Transcript

BIOMEDICAL CHROMATOGRAPHY, VOL. 10, 161-166 (1996)

Quantification of 9-Carboxy-ll-nor-A’- tetrahydrocannabinol in Urine Using Brominated 9- Carboxy - 11 -nor=A’= tetrahydrocannabinol as the Internal Standard and High-Performance Liquid Chromatography with Electrochemical Detection

Daniel H. Fisher* and Marc I. Broudy Department of Medical Laboratory Science, 206 Mugar Building, Northeastern University, Boston. MA 021 15. USA

L. Megan Fisher Department of Biology, 414 Mugar Building, Northeastern University, Boston, MA 02115. USA

A method was developed for quantitating 9-carboxy-ll-nor-A’-tetrahydrocannabinol in human urine as part of the process for validating an automated enzyme immunoassay for marijuana metabolites. Sample cleanup was accomplished using a mixed-mode solid-phase extraction. 9-Carboxy-ll-nor-A’-tetrahydrocannabinol and the internal standard, brominated 9-carboxy-ll-nor-A9-tetrahydrocannabinol, were quantified using high-perform- ance liquid chromatography with electrochemical detection (+0.85 V). The linear range for this method is 0.0124.20 pglmL. No interference was seen for 22 drugs and metabolites. The pooled relative standard deviation is 4.1% (n=27) for the quality control samples. This method was compared to gas chromatography with mass spectrometry by linear regression analysis. The slope of the line is 1.00*0.05 (standard error), the intercept is approximately zero, the coefficient of determination is 0.994, and the standard error of the estimate is 0.006 pglmL.

~

INTRODUCTION

Our goal was to develop a high-performance liquid chromatographic (HPLC) method for determining 9-car- ~ x y - 1 1 -nor-A9-tetrahydrocannabinol (THC-COOH) con- centrations in human urine. This was part of the process of adapting immunoassay reagents to an automated chemistry analyser. The following constraints had to be met: (1) the urine volume must not exceed 2 mL so that duplicates of expensive immunoassay urine standards could be analysed (2) the limit of quantification for THC-COOH must be approximately 10ng/mL of human urine in order to quantify the lowest expected concentrations in the urine samples; (3) the procedure must be amenable to batch analysis because of the large number of urine samples to be analysed; (4) the method must be practical and cost effective.

Gas chromatography with mass spectrometry (GC/MS) is an excellent method for analysis of THC-COOH (Bronner and Xuy, 1992; Dixit and Dixit, 1991; Jenkins et al., 1995; Joern, 1992; Lisi et al., 1993; Liu et al., 1994; Moody et af. 1992 a, b; Wu et al., 1993, but was not selected for this application because the cost of sample analysis using MS is relatively high compared to other forms of quantitation. THC-COOH has also been quantified by high-performance liquid chromatography With electrochemical detection ( H P L a C D ) or gas chromatography with electron capture detection (GCECD). HPLCECD is a less time consuming method than GCECD because derivatization is required in

*Author to whom correspondence should be addressed at: Center for Bioandytical Research, University of Kansas, Lawrence, KS 66047, USA.

order to improve the signal and reduce chromatographic tailing (Maseda et a!.. 1986; Rosenfield et al., 1986, 1989). No derivatization is required prior to HPLCECD because an acidic pH is used to suppress ionization and the signal is generated by oxidation of the phenol moiety of THC- COOH.

A significant limitation associated with previously reported HPLCECD or GCECD methods is the lack of an ideal internal standard. A variety of compounds have been used as internal standards. For example, p,p’-DDT (Maseda et af., 1986), tetracosanoate (Rosenfeld ef af., 1986, 1989). cannabinol (Bourquin and Brenneisen, 1987), 4dodecylre- sorcinol (Thompson and Cone, 1987), 1 l-hydr~xy-A’-tetrahydrocannabinol (Craft et al.. 1989). and n-octyl-p-hydroxybenzoate (Nakahara and Cook, 1988; Nakahara et al., 1989). These internal standards are not ideal (Nakamura et al. 1990) because they either do not have the Same basic ring structure as THC-COOH, or they are present in marijuana.

The best internal standard for HPLC/ECD quantification would be a chemical derivative of THC-COOH that does not modify the carboxyl moiety or phenol moiety and results in a final structure identical to a compound present in marijuana. Retaining the acidic and lipophilic properties of THC-COOH would allow the use of a mixed-mode solid- phase extraction (SPF) column. The mixed-mode SPE employed here was designed to extract THC-COOH based on hydrophobic, polar and anionic interactions, and this type of SPE provides enhanced sample clean-up when compared to liquid-liquid extraction or reversed-phase SPE (Pocci er af., 1992).

We report here the synthesis of brominated THC-COOH for use as an internal standard and validation of a HPLC/

Received 14 November 1995 Accepted 22 February 19%

I62 D. H. FISHER ETAL.

ECD method using brominated THC-COOH for internal standard quantification of THC-COOH in human urine.

dimethyldichlorosilane in toluene for 30 rnin (Knapp, 1979). The surface was then rinsed with toluene followed by methanol. The glassware was allowed to dry at room temperature.

~~ ~

EXPERIMENTAL

Chemicals. Chloracetic acid (99%) and triethylamine (99%) were obtained from Aldrich (Milwaukee, WI, USA). HPLC grade water, methanol, acetonitrile and tetrahydrofuran were supplied by Fisher Scientific (Pittsburgh, PA, USA). High purity helium (99%) was prepared by Medical Technical Gases (Medford, MA, USA). All other chemicals were reagent grade. d-9-Carboxy-l l -nor-A9- tetrahydrocannabinol was purchased from Alltech-Applied Science Labs, State College, PA, USA.

HPLC system. The HPLC system consisted of a SP8800 ternary pump (Spectra-Physics. San Jose, CA, USA), a 7125 injector with a 200pL sample loop (Rheodyne, Cotati, CA, USA), a Bio- analytical Systems column (3.2 x 100 mm) packed with octadecyldimethylsilyl 3 pm silica gel (BAS, West Lafayette, IN, USA), a BAS LC4A amperometric detector, a BAS LC-44 thin- layer cell housed in a BAS CC-5 cabinet, and an Eldex CH-150 column heater (San Carlos, CA, USA), Chromatograms were recorded with a Spectra-Physics SP4270 integrator.

Mobile phases. Mobile phase A was prepared by adding 100 mL of a 0.5 M, pH 3.0 monochloroacetic acid buffer and 1.4 mL of triethylamine to a 1 L volumetric flask. The flask was filled to the mark with a solution of 30% H,O:70% CH,OH (v/v). Mobile phase B was prepared by adding 100 mL of a 0.5 M, pH 3.0 monochloroacetic acid buffer and 1.4 mL of triethylamine to a 1 L volumetric flask. The flask was then filled to the mark with a solution of 5 % H,O:15% THF8O% CH30H (v/v/v). The mobile phases were filtered using an all-glass filtration apparatus and a Gelman Nylaflo membrane (0.20 pm x 45 mm, Fisher Scientific). Mobile phase A was stored in a tightly capped bottle in the refrigerator. Mobile phase B was partitioned equally into three storage bottles, tightly capped and stored in the refrigerator. Both mobile phases were prepared weekly.

Chromatographic conditions. A two-foot long coil of stainless steel tubing connecting the pump to the injector, the injector and the analytical column were maintained at 40°C in the column heater. The tubing connecting the analytical column to the detector was insulated with glass wool. A BAS L C 4 A amperometric detector was connected electrically in accordance with manu- facturers directions to a BAS LC-44 thin-layer cell with a dual glassy-carbon electrode housed in a BAS CC-5 cabinet. The dual glassy-carbon electrode was polished weekly in accordance with manufacturers directions. The LC4A amperomehc detector was set to +0.85 V relative to a Ag/AgCI reference electrode. The mobile phases were degassed for 10 rnin with helium before use and continuously thereafter. The mobile phase flow was set at 1.0 mllmin. The initial mobile phase composition was 90% A: 10% B. The SP8800 was programmed to linearly decrease the concentration of A to 70% in the first 15 rnin and then to 40% A in the next 15 min where the mobile phase composition was held for 5 min. At the end of each day, the column was washed with 40% H,O:60% CH,OH (v/v). the cell was washed with water and methanol and the reference electrode was stored in 3 M NaCI. The mobile phase reservoirs were tightly capped and stored overnight in the refrigerator.

Silanization. Borosilicate glassware was silanized by allowing the surface to remain in contact with a 5% (v/v) solution of

Standards. The following solutions were prepared in freshly voided drug-free urine preserved with 0.05% (w/v) sodium azide. The 2.0 pg/mL working solution was prepared by adding a 200 pL aliquot of a 100 pg/mL stock solution of d-9-carboxy- 1 1 -nor-A9-tetrahydrocannabinol to a silanized 10-mL volumetric flask and filling to the mark with urine. Six standards were prepared in 16 x 125 mm borosilicate test tubes before each use by diluting the2.0 Fg/mL working solution 1 5 , l:lO, 1:20, 1:40. 1:80 and 1 : 160 with an aliquot of urine such that the final total volume of each standard was 2.0 mL.

Internal standard. The internal standard was synthesized by brominating THC-COOH in carbon tetrachloride. A 0.58 mmoV mL stock solution of Br, in CCI, was prepared before each synthesis by adding 150 p L of neat Br, to 5.0 mL of CCI, in an amber vial. A 1.16 pmoVmL solution of Br, in CCI, was then prepared by adding 10.0 p L of the 0.58 mmoVmL stock solution to 5.0 mL of CCl, in an amber vial. A 100 p L aliquot of a 100 pg/mL (0.29 pmoVmL) solution of THC-COOH was added to a 1.5 mL silanized amber vial and the methanol was removed by evapora- tion under a gentle stream of nitrogen at ambient temperature. A 100 p L aliquot of the 1.16 FmoVmL solution of Br, in CCI, was added to the vial containing the THC-COOH and allowed to react for exactly 2 min at room temperature in reduced light. The reagent and solvent were evaporated under a gentle stream of nitrogen. The product was dissolved in 1.00 mL of HPLC grade CH,OH. The vial was capped and stored at 4°C. A new batch of internal standard was prepared when approximately 100 pL was left or after one month, whichever came first.

Quality controls. The DAU-I, DAU-11, DAU-111 and DAU-IV quality controls were prepared using unpreserved human urine, which was then lyophilized (Ciba-Corning Diagnostics. Medfield, MA, USA). The Ciba-Corning controls were reconstituted in 25.0 mL of HPLC grade water and stored at 4°C for up to 7 days. The above threshold and high positive controls were prepared in drug-free human urine and preserved with sodium azide (Hycor Biomedical, Garden Grove, CA, USA). The Hycor quality controls were stored at 4°C until the expiration date. All quality controls were discarded if there were signs of microbial growth. All quality controls were gently swirled before each use.

Protocol. The protocol is summarized in Fig. 1. For either the quality control or patient samples, a 2.00 mL aliquot was added to a silanized 16 x 125 mm borosilicate glass test tube. If required samples were diluted 1:2 with drug-free human urine. A 3.0 mL aliquot of a 95% 0.10 M, pH 6.0 sodium acetate: 5% methanol buffer and a 20.0 p L aliquot of the internal standard solution were added to each test tube containing the standard, quality control or sample. The pH of each solution was checked using pH paper. If the pH was not between 4.5 and 6.5, it was adjusted using 0.1 M NaOH or 0.1 M HCI.

The metabolite of interest and internal standard were extracted from urine using a Supelco solid-phase extraction manifold (Supelco, Bellefonte, PA, USA) and Bond-Elut Certify I1 solid- phase extraction columns (Varian, Harbor City, CA, USA). Each 130-mg Bond Elut Certify I1 column was conditioned at a rate of 1 drop/s under ca. 3-5 mm of Hg with 2 mL of HPLC grade methanol followed by 2 mL of a 95% 0.10 M, pH 6.0 sodium acetate: 5% methanol (The vacuum was turned off just before the methanol or buffer reached the top of the sorbent bed to prevent

HPLCECD OF TCH-COOH IN HUMAN URINE 163

the column from drying). The treated sample was transferred to the top of the solid-phase extraction column with a Pasteur pipette and extracted at a rate of 1 drop every 4-5 s under 3-5 mm of Hg. The column was washed with 5 mL of CH,OH:H,O (1 : 1, v/v) at a rate of 1 drop/s under ca. 5 mm of Hg. The waste collecting tubes were replaced with clean silanized disposable borosilicate glass test tubes (16 x 125 mm). The metabolite of interest and internal standard were eluted with 5 mL of a solution containing 75 parts hexane:25 parts ethy1acetate:l part acetic acid at a rate of 1 drop every 1.5-2 s under ca. 3-5 mm of Hg.

I

I -

+ I

Y;.

I -- I I Enponlrs*al

Rca*lltrrtrk M N e P h u A

w HPLCIECD auur(iliuti

Figure 1. Method flow chart for quantification of THC-COOH in human urine using SPE and HPLCECD.

C02H

I

The solvent was removed by evaporation under a gentle stream of nitrogen at ambient temperature. The residue was immediately reconstituted in 200 FL of Mobile Phase A and stored at 4°C. A 20-40 FL aliquot was injected into the HPLC.

Method evaluation. The method evaluation protocol of Peters et al. (1986) was followed. The peak-height ratio was calculated by dividing the peak-height of THC-COOH by the peak-height for the internal standard. The coefficient of determination (R’), correlation coefficient (r). slope (m), standard error of the slope (se,), intercept (b), standard error of the intercept (seb) were determined by linear regression analysis of each standard concentration and the corresponding peak-height ratio. The limit of quantification was set equal to the concentration of the lowest standard (Talyor, 1987). The percent interday relative standard deviation (%RSD) and pooled %RSD was calculated following the method of Anderson (1987). The concentration of THC-COOH for each of the five positive quality samples obtained using the method reported here was compared to the concentrations obtained by GC/ MS using linear regression analysis. The standard error of the estimate (se,) was also calculated for the HPLC method using the GC/MS method as the reference method.

RESULTS AND DISCUSSION

Internal standard The internal standard was synthesized by brominating THC- COOH. In the absence of light, bromine reacts by addition with an alkene in CCI, while the addition of bromine to benzene will not react in CCI, unless a Lewis acid is added. Phenol reacts rapidly with bromine in water to form a tribromophenol. Monobromination of phenol can be achieved by reacting phenol and bromine at 4°C in carbon disulphide.

Based on the reactions of cyclohexene and phenol with bromine, the reaction conditions were optimized to favour addition of bromine across the double bond between carbons 8 and 9 of THC-COOH. The optimum molar ratio of bromine (BrJ to THC-COOH is 4:l. The optimum reaction conditions are 2min at room temperature in the dark when CCl, is used as the solvent. These reaction conditions result in the formation of a single product as determined by reversed-phase HPLC with UV detection. A bromine-to-THC-COOH ratio greater than 4: 1 resulted in multiple products, presumably from electrophilic substitu- tion of bromine at the phenol moiety of THC-COOH.

Proton NMR characterization or elemental analysis would be prohibitively expensive given the Cost of the starting material ($50 for l00p.g) and the amount of

OH Brx Figure 2. Proposed reaction scheme for the synthesis of the internal standard structure.

D. H. FISHER E T A L . 164

material required for analysis by these two techniques. The fast atom bombardment MS fragmentation pattern is consistent with a dibrominated product. The prominent peaks show the isotopic clustering patterns expected for the addition of bromine. This is consistent with the product being 9-carboxy- 1 l-nor-8,9-dibromotetrahydrocannabinol (Fig. 2).

Chromatography

The electrochemical detector background current is a function of organic solvent composition and ionic strength. Therefore, a mobile phase gradient will change the background current. However, this can be minimized by making the ionic strengths of the mobile phases A and B

Time (mi.)

10 nA

t S 10 1s 20 2s 30

Tim. (mi-)

Figures. HPLCECD Chromatographs: A, 0.100 pg of THC-CCOH per mL of drug-free human urine; 8, DAU-Ill (0.125 Fg/mL of THC-COOH).

HPLCECD OF TCH-COOH IN HUMAN URINE I65

Table 1. Method comparison: the average concentration for each quality control sample determined by this HPLCIECD method is compared to the expected values as determined by CC/MS with deuterated internal standard

HPLC/ECD [TCH - COOHI.,,.,

Quality control sample In') pglrnL %RSW

Ciba-Corning' DAU-I (3) < LOD' NAg Ciba-Corning' DAU-I1 (3) 0.015 6.1 Ciba-Corning' DAU-Ill (6) 0.132 3.3 Ciba-Corning' DAU-IV (7) 0.264 6.2 Hycor" above threshold (4) 0.121 0.5 HycoP high positive (7) 0.291 0.3

GUMS' [THC - COOHI,,,& pg/rnL %Biasd < LOD' NA'J 0.015 +0.3 0.125 +5.4 0.250 +5.7 0.125 -3.2 0.300 -3.1

'The number of replicates of each sample. Ciba-Corning determined the urine concentrations at their

facility using GUMS. Hycor Biomedical contracted Research Triangle Institute to determine the concentrations using GC/MS (Research Triangle Park, NC, USA).

lnterday relative standard deviation. The interday per cent bias of this HPLC/ECD method compared

to GUMS at that level. The quality control samples are described in the experimental

section. 'Limit of detection. Q Not applicable.

identical. The shift in the baseline is about 1-2 nA under the chromatographic conditions reported here (Fig. 3).

Analysis of drug-free human urine shows no interfering peaks. The standard, THC-COOH, has a retention time of 8.7 f i n and the internal standard has a retention time of 22.5 min (Fig. 3A). The chromatograms for a positive human urine control are shown in Fig. 3B.

Linearity and precision

For the protocol described here, the linear range is from 0.012 to 0.40 pg of THC-COOH per mL of human urine as established by linear regression analysis (r=0.999, m=13.9+-0.1 (se,), b=-0.07*0.01 (seb)). The limit of quantification for THC-COOH is 0.012 kg/mL of human urine. Similar regression analysis results were obtained for six other standard curves prepared separately over a 6-month period. The interday %RSD for each quality control sample ranges from 0.3-6.2% (Table 1) and the pooled interday %RSD for all the quality controls is 4.1% (n=27, Table 1). (Note, DAU-I was not included in the

%RSD calculations because THC-COOH was not added to this quality control specimen).

Interference

No endogenous interfering peaks were observed in pooled Ciba-Corning DAU-I. None of the 22 additional drugs in the Ciba-Corning DAU 11-IV controls co-eluted with THC- COOH or the internal standard. These were amobarbital, amphetamine, benzoylecgonine, cocaine, codeine, ecognine methyl ester, gluthethimide, imipramine, meperidine, meth- adone and its metabolite, methamphetamine, methaqualone, morphine, morphine-3-glucuronide, nortriptyline, oxaze- pam, phencyclidine, phenobarbital, propoxyphene, quinine and secobarbital.

Method comparison

The average concentration of the quality control samples as determined by HPLCECD was compared to the ex- pected concentration for these samples determined by GC/MS (Table 1). The GC/MS method used a deuterated internal standard. The data in Table 1 was also compared using linear regression. The resulting equation is THC- COOH,,,,,,= l.OO*THC-COOH~, (R2=0.994, se,= 0.05, b S 0 , se,=O.Ol). The standard error of the estimate (se,) is =0.006 kg/mL and the probability (P) that the rela- tionship occurred by chance is <0.05.

Conclusion

The HPLCECD method reported here is accurate, precise, practical and cost-effective. The method was successfully applied to validating an automated enzyme immunoassay for marijuana metabolites.

Acknowledgements

This work was supported by a grant from Instrumentation Laboratory, Lexington, MA, USA. The authors thank Dr Gerald Sheys at Bioran Laboratories, Cambridge, MA, USA for supplying us with positive human urine samples.

REFERENCES

Anderson, R. L. (1987). Practical Statistics for Analytical Chem-

Bourquin, D. and Brenneisen. R. (1987). J. Chromatogr. 414,

Bronner, W. E. and XU. A. s. (1992). J. Chromatogr. 58011-2).

Craft, N. E., Byrd, G. D. and Hilpert, L. R. (1989). Anal. Chem. 61,

Dixit, V. and Dixit, v. M. (1991). J. Chromatogr. 567(1), 81. Jenkins, A. J., Darwin, w. D., Huestis, M. A., Cone, E. J. and

Joern, W. A. (1992). Clin. Chem. 38(5), 717. KnapP, D: R. (1979). Handbook of Analytical Derivatization

Lisi. A. M.. Kazlauskas, R. and Trout, G. J. (1993). J. Chromatogr.

ists, pp. 44-45. Van Nostrand Reinhold, New York.

187.

63.

540.

Mitchell, J. M. (1995). J. Anal. Toxjco/. 1g(1), 5.

ReactJons, P. 20. John Wiley &Sons, New York.

617(2). 265. LiU, R. H.. Edwards, C., Baugh, L. D., Weng, J. L., Fyfe, M. J. and

Walia* A- so (1994). J. Anal. Toxicol. 18(2), 65.

Maseda, c., Hama, K., Fukia, Y., Matsubara, K., Takahashi, s. and

Moody, D. E., Rittenhouse, L. F. and Monti, K. M. (1992a). J. Anal.

Moody, D. E., Monti, K. M. and Crouch, D. J. (1992b). J. Anal.

Nakahara, Y. and Cook, C. E. (1988). J. Chromatogr. 434,247. Nakahara, Y., Sekine, H. and Cook, C. E. (1989). J. Anal. Toxicol.

13, 22. Nakamura, G. R., Meeks, R. D. and Stall, W. J. (1990). J. forensic

Sci. 35(4), 792. Peters, T. Jr., and Westgard, J. 0. (1986). Evaluation of methods.

In Textbook of Clinical Chemistry, (Tetz, N. W. ed.), pp 410-423. W. 6. Saunders, Philadelphia, PA.

Pocci, R., Dixit, V. and Dixit, V. M. (1992). J. Anal. Toxicol. 16, 45.

Rosenfeld, J. M., McLeod, R. A. and Foltz, R. L. (1986). Anal. Chem. 58,7 16.

Akane, A. (1986). ~ O f e n S k SCi lnt32.259.

Toxicol. 16(5), 297.

Toxicol. 16(5), 302.

I66 D. H. FISHER ETAL.

Rosenfeld, J. M., Moharir, Y. and Sanler, S. D. (1989). Anal.

Talyor, J. K. (1987). Quality Assurance of Chemical Measure-

Thompson. L. I. and Cone, E. J. (1987). J. Cbromatogr. 421,91. Wu, A. H., Liu, N.. Cho, Y. J., Johnson, K. G. and Wong, S. S. Cbem. 61,925.

ments, pp. 79-43. Lewis Publisher, Chelsa, MI. (1993). J. Anal. Toxicol. 17(4), 215.


Recommended