Jacob Bell
New Member
Richard Rodgers, Catherine P. Crowl, Wendy M. Eimstad, Mae W. Hu, Jean K. Kam, Robert C. Ronald,1
Gerald 1. Rowley, and Edwin F. Uliman
We describe a homogeneous enzyme immunoassay for
measurement of cannabinoid metabolites, as well as
ı9-tetrahydrocannabinol (I) in urine. Malate dehydrogenase
from pig heart mitochondria was labeled with a derivative
of I. The compound used to calibrate the assay was the I
metabolite, 1 1-nor-ı9-tetrahydrocannabinol-9-carboxylic
acid (II). With 15 ıg of II per literof urine as the cutoff
concentration, the assay can detect 25 ıtg of II per literwith
>95% confidence. A positive response was obtained for
urine specimens assayed within 30 mm after exposure to
cannabinoids. However, the persistence of metabolites
of I in urine indicates that assay of this fluid is useful as an
indicator of cannabinoid use but not as an indicator of intoxication.
Additional Keyphrases: marijuana drug abuse - toxicology
. screening
Several years ago a homogeneous enzyme immunoassay
was described (1) that made use of the observation
that certain enzymes, when covalently bound to
haptens, could be inhibited by antibodies directed
against the hapten. This technique, called EMIT,2 has
been used clinically with three different enzymes, lysozyme
(EC 3.2.1.17) (2), m-MDH3 (3, 4), and glucose-
6-phosphate dehydrogenase (EC 1.1.1.49) (5, 6), to assay
various compounds, including opiates, thyroxine,
anti-epileptic drugs, and digoxin.
Recently, using pig heart m-MDH conjugated with
a derivative of ı9-THC, we demonstrated that this
method was applicable to the assay of cannabinoids in
water. The antibodies we used were raised in sheep alter
injection of a bovine 'y-globulin conjugate of the same
ı9-THC derivative (7). We report here on the usefulness
of this procedure in determination of cannabinoids in
urine.
Materials and Methods
Instruments
Spectrophotometer: For the assay we used a Stasar
III spectrophotomer (Gilford Instrument Labs., Oberlin,
Ohio 44074) equipped with a 30 #{176t}hCermally regulated
(±0.05 #{176}Cm)icroflow cell. For this assay, the
instrument was calibrated so that the readout from the
instrument was equal to twice the absorbance of the
sample. The enzyme reaction was monitored by the
change in absorbance at 340 nm.
Printer: We used a Syva Timer Printer (Model 2400)
for these studies. The instrument interfaces with the
spectrometer and is programmed to assign a sample
number, provide a printed record of the initial absorbance
reading of the sample, and print the change in
absorbance as a function of time. The instrument controls
the timing of the assay, and was set to provide a
13-s delay for temperature equilibration after the
sample is aspirated into the flow cell, followed by two
absorbance readings timed 60 s apart.
Pipetter-diluter: A Syva pipetter-diluter (Model
1500) was used. The instrument was set to aspirate 50
ı of sample and to deliver the sample plus 250 ıil of
buffer.
Reagents
Antibodies. Preparation of the ı9-THC derivative
and its m-MDH conjugate have been described previously
(7). An aldehyde derivative of zı9-THC, 11-
oxo-ı8-tetrahydrocannabinol (3), was prepared according to the scheme illustrated below:
Lithium p -chlorophenylsulfinylchloromethide was
prepared by slowly adding 7.5 ml of a 2 mol/liter solution
of n-butyllithium in hexane to 3.15 g of p-chlorophenylchloromethylsulfoxide
in 70 ml of dry tetrahydrofuran
at -89 #{176(}8C). A solution of 0.34 equivalents
of 1 in 30 ml of tetrahydrofuran, at -78 #{176}wCa,s added
under N2, with stirring, during 60 mm. The reaction
mixture was then quenched with 50 ml of 2 mol/liter
NaOH and allowed to warm to ambient temperature.
The mixture was diluted with water, neutralized with
dilute HC1, and extracted three times with diethyl ether.
The combined extracts were concentrated under reduced
pressure to yield an oily product. Chromatography
on silica gel-PF with 1/9 (by vol) ether/dichloromethane
yielded a mixture of stereoisomers of 2 in a 59%
yield, which was used directly in the next reaction.
A solution of 506 mg of 2 in 5 ml of xylene was heated
for 4.5 h at 170 #{176(}9C), then cooled, diluted with 5 g/liter
aqueous NaHCO3, and extracted twice with ether. The
combined extracts were concentrated under reduced
pressure to yield a yellow oil. Thin-layer chromatography,
first with 4/6 (by vol) ether/petroleum ether
(boiling range 35-60 #{176}aCn)d then with 0.5/9.5 (by vol)
ether/chloroform, separated aldehydes 3 and 4. The
structure of each was confirmed by the chemical shifts
of the vinyl protons [H8 in the ı8 isomer (#{2=44}6.9) and
H10 in the ı isomer (#{2=44}7.9)] in nuclear magnetic resonance.
The positions of these protons in ıX8- and ı-
THC are 5.45 and 6.33, respectively (10).
Compound 3 was conjugated to bovine serum albumin
by slowly adding 28 mg of 3 in 4 ml of methanol to 110 mg of bovine serum albumin and 10 mg of sodium
cyanoborohydride in 15 ml of 50 mmol/liter phosphate
buffer, pH 7.0, at 4 #{176}ACf.ter stirring at 4 #{176f}oCr four
days and at room temperature for two days, centrifugation
at 17 500 X g for 20 mm yielded a clear solution.
Dialysis against de-ionized water, centrifugation, and
lyophilization yielded 93 mg of protein conjugate, estimated
from the ultraviolet spectrum to contain four to
five hapten molecules per molecule of albumin.
Sheep were immunized every two weeks with an
emulsion of 0.5 mg of the conjugate, saline, and complete
Freund's adjuvant. For the third and subsequent
immunizations we used incomplete Freund's adjuvant.
Animals were bled every two months.
The antisera were precipitated by half-saturation
with (NH4)2S04, resuspended in 55 mmol/liter
tris(hydroxymethyl)aminomethane, pH 8.1, containing
100 mg of NaN3 per liter, and dialyzed against the same
buffer.
The "antibody reagent" was prepared by diluting
this material 38.5-fold with glycine buffer (153 mmol/
liter, pH 5.0) containing 0.1 mol of NADı per liter. This
reagent, lyophilized, is stable for at least one year.
The "enzyme reagent" was prepared by diluting the
stock conjugate (1.15 g/liter in 0.5 mol/liter phosphate,
containing 100 mg of disodium ethylenediaminetetraacetate
per liter, pH 7.6) 163-fold with a mixture of
phosphate buffer, pH 7.4, 0.5 mol/liter, and containing,
per liter, 100 mg of disodium ethylenediaminetetraacetate,
100 mg of NaN3, and glycerol (300 mI/liter).
This reagent is stable for about six months.
The assay buffer (pH 9.75) contained, per liter, 0.1
mol of glycine, 375 mmol of phosphate, 143 mmol of
L-malate, 100 mg of disodium ethylenediaminetetraacetate,
and 100 mg of NaNı3.
Calibration Materials
Urine was collected from 60 reliable individuals who
responded to a request for donors who had not been
exposed to cannabinoids for six months. The urine was
pooled and filtered.
Initial attempts to prepare urine calibration material
with ı9-THC were unsuccessful, because this compound
quickly adsorbs onto the walls of either glass or plastic
containers. A urine pooi containing an added 25 ısg of
ıX9-THC per liter lost about 80% of the drug when stored
in glass at 4 #{176f}oCr 24 h. Similar behavior was previously
observed by Garrett and Hunt (11). For this reason the
more soluble THC-9-acid (>95% pure, obtained from
Research Triangle Institute, N. C.) was used as an assay
calibrator. This material cross reacts well with the antibodies
(7) and is a major urinary metabolite of ı-
THC (12). Samples of a negative urine pooi with this
compound added were stable at 4 #{176f}oCr 24 h. Aliquots
could be frozen and lyophilized for long-term storage.
The assay response of reconstituted lyophilized calibrators
containing 15, 25, and 75 ısg of THC-9-acid per
liter compared favorably to liquid, nonlyophilized
samples. However, the enzyme activity obtained with
the negative urine pool significantly increased after
lyophilization. The reason for this increase was not
identified, but it was only apparent in negative urine
pools. Because it was essential that the reconstituted
negative calibrator give an assay response identical to
that of an unlyophilized negative pool, a "synthetic
urine" was prepared for use as a negative calibrator.
This solution (pH 6.1) contained, per liter, 22 g of urea,
1.1 g of Na2HPO4, 1.4 g of NaH2PO4, 8.25 g of NaC1, 5.2
g of KC1, and 1.5 g of creatinine, and yielded the same
assay response after lyophilization as did an unlyophilized
pool of zı9-THC-negative urine.
Test Procedure
Add a 50-ısl urine sample and 250 ısl of assay buffer
to a 10 X 75 mm disposable glass test tube with the pipetter-
diluter. Then add 50 ıil of antibody reagent,
followed by 50 ıil of enzyme reagent, each with 250 ıl of
buffer, to give a total assay volume of 900 ısl (pH 9.5).
Immediately after the last addition, vortex-mix the
assay mixture for 2 or 3 s while purging the spectrophotometer
cell with air, and promptly aspirate the
reaction mixture into the instrument. This action automatically
activates the timer-printer.
After the 13-s delay for thermal equilibration, an
initial reading is printed and 60 s later the difference
between the initial and final readings is printed. Determine
the amount of drug in the sample by reference
to a standard curve prepared by plotting the calibrator
readings vs. concentrations (Figure 1).
About 1% of the urine specimens will give initial
readings >2.7 (an absorbance of 1.35) and cannot be
accurately analyzed by this procedure, because the
readings are no longer in the range of linear instrumental
response.
Results
Analytical Variables
Specificity. The antibodies used in this assay were
specifically intended to permit detection of various
ı9-THC congeners and metabolites. Figure 1 shows the
response of the assay to five different cannabinoids. The
assay is most sensitive to THC-9-acid and to 11 -hydroxy-
ı9-THC; the response to cannabinol and to
ı9-THC itself is about 30% less. By contrast, cannabidiol
was much less cross reactive, 47-fold as much being
required to give the same response as did THC-9-acid.
Thus an intact B ring (the heterocyclic ring) appears to
be important for good antibody binding.
We examined the cross reactivity of the assay to
various natural hormones, drugs, and their metabolites.
Each compound was added to synthetic urine to give a
concentration of either 1000 mg/liter or the highest
concentration at which it was soluble. None of the
compounds cross reacted significantly, including steroid
hormones and cholesterol, which were tested at concentrations
much higher than amounts normally
present in urine (Table 1).
Sensitivity and interpretation of data. To monitor
for the presence or absence of cannabinoids in urine, it
is necessary to select a specific minimum "cutoff"
reading above which samples will be identified as positive.
For the selection of appropriate cutoff values,
urine was collected from 60 individuals who responded
to a request for donors who had had no exposure to
cannabinoids for the previous six months. Portions of
these samples were supplemented with 25 and 45 ıig of
THC-9-acid per liter. The former gave a mean assayed
value of 25.5 ± 4.3 ızg/liter, the latter a mean value of
45.3 ± 5.2 ıgfliter. The average observed with the negative
samples was nearly the same as the negative calibrator
rate, and all but one negative sample gave assay
values of less than 15 ısg/liter. These results, displayed
in the form of a histogram in Figure 2, illustrate the
ability of the assay to discriminate among these concentrations.
Some of the scatter was found to be associated
with endogeneous m-MDH activity in the samples,
but the data suggest that a 15 ıtg/liter cutoff value
provides a practical distinction between positive and
negative samples without correction for endogenous
enzyme activity. This cutoff is expected to produce <5%
false positives and to assure that >95% of all samples
containing 25 ıtg/liter will be correctly identified as
positive. By using 25 ısg/liter as a cutoff, false positives
are virtually eliminated, while providing a >95% probability
of detection of samples containing at least 45
zgfliter.
Precision. The precision of the cannabinoid enzyme
immunoasasy was determined by supplementing pooled
normal urine with 15, 25, and 75 ısg of THC-9-acid per
liter. Each portion was then analyzed repetitively (20
times) by a single operator. The results (Table 2) show
high precision at all three concentrations.
Between-sample precision of the assay, derived from
the data given in Figure 2, is summarized in Table 3.
Variations in urine composition, particularly variations
in the amount of m-MDH activity, contribute significantly
to the coefficients of variation.
Clinical Results
"Blind" clinical studies were done on urine samples
provided by Dr. L. E. Hollister and S. Kanter (Veterans
Administration Hospital, Palo Alto) and by Dr. K.
Dubowski (University of Oklahoma Health Sciences
Center). The analyst was provided only an identifying
number with each sample and was unaware of the
sample history.
The subjects tested at the V.A. Hospital, Palo Alto,
were administered, orally, 30-mg ı9-THC, 100 mg of
cannabinol, or 100 mg of cannabidiol. Urine samples
were obtained before the dose and at various intervals
afterward. The results are presented in Table 4. Urine
from all of the subjects who received zı9-THC or cannabinol
was strongly positive for one or two days after
the dose, but urine samples taken after the oral dose of
cannabidiol tested negative (<15 ıg of THC-9-acid
equivalents per liter) by our assay, a result that was to
be expected because 470 ısg of cannabidiol per liter is
required to give the same response as is obtained with
15 ıtg of THC-9-acid per liter. One subject (VA-I)
showed the presence of cannabinoids before doses of
zı9-THC and cannabidiol. This subject reported frequent
use of cannabinoids before the experiment and
appears to be continually excreting metabolites. In
addition to these positive samples, 12 negative urine
samples were included and were correctly identified as
negative.
Because the volume of urine produced changes during
the day, the assay results are also reported as micrograms
of THC-9-acid equivalents per gram of creatinine
(creatinine values for the samples were kindly provided
by S. Kanter). The adjusted data show a maximum for
metabolites excreted at about 6 h after the dose, followed
by a very slow decline in values for THC-9-acid
equivalents.
A set of 198 urine specimens from the University of
Oklahoma Health Sciences Center were from subjects
who had taken part in a controlled study of the effects
of smoking cigarettes containing zı9-THC. The samples
were analyzed both in our laboratory and at the University
of Oklahoma, using similarly prepared reagent.
Table 5 compares our results with those obtained at the
University of Oklahoma (K. M. Dubowski, personal
communication, May 2, 1977); there is 95.5% agreement
between results. Of the nine discrepant samples, six
were borderline positives (15-19 ıg of THC-9-acid
equivalents per liter). For the remaining three samples
the results were negative at Syva but were reported by
the University of Oklahoma to contain 24-28 ıig of
THC-9-acid equivalents per liter.
A Pearson correlation coefficient of 0.97 1 was calculated
calculated
for the values of the 108 samples that both laboratories
reported as positive. Despite this good correlation,
a positive bias of 5.4 ızg/liter in the University of
Oklahoma results was shown by the position of the intercept.
This result appears to be attributable to a decrease
in the concentration of the calibrators used at the
University of Oklahoma; their reagents were stored for
several months before use and this particular preparation
was subsequently found to have limited stability.
More detailed results of this study are to be published
(K. M. Dubowski, manuscript in preparation).
In a third clinical study (Figure 3), urine samples were
collected from reliable volunteers at various times after
each had smoked a single marijuana cigarette of unknown
origin and cannabinoid content. Subject A, a
moderately frequent user of cannabinoids, showed a
higher cannabinoid excretion before smoking than the
peak values of the other three subjects. Subjects B and
C reached higher peak cannabinoid values than did D
and also peaked earlier (2 h vs. 6 h) after smoking. These
differences probably represent differences in absorption
among the various individuals as well as differences in
potency of the marijuana cigarettes. The values were not
corrected for changes in urinary volume.
Discussion
The homogeneous enzyme immunoassay technique
(2-6) as applied to cannabinoids combines rapid measurement
with ease of sample manipulation.
Our results with the clinical specimens show that the
peak excretion of cannabinoids occurs from 2-6 h after
exposure. It is particularly important to note that urinary
cannabinoid excretion can remain high for more
than 24 h. It was also observed that frequent users of
this drug (several exposures per week) have basal values
for metabolites in their urine that exceed the peak
values attained by relatively infrequent users. These
data, and the fact that the period of intoxication lasts
only from 1 to 4 h (13), indicate that assay of zı9-THC
metabolites in urine is useful only as an indicator of the
use of cannabinoids, not as a measure of intoxication.
Quite possibly, concentrations of zı9-THC and its metabolites
in serum or saliva may more reliably indicate
intoxication.
We are grateful to Toni Armstrong and Bernard Sheldon for fine
technical assistance. We also thank the National Institute on Drug
Abuse for financial support of a portion of this study.
References
1. Rubenstein, K. E., Schneider, R. S., and Uliman, E. F., "Homogeneous"
enzyme immunoassay. A new immunochemical technique.
Biochern. Biophys. Res. Commun. 47,846 (1972).
2. Schneider, R. S., Lindquist, P., Wong, E. T., et al., Homogeneous
enzyme immunoassay for opiates in urine. Clin. Chem. 19, 821
(1973).
3. Rowley, G. L., Rubenstein, K. E., Huisjen, J., and Ullman, E. F.,
Mechanism by which antibodies inhibit hapten-malate dehydrogenase
conjugate. An enzyme immunoassay for morphine. J. Biol. Chem.
250, 3759 (1975).
4. Ullman, E. F., Blakemore, J., Leute, R. K., et al. Homogeneous
enzyme immunoassay for thyroxine. Gun. Chem. 21, 1011 (1975).
5. Chang, J. J., Crowl, C. P., and Schneider, R. S., Homogeneous
enzyme immunoassay for digoxin. Clin. Chem. 21, 967 (1975).
6. Schneider, R. S., Bastiani, R. J., Rubenstein, K. E., et al., Homogeneous
enzyme immunoassay for diphenylhydantoin and phenobarbital
in serum. Gun. Chem. 20, 869 (1974).
7. Rowley, G. L., Armstrong, T. A., Crowl, C. P., et al., Determination
of THC and its metabolites by EMITS homogeneous enzyme immunoassay:
A summary report. In Cannabinoid Assays in Humans,
NIDA Research Monograph 7, R. E. Willette, Ed., 1976, p 28.
8. Durst, T., Stereospecific hydroxalkylation of chloromethylphenylsulfoxide.
J. Am. Chem. Soc. 91, 1034 (1969).
9. Durst, T., and Tin, K. C., Thermal and acid-catalyzed rearrangements
of a-epoxy sulfoxides and sulfones. Tetrahedron Lett., 2369
(1970).
10. Archer, R. A., Boyd, D. B., Demarco, P. V., et al., Structural
studies of cannabinoids. A theoretical and proton magnetic resonance
analysis. J. Am. Chem. Soc. 92,5200 (1970).
II. Garrett, E. R., and Hunt, C. A., Physicochemical properties,
solubility, and protein binding of ı19-tetrahydrocannabinol. J. Pharm.
Sci. 63, 1056 (1974).
12. Lemberger, L., Axelrod, J., and Kopin, I. J., Metabolism and
disposition of tetrahydrocannabinols in naive subjects and chronic
marijuana users. Ann. N.Y. Acad. Sci. 191,142(1971).
13. Patton, W. D. M., and Pertwee, R. G., The actions of cannabis in
man. In Marijuana: Chemistry, Pharmacology, Metabolism and
Clinical Effects, R. Mechoulam, Ed., Academic Press, New York, N.
Y., 1973.
Source: Homogeneous enzyme immunoassay for cannabinoids in urine
Gerald 1. Rowley, and Edwin F. Uliman
We describe a homogeneous enzyme immunoassay for
measurement of cannabinoid metabolites, as well as
ı9-tetrahydrocannabinol (I) in urine. Malate dehydrogenase
from pig heart mitochondria was labeled with a derivative
of I. The compound used to calibrate the assay was the I
metabolite, 1 1-nor-ı9-tetrahydrocannabinol-9-carboxylic
acid (II). With 15 ıg of II per literof urine as the cutoff
concentration, the assay can detect 25 ıtg of II per literwith
>95% confidence. A positive response was obtained for
urine specimens assayed within 30 mm after exposure to
cannabinoids. However, the persistence of metabolites
of I in urine indicates that assay of this fluid is useful as an
indicator of cannabinoid use but not as an indicator of intoxication.
Additional Keyphrases: marijuana drug abuse - toxicology
. screening
Several years ago a homogeneous enzyme immunoassay
was described (1) that made use of the observation
that certain enzymes, when covalently bound to
haptens, could be inhibited by antibodies directed
against the hapten. This technique, called EMIT,2 has
been used clinically with three different enzymes, lysozyme
(EC 3.2.1.17) (2), m-MDH3 (3, 4), and glucose-
6-phosphate dehydrogenase (EC 1.1.1.49) (5, 6), to assay
various compounds, including opiates, thyroxine,
anti-epileptic drugs, and digoxin.
Recently, using pig heart m-MDH conjugated with
a derivative of ı9-THC, we demonstrated that this
method was applicable to the assay of cannabinoids in
water. The antibodies we used were raised in sheep alter
injection of a bovine 'y-globulin conjugate of the same
ı9-THC derivative (7). We report here on the usefulness
of this procedure in determination of cannabinoids in
urine.
Materials and Methods
Instruments
Spectrophotometer: For the assay we used a Stasar
III spectrophotomer (Gilford Instrument Labs., Oberlin,
Ohio 44074) equipped with a 30 #{176t}hCermally regulated
(±0.05 #{176}Cm)icroflow cell. For this assay, the
instrument was calibrated so that the readout from the
instrument was equal to twice the absorbance of the
sample. The enzyme reaction was monitored by the
change in absorbance at 340 nm.
Printer: We used a Syva Timer Printer (Model 2400)
for these studies. The instrument interfaces with the
spectrometer and is programmed to assign a sample
number, provide a printed record of the initial absorbance
reading of the sample, and print the change in
absorbance as a function of time. The instrument controls
the timing of the assay, and was set to provide a
13-s delay for temperature equilibration after the
sample is aspirated into the flow cell, followed by two
absorbance readings timed 60 s apart.
Pipetter-diluter: A Syva pipetter-diluter (Model
1500) was used. The instrument was set to aspirate 50
ı of sample and to deliver the sample plus 250 ıil of
buffer.
Reagents
Antibodies. Preparation of the ı9-THC derivative
and its m-MDH conjugate have been described previously
(7). An aldehyde derivative of zı9-THC, 11-
oxo-ı8-tetrahydrocannabinol (3), was prepared according to the scheme illustrated below:
Lithium p -chlorophenylsulfinylchloromethide was
prepared by slowly adding 7.5 ml of a 2 mol/liter solution
of n-butyllithium in hexane to 3.15 g of p-chlorophenylchloromethylsulfoxide
in 70 ml of dry tetrahydrofuran
at -89 #{176(}8C). A solution of 0.34 equivalents
of 1 in 30 ml of tetrahydrofuran, at -78 #{176}wCa,s added
under N2, with stirring, during 60 mm. The reaction
mixture was then quenched with 50 ml of 2 mol/liter
NaOH and allowed to warm to ambient temperature.
The mixture was diluted with water, neutralized with
dilute HC1, and extracted three times with diethyl ether.
The combined extracts were concentrated under reduced
pressure to yield an oily product. Chromatography
on silica gel-PF with 1/9 (by vol) ether/dichloromethane
yielded a mixture of stereoisomers of 2 in a 59%
yield, which was used directly in the next reaction.
A solution of 506 mg of 2 in 5 ml of xylene was heated
for 4.5 h at 170 #{176(}9C), then cooled, diluted with 5 g/liter
aqueous NaHCO3, and extracted twice with ether. The
combined extracts were concentrated under reduced
pressure to yield a yellow oil. Thin-layer chromatography,
first with 4/6 (by vol) ether/petroleum ether
(boiling range 35-60 #{176}aCn)d then with 0.5/9.5 (by vol)
ether/chloroform, separated aldehydes 3 and 4. The
structure of each was confirmed by the chemical shifts
of the vinyl protons [H8 in the ı8 isomer (#{2=44}6.9) and
H10 in the ı isomer (#{2=44}7.9)] in nuclear magnetic resonance.
The positions of these protons in ıX8- and ı-
THC are 5.45 and 6.33, respectively (10).
Compound 3 was conjugated to bovine serum albumin
by slowly adding 28 mg of 3 in 4 ml of methanol to 110 mg of bovine serum albumin and 10 mg of sodium
cyanoborohydride in 15 ml of 50 mmol/liter phosphate
buffer, pH 7.0, at 4 #{176}ACf.ter stirring at 4 #{176f}oCr four
days and at room temperature for two days, centrifugation
at 17 500 X g for 20 mm yielded a clear solution.
Dialysis against de-ionized water, centrifugation, and
lyophilization yielded 93 mg of protein conjugate, estimated
from the ultraviolet spectrum to contain four to
five hapten molecules per molecule of albumin.
Sheep were immunized every two weeks with an
emulsion of 0.5 mg of the conjugate, saline, and complete
Freund's adjuvant. For the third and subsequent
immunizations we used incomplete Freund's adjuvant.
Animals were bled every two months.
The antisera were precipitated by half-saturation
with (NH4)2S04, resuspended in 55 mmol/liter
tris(hydroxymethyl)aminomethane, pH 8.1, containing
100 mg of NaN3 per liter, and dialyzed against the same
buffer.
The "antibody reagent" was prepared by diluting
this material 38.5-fold with glycine buffer (153 mmol/
liter, pH 5.0) containing 0.1 mol of NADı per liter. This
reagent, lyophilized, is stable for at least one year.
The "enzyme reagent" was prepared by diluting the
stock conjugate (1.15 g/liter in 0.5 mol/liter phosphate,
containing 100 mg of disodium ethylenediaminetetraacetate
per liter, pH 7.6) 163-fold with a mixture of
phosphate buffer, pH 7.4, 0.5 mol/liter, and containing,
per liter, 100 mg of disodium ethylenediaminetetraacetate,
100 mg of NaN3, and glycerol (300 mI/liter).
This reagent is stable for about six months.
The assay buffer (pH 9.75) contained, per liter, 0.1
mol of glycine, 375 mmol of phosphate, 143 mmol of
L-malate, 100 mg of disodium ethylenediaminetetraacetate,
and 100 mg of NaNı3.
Calibration Materials
Urine was collected from 60 reliable individuals who
responded to a request for donors who had not been
exposed to cannabinoids for six months. The urine was
pooled and filtered.
Initial attempts to prepare urine calibration material
with ı9-THC were unsuccessful, because this compound
quickly adsorbs onto the walls of either glass or plastic
containers. A urine pooi containing an added 25 ısg of
ıX9-THC per liter lost about 80% of the drug when stored
in glass at 4 #{176f}oCr 24 h. Similar behavior was previously
observed by Garrett and Hunt (11). For this reason the
more soluble THC-9-acid (>95% pure, obtained from
Research Triangle Institute, N. C.) was used as an assay
calibrator. This material cross reacts well with the antibodies
(7) and is a major urinary metabolite of ı-
THC (12). Samples of a negative urine pooi with this
compound added were stable at 4 #{176f}oCr 24 h. Aliquots
could be frozen and lyophilized for long-term storage.
The assay response of reconstituted lyophilized calibrators
containing 15, 25, and 75 ısg of THC-9-acid per
liter compared favorably to liquid, nonlyophilized
samples. However, the enzyme activity obtained with
the negative urine pool significantly increased after
lyophilization. The reason for this increase was not
identified, but it was only apparent in negative urine
pools. Because it was essential that the reconstituted
negative calibrator give an assay response identical to
that of an unlyophilized negative pool, a "synthetic
urine" was prepared for use as a negative calibrator.
This solution (pH 6.1) contained, per liter, 22 g of urea,
1.1 g of Na2HPO4, 1.4 g of NaH2PO4, 8.25 g of NaC1, 5.2
g of KC1, and 1.5 g of creatinine, and yielded the same
assay response after lyophilization as did an unlyophilized
pool of zı9-THC-negative urine.
Test Procedure
Add a 50-ısl urine sample and 250 ısl of assay buffer
to a 10 X 75 mm disposable glass test tube with the pipetter-
diluter. Then add 50 ıil of antibody reagent,
followed by 50 ıil of enzyme reagent, each with 250 ıl of
buffer, to give a total assay volume of 900 ısl (pH 9.5).
Immediately after the last addition, vortex-mix the
assay mixture for 2 or 3 s while purging the spectrophotometer
cell with air, and promptly aspirate the
reaction mixture into the instrument. This action automatically
activates the timer-printer.
After the 13-s delay for thermal equilibration, an
initial reading is printed and 60 s later the difference
between the initial and final readings is printed. Determine
the amount of drug in the sample by reference
to a standard curve prepared by plotting the calibrator
readings vs. concentrations (Figure 1).
About 1% of the urine specimens will give initial
readings >2.7 (an absorbance of 1.35) and cannot be
accurately analyzed by this procedure, because the
readings are no longer in the range of linear instrumental
response.
Results
Analytical Variables
Specificity. The antibodies used in this assay were
specifically intended to permit detection of various
ı9-THC congeners and metabolites. Figure 1 shows the
response of the assay to five different cannabinoids. The
assay is most sensitive to THC-9-acid and to 11 -hydroxy-
ı9-THC; the response to cannabinol and to
ı9-THC itself is about 30% less. By contrast, cannabidiol
was much less cross reactive, 47-fold as much being
required to give the same response as did THC-9-acid.
Thus an intact B ring (the heterocyclic ring) appears to
be important for good antibody binding.
We examined the cross reactivity of the assay to
various natural hormones, drugs, and their metabolites.
Each compound was added to synthetic urine to give a
concentration of either 1000 mg/liter or the highest
concentration at which it was soluble. None of the
compounds cross reacted significantly, including steroid
hormones and cholesterol, which were tested at concentrations
much higher than amounts normally
present in urine (Table 1).
Sensitivity and interpretation of data. To monitor
for the presence or absence of cannabinoids in urine, it
is necessary to select a specific minimum "cutoff"
reading above which samples will be identified as positive.
For the selection of appropriate cutoff values,
urine was collected from 60 individuals who responded
to a request for donors who had had no exposure to
cannabinoids for the previous six months. Portions of
these samples were supplemented with 25 and 45 ıig of
THC-9-acid per liter. The former gave a mean assayed
value of 25.5 ± 4.3 ızg/liter, the latter a mean value of
45.3 ± 5.2 ıgfliter. The average observed with the negative
samples was nearly the same as the negative calibrator
rate, and all but one negative sample gave assay
values of less than 15 ısg/liter. These results, displayed
in the form of a histogram in Figure 2, illustrate the
ability of the assay to discriminate among these concentrations.
Some of the scatter was found to be associated
with endogeneous m-MDH activity in the samples,
but the data suggest that a 15 ıtg/liter cutoff value
provides a practical distinction between positive and
negative samples without correction for endogenous
enzyme activity. This cutoff is expected to produce <5%
false positives and to assure that >95% of all samples
containing 25 ıtg/liter will be correctly identified as
positive. By using 25 ısg/liter as a cutoff, false positives
are virtually eliminated, while providing a >95% probability
of detection of samples containing at least 45
zgfliter.
Precision. The precision of the cannabinoid enzyme
immunoasasy was determined by supplementing pooled
normal urine with 15, 25, and 75 ısg of THC-9-acid per
liter. Each portion was then analyzed repetitively (20
times) by a single operator. The results (Table 2) show
high precision at all three concentrations.
Between-sample precision of the assay, derived from
the data given in Figure 2, is summarized in Table 3.
Variations in urine composition, particularly variations
in the amount of m-MDH activity, contribute significantly
to the coefficients of variation.
Clinical Results
"Blind" clinical studies were done on urine samples
provided by Dr. L. E. Hollister and S. Kanter (Veterans
Administration Hospital, Palo Alto) and by Dr. K.
Dubowski (University of Oklahoma Health Sciences
Center). The analyst was provided only an identifying
number with each sample and was unaware of the
sample history.
The subjects tested at the V.A. Hospital, Palo Alto,
were administered, orally, 30-mg ı9-THC, 100 mg of
cannabinol, or 100 mg of cannabidiol. Urine samples
were obtained before the dose and at various intervals
afterward. The results are presented in Table 4. Urine
from all of the subjects who received zı9-THC or cannabinol
was strongly positive for one or two days after
the dose, but urine samples taken after the oral dose of
cannabidiol tested negative (<15 ıg of THC-9-acid
equivalents per liter) by our assay, a result that was to
be expected because 470 ısg of cannabidiol per liter is
required to give the same response as is obtained with
15 ıtg of THC-9-acid per liter. One subject (VA-I)
showed the presence of cannabinoids before doses of
zı9-THC and cannabidiol. This subject reported frequent
use of cannabinoids before the experiment and
appears to be continually excreting metabolites. In
addition to these positive samples, 12 negative urine
samples were included and were correctly identified as
negative.
Because the volume of urine produced changes during
the day, the assay results are also reported as micrograms
of THC-9-acid equivalents per gram of creatinine
(creatinine values for the samples were kindly provided
by S. Kanter). The adjusted data show a maximum for
metabolites excreted at about 6 h after the dose, followed
by a very slow decline in values for THC-9-acid
equivalents.
A set of 198 urine specimens from the University of
Oklahoma Health Sciences Center were from subjects
who had taken part in a controlled study of the effects
of smoking cigarettes containing zı9-THC. The samples
were analyzed both in our laboratory and at the University
of Oklahoma, using similarly prepared reagent.
Table 5 compares our results with those obtained at the
University of Oklahoma (K. M. Dubowski, personal
communication, May 2, 1977); there is 95.5% agreement
between results. Of the nine discrepant samples, six
were borderline positives (15-19 ıg of THC-9-acid
equivalents per liter). For the remaining three samples
the results were negative at Syva but were reported by
the University of Oklahoma to contain 24-28 ıig of
THC-9-acid equivalents per liter.
A Pearson correlation coefficient of 0.97 1 was calculated
calculated
for the values of the 108 samples that both laboratories
reported as positive. Despite this good correlation,
a positive bias of 5.4 ızg/liter in the University of
Oklahoma results was shown by the position of the intercept.
This result appears to be attributable to a decrease
in the concentration of the calibrators used at the
University of Oklahoma; their reagents were stored for
several months before use and this particular preparation
was subsequently found to have limited stability.
More detailed results of this study are to be published
(K. M. Dubowski, manuscript in preparation).
In a third clinical study (Figure 3), urine samples were
collected from reliable volunteers at various times after
each had smoked a single marijuana cigarette of unknown
origin and cannabinoid content. Subject A, a
moderately frequent user of cannabinoids, showed a
higher cannabinoid excretion before smoking than the
peak values of the other three subjects. Subjects B and
C reached higher peak cannabinoid values than did D
and also peaked earlier (2 h vs. 6 h) after smoking. These
differences probably represent differences in absorption
among the various individuals as well as differences in
potency of the marijuana cigarettes. The values were not
corrected for changes in urinary volume.
Discussion
The homogeneous enzyme immunoassay technique
(2-6) as applied to cannabinoids combines rapid measurement
with ease of sample manipulation.
Our results with the clinical specimens show that the
peak excretion of cannabinoids occurs from 2-6 h after
exposure. It is particularly important to note that urinary
cannabinoid excretion can remain high for more
than 24 h. It was also observed that frequent users of
this drug (several exposures per week) have basal values
for metabolites in their urine that exceed the peak
values attained by relatively infrequent users. These
data, and the fact that the period of intoxication lasts
only from 1 to 4 h (13), indicate that assay of zı9-THC
metabolites in urine is useful only as an indicator of the
use of cannabinoids, not as a measure of intoxication.
Quite possibly, concentrations of zı9-THC and its metabolites
in serum or saliva may more reliably indicate
intoxication.
We are grateful to Toni Armstrong and Bernard Sheldon for fine
technical assistance. We also thank the National Institute on Drug
Abuse for financial support of a portion of this study.
References
1. Rubenstein, K. E., Schneider, R. S., and Uliman, E. F., "Homogeneous"
enzyme immunoassay. A new immunochemical technique.
Biochern. Biophys. Res. Commun. 47,846 (1972).
2. Schneider, R. S., Lindquist, P., Wong, E. T., et al., Homogeneous
enzyme immunoassay for opiates in urine. Clin. Chem. 19, 821
(1973).
3. Rowley, G. L., Rubenstein, K. E., Huisjen, J., and Ullman, E. F.,
Mechanism by which antibodies inhibit hapten-malate dehydrogenase
conjugate. An enzyme immunoassay for morphine. J. Biol. Chem.
250, 3759 (1975).
4. Ullman, E. F., Blakemore, J., Leute, R. K., et al. Homogeneous
enzyme immunoassay for thyroxine. Gun. Chem. 21, 1011 (1975).
5. Chang, J. J., Crowl, C. P., and Schneider, R. S., Homogeneous
enzyme immunoassay for digoxin. Clin. Chem. 21, 967 (1975).
6. Schneider, R. S., Bastiani, R. J., Rubenstein, K. E., et al., Homogeneous
enzyme immunoassay for diphenylhydantoin and phenobarbital
in serum. Gun. Chem. 20, 869 (1974).
7. Rowley, G. L., Armstrong, T. A., Crowl, C. P., et al., Determination
of THC and its metabolites by EMITS homogeneous enzyme immunoassay:
A summary report. In Cannabinoid Assays in Humans,
NIDA Research Monograph 7, R. E. Willette, Ed., 1976, p 28.
8. Durst, T., Stereospecific hydroxalkylation of chloromethylphenylsulfoxide.
J. Am. Chem. Soc. 91, 1034 (1969).
9. Durst, T., and Tin, K. C., Thermal and acid-catalyzed rearrangements
of a-epoxy sulfoxides and sulfones. Tetrahedron Lett., 2369
(1970).
10. Archer, R. A., Boyd, D. B., Demarco, P. V., et al., Structural
studies of cannabinoids. A theoretical and proton magnetic resonance
analysis. J. Am. Chem. Soc. 92,5200 (1970).
II. Garrett, E. R., and Hunt, C. A., Physicochemical properties,
solubility, and protein binding of ı19-tetrahydrocannabinol. J. Pharm.
Sci. 63, 1056 (1974).
12. Lemberger, L., Axelrod, J., and Kopin, I. J., Metabolism and
disposition of tetrahydrocannabinols in naive subjects and chronic
marijuana users. Ann. N.Y. Acad. Sci. 191,142(1971).
13. Patton, W. D. M., and Pertwee, R. G., The actions of cannabis in
man. In Marijuana: Chemistry, Pharmacology, Metabolism and
Clinical Effects, R. Mechoulam, Ed., Academic Press, New York, N.
Y., 1973.
Source: Homogeneous enzyme immunoassay for cannabinoids in urine
