1 mg in 10 ml); 10 pl (0.25 nmol) of a freshly prepared chloramine-T solution was added and the reaction mixture was vortex-mixed.After 4 and 8 min, a further 10 pl of the chloramine-T solution was added. The chloramine-T was neutralized after 10 min with 10 pl of a sodium metabisulfite solution (1 mg in 10 ml of distilled water). Assessment of the Optimum Conditions for the Assay Pro- cedure In order to optimize the immunoassay conditions, several methods for the separation of bound and free antigen were examined. Other parameters that influence the immunoassay, such as pH, incubation time and Concentrations of BSA and tracer, were also tested. No remarkable effects were observed when the pH of the reaction mixture ranged between 6 and 8 and when the final concentration of BSA was kept between 0.2 and 5%.The equilibrium between antigen and antibody was reached after 90 min. These optimum incubation condi- tions were kept constant for all remaining experiments. To 300 ELI of BSA solution (2% BSA in phosphate buffer, 0.1 moll-1, pH 7.4) were added, 100 pl of tracer (diluted in 2% BSA solution to approximately 30000 counts min-1) and 100 pl of antibody (dilution 1 + 199 in phosphate buffer). The mixture was allowed to equilibrate at room temperature for 90 min. Only for the separation of bound and free ligand with goat anti-rabbit y-globulin (GARGG) was an antiserum dilution of 1 + 399 used. Different separation techniques were tested and the effects o f the separation procedures on the apparent titre of the antiserum were examined.First, the fractional precipitation of bound and free antigen with ethanol, ammonium sulfate and PEG 6000 was attempted. Different concentrations of ethanol (10,20,50,80 and 90%), ammonium sulfate (0.5, 1, 1.5,2 and 2.5 mol 1 - 1 ) and PEG 6000 (2.5, 5, 10, 15 and 20%) were applied. The NRS was used as carrier protein for the PEG precipitation, and different concentrations of the carrier protein were checked. Second, the non-specific adsorption of the tracer to charcoal particles, as described by Herbert et al. , I 9 was attempted. Three types of charcoal (Norit Supra A, Norit Supra B and Norit Extra C), were examined. For the Separation of bound and free buprenorphine, 200 pl aliquots (1% suspension) of the three types of charcoal were added to the incubation mixture.After different time intervals, the matrix was centrifuged at 188Og for 15 min, and the supernatant phase was counted for 1 min. Furthermore, the influence of coating the charcoal with dextran was tested. As a third separation technique, a second antibody method using GARGG was evaluated. This technique was first introduced by Utigcr et ~ 1 . 2 ~ ) and Morgan and Lazarow." To optimize this procedure, several parameters were investi- gated, such as the incubation time and the optimum amount of GARGG and NRS. After the initial incubation (90min) at room temperature, bound and free buprenorphine could be separated by adding 50 p1 of NRS (1 + 100 in distilled water) and 50 pl of GARGG (1 + 11.5 in phosphate buffer, 0.1 mol I-', pH 7.4). After vortex-mixing, the mixture wasANALYST, FEBRUARY 1993, VOL.118 139 incubated for a further 6 h. The precipitate was centrifuged for 15 min at 300013 and counted for 1 min. Finally, a combination of the second antibody technique and fractional precipitation was examined. Different concen- trations of ammonium sulfate (0.5, 1.0 and 2.0 mol I-*), ethanol (12.5, 25 and 50%) and PEG 6000 (1, 4.5 and 9%) were added and mixed with 50 pl of GARGG and 50 p1 of NRS. At different time intervals (0.25,1,2,6,12 and 24 h) the precipitates were centrifuged and counted. The second antibody separation technique (GARGG) was selected for the derivation of the calibration graph. The logit-log transformation22 was used to linearize the calibration graph. For the derivation, buprenorphine hydrochloride was diluted in drug-free urine to produce a concentration range of 100-25000 pg ml-1, and 100 p1 of the spiked urine samples were analysed by the RIA procedure.When the determina- tion of very low buprenorphine concentrations is necessary, a calibration graph between 10 and 10000 pg ml-l is more appropriate. Non-specific binding (NSB) was determined by replacing the antiserum by an equal volume of phosphate buffer. For each batch of samples a calibration graph was obtained within the same day. Human Samples To one healthy human volunteer (male, 27 years), 0.3 mg of buprenorphine (Temgesic injection formulation) was ad- ministered by intramuscular injection. Urine samples were collected over 72 h. The same volunteer was treated several months later with 0.2 mg of buprenorphine sublingually (Temgesic sublingual compr.0.2 mg). For the sublingual administration, a 0.2 mg tablet was placed under the tongue until dissolved (8-10 min). The urine was collected over 10 d. Several urine samples from persons suspected of Temgesic misuse were also analysed by the immunoassay. The RIA has also been applied to the analysis of plasma samples, using a slightly modified procedure. The standards are dissolved in blank drug-free plasma, and the reaction matrix consists of a 0.5% (instead of 2%) BSA solution in buffer. Results Preparation of the Tracer, Hapten and Immunogen After isolation by solid-phase extraction (reversed-phase C-18) and determination by HPLC with electrochemical detection (ECD), the specific activity of the tracer was counted to be 34 TBq mmol--l. The label could be used for 3 or 4 weeks without any loss of binding to the antiserum. Thereafter, the antibody-binding capacity declined rapidly.The hapten-BSA conjugate was prepared by the carbodi- imide technique, which was first described by Sheehan and Hess.2’ The spacer (diazobenzoic acid) was placed in the ortho position to the phenolic group. It was expected that the CH 5 18 - 14 N- 6 -J;9 8 CH,- -OH 20; 3 0 -CH2 24 26 27 Fig. 1 buprenorphine Structural formula of the hapten used for the immunization of antibodies would recognize the most important functional groups of buprenorphine, such as the phenolic, methylcyclo- propyl and tert-butyl groups. The HPLC procedure described in the preparation of the antigen allowed the isolation of the hapten (Fig.1). The collected fraction was dried with sodium sulfate, and the solvent was evaporated at 40°C under a stream of nitrogen. The structure of the hapten was confirmed by NMR and MS. The major fragments (mlz) of the methylated hapten are: 55,57,83, 101, 135,542,554,586 and 643. ‘3C NMR (CDC13/DMSO): 6 166.8 (C-7’), 151.4 (C-l’), 144.6 (C-4), 138.0 (C-2), 136.8 (C-3), 135.1 (C-12), 132.7 (C-4’), 129.4 (C-3’ + C-5’), 125.5 (C-ll), 121.3 (C-l), 120.3 (C-2’ + C-6‘), 95.4 (C-5), 79.2 (C-6), 77.7 (C-20), 57.9 (C-24), (C-22), 34.4 (C-14) 34.1 (C-15), 32.0 (C-19), 28.1 (C-lo), 25.1 56.6 (C-9), 51.1 (OMe), 45.1 (C-13), 42.1 (C-7 + C-16), 38.9 (3 x Me), 21.2 (C-8), 18.7 (C-21), 16.7 (C-18), 8.1 (C-25), 2.8 (C-26), and 2.0 (C-27).The values of the aromatic ring indicate that the diazo substituent is attached at the 2-position as could be expected from the reaction conditions. It was observed that the sequence of adding the hapten solution to a mixture of BSA and carbodiimide was important because of the poor solubility of the hapten. A precipitation was observed if carbodiimide was first added to the hapten solution. By using the latter sequence of mixing, almost no hapten was coupled with BSA. Hapten-BSA Ratio By using UV spectroscopy, a hapten-protein ratio (pmol) of 4.3 : 0.214 was calculated, corresponding to the coupling of about 20 mol of hapten with 1 mol of BSA. An average of 15 molecules of hapten per molecule of BSA has been recom- mended for an optimal immuno-resp0nse.2~ Titre of the Antibodies Results for serial dilutions of the antiserum, examined for their ability to bind 50% of the tracer, are shown in Fig.2. It was observed that rabbit 1 was likely to produce sera with a titre; after 2 months however, the animal generated ‘elephant teeth’ and had difficulties in eating. From that moment on, the titre decreased and stabilized at Ifi200. Although the titre of the antibodies was sufficient (600) 2 months after the first injection of the conjugate, the cross-reaction of the antibodies towards N-dealkylbuprenorphine and etorphine was high (100 and 12%, respectively) in comparison with the selected antibodies of rabbit 2 after the third bleeding. The titre of the latter antibodies reached only 400, but the specificity towards the two compounds was superior when compared to the antibodies of rabbit 1 (90 and +4%, respectively).The quality of the antibodies remained constant while the titre fluctuated around 300. For their superior cross-reactivity characteristics, the antisera of rabbit 2 (third collection) were lyophilized and used for the development of the immunoassay. They were stored at 4°C. .- r) c 300 - 0 50 100 150 Timeld Fig. 2 Antiscra dilutions necessary to obtain 50% binding of the tracer. The arrows indicate the time intcrvals where the BSA-hapten conjugate was injectcd140 ANALYST, FEBRUARY 1993, VOL. 118 Determination of Optimum Conditions for the RIA Precipitation with salts or organic solvents was among the earliest methods of fractionation of biological molecules.The major problem with fractional precipitation methods is that they tend to yield a high assay blank. In the buprenorphine assay, comparable results were obtained for the precipitation with ammonium sulfate and ethanol in a final optimum concentration of 1.5 and 7.4 mol I - l , respectively; 4045% of the tracer was bound with an antiserum dilution of 1 + 199. The blank values ranged from 12 to 17%. At least 40 pl of carrier protein (NRS) were necessary to precipitate the antibodies with PEG 6000. The blank value with this method reached 18%. The precipitation with ammonium sulfate and ethanol occurred instantly, whereas the PEG 6000 precipita- tion was only complete after 30 min of incubation at 4 "C. Different types of charcoals of the Norit range were examined for their separation properties.The types Supra A and Supra B have similar adsorption characteristics. When the separation step was carried out with the Extra C quality, however, a lower value for the percentage binding of the tracer was obtained. Coating the charcoal with dextran ( M , 60 000-90 000) had n o influence on the adsorption characteris- tics of the charcoal. It was shown that incubation for 15 min with 200 p1 of a 1% freshly prepared charcoal suspension was sufficient to capture the free tracer. The NSB value was high, but acceptable (8.5%), and the percentage binding of the tracer reached 45% (antiserum dilution 1 + 199). In comparison with the results of the adsorption and precipitation techniques, the GARGG second antibody method resulted in a lower NSB value and in higher dilutions of the antiserum for a 50% binding of the tracer.An antiserum dilution of 1 + 399 could bind 50% of the tracer. To separate bound from free label, the optimum ratio of GARGG antiserum to NKS was investigated. Final dilutions of GARGG (1 + 124) and NRS (1 + 999) resulted in the highest percentage binding of the tracer. The optimum second incubation time for the precipitation of the antibody bound tracer was 6 h at room temperature. The rate of immunoprecipitation could be enhanced by addition of precipitation reagents to the second antibody.25 It was possible to reduce the total analysis time to about 2 h by using the combination method (fractional precipitation and GARGG). Although a higher NSB value and a lower maximum binding (+40%) were observed, it was still possible to obtain reproducible results.The three precipitation rcagents were mixed in different concentrations for the second antibody separation technique. The initial incubation condi- tions (titre 1 + 399) were the same as for the second antibody method. The best results were obtained with the combination of PEG (final concentration 9%) and GARGG. The major advantage of this method is the short second incubation time (15 min), permitting substantial reduction of the total analysis time. From the different methods studied, the second antibody method was selected for the assessment of the performance 95 2: 90 - 10 - 5 - I I I I I I I I I -1-4 0.1 0.25 0.50 1 2.5 5 10 25 50 Dose/ng rn I ~~ 1 Fig. 3 regression Calibration graph for buprenorphine based o n logit-log characteristics of the immunoassay , the longer analysis time not being a major criterion for the laboratory tests.The immunoassay with the latter separation technique resulted, after logit-log transformation, in a linear calibration graph for the concentration range 100-25000 pg ml-* (Fig. 3). The parameters of five RIA calibration graphs are summarized in Table 1. The calibration graph is expressed as: (Y) = PO + PI ( X ) with Y = log, y/l - y ( y = R/Bo; counts bound for arbitrary dose relative to that for zero dose); X = log, x (x = buprenorphine standard concentration); PO = intercept of the curve; and PI = slope of the curve. The 95% confidence limits for and PI are, respectively, 0.118 and 0.059. Minimum Detectable Dose (MDD) The MDD of the RIA can be defined as the dose level that results in an expected response, which is significantly different from the expected responsc for a zero-dose tube.26 Two populations (zero-dose and 10 pg ml-1 tubes) were compared with a 'one-sided' Student's t-distribution (p = 0.01, degrees of freedom = 8).The two populations were first checked for the validity of the equal variance distribution with an F-test. It was found that a concentration of 10 pg ml-1 is, for 99% of the examples, significantly different from the blank value. For routine analysis, however, only the concentration range between 0.1 and 25 ng ml-1 is considered. Precision The intra- and inter-assay RSDs for the immunoassay are illustrated in Table 2. The results are the mean of five replicate determinations.The intra- and inter-assay precision was calculated after the analysis of six different positive urine samples. Spiking Recovery To determine the recovery of buprenorphine, 900 p1 of a randomly chosen positive urine sample with an apparent Table 1 Pararncters o f the RIA calibration graph ( n = 5 ) Parameter Mean S Total countsldisintegrations min-' 28 050 384 Maximum binding, Po (%) 45 2.8 Intercept (Po) 0.383 0.059 Slope (PI) -0.535 0.025 Mid-point ((3501K)/ng ml- 1 2.3 0.30 Non-specific binding 6.6 0.8 r 0.997 0.001 Table 2 The intra- and intcr-assay prccision for six randomly selected urine samples Buprenorphine concentrationhg ml-' RSD" (%) n I n tra-assay- 5. 10 2.55 5 9.80 3.06 5 20.29 4.34 5 In fer-assay- 2.18 10.66 24.11 5.02 5 5.16 5 6.35 5 * RSD = relative standard deviation. Table 3 Spiking recovery of buprenorphine in the assay Addedlng Observedlng m l ~ 1 Expected1 Observed1 1 5.6 f 0.16 5.77 97 5 0.6 f 0.25 9.77 98 10 13.2 t 0.139 14.77 89 (n = 3) (mean f s) ng m l - 1 expected (%)ANALYST, FEBRUARY 1993, VOL.118 141 Table 4 The specificity of the RIA method for the major metabolite and the structurally related compounds etorphine and diprenorphine Compound Chemical structure Apparent added/ concentration/ reactivity Concentration buprenorphine Cross- ng ml-1 ng ml-1 (Yo) Etorphinc Diprenorphine 100 10 1 93 9 0.9 93 90 90 250 'CH2-CH3 25 2.5 6 1.1 0.15 2.4 4.4 6.0 250 0.8 0.3 25 0.3 1.2 2.5 0.0s 2.0 buprenorphine concentration of 5.3 ng ml-1 was spiked (n = 3) with 100 pl of three standard dilutions of buprenorphine ( I , 5 and 10 ng in 100 vl) in order to obtain a 1 : 10 spiking ratio, leaving the matrix of the spiked samples relatively intact.A 100 pl aliquot of the mixture was analysed by the assay procedure. To calculate the expected value, 90% of the non-spiked value was added to the added amount of buprenor- phine standard (Table 3). This recovery study resulted in the following linear regression equation: Y (observed) = 0.94 + 0.84 X (expected); Y = 0.997. Specificity The antibody specificity was assessed by measuring the cross-reactivity to other compounds that could be present in urine samples from persons suspected of drug misuse. The following compounds were found to be not detectable by the immunoassay procedure at a level of 10 pg ml-1: acetylsalicy- lic acid, amobarbital, Dr*-amphetamine, barbital, bezitramide, caffeine, chlorpromazine, cocaine, codeine, cotinine, dextro- moramide, diazepam, ethylmorphine, fentanyl, flunitraze- Pam, haloperidol, heroin, imipramine, lidocaine, lysergide (LSD), metamphetamine, methaqualone, morphine, nalox- one, normethadone, pentazocine, pethidine, phencyclidine, phenobarbital , pholcodine, propoxyphene and zolpidem.The specificity of the antiserum towards the major metabol- ite, N-dealkylbuprcnorphine, and towards t h c structural analogues etorphine and diprenorphine, is demonstrated in Table 4. The intermediate products of the synthesis of buprenor- phine, together with the 3-O-methyl derivative of buprenor- phine, N-ethylbuprenorphine and N-propylbuprenorphine, were also tested for their cross-reaction with the antibodies.The results are shown in Table 5 . Although the spacer was placed in the C-2 position, the antibodies were not able to recognize the alkyl group on the nitrogen. The major metabolite, N-dealkylbuprenorphine cross-reacts at the 90% level over a large concentration range, while the N-ethyl- and N-propyl-derivatives of buprenorphine show an even larger cross-reactivity (106 and 197%, respec- tively). This pattern of cross-reactivity was also observed for morphine by Catlin et al. 1s In this study, rabbits were injected with the 2-diazomorphine-BSA conjugate. Normorphine, the N-dealkylation product of morphine, was observed to be the most effective inhibitor of the 3H-labelled morphine.It was observed that the specificity of the antibodies was directed particularly to the region adjacent to the point of attachment to the carrier. This was indeed confirmed by the low cross-reaction of codeine and morphine 3-glucuronide and the high cross-reaction of normorphine. A 50% reduction in [3H]morphine binding with 4.4 pmol of morphine was obtained, whereas 225 and 350 pmol of morphine glucuronide and codeine, respectively, were required for a SO% binding reduction of the tracer. In the present study, the 3-0-methyl derivative of buprenorphine also shows a relatively low cross-reaction (12%). This substituent is positioned adjacent to the point of attachment of the hapten to BSA. Further- more, when urine samples were hydrolysed with (3-glucoroni- dase, a remarkable increase in apparent buprenorphine concentration was obtained, indicating a low cross-reaction of the glucuronides.Indeed, when the same urine samples werc analysed with use of the DPC RIA kit for buprenorphine, roughly the same results as those obtained by Hand et ~ 1 . 2 7 were observed. Hand etal. found an average increase in apparent buprenorphine concentration of only 50% after enzymic hydrolysis. With the newly developed assay, the apparent buprenorphine concentration increased 20-fold after enzymic hydrolysis. This clearly demonstrates a specificity of the antibodies towards the substituent at the 3-0-position higher than that obtained with the antibodies of the DPC kit.142 ANALYST, FEBRUARY 1993, VOL. 118 Table 5 Specificity of thc antibodies towards synthesized analogues of buprenorphine Concentration Cross- Chemical structure added/ ngml-L reactivity (Yo 1 Chemical Structure Concentration Cross- added/ reactivity ng ml-1 (Yo) 20 12 r Y-cH7-cH3 20 106 20 197 1000 0.02 ‘COCH3 1000 0.008 1000 0,009 1000 0.5 1000 0.04 Determination of Unchanged Buprenorphine in Urine Samples Several urine samples (n = 32) were analysed by both the RIA (Y) and HPLC-ECD ( X ) methods.The HPLC method allows the detection of 0.2 ng ml-1 of both unchanged buprenorphine and N-dealkylbuprenorphine.28 Results by the two methods were related by the regression line Y = 1.44 + 1.64 X (Y = 0.910). Determination of Unchanged Buprenorphine in Plasma Samples The immunoassay was also used for the determination of buprenorphine in plasma samples.Quantitative results for the analysis of 10 plasma samples by both RIA (Y) and HPLC- ECD (X) are correlated by the following regression line: Y = 0.007 + 1.58 X (Y = 0.930). Parallelism The logit-log method for RIA dose interpolation permits testing of parallelism of dose-response curves for the standard and unknown.29 An unknown urine sample, contain- ing 16.1 ng ml-1, of the drug, was assayed three times, both undiluted and diluted with blank urine (1 + 2, 1 + 4 and 1 + 8). The observed (Y) and expected ( X ) values yielded the following linear regression equation: Y = 0.72 + 0.989 X (Y = 0.996). Analysis for Buprenorphine-like Material in Human Samples Apparent buprenorphine concentrations in urine samples collected from a volunteer, following intramuscular injection of 0.3 mg of buprenorphine, ranged from 2.5 ng ml-1 (10 h after administration) to 0.3 ng ml-I (48 h).Lower values were determined in urine samples after sublingual administration of 0.2 mg of buprenorphine (0.9 ng mi-I 3 h after intake of the tablet and 0.03 ng ml-1 after 4 d). When the urine samples were submitted to enzymic hydrolysis prior to analysis, it was still possible to detect buprenorphine up to 7 d after administration. In urine samples of persons suspected of Temgesic misuse, buprenorphine concentrations as high as 50 ng ml-1 were measured. These high figures, when com- pared with those obtained after therapeutic administration, clearly suggest a misuse of the drug. Discussion This paper describes the development of an RIA for bupre- norphine, which depends on the synthesis of [ 1”IJiodobupre-ANALYST, FERKUARY 1993, VOI,.118 143 norphine and of the diazobenzoic acid derivative of buprenor- phine. The diffcrcnt results obtained with the four separation techniques clearly demonstrate the importance of the separa- tion procedure. The best results were obtained with the second antibody method (GARGG). Howcver, a long incuba- tion time is a disadvantage of this method. On the other hand, thc second incubation time could bc reduced to 15 min, when using the combination method of second antibody and PEG It is known that the immunological specificity of the hapten is determined by the position whcre the spaccr is attached to the antigen.15.10-7’ I t was therefore expected that, by attaching the spacer at the C-2 position of buprenorphine, the most important antigenic determinants of buprenorphine (the N - 17-methylcyclopropyl, the substituent on C-7 and the hydroxyl group on position 3) would still be accessible to the immunogenic system. The results, however, have shown that the antibodies obtained can recognizc thc hydroxy group and the substitutent on C-7, but cannot distinguish buprenorphinc from N-dealkylbuprenorphine.A comparison of the results on urine samples, obtained by the assay developcd ‘in-house’ and using a commercially available kit from DPC, indirectly shows a remarkably lower cross-reactivity of the formcr assay towards the 3-0 glucuronide of buprenorphine. This can also be concluded from the regression line obtained by comparison of HPLC and RIA results for the same urine samples.The exact value for the cross-reactivity, however, could not be determined, thc glucuronidc not being available. The purification of [ ~~~I]iodobuprenorphine by HPLC, as described previously,l3 resulted in a high radiochemical purity and specific activity of the tracer, thereby allowing a very sensitive immunoassay. The sensitivity of the assay was superior to that of the recently introduced commercial RIA for buprenorphinc. The latter assay had a limit of detection of 1 ng ml- 1. The proposed RIA has been used in the first place to screen a large number of urine samples from persons suspected of Temgesic misuse. Furthermore, thc test allows the determination of low levels of buprenorphine in plasma samples.Its applicability to the analysis of forensic spccimens (also containing other drugs) and to the analysis of plasma samples, collected from addicts during a detoxification pro- gramme with buprcnorphine, is under study. 6000 (9%). We thank Dr. J . Heykants, R. Woestenborghs and 1. Geuens of Janssen Pharmaceutica (Beerse, Belgium) for their help and interesting discussions. Dr. R. Busson and L. Laruclle arc gratefully acknowledged for recording the 13C NMR and mass spectra. References Cowan, A , . Lewis, J . W.. and McFarlane, J . R., Br. J . Pharmucol., 1977, 60, 537. Dum. J . E., Herz, A , , Br. J . Plzurrnucd., 1981, 74, 627. Hanbrock, J . M., and Rance. M. J.. in OpiatesundEncl‘ogenouJ Opioid Peptidos. ed. Kosterlitz, H. W. Elsevicr/North Holland Biomedical Press, Amsterdam, 1976, p.295. 4 5 6 7 8 9 10 11 12 13 14 IS 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Heel, R. C., Brodgen, R. M., Speight, T. M., and Avery, G. S . , Drugs, 1979, 17, 81. Strang, J., I,uncet, 1985, 25. 725. Rainey, H. B., N . 2. Med. J.. 1986, 134, 72. Robertson, J. R., and Bucknall, A. B. V., Br. Med. J . , 1986, 292, 1465. Debrabandere, L . , Van Boven, M., and Dacnens, P.. .I. Forensic. Sci., 1992, 38, 82. Chowdhury, A. N.. and Chowdhury, S . , Br. J . Addict., 1990, 85, 1349. O’Connor, J . J . , Moloney, E., Travers, R . , and Campbell, A . , Br. J . Addict.. 1988, 83, 1085. Strang, J., Rr. Med. J . , 1991, 302, 969. Quigley, A. J.. Bredemeyer, L). E., and Scow, S. S . , Med. J . Aust., 1084, 31, 425. Harper, I . , N.Z. Med. J . , 1983, 96. 777. Bartlett, A. J., Lloyd-Jones, J . G., Ranee. M. J . . Flockhart, 1. R., Dockray, G. J., Bennett, H. R. D., and Moore, K. A.. Eur. J . Clin. Pharmucol., 1980, 18, 339. Catlin, D. H.. Schaeffcr, J . C., and Liewen, M. B., Life Sti., 1977, 20, 123. Kleemann. A., and Engel. J., Plrarrnuzeutisc/ie Wirkstoffe, Ergangzungsband 1982-IO87, Cieorg Thieme, Stuttgart. New York. 1987. Erlanger, B. F., Borik, F., Beiser. S . M., and Licberman, S . , J. Riol. Chern., 1957, 228, 713. Debrabandcre, L.. Van Bovcn, M., and Daenens, P., J . Radiolabeled Comp. and Radiopharrn., 1992, 31, 575. Herbert, V., Lan, K. S . . and Gottlieb, C. W., and Bleichcr, S . J., J. Endocrinol., 1965, 25, 1375. Utiger, C. R., Parker, M. L., and Daughaday, W. H., J. Clin. Invest., 1962, 41, 254. Morgan, C. R., and Lazarow, A., Diubetes, 963, 12, 115. Rodbard, D., Clin. Chem., (Winston-Salem, N.C.), 1974. 20, 1255. Sheehan, .I. C.. and Hcss, G. P., J . Am. Cirern. Soc.. 1955, 20, 1067. Erlanger, €3. F., Phurrnacol. Rev., 1973, 25, 271. Martin, M. J . , and Landon, J., in Rmdioimrnunoassay in Clinical Biochemistry, ed. Pasternark, K. Heyden, London, 1075, p. 269. Rodbard, D., Anal. Biocizem., 1978, 90. 1. Hand, C . W.. Ryan, K. E., Dutt, S . K., Moore, R. A., O’Connor, J . , Tabbot, D., and McUuay, H. J . , J . Anal. Toxicol., 1989, 13, 100. Debrabandere, L.. Van Boven, M., and Dacncns, P . , J . Chrornutogr., 1991, 564, 557. Rodbard. D., Munson. P. J . , and De Lean, A. K., in Medicine, International Atomic Energy Agency, Vienna, 1978, vol. I , Koida. M., Takahashi, M., Muraoka, S . , and Kaneto, H., Jup. J. Phurrnacol., 1974, 24, 165. Findley. J . W. A . , Butz, R . T., and Jones, E. C . , Clin. Chem. (Winston-Salem, N . C . ) , 1981, 27. 1524. Aherne, G. W., International CongreJJ and Symposium Series, Volume 58, ed. Stuart, J . F. R., Academic Press and the Royal Society of Medicine, London, 1983, p. 21. p. 469. Paper 2l03772F Received July 15, 1992 Accepted September 10, 1992