BAM
A New Caffeine Immunoassay:
Clinical and Environmental Application
J. J. Carvalho, R. J. Schneider*, M. G. Weller, U. Panne
BAM Federal Institute for Materials Research and Testing, Richard-Willstätter-Str. 11, D-12489 Berlin
Introduction
CAFFEINE (1,3,7-trimethylxanthine) was discovered by the German chemist Friedrich
[1]
Ferdinand Runge in 1819. He coined the term "kaffein", a chemical compound in “kaffee” .
Ever since caffeine has been used with several purposes: as pesticide in agriculture,
medicine in human therapeutics, additive in cosmetics, as a marker of human liver function[2]
[3]
and. Lately, as anenvironmental marker .
Several analytical methods, mostly based on chromatographic techniques, are
currently in use for caffeine analysis. Besides the chromatographic analysis time
itself, such techniques are usually preceded by sample preparation/concentration
steps, making the entire process costly and time-consuming.
An ELISA (Enzyme-Linked Immunosorbent Assay)
was developed to overcome such drawbacks.
Caffeine is predominantly eliminated
via N-3
demethylation to paraxanthine (1,7-dimethylxanthine). The
reaction is catalyzed by cytochrome P450 (CYP1A2) and
caffeine clearance is considered a “gold standard” for the
enzyme’s activity in humans.[4]
Caffeine is one of the most widely and frequently ingested
compound throughout the world. Its presence in environmental
water has been already shown, including in treated drinking
water, indicating contamination of human origin.[3]
Results and Discussion
Experimental
ELISA vs LC-MS/MS
ELISA
LC-MS/MS
12000
Surface waters
Salivas
Assay parameters
Surface waters
Sample volume
0.5 mL
0.2 mL
Sample volume
500 mL
Protein precipitation
Sample preparation
Either Solid Phase
Extraction (SPE) or direct
analysis
Sample preparation
Solid Phase Extraction
(SPE)
Calibration range
(n=8)
Monoclonal antibody dilution
(anti-caffeine IgG mouse)
1: 100 000
1: 10 000
1: 75 000
1: 10 000
10 µg L -1 - 175 µg L -1
Tracer dilution
(Horseradish peroxidase conjugate)
0.02 µg L
Calibration range
(n=8)
-1
- 100 µg L
-1
2.5 µgL
-1
- 125 µg L
-1
y = 0.9434x + 176.08
2
R = 0.902
10000
ELISA conc. (ug/L)
Assay parameters
Table 1 and 2. Summary table regarding
ELISA and LC-MS/MS methods.
8000
6000
4000
2000
0
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
LC-MS/MS conc. (ug/L)
Calibration curve
Cross Reactivity
Figure 3. Caffeine concentrations in sixty saliva samples analyzed by both methods. The samples were
provided by ten regular caffeine consumers throughout one day, before caffeine intake and thereafter at
regular intervals. The red lines represent ± 20% of deviation from the ideal curve with slope one ( dashed line),
and the bold line represents the trendline of all values (navy-blue rhombi).
1.2
Theobromine (1.2 %)
1.0
Bland-Altman analysis and independent two-sided t -test (equal
variances, p=0.05, 118 degrees of freedom) were also performed and
confirmed the outcome visualized in the figure above.
H
H
0.8
A.U.
Theophylline (7.1%)
0.6
0.4
C affeine in B erlim S urface W aters (n=10)
C affeine in salivas (n=60)
160
H
Paraxanthine (0.01 %)
0
20
40
60
80
100
120
140
Conc. (ug.L-1)
Figure 1. Calibration curve used for quantitation of
salivas by ELISA. A third order exponential decay
fitting was selected after residues evaluation in
comparison with other models.
Figure 2. Antibody cross reactivity against caffeine
demethylated metabolites. Paraxanthine is the most
abundant metabolite produced in vivo (90%) and
advantageously presents the lowest value.
C oncentration (n g /L )
0.0
140
120
116
80
40
66
38
10000
109
100
60
12000
138
69
64
M ax
M in
A verage
44
31
C oncentration (ug/L)
0.2
6000
4000
2000
0
0
E LIS A
S P E -E LIS A
9781
8000
20
S P E -LC -M S /M S
9503
3925
3879
651
814
LC -M S /M S
E LIS A
* Corresponding author: rudolf.schneider@bam.de
Figure 4. Variation between the methods found for saliva and surface water samples. The surface
waters were analyzed directly by ELISA (1000fold less concentrated than the extracts injected into
the LC-MS/MS). The SPE extracts were additionally analyzed by ELISA to provide an easier
comparison between methods in the same concentration range.
Outlook
Is the assay selective enough to distinguish caffeine in other matrices?
Can the assay sensitivity be further improved by means of a different detection
mode, like fluorescence spectroscopy?
Literature
[1] B.A. Weinberg, B.K. Bealer, The world of caffeine The science and culture of the world's most popular drug,
in Prologue, Routledge, 1st Edition, 2001, London, UK
[2] T. Wang et al., Klinische Wochen-Schrift, 1985, 63, 1124-1128
[3] I. J. Buerge et al., Environmental Science & Technology, 2003, 37, 691-700
[4] J.A. Carrillo et al., Therapeutic drug monitoring, 2000, 22, 409-417
The immunoassay provides results equivalent to those from the LC-MS/MS
method regarding the studied matrices and the intended purpose.
On a single ELISA plate 72 samples are analyzed within 3 hours. Even excluding the SPE procedure, the same number of samples requires 64 hours using
LC-MS/MS.
Acknowledgement
We would like to thank all colleagues who provided the saliva samples and Kristin Petsch
of WG Immunochemical Methods for repeating the analyses, thus independently
confirming the results.