BIOMEDICAL CHROMAlOGRAPHY, VOL. 6, 300-304 (1992)

Solid Phase Extraction and Simultaneous High Performance Liquid Chromatographic Determination of Antipyrine and its Major Metabolites in Urine Mohamadi A. Sarkar* West Virginia University, School of Pharmacy, Department o f Basic Pharmaceutical Scicnccs, Morgantown, WV 26506, USA

Clark March and H. Thomas Karnes Virginia Commonwealth University, Department of Pharmacy & Pharmaceutics, Box 533, MCV Station. Richmond, VA 23298-0533, USA

A reversed phase gradient high performance liquid chromatographic method utilizing solid phase extraction has been described for the simultaneous determination of antipyrine (AP), 4-hydroxyantipyrine (COHAP), norantipyrine (NorAP) and 3-hydroxymethylantipyrine (3-OHMAP) in human urine after hydrolysis with pglucuronidase. The C-18 sorbent cartridges were conditioned and urine samples were applied, washed with 1 X 4 mL of phosphate buffer and eluted with 3 X 100 pL of 20% v/v of acetonitrile in methylene chloride. The eluent was evaporated to dryness, reconstituted in 100 pL phosphate buffer and injected. The calibration ranges were 2.0-250 pg/mL (AP), 2.5-250 pg/mL (NorAP), 2.0-250 pg/mL (3-OHMAP) and 5.0-500 pg/mL (4-OHAP) with regression coefficients of 0.998 or greater. Specificity was indicated by the absence of interferences in chromatogram of blank urine from normal as well as cirrhotic patients. The average recovery was 86.7% for AP, 90.5% for NorAP, 85.2% for 4-OHAP and 74.2% for 3-OHMAP. The within-assay precision as indicated by the reproducibility of the assayed spiked urine was less than 9% in all cases and the between-assay precision was less than 12%. The method was applied to studies on antipyrine metabolism in stable cirrhotic patients. Following administration of a single oral dose of about 1000 mg to nine stable cirrhotic patients and eight age-matched healthy volunteers, the cumulative account excreted in the urine up to 48 h for AP and the three metabolites was comparable to other literature reports.

INTRODUCTION Elimination of drugs from the body is mainly mediated by the hepatic oxidative Phase I enzymes. The Phase I cytochrome P-450 enzyme family consists of several isoforms with different substrate selectivity and metabolic capacity (Guengerich, 1989). The constitution and activity of these cytochrome P-450 isozymes differ based on genetic and environmental factors, like concomitant drug intake, smoking, liver disease, etc. leading to wide intersubject variability in the metabolic oxidation of drugs (Breimer, 1983; Vesell, 1984). The variability in drug metabolism could best be understood and predicted if the metabolic activity of the principal metabolizing enzymes could be characterized. Characterization of the multiple forms of the P-450 enzymes can be established by using several probe drugs in a “cocktail” or a single drug with several metabolites formed by different P-450 isozymes (Schellens et al., 1989). Antipyrine (AP) is now finding widespread application in estimating the oxidative metabolism capacity in liver disease patients (Wensing et al., 1990). Although A P elimination half-life by itself has often been used for investigating hepatic oxidation, a more detailed fingerprinting of the Phase I oxidative enzymes can be obtained by also measuring the major * Author to whom correspondence should be addressed. 026Y-387Y/92/060300-0S $07.SO (Q 1992 by John Wiley & Sons, Ltd

metabolites of AP. It has been shown that Phase I metabolism of AP yields three major oxidative products, norantipyrine (NorAP), 4-hydroxyantipyrine (4OHAP) and 3-hydroxymethylantipyrine (3-OHMAP), which are formed by three different hepatic cytochrome P-450 isozymes (Danhof et al., 1982; Inaba et nl., 1980, Pelkonen et al., 1980). Thus a rapid and easy method to measure the formation rate of these metabolites in urine (non-invasive sampling) would enhance the practicality of using A P as a probe for hepatic metabolic capacity. Several methods have been reported for the assay of AP and its metabolites in urine (Teunissen et al., 1983; De Vries et al., 1984; and Bottcher ef al., 1982). The results obtained with these methods have been very variable, mainly due to the instability of the metabolites at pH6, or oxidation of NorAP and 4-OHAP. One of these methods (Tang et al., 1982) uses gas chromatographychemical ionization mass spectrometry (expensive instrumentation found in only a few research laboratories) , others are extremely tedious and cannot measure all the three metabolites simultaneously (Danhof et al., 1979). Newer methods have been reported for simultaneous determination of A P and its metabolites, but these methods involve time-consuming liquid-liquid extraction (Eichelbaum et al., 1981). We have already established that solid phase extraction of AP offers the advantage of not only superior recovery and reproducibility to liquid-liquid extraction, but also rapidity and Received 20 December 1991 Accepted 22 June 1992

HPLC OF ANTIPYRINE AND I1 S MAJOR MEI'AHOLI rES

convenience (Sarkar and Karnes, 1988). Only one researcher (Moncrieff, 1986) has attempted to perform a solid phase sample clean-up; however. this method suffers from the disadvantage of being an on-line method that requires manual injection and does not allow automated sample injection. Also it is unclear how well this method could stand up to the rigours of a large number of samples, due to the obvious problem of back-pressure resulting from the multiple direct injection of urine. In this paper, a reversed phase automated gradient high performance liquid chromatographic (HPLC) method is presented for the simultaneous assay of A P and its metabolites by solid phase extraction from urine. The method was then applied to a metabolic study. A P and its metabolites were analysed in nine stable cirrhotic patients and eight healthy volunteers matched for age with the cirrhotic patients. The method proved to be robust and reproducible.

~~

EXPERIMENTAL Materials. AP, 4-dimethylaminoantipyrine (internal standard, IS), NorAP and 4-OHAP were purchased from Aldrich Chemical Company (Milwaukee, WI 53233, USA). 3-OHMAP was synthesized according to Yoshimura (1971), starting from AP. The melting point of the recrystallized product was 142-144 "C, analytical data (NMR, UV and IR) were in keeping with the literature reports and the final product was pure as judged by thin layer chromatography assay. Limpet acetone powder, type I (contains /3glucuronidase and sulphatase), was obtained from Sigma Chemical Company (St. Louis, MO 63178, USA). The C-18 solid phase extraction cartridges (AASP@)were from Varian Sample Preparation (Harbor City, CA 90710, USA). All chemicals used were HPLC or analytical grade. Water was deionized and distilled. Apparatus. The liquid chromatograph consisted of two Gilson Model 302 pumps with a Model 802B manornctric module, Model 811 dynamic mixer, Model 231 autosanipler and Model 714 system controller with automated data acquisition through a Tandy 3000 personal computer (Gilson Medical Electronic, Middleton, W1 53562, USA). Detection was with a Model SPD6A UV detector (Shimadzu, Columbia, MD 21046, USA) monitoring at 254nm at 0.08 aufs. An in-line 0.5 micron filter was used before a 2 mm i.d. x 20 mm guard column dry packed with Corasil phenyl packing, 40 micron particle size (Waters Associates, Milford, MA 01757, USA). The analytical column was a 4.6mm i.d. X 100 mm cartridge column packed with Spherisorb phenyl packing, 5 micron particle size (Alltech Associates, Deerfield, IL 60015. USA). The injection volume was 20 pL and the How rate was 1.0 mumin. The binary gradient consisted of solvent A (0.05 M potassium dihydrogen phosphate, pH 4.5) and solvent B (methanol). The gradient profile was a linear increase in solvent B from 10% at time 0 min, to 40% at 14min, to 60% at 21 min and a decrease to initial conditions at 21.1 min. The entire run time was 29 min. Standards. Stock solutions of AP, NorAP, 3-OIlMAP, 4OHAP and 4-dimethylaminoantipyrine (IS) were made up in methanol to 10 mg/mL. The four analyte solutions (first four) were diluted to 250 yg/mL using methanol and the IS to 1 mg/

30 1

mL using water. The two solutions of AP and metabolites were used to prepare standards in blank urine at the following concentrations: 2 , 5 , 10,25,50, 100 and 250 pg/mL for AP, 3OHMAP; 2, 5, 10, 25, 50, 100 and 250 WghL for NorAP; 5, 10, 20, 25, 50, 100, 250 and 500pg/mL for 4-OHAP. The quality control samples consisted of blank urine spiked to concentrations of 15,75 and 200 pg/mL for AP, NorAP and 3OHAP; 15, 75 and 350 pg/mL for 4-OHAP. Standards and controls were aliquoted and frozen at - 20 "C until analysis. Urine sample treatment. To a 12 X 75 mm borosilicate tube containing 100 pL of urine and 100 yL of 0.05 M potassium dihydrogen phosphate (pH 4 . 9 , 20 mg of limpet acetone powder (for enzyme hydrolysis of conjugated metabolites) and 16mg of sodium metabisulphite were added and the mixture vortexed for 15 s. The tube was indicated at 40°C with mild agitation for 3 h. After cooling to room temperature, 10 pL of the 1 mg/mL solution of IS in water was added, vortexed and centrifuged at SOOxg for 20 min. Supernatant (100 pL) was transferred to a clean tube containing 0.9 mL of 0.1 M tris (pH4.5). The C-18 solid phase solvent was conditioned with 1 mL of methanol followed by 1 mL of water. To the top of the sorbent bed, 0.5 mL of the diluted urine was added along with 0.5mL of 0 . 0 5 ~potassium dihydrogen phosphate. The mixture was slowly forced through the cartridge and the bed was washed four times with 1 mL of phosphate buffer. The sorbent bed was allowed to completely dry before the analytes were eluted with thrce 100pL volumes of acetonitri1e:methylene chloride (20:80). The eluate was evaporated at 40 "C under vacuum while vortexing. The residue was reconstituted in 1OOpL of phosphate buffer and 20 pL was injected. Recovery. The recovery of the assay was determined at the low and high control concentrations for each analyte and for the IS by comparing the mean of six extracted determinations to the response of the compounds injected directly in phosphate buffer.

RESULTS AND DISCUSSION

Method development HPLC conditions. The main problem associated with the analysis of parent compound and metabolite is that significant differences in polarity exist between the analytes due to the structural differences resulting from metabolism. The metabolic pathways for A P are shown in Scheme 1. NorAP ( ~ K ~ 9 .and 2 ) 4-OHAP (pK, 10.3) are not as non-polar as 3-OHMAP ( ~ K ~ 2 - 3and ) AP (pKd1.4) (Danhof et al., 1979). In the past researchers have always been faced with the challenge of attempting to analyse all the three major metabolites of A P and AP itself under the same HPLC conditions in one injection. An initial attempt was made to measure all the four analytes by an isocratic method. Several organic modifiers were tried to achieve optimal resolution within the shortest analysis time. Acetonitrile in phosphate buffer at pH6.S or pH4.5, on a C-18 column, did not give good resolution of the metabolites. Switching to tetrahydrofuran as the organic modifier gave better resolution but the peak shape was poor. A mixture of methanol and tetrahydrofuran in phosphate buffer at pH 6.5 gave good peak shape but the resolu-

MOHAMADI A. SARKAR. CLARK MARCH AND H. THOMAS KARNES

302

ANTIPYRINE (AP)

I

4-Hydroxy Antipyrine (4-OHAP)

\

3-Hydroxymethy1 Antipyrine (3-OHNAP)

Norantipyrine (NorAP)

Scheme 1. Oxidative biotransformation pathways of AP in humans. Although 3-OHMAP and 4-OHPA further undergo conjugative Phase II metabolism, since the method reported in this paper measured the total metabolites (conjugated + free) after enzyme hydrolysis, this schematic only includes the Phase I oxidative metabolites of AP.

tion of NorAP and 4-OHAP was inadequate. A C-8 column did not correct the problems observed on the C-18 column. However, a phenyl column gave promising resolution and good peak shape with 45% (vh) methanol in 0.05 M phosphate buffer, pH 6.5. Since the peaks were eluting very close to each other, lower concentrations of methanol, 20% v/v, were tried but NorAP eluted very quickly. Changing the p H of the phosphate buffer to 4.5 from 6.5 gave reasonable resolution of all four analytes; however, 4-OHAP and AP were late eluting and had poor peak shape. Thus the best recourse was to use a gradient system to elute these two peaks sooner. The final IIPLC condition giving optimal resolution was a binary gradient consisting of 0.05 M potassium dihydrogen phosphate, p H 4.5 and methanol. The retention times were 9.1 min for NorAP, 11.0min for 3-OHMAP, 14.8min for 4OHAP, 17.8 min for AP and 19.9 min for IS. Solid phase extraction. Although a prospective compara-

tive evaluation was not carried out between the existing liquid-liquid extraction and solid phase extraction, a liquid-liquid extraction method for the analysis of AP and its three metabolites was first explored. Various proportions of chloroform and ethanol, methylene chloride and pentane, or methylene chloride by itself were tried for extracting AP and all the three metabolites simultaneously. It was observed that extraction of analytes was unpredictable and erratic even under controlled conditions, since metabolites are unstable at pH6 and also are prone to oxidative degradation. Thus the attention was focussed on developing a solid phase extraction method. The fundamental principle of solid phase extraction is to load the analytes on a packing which specifically and strongly binds the analytes, wash away the interfering organics and selectively elute the analytes. The first step was to identify the solid phase packing for which all the four analytes had strong affinity. This was investigated by conditioning (1 mL MeOH and 1 mI, water) and loading C-18, C-8, C-2, cyano and phenyl cartridges with aqueous solutions of the four analytes

and IS in phosphate buffer, ( 0 . 0 5 ~ ,pH4.5). The buffer was collected and evaporated to dryness, reconstituted and injected to measure any of the analytes that could have passed through without adhering to the cartridges. Also the analytes bound to the cartridges were then eluted with 100”/0 methanol and assayed for their concentration. It was found that the C-18 cartridge had optimal affinity for all the four analytes. Tris buffer (0.1 M , pH 4.5) allowed better loading of the analytes on the C-18 cartridge than phosphate buffer alone, probably due to the polarity differences of the ions in the two buffers. Several washings using phosphate buffer (0.05 M , pH 4.5) did not elute the analytes and this property was helpful in cleaning up the urine samples, as four 1 mL washings of phosphate buffer were sufficient to remove interfering peaks co-eluting with the analytes. The optimal eluting solvent was 3 x 100 pL of 20% v/v acetonitrile in methylene chloride. This solution was most effective if used after thoroughly drying the sorbent bed for at least 60 s. This solid phase extraction procedure proved to be very robust since numerous healthy blank urine samples were analysed with or without the analytes and no interfering peaks coeluted at any of the retention times of the four analytes or IS. The C-18 cartridges provided adequate affinity for all the four analytes despite the significant differences in the polarity. The phosphate buffer washing step (1x 4 mL) was sufficient to remove any interfering peaks co-eluting with the analytes and provided excellent specificity for simultaneous analysis of all the four analytes. The only previous report of using an on-line solid phase sample cleanup may not be very applicable for large volumes of samples since it involves direct injection of urine. The method reported in this paper utilizes off-line sample preparation, making it possible to automate the whole analysis process. The principal advantage offered by this solid phase extraction procedure is rapidity of analysis, unlike the inconvenience of handling organic solvents and two separate extractions (Danhof et al., 1979) or the use of a normal phase (Eichelbaum et al., 1981) HPLC method. Stability of metabolites. Since 4-OHAP is unstable, an

antioxidant (sodium metabisulphite) was added during the 3 h enzymatic incubation period. The enzymatic hydrolysis was performed with /3-glucuronidasesulphatase and the 3 h incubation was optimized by comparing the concentrations obtained after 1 h, 2 h, 3 h and 4 h incubations of three of the actual study samples. No increase was observed in the metabolite concentrations after 3 h.

Specificity

The assay method reported in this article has good specificity. A blank urine sample is shown in Fig. 1 along with the blank urine with spiked standards at 75 pg/mL in Fig. 2, comparison of the two chromatograms shows absence of interfering peaks at the retention window of interest. However, during the analysis of samples in the pharmacokinetic study it was observed that 34% (12 of 35 subjects) of the first-

HPLC OF ANTIPYRINE AND ITS MAJOR METABOLITES

303

Table 1. Accuracy and precision of AP and metabolite assay. Intraday (within-assay) precision Analyte

AP

NorAP

30

3-OHMAP

T I W E Imin)

4-OHAP

Figure 1. Blank chromatogram of urine (cirrhotic patient) processed through the solid phase extraction.

morning urine samples had a peak that co-eluted with NorAP. The maximum apparent NorAP concentration produced by this peak was 8.5 &mL in one sample, but the remainder of the samples were all less than 3.5 pg/mL. This interferant peak was absent in samples collected later in the day as well as in all the normal volunteer blank urine samples used during the assay validation study.

Pred. conc. ipg/mL)

Mean conc. (pg/mLI

Standard deviation

Relative

SDI%l

Relative error (%)

15 75 200 15 75 200 15 75 200 15 75 350

14.7 76.0 202.3 15.6 76.9 203.0 15.8 75.4 206.6 15.0 75.2 353.3

0.73 1.89 8.58 0.69 3.89 10.33 1.14 6.26 18.37 1.04 2.38 15.20

4.97 4.49 4.24 4.42 5.06 5.09 7.22 8.30 8.89 6.93 3.16 4.30

-2.00 1.33 +1.15 +4.00 +2.53 + 1.50 t5.33 +0.53 +3.30 0.00 +0.27 +0.94

+

Recovery

The absolute recovery for the analytes and IS were assessed with the two quality control samples ( n = 6). A comparison of the absolute area counts, represented as percent of direct injection, is shown in Table 3 . The recovery for the IS was 87.6%. Limit of quantitation

Accuracy and precision

The three quality control samples were used to evaluate the within-run and intraday precision. The standard curves were linear within the calibration range and were calculated using least-squares linear regression (weighting of Uconcentration) for all analytes. The mean correlation coefficient was at least 0.998. The control samples ( n = 7) were analysed between the standards used in the standard curve concentrations. The linear regression equation for the calibration curve was used to calculate the observed concentrations. Table 1 shows the results of the intraday precision. The interday precision results are expressed in Table 2. The precision is indicated as the percent relative standard for spiked controls. The precision of the method was within