ANALYTICAL

BIOCHEMISTRY

202,256-261

(1992)

An Assay for Cocaethylene and Other Cocaine Metabolites in Liver Using High-Performance Liquid Chromatography Stephen M. Roberts,*>? John W. Munson,* Robert C. James,* and Raymond D. Harbison*vt *Center for Environmental & Human Toxicology, and TDepartments of Physiological Sciences and Pharmacology Therapeutics, J. Hillis Miller Health Science Center, University of Florida, Gainesville, Florida 32615

&

Received June 11,199l

Cocaethylene (benzoylecgonine ethyl ester or ethyl cocaine) is a transesterification product of cocaine and ethanol that has been observed in the urine of individuals using these drugs in combination. There is evidence that cocaethylene is pharmacologically active, and its formation in vivo may contribute to the toxicity of cocaine. A new method is presented here which enables the quantification of cocaethylene and cocaine, as well as the cocaine metabolites benzoylecgonine and norcocaine in liver tissue. This method utilizes highperformance liquid chromatography with uv detection (235 nm), and the propyl ester of cocaine is used as an internal standard. Liver homogenates are first buffered with 0.1 N dibasic potassium phosphate (pH 9.1) and then extracted with methylene chloride:isopropanol (9: 1). Extraction efficiencies were approximately 7545% for the compounds of interest. The coefficient of variation for replicate determinations (IV = 10) of cocaethylene concentration was 5.75%, with comparable values obtained for cocaine, norcocaine, and benzoylecgonine. The detection limit for cocaethylene, based on a peak height threefold greater than background noise, was approximately 1.7 ng of injected compound. Using this method, it was demonstrated that cocaethylene is present in mouse liver following cocaine and ethanol administration, with an apparent rapid rate of formation and elimination. o 1992 Academic Press,

Inc.

Cocaine, in the presence of ethanol, can undergo a transesterification reaction producing a derivative known as cocaethylene (benzoylecgonine ethyl ester or ethyl cocaine). Evidence that cocaethylene might be a

metabolite of cocaine was first reported by Rafla and Epstein (l), who found measurable levels in urine obtained from three individuals who had used a combination of cocaine and ethanol. Smith (2) also reported the presence of cocaethylene, as well as arylhydroxyand arylhydroxymethoxycocaethylene derivatives, in the urine of subjects using cocaine and ethanol together. The conditions required for cocaethylene formation in uiuo, other than the presence of cocaine and ethanol, have not been well characterized. The transesterification reaction is catalyzed by human liver homogenates (3), but the role and identity of hepatic enzymes in this process have not been defined. It appears that under favorable circumstances the extent of formation of cocaethylene in uivo can be significant. Hearn et al. (3) examined postmortem blood, liver, and brain samples from seven medical examiner cases involving cocaine and ethanol use. Cocaethylene was found in each of the tissues in measurable levels in nearly all cases, and liver cocaethylene concentrations typically exceeded those of cocaine. The formation of cocaethylene may be important in that it appears to possess pharmacologic activity. It is more potent than cocaine in producing acute lethality in mice (4), and has affinity for dopamine transporter sites in human brain comparable to that of cocaine (3). In order to facilitate the study of the hepatic formation of this potentially toxic metabolite of cocaine, there was a need for a simple, yet sensitive and selective assay in liver tissue. An HPLC method for measuring cocaethylene and cocaine in plasma has been previously reported (5), with sensitivity comparable to that achieved in assays using gas chromatography with nitrogen detection. However, the applicability of this assay to tissue samples was uncertain, and it was desirable to be able to

256 All

Copyright 0 1992 rights of reproduction

0003~2697/92 $3.00 by Academic Press, Inc. in any form reserved.

CHROMATOGRAPHIC

ASSAY

CHs /

h

cocshe

COOCH3CH3

COOCH3 -a=&

ooc(3

O=c) CW.Mhyhe

\

lH

a=& COOCH3

norcocaine

FIG. 1. lites

Metabolic relationship measured in this assay.

between

OOCQ

cocaine

and the

metabo-

quantitate in the same assay the metabolism of cocaine by other pathways. Consequently, an HPLC assay using uv detection was developed for cocaethylene in liver tissue which included also cocaine, the esteratic metabolite benzoylecgonine, and the N-oxidative metabolite norcocaine (Fig. 1). The inclusion of norcocaine in the assay was particularly valuable from our perspective since it is pharmacologically active (6) and represents the first step in a sequential oxidative pathway leading to reactive metabolite formation from cocaine in mice and humans (7-9).

MATERIALS

AND

METHODS

Cocaethylene fumarate and norcocaine were obtained from the National Institute on Drug Abuse (Rockville, MD), and cocaine propyl ester was purchased from Research Biochemicals Inc. (Natick, MA). Cocaine hydrochloride, benzoylecgonine, and ecgonine methyl ester were purchased from Sigma Chemical Co. (St. Louis, MO). N-Hydroxynorcocaine was synthesized as described by Thompson et al. (lo), with the structure verified by mass spectrometry. Acetonitrile (Optima grade) and other laboratory chemicals were obtained from Fisher Scientific (Orlando, FL). Instrumentation consisted of a Spectra-Physics (San Jose, CA) Isochrom pump and Spectra 100 variable wavelength uv detector, a Rheodyne (Cotati, CA) Model 7125 injector, and a Hewlett-Packard (Palo Alto, CA) Model 3396A integrating recorder. Chromatographic separation was accomplished using a deactivated re-

FOR

COCAINE

METABOLITES

257

verse-phase (octadecyldimethylsilyl) column, 15 cm X 4.6 mm, 5 pm packing (Suplex pKb-100, Supelco, Bellefonte, PA). The mobile phase was acetonitri1e:O.l N monobasic potassium phosphate aqueous buffer, pH 5.5, 17:83, at a flow rate of 2.0 ml/min. Effluent was monitored at 235 nm. Cocaine and its metabolites were extracted from liver tissue samples prior to analysis. Liver samples (about 2 g) were weighed and homogenized in 5 ml of ice-cold saline using a motorized tissue grinder (Tekmar, Cincinnati, OH). A l.O-ml aliquot of tissue homogenate was added to 10 ml of methylene chloride:isopropanol (90:10), along with 2.0 ml of 0.1 N dibasie potassium phosphate (pH 9.1) and 50 ~1 of cocaine propyl ester (20 pg/ml in water) as an internal standard. After vortex mixing for 30 s, the extraction mixture was allowed to stand for 5 min and then vortex mixed for an additional 30 s. Phases were separated by centrifugation at 2000g for 5 min, and 5 ml of the organic (bottom) phase was then removed and evaporated to dryness under a stream of nitrogen. All extraction steps were conducted at 4°C. The sample was redissolved in 0.5 ml of mobile phase, filtered using a 0.22-pm centrifugal filtration unit (Bioanalytical Systems, West Lafayette, IN), and injected onto the HPLC column using a 50-~1 sample loop. Freshly prepared standard solutions of cocaine, cocaethylene, benzoylecgonine, and norcocaine were used to generate standard curves in liver homogenate from untreated mice. In order to evaluate the effectiveness of this assay in measuring hepatic concentrations of cocaethylene formed in uiuo, cocaine and ethanol were administered to ICR male mice. Mice were obtained from Harlan Sprague-Dawley (Indianapolis, IN) with a body weight of 20-24 g. The animals were housed in polycarbonate and stainless steel cages on corn cob bedding with free access to food (Purina 5001, Ralston Mills, St. Louis, MO) and water. Ethanol (3 g/kg, i.p.) was administered 1 h before a dose of cocaine (50 mg/kg, i.p.). Mice were euthanized by carbon dioxide asphyxiation at varying intervals following the cocaine dose and the livers removed for analysis. Assays were performed immediately following harvest of the tissues. RESULTS

AND

DISCUSSION

Assay characterization. Under the isocratic conditions of the assay, satisfactory resolution was obtained for cocaethylene, cocaine, benzoylecgonine, norcocaine, and the internal standard cocaine propyl ester. Using liver homogenates spiked with analytical standards, retention times for compounds of interest were, in order of elution: benzoylecgonine, 2.0 min; cocaine, 3.2 min; norcocaine, 3.9 min; cocaethylene, 5.4 min; and cocaine propyl ester, 11.0 min. Figure 2 shows chromatograms

258

ROBERTS -

ET

AL.

Cocaine

-

Benzoylecgonine

Benzoylecgonine /

/

I

s Cocaethylene

Cocaine

f /

Cocaine

Propyl

Ester

Norcocaine

Cocaine

15.0 min

16.0 min

Ester

I

1

FIG. 2. Representative chromatograms of liver samples from cocaineand ethanol-treated mice. Complete ethanol (3 g/kg, i.p.) administered 1 h before cocaine (50 mg/kg, i.p.). The sample shown was taken 15 min after illustrated are chromatograms taken at the same relative time from mice treated with the same dose of cocaine and saline-treated controls (D).

obtained from mice treated with saline, cocaine alone, ethanol alone, or cocaine plus ethanol. Only the mouse treated with cocaine plus ethanol had a chromatographic peak with the retention time of cocaethylene, consistent with the reported requirement of both cocaine and ethanol for its formation. An unidentified endogenous compound with a retention time of 6.6 min was present in each chromatogram, but was distinctly separated from the cocaethylene and cocaine propyl ester eluting before and after it, respectively. No other peaks which might interfere with quantitation were observed in chromatograms from either saline- or ethanol-treated controls. Metabolism of cocaine in uiuo might be expected to also generate significant levels of the esteratic metabolite ecgonine methyl ester. However, ecgonine methyl ester was not detected under these chromatographic conditions using concentrations up to 50 pg/ml. Its poor absorbance at the wavelength used for detection (235

Propyl

treatment (A) consisted of cocaine administration. Also alone (B), ethanol alone (C),

nm), and perhaps the chromatography conditions, prevented it from appearing within (or interfering with) the chromatogram. Using the specified mobile phase, the ZV-hydroxy metabolite of norcocaine eluted with a relatively long retention time (ca. 25 min) and poor peak shape. This metabolite could be quantitated by increasing the acetonitrile content of the mobile phase to 20%, resulting in a retention time of 16 min and much improved peak symmetry. Under these modified conditions, correspondingly shorter retention times were observed for cocaine (2.2 min), norcocaine (2.7 min), cocaethylene (3.3 min), and the internal standard (5.9 min), with adequate resolution maintained. However, the rapidly eluting benzoylecgonine (1.6-min retention time) was poorly separated from the solvent front under these conditions. To evaluate reproducibility and precision of the assay, standards for each of the compounds were added to aliquots of liver homogenate from untreated mice to a

CHROMATOGRAPHIC

ASSAY

FOR

COCAINE

259

METABOLITES

C-

I I > , , ,

Cocaine

Propyl

Ester Cocaine

/

-

-

15.0 min

FIG.

Precision,

Variability,

Extraction

Assay precision” h&d)

Compound Cocaethylene Cocaine Norcacaine Benzoylecgonine Cocaine propyle

and

0.98 1.08 1.01 1.05 ester

+ f * + -

0.06 0.06 0.05 0.09

1 Efficiency Coefficient of variation* W) 5.8 5.2 4.8 8.2 -

of the

Assay

Extraction efficiency’ (%) 83.6 78.1 72.4 85.4 74.2

I

Z-Continued

concentration of 1 pg/ml. Ten of these aliquots were then assayed. The resulting measurements of assay precision and coefficient of variation appear in Table 1. Extraction efficiencies from liver were determined by TABLE

Ester

LJLL 15.0 min

I

Propyl

+ -c + + f

3.1 3.4 1.6 5.9 2.3

a Measured values in liver homogenate containing standards with concentrations set at 1.0 pglml, mean + SD, N = 10. b Based upon the mean and SD of IO replicate samples. ’ Mean * SD, N = 5, for concentrations ranging from l-5 @g/ml. Extraction efficiencies were based upon comparison of absolute peak height values from spiked liver homogenates following extraction versus direct injection of appropriately diluted standards.

adding cocaine and metabolite standards to l-ml liver homogenate aliquots to achieve concentrations ranging from 1 to 5 pg/ml. So that the extraction efficiency of the internal standard could also be evaluated, peak heights were measured in absolute terms rather than relative to the internal standard. Individual peak heights from the spiked samples following the extraction procedure were compared with those obtained from direct injection of standards prepared in appropriate volumes of water. There was no evidence of a trend of increasing or decreasing extraction efficiencies over the limited range of concentrations tested for any of the compounds (data not shown), and the results were consequently averaged (Table 1). Relatively consistent extraction efficiencies were obtained among the various cocaine metabolites and internal standard. The limit of sensitivity for detection was defined as the smallest amount injected that yielded a peak that was threefold greater than the background noise. The limits of detection (ng) and minimum concentrations detectable in tissue samples (rig/g tissue) were 1.71 and 111 for cocaethylene, 1.45 and 94 for cocaine, 1.46 and 95 for norcocaine, and 1.12 and 71 for benzoylecgonine.

260

ROBERTS 4

3

a

A

benzoylecgonine

0

cocaine

0

norcocaine

l

cocaethylene

2

4

6

Concentration (&g liver) FIG. 3.

Standard curves for cocaethylene, cocaine, norcocaine, and benzoylecgonine. Peak height ratio versus concentration relationships were established using varying concentrations of standards in liver homogenate from untreated mice. Peak height ratio is defined as the ratio of the height of the cocaine or metabolite peak to that of the internal standard, cocaine propyl ester. Linear regression analysis yielded the lines shown (F’ 3 0.997).

Standard curves derived in liver homogenate are shown in Fig. 3. The relationship between concentration and peak height ratio was linear (r2 2 0.997) for each compound over the range of interest, in this case up to 6 pgfg liver. At the wavelength for detection (235 nm), the molar extinction coefficient ratios, relative to the cocaine propyl ester internal standard, were 1.75,1.87,1.20, and 1.21 for benzoylecgonine, cocaine, norcocaine, and cocathylene, respectively. Measurement of cocaethylene and other cocaine metabolites in mice treated with cocaine ano! ethanol. While

cocaethylene has been reported in the urine and tissues of humans using cocaine and ethanol, we could find no reports of cocaethylene formation in an experimental animal such as the mouse. As an initial step in evaluating the usefulness of the mouse as a model for examining potential toxicity from cocaethylene formed in uiuo, evidence for its formation was sought in liver tissues from mice treated with cocaine (50 mg/kg, i.p.) and ethanol (3 g/kg, i.p. as a 1 h pretreatment). As shown in Fig. 4, cocaethylene was found in measurable quantities in these animals. Cocaethylene concentrations were greatest 15 min after the cocaine dose and declined to nondetectable levels after 90 min. Although the number of sampling points was limited, the decline in cocaethylene concentrations from the peak appeared to be monoexponential with a slope corresponding to an apparent half-life of approximately 16 min. These observations

ET

AL.

suggest a rapid formation of cocaethylene from cocaine and ethanol in uivo and also a rapid rate of elimination -perhaps through further metabolism. Unlike the postmortem observations of Hearn and co-workers (3) in humans, hepatic cocaethylene concentrations in these mice did not exceed the cocaine concentrations. Strict comparisons of cocaethylene to cocaine concentration ratios between these results and the limited observations in humans are probably not meaningful, however, since they could conceivably be influenced by the relative doses of cocaine and ethanol, as well as the time following dose at which the concentrations are measured. The formation of cocaethylene from cocaine and ethanol in situ has interesting toxicologic implications. Activity of cocaethylene in the central nervous system could contribute to exacerbated acute toxicity of cocaine when used in combination with ethanol (3). Depending upon cocaethylene’s effects on the liver, it might also cause or contribute to the increased hepatotoxicity from cocaine that has been observed with either acute or subacute pretreatment of mice with ethanol (11,12). Though our focus in developing an HPLC assay for cocaethylene was the measurement of concentrations in the liver, the assay could perhaps be adapted to measure cocaethylene concentrations in other potential target tissues such as the brain, heart, etc. As such, this assay may be a useful tool in exploring possible roles for cocaethylene in the toxicity of cocaine.

25

c g! ‘0) t3J .-6 H E

20

15

10

z 0” 5

0 30

45

64

75

90

Time (min.) FIG. 4. Time course of hepatic concentrations cocaethylene (O), cocaine (0), norcocaine (0) and benzoylecgonine (A) in mice treated with cocaine and ethanol. Results are expressed as means + SD, N = 3. Inset is cocaethylene concentrations on an expanded concentration scale.

CHROMATOGRAPHIC

ASSAY

ACKNOWLEDGMENT This Institute

research was supported in part on Drug Abuse (DA-06601).

by a grant

from

the National

REFERENCES 1. Rafla, 2. Smith,

F. K., and Epstein, R. M. (1984)

R. L. (1979)

J. Anal.

Toxicol.

J. And.

ToricoE.

3. Hearn, W. L., Flynn, D. D., Hime, G. W., Rose, Mantero-Atienza, E., Wetli, C. V., and Mash, Neurochem. 56,698-701. 4. Hearn, (1991) 5. Jatlow,

3,59-63.

8, 38-42. S., Cofino, J. C., D. C. (1991) J.

W. L., Rose, S., Wagner, J., Ciarleglio, A., and Mash, Phnrmucol. Biochem. Behau. 39,531-533. P., and Nadim,

H. (1990)

Clin.

Chem.

36,

1436-1439.

D. C.

FOR

COCAINE

261

METABOLITES

6. Hawks, R. L., Kopin, I. J., Colburn, R. W., and Thoa, N. B. (1975) Life Sci. 15,2189-2195. 7. Kloss, M. W., Rosen, G. M., and Rauckman, E. J. (1984) Biochem. Pharmacol. 33,169-173. 8. Shuster, L., Garhart, C. A., Powers, J., Grunfeld, Y., and Kanel, G. (1988) in Mechanisms of Cocaine Abuse and Toxicity (Clouet, D., Asghar, K., and Brown, R., Eds.), pp. 250-275, NIDA Research Monograph 88. 9. Roberts, Dispos. 10. Thompson, Pharmacol.

S. M., Harbison, 19,1046-1051. M. L., Shuster, 28,2389-2395.

R. D., and James, L., and

Shaw,

R. C. Drug K. (1979)

11. Smith, A. C., Freeman, R. W., and Harbison, them. Pharmacol. 30,453-458. 12. Boyer, C. S., and Petersen, 14,28-31.

D. R. (1990)

Alcohol.

Metab. Rio&em.

R. D. (1981) Clin.

Exp.

BioRes.