Histamine and tele-methylhistamine quantification in cerebrospinal fluid from narcoleptic subjects by liquid chromatography tandem mass spectrometry with precolumn derivatization

Article abstract of DOI:10.1016/j.ab.2010.09.045

An ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) assay was developed for the simultaneous analysis of histamine, its major metabolite tele-methylhistamine, and an internal standard (N-tele-(R)-α-dimethylhistamine) from huma

Full text of DOI:10.1016/j.ab.2010.09.045

Analytical Biochemistry 409 (2011) 28–36
Contents lists available at ScienceDirect
Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio
Histamine and tele-methylhistamine quantification in cerebrospinal fluid
from narcoleptic subjects by liquid chromatography tandem mass spectrometry
with precolumn derivatization
Mikaël Croyal a, Yves Dauvilliers b, Olivier Labeeuw a, Marc Capet a, Jean-Charles Schwartz a,
Philippe Robert a,
⇑
a Bioprojet Biotech, Saint Grégoire 35762, France
b Department of Neurology, Gui de Chauliac Hospital, CHU Montpellier, INSERM U888, National Reference Network for Orphan Diseases
(Narcolepsy, Hypersomnia, and Kleine–Levin Syndrome), Montpellier 34295, France
a r t i c l e i n f o
a b s t r a c t
Article history:
An ultra-performance liquid chromatography tandem mass spectrometry (UPLC™–MS/MS) assay was
Received 21 May 2010
Received in revised form 29 September
2010
Accepted 29 September 2010
Available online 7 October 2010
developed for the simultaneous analysis of histamine, its major metabolite tele-methylhistamine, and
an internal standard (N-tele-(R)-a-dimethylhistamine) from human cerebrospinal fluid (CSF) samples.
The method involves derivatization of primary amines with 4-bromobenzenesulfonyl chloride and sub-
sequent analysis by reversed phase liquid chromatography with mass spectrometry detection and posi-
tive electrospray ionization. The separation of derivatized biogenic amines was achieved within 3.5 min
on an AcquityÒ BEH C18 column by elution with a linear gradient of acetonitrile/water/formic acid (0.1%).
The assay was linear in the concentration range of 50–5000 pM for each amine (5.5–555 pg/ml for hista-
mine and 6.25–625 pg/ml for tele-methylhistamine). For repeatability and precision determination, coef-
ficients of variation (CVs) were less than 11.0% over the tested concentration ranges, within acceptance
criteria. Thus, the developed method provides the rapid, easy, highly sensitive, and selective requirement
to quantify these amines in human CSF. No significant difference was found in the mean standard error
levels of these amines between a group of narcoleptic patients (histamine = 392 64 pM, tele-methylhis-
tamine = 2431 461 pM, n = 7) and of neurological control subjects (histamine = 402 72 pM, tele-meth-
ylhistamine = 2209 463 pM, n = 32).
Keywords:
Histamine
tele-Methylhistamine
Derivatization
Mass spectrometry
Liquid chromatography
Cerebrospinal fluid
Narcolepsy
Ó 2010 Elsevier Inc. All rights reserved.
Histamine (HA)1 is released all over the brain by tuberomamm-
illary nucleus neurons. A number of data in animals indicate that his-
taminergic neurons have a major role in wakefulness control [1]. For
instance, HA neuron firing and HA release are maximal during wake-
fulness, whereas wakefulness is diminished as a consequence of
destruction of HA-releasing neurons and/or inhibition of HA synthe-
sis, release, or action (via blockade of the H1 receptor subtype).
Hence, H1 receptor blockade decreases wakefulness in humans,
whereas activation of histaminergic neurons by an inverse agonist
of the H3 receptor enhances wakefulness in normal animals and de-
creases excessive daytime sleepiness (EDS) in both orexin (ꢀ/ꢀ)
mice and human narcoleptics [2]. HA is metabolized by histamine-
N-methyltransferase to tele-methylhistamine (t-MHA) in certain
tissues, including brain. Whereas the positive clinical result in narco-
leptic patients [2] suggests a role of endogenous HA in this disease, it
is not clearly established whether a deficit in this waking amine par-
ticipates in the pathophysiology of narcolepsy, as is the case for
orexins [3].
⇑
Various assay methods for HA and/or its metabolite in biological
samples have been developed, including radioenzymatic assay,
immunoassay, gas chromatography (GC), liquid chromatography
(LC), capillary electrophoresis, and capillary electrochromatogra-
phy [4–7]. HA and t-MHA quantification in human cerebrospinal
fluid (CSF) were first reported by Autio and Ermala [8], Jackson
and Rose [9], and Swahn and Sedvall [10], respectively.
Two studies recently reported the first measurements of CSF
HA in patients affected with primary hypersomnias using high-
performance liquid chromatography (HPLC)-fluorometry with
Corresponding author. Fax: +33 0 299 280 444.
E-mail address: p.robert@bioprojet.com (P. Robert).
Abbreviations used: HA, histamine; EDS, excessive daytime sleepiness; t-MHA,
1
tele-methylhistamine; GC, gas chromatography; LC, liquid chromatography; CSF,
cerebrospinal fluid; HPLC, high-performance liquid chromatography; UPLC™–MS/MS,
ultra-performance liquid chromatography tandem mass spectrometry; ESI, electro-
spray ionization; LOD, limit of detection; 4-BBS, 4-bromobenzenesulfonyl chloride;
BSA, bovine serum albumin; HLA, human leukocyte antigen; REM, rapid eye
movement; MSLT, multiple sleep latency test; IS, internal standard; DMSO, dimethyl
sulfoxide; UV, ultraviolet; QC, quality control; LOQ, limit of quantification; MRM,
multiple reaction monitoring; SEM, standard error of the mean; RT, retention time;
CV, coefficient of variation.
0003-2697/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.ab.2010.09.045

Histamine and tele-methylhistamine quantification / M. Croyal et al. / Anal. Biochem. 409 (2011) 28–36
29
post-column derivatization, a method initially developed for the
assay of this amine in human serum and mouse tissues [11]. De-
spite the absence of a clear-cut threshold for normal and abnor-
mal CSF HA levels, both studies reported lower HA levels in
patients with narcolepsy or idiopathic hypersomnia as compared
with controls [12,13].
7 patients with narcolepsy–cataplexy (1 with 126 pg/ml) and
showed levels above 200 pg/ml in all neurological controls except
2 (129 and 165 pg/ml, respectively). CSF sampling in patients and
related analyses was approved by the ethics committee of the Hos-
pital of Montpellier and complied with national rules on the in-
formed consent of patients.
Nevertheless, a reliable measurement of HA, which occurs in
low levels in CSF compared with other biological fluids, has re-
mained a difficult problem, as judged by reported normal values
differing by more than 100-fold according to the method of assay
used [10–16].
Synthesis of N-tele-(R)-a-dimethylhistamine as internal standard
Stable isotope-labeled analogs of HA are commercially avail-
able. N-tele-(R)- -dimethylhistamine, present in our medicinal
a
In view of these difficulties, we have developed and validated a
very sensitive ultra-performance liquid chromatography tandem
mass spectrometry (UPLC™–MS/MS) technique with positive elec-
trospray ionization (ESI) for the assay of HA and t-MHA, its major
metabolite, after precolumn derivatization of both amines, and
we applied it to the CSF of patients with narcolepsy–cataplexy.
A similar approach, based on LC–MS/MS, was recently devel-
oped to quantify HA in cell culture medium [17]. However, the lim-
it of detection (LOD) targeted in this assay was not sensitive
enough to be applied to CSF samples (LOD ꢁ2727 pM) and not able
to simultaneously quantify HA and its metabolite. HA levels, quan-
tified in basophilic derivative cell line medium KU812, were sev-
eral orders of magnitude higher.
chemistry compounds collection, was selected as internal standard
(IS) according to its similar physicochemical properties with the
biogenic amines HA and t-MHA.
N-tele-(R)-a-Dimethylhistamine was synthesized from (R)-a-
methylhistamine analogously to t-MHA according to Durant and
coworkers [19]. Colorless oil. 1H NMR (250 MHz, dimethyl sulfox-
ide [DMSO]-d6): d = 7.56 (s, 1H), 6.95 (s, 1H), 3.59 (s, 3H), 3.38
(m, 1H), 2.64 (m, 2H), 1.12 (d, J = 6.5 Hz, 3H) (Scheme 1).
Synthesis of 4-BBS-derivatized t-MHA, HA, and IS
4-BBS-derivatized amines (4-BBS t-MHA, 4-BBS HA, and 4-BBS
IS) were synthesized in our laboratories and used for determina-
tion of the derivatization yield of each amine.
Materials and methods
4-Bromo-N-[2-(1-methyl-1H-imidazol-4-yl)ethyl]benzenesulfonamide
synthesis
Materials
To a solution of t-MHA, dihydrochloride (33 mg, 0.167 mmol),
and triethylamine (70 ll, 3 eq.) in chloroform (2 ml) was added
Histamine dihydrochloride, tele-methylhistamine (1-methyl-
histamine) dihydrochloride, (R)-a-methylhistamine dihydrochlo-
4-BBS (48 mg, 1.1 eq.) at 0 °C. After stirring at room temperature
for 15 h, the mixture was diluted with dichloromethane, washed
with water, dried over MgSO4, and concentrated under reduced
pressure. Purification of the residue by flash chromatography (sil-
ica gel, CH2Cl2/MeOH/NH4OH, 95:5:0.5) afforded the desired prod-
uct (36 mg, 63%) as a white solid with a melting point of 115 °C. 1H
NMR (250 MHz, CDCl3): d = 7.72 (d, J = 8.5 Hz, 2H), 7.60 (d,
J = 8.5 Hz, 2H), 7.30 (s, 1H), 6.60 (s, 1H), 6.12 (s, 1H), 3.63 (s, 3H),
3.22 (m, 2H), 2.67 (t, J = 6.1 Hz, 2H). UPLC™–ultraviolet (UV) purity
was found to be higher than 99%.
ride, 4-bromobenzenesulfonyl chloride (4-BBS), triethylamine,
and bovine serum albumin (BSA) were purchased from Sigma–Al-
drich (France). Artificial CSFs were obtained from Harvard Appara-
tus (France). Acetonitrile, water (LC–MS grade), and formic acid
were obtained from Fischer (France).
Analyses were performed on a UPLC™–MS/MS system consist-
ing of a Quattro PremierÒ XE triple-quadrupole mass spectrometer
equipped with an ESI interface and an AcquityÒ UPLC™ device as
LC system (Waters, France). Data acquisition and analysis were
performed using MassLynxÒ software (version 4.1, Waters).
4-Bromo-N-[2-(1H-imidazol-4-yl)ethyl]benzenesulfonamide synthesis
To a solution of histamine dihydrochloride (500 mg, 2.72 mmol)
and sodium hydroxide (1.31 g, 12 eq.) in water (500 ml) was added
a solution of 4-BBS (2.08 g, 3 eq.) in acetonitrile (125 ml) at room
temperature. After stirring for 21 h, the mixture was partially con-
centrated to remove acetonitrile and was extracted with ethyl ace-
tate (three times). The combined organic layers were then dried
over MgSO4 and concentrated under reduced pressure. Purification
of the residue by flash chromatography (silica gel, CH2Cl2/MeOH/
NH4OH, 95:5:0.5–90:10:0.5) afforded the desired product
(225 mg, 25%) as a white solid with a melting point of 156 °C. 1H
NMR (250 MHz, DMSO-d6): d = 11.76 (s, 1H), 7.83 (s, 1H), 7.79
(d, J = 8.6 Hz, 2H), 7.69 (d, J = 8.6 Hz, 2H), 7.46 (s, 1H), 6.74 (s,
1H), 2.92 (m, 2H), 2.55 (m, 2H). UPLC™–UV purity was found to
be higher than 99%.
Subjects
Subjects in the study were 7 patients with sporadic narcolepsy–
cataplexy between 8 and 57 years of age (median age = 28 years).
Inclusion criteria for narcolepsy were the presence of EDS, cata-
*
plexy, human leukocyte antigen (HLA) DQB1 0602 positivity
(genotype known to be associated with the clinical expression of
narcolepsy) [18], at least two sleep onset rapid eye movement
(REM) periods, and a mean sleep latency below 8 min during the
multiple sleep latency test (MSLT). None of the patients was taking
any medication for at least 1 month prior to the lumbar puncture.
In addition, 32 neurological control subjects, between 20 and
86 years of age (median age = 52 years), were also included.
Patients were diagnosed as headache, multiple sclerosis, neuropa-
thy, and glioma. None of them reported any symptom of narcolepsy.
A lumbar puncture was performed in all subjects (patients and
neurological controls), and CSF samples were collected during a 2-
h period (5–7 p.m.) and stored immediately at ꢀ80 °C until use.
CSF hypocretin-1 level was determined in duplicate from CSF
samples in all patients using 125I radioimmunoassay kits (Phoenix
Peptide). Levels are considered as low when below 110 pg/ml,
intermediate when between 110 and 200 pg/ml, and normal when
over 200 pg/ml. Results showed levels below 110 pg/ml in 6 of the
(R)-4-Bromo-N-[1-methyl-2-(1-methyl-1H-imidazol-4-yl)ethyl]benzene-
sulfonamide synthesis
To
a
solution of N-tele-(R)-
a
-dimethylhistamine (51 mg,
l) in chloroform (4 ml) was
0.366 mmol) and triethylamine (100
l
added 4-BBS (105 mg, 1.1 eq.) at 0 °C. After stirring at room tem-
perature for 15 h, the mixture was diluted with dichloromethane,
washed with water, dried over MgSO4, and concentrated under
reduced pressure. Purification of the residue by flash chromatogra-
phy (silica gel, CH2Cl2/MeOH/NH4OH, 95:5:0.5) afforded the

30
Histamine and tele-methylhistamine quantification / M. Croyal et al. / Anal. Biochem. 409 (2011) 28–36
O
N
N
N
N
NH2
NH2
1) CH3I
2) HCl
N
N
N
N
NH
HN
N
O
Scheme 1. Synthesis reaction of N-tele-(R)-a-dimethylhistamine as IS.
desired product (57 mg, 43%) as a white solid with a melting point
of 119 °C. 1H NMR (250 MHz, CDCl3): d = 7.71 (d, J = 6.8 Hz, 2H),
7.56 (d, J = 6.8 Hz, 2H), 7.31 (s, 1H), 6.53 (s, 1H), 6.48 (m, 1H),
3.61 (s, 3H), 3.52 (m, 1H), 2.59 (dd, J = 4.3 and 14.5 Hz, 1H), 2.46
(dd, J = 4.3 and 14.5 Hz, 1H), 1.15 (d, J = 6.5 Hz, 3H). UPLC™–UV
purity was found to be higher than 99%.
followed by a reequilibration at 5% acetonitrile in water with
0.1% formic acid in 0.5 min.
A triple-quadrupole mass spectrometer equipped with an ESI
interface was operated in the positive ion mode. The multiple reac-
tion monitoring (MRM) mode was applied in MS detection.
The conditions for the UPLC™–MS/MS are presented in Table 1.
Method validation
Calibration curves and quality control samples for HA and t-MHA
assays
The method was validated by carrying out tests of linearity, in-
tra- and inter-assay precision, and derivatization yield for each
amine as well as LOQ and LOD. The stability of amines after 4 h
at room temperature and after three freeze–thaw cycles was estab-
lished, as well as the stability of the derivatized amine extracts
after 72 h to the UPLC™ autosampler (4 °C). Long-term reproduc-
ibility was also studied over 1 year on both native human CSF
and artificial CSF samples.
A mixed standard stock solution of 10 lM HA and t-MHA was
prepared in water and diluted to the required 10-fold physiological
range concentration in water containing BSA at a concentration of
3 g/L.
The calibration curves consisted of six single points and were
generated by adding 10
ll of 10-fold concentrated solutions of
mixed HA and t-MHA in BSA (3 g/L) to 90
ll of artificial CSF so as
to obtain final concentrations of 5000, 2500, 1000, 500, 100, and
50 pM for each amine. Dilution steps in artificial CSF of 3 g/L albu-
min solutions leads to both physiological salts and albumin
(0.3 g/L) levels in calibrators. This dilution was performed directly
into a 1-ml recovery 96-well microplate (Waters).
Quality control (QC) samples were prepared in the same way at
three concentration levels: one corresponding to the limit of quan-
tification (LOQ, 50 pM), one in the midrange concentration
(500 pM), and one at the highest concentration of the calibration
curve (5000 pM). Finally, QCs were aliquoted into polypropylene
tubes and stored under the same conditions as the samples at
Matrix effect evaluation
The matrix effect was evaluated in artificial CSF at the three
concentration levels of QC samples in five replicates and in native
human CSF in 15 individual samples (at three concentration levels,
5 individuals per level) chosen arbitrarily and spiked with 10 ll of
10ꢂ standard solutions corresponding to the three concentration
levels of QC samples. The peak area ratio of the analyte to the IS
was calculated for the two biogenic amines, HA and t-MHA, in each
sample. In the case of native human CSF, samples were analyzed
with and without spiking.
ꢀ80 °C. The day of the assay, 100
into a 1-ml recovery 96-well microplate.
A solution of 10 M of IS was prepared in water and diluted in
ll of QC samples was aliquoted
The increases or decreases in the peak area ratios of the com-
pounds were compared with the respective area ratio measured
l
artificial CSF to obtain a final concentration of 10,000 pM.
Table 1
Summary of UPLC™–MS/MS conditions.
4-BBS derivatization of HA, t-MHA, and IS
Column
AcquityÒ UPLC™ BEH C18
(1.7
0.1% formic acid in acetonitrile/water
A 2.5-mM solution of 4-BBS was prepared extemporaneously
l
m, 2.1 ꢂ 50 mm i.d.)
Mobile phase A
(because of its low chemical stability, ꢁ1 h at room temperature)
(5:95, v/v)
in acetonitrile added to triethylamine (0.1 M). Then 100 ll of this
Mobile phase B
Linear gradient
0.1% formic acid in acetonitrile
0–20% of mobile phase B in 3 min
A% = 100% (3–3.5 min) for reequilibration
reagent mixture was added to 100
ll of the test and standard sam-
ples. After that, 10
added.
ll of IS solution (10,000 pM) in artificial CSF was
Flow rate
Injection volume
Column temperature
600
ll/min
5 ll in partial loop mode
50 °C
After a 10-min incubation at 45 °C under orbital agitation
(150 rpm), samples were evaporated to dryness and reconstituted
with 100 ll of mobile phase consisting of acetonitrile and water
(20:80, v/v) added to 0.1% of formic acid. Reaction temperature
was optimized to restrain derivatization reaction to the aliphatic
primary amine moiety.
Polarity
Positive ESI (V mode)
3 kV
1000 L/h
50 L/h
140 °C
450 °C
Capillary voltage
Desolvatation gas flow
Cone gas flow
Source temperature
Desolvatation temperature
Desolvatation gas
Collision-induced
dissociation gas
Dinitrogen N2
Argon
Chromatographic and mass spectrometric conditions for HA and t-
MHA assays
MRM transitions
4-BBS HA: 330.2 > 95.1 + 332.2 > 95.1
4-BBS t-MHA: 344.2 > 109.1 + 346.2 > 109.1
4-BBS IS: 358.2 > 123.1 + 360.0 > 123.1
45 V
Derivatized HA, t-MHA, and IS were separated on an AcquityÒ
BEH C18 column (2.1 ꢂ 50 mm, 1.7
with a linear gradient of acetonitrile in water (5:95–20:80, v/v) in
3 min with 0.1% formic acid at a flow rate of 600 l/min and
lm, Waters) at 50 °C by elution
Cone voltage
Collision energy
25 eV
l

Histamine and tele-methylhistamine quantification / M. Croyal et al. / Anal. Biochem. 409 (2011) 28–36
31
in each aqueous calibrator to which the same concentrations of
Validation parameters
analytes had been added. The matrix effect was calculated as
follows:
For validation of the assays, the mean back-calculated concen-
trations were processed with the 1/x weighted quadratic regression
model after analysis of six calibration curves with six concentra-
tions each obtained over three distinct experiments. The mean r2
values of the six calibration cures were 0.9990 0.0004 for deriva-
tized HA and 0.9992 0.0001 for derivatized t-MHA.
For QC samples,
matrix effect ð%Þ ¼ 100 ꢂ ½1 ꢀ peak area ratioðaddedÞ=
peak area ratioðaqueous calibratorÞꢃ:
The linearity of the method was successfully tested from 50 to
5000 pM for each derivatized amine. At the LOQ, the coefficients
of variation (CVs) were 5.4% for HA and 9.4% for t-MHA. For other
concentration levels, the CVs obtained were lower than 7.2% for HA
and 6.5% for t-MHA. The mean absolute calculated concentrations
(mean bias) did not deviate by more than 5.7% for HA and ꢀ4.8%
for t-MHA at each concentration level (Table 2).
For native human CSF samples,
matrix effect ð%Þ ¼ 100 ꢂ ½1 ꢀ ðpeak area ratioðaddedÞ
ꢀ peak area ratioðnativeÞÞ=
peak area ratioðaqueous calibratorÞꢃ:
Intra- and interassay precision and accuracy parameters were
assessed using freshly prepared calibration curves and QC samples,
obtained as described in Materials and methods, and were calcu-
lated from three assays of five replicates at three concentration lev-
els (50, 500, and 5000 pM).
Results
Evaluation and validation of HA and t-MHA assays
At the LOQ, the intraassay CVs were 10.6% for HA and 8.1% for
t-MHA, and the interassay CVs were 10.9% for HA and 11.0% for
t-MHA. For other concentration levels, intra- and interassay CVs
did not exceed 3.7% for HA and t-MHA (Table 3).
Precision and accuracy were also evaluated by spiking native
human CSF samples with three concentration levels (50, 500,
and 5000 pM). The CVs calculated between the initial results
(before spiking) and the spiked values were included in the
following range: 0.7–2.9% and 0.9–4.1% for HA and t-MHA,
respectively. This test was performed in five replicates per level
(Table 4).
The derivatization conditions led to a single derivative on the
primary amine group and not the imidazole nucleus (Scheme 2).
Derivatization yields, as determined over three distinct experi-
ments by comparing areas of HA and t-MHA (at six concentrations
of a calibration curve) that had been derivatized during the assay
to corresponding areas of authentic derivatized amines, were
103.0 2.0 and 99.5 1.2% for HA and t-MHA, respectively
(means standard errors [SEMs], n = 36). For IS, derivatization
yield was determined in an identical way in the same calibrator
samples (n = 36) and was calculated at 99.5 0.9%.
The selected mass spectrometric conditions led to the produc-
tion of specific daughter ions for each derivatized biogenic amine
by a four-center rearrangement leading to a common daughter
ion (Scheme 3). (We cannot exclude an alternative mechanism
for productions with the formation of azetidino [1,2-c] imidazole
quaternary ammonium ions, as suggested by a referee.)
4-BBS derivatization reagent was preferred to other benzene-
sulfonyl chlorides. Indeed, the two isotopic forms (79Br and 81Br)
of each derivatized entity present in equal proportion were
scanned, leading to a strong MS signal and a lower risk of interfer-
ent ions (Fig. 1).
The selected UPLC™ conditions allowed a rather clear-cut sep-
aration of the two derivatized biogenic amines and their IS, as
shown by their retention times (RTs, 2.05, 2.13, and 2.30 min for
HA, t-MHA, and IS, respectively) (Fig. 2).
Chromatograms obtained with human CSF samples showed an
absence of interfering peaks using the selected reaction monitoring
mode (Fig. 2C).
The LOQ was targeted at 50 pM for each derivatized amine and
validated as the lowest point of the calibration curve with an ana-
lyte response of at least 10 times compared with the blank re-
sponse. The mean SEM signal-to-noise ratios (peak-to-peak
calculation mode) obtained from six calibration curves were
34.2 3.9 and 31.1 5.1 for HA and t-MHA, respectively, thereby
validating the LOQ of the assay.
The LOD was determined at 12.5 pM for each derivatized amine
with an analyte response of at least three times compared with the
blank response. The signal-to-noise ratios (peak-to-peak calcula-
tion mode) obtained in an artificial CSF sample containing
12.5 pM of each amine were 11.4 and 10.7 for HA and t-MHA,
respectively (Fig. 3).
The stability of biogenic amines was evaluated in QC samples.
After 4 h at room temperature or three freeze–thaw cycles, QC
samples were treated as described in Materials and methods and
quantified with two freshly prepared calibration curves. Then the
Br
Br
Br
Br
NH2
H
N
Cl
Cl
Cl
+
+
+
S
A
HN
N
N
N
N
S
O O
HN
N
N
N
N
O O
Br
Br
NH2
NH2
H
N
S
B
C
S
O O
O O
H
N
S
N
S
O O
O O
N
Scheme 2. 4-BBS derivatization reactions of HA (A), t-MHA (B), and IS (C).

32
Histamine and tele-methylhistamine quantification / M. Croyal et al. / Anal. Biochem. 409 (2011) 28–36
Putative neutral
Compound
Parent ion
Daughter ion
loss
Br
Br
H
N
S
N
N+
4-BBS HA
H
H
O O
H
m/z 330.2 / 332.2
N
N
N+
H
m/z 95.1
Br
H
N
S
N+
H2N
N
4-BBS t-MHA
4-BBS IS
S
N+
H
O O
O O
H
m/z 109.1
m/z 344.2 / 346.2
Br
N+
N
N+
HN
N
S
H
H
O O
m/z 123.1
m/z 358.2 / 360.0
Scheme 3. Formation of product ions in mass spectrometer.
Fig.1. (A) Mass spectrum of 4-BBS-derivatized HA (m/z 330.2/332.2), 4-BBS-derivatized t-MHA (m/z 344.2/346.2), and 4-BBS-derivatized IS (m/z 358.2/360.0). (B–D) Daughter
ion spectra of 4-BBS-derivatized HA (B), t-MHA (C), and IS (D) with the 79Br isotopic form.
mean recovery was calculated for each compound at each concen-
tration level. The mean recovery ranges obtained for each deriva-
tized amine at each level were 85.8–103.9% and 97.1–107.8% for
HA and t-MHA, respectively. Both amines were stable in the tested
conditions, as shown in Table 5.
The long-term stability of both amines at ꢀ80 10 °C in artifi-
cial CSF was studied in QC samples prepared 1 year later. The mean
recovery ranges obtained for each derivatized amine at each level
were 92.6–100.0% and 96.3–105.1% for HA and t-MHA, respec-
tively. Both amines were considered as stable in our storage condi-
tions for up to 1 year (Table 5).
The stability of the derivatized biogenic amine extracts was
evaluated by redosing QC samples after 72 h of storage in the
UPLC™ autosampler at 4 °C with two freshly prepared calibration
curves prepared as described in Materials and methods. The mean
recovery SEM ranges obtained for each derivatized amine at each
level were 88.3 8.5 to 101.0 1.4% and 99.0 5.5 to 99.9 1.0%
for HA and t-MHA, respectively. Thus, the derivatized compounds
were considered as stable in the analytical conditions of the assay.
For long-term reproducibility, the same set of samples, stored at
ꢀ80 10 °C for 1 year (7 narcoleptic patients and 32 neurological
control subjects), was assayed using freshly prepared calibration

Histamine and tele-methylhistamine quantification / M. Croyal et al. / Anal. Biochem. 409 (2011) 28–36
33
2.30
3.42e+5
A
B
C
3.83e+3
3.72e+3
Magnified chromatogram
2.05
2.13
2.45
2.50
2.63
3.19
1
0.00
0.50
0.50
0.50
1.00
1.00
1.00
1.50
2.80e+5
2.00
3.00
3.50
2.30
2.13
2.05
3.62e+5
2.31e+5
0
0.00
1.50
2.00
2.50
3.00
3.50
2.30
9.19e+4
3.21e+5
2.13
2.02e+4
2.05
2.00
0
0.00
Time
3.50
1.50
2.50
3.00
Time (min)
Fig.2. (A) Representative chromatogram of 4-BBS-derivatized HA (RT = 2.05 min) and t-MHA (RT = 2.13 min) standards and IS (RT = 2.30 min) in artificial CSF at 50 pM (LOQ).
The chromatogram was magnified for better HA and t-MHA peak visualization at the LOQ level. (B) Representative chromatogram of 4-BBS-derivatized HA (RT = 2.05 min)
and t-MHA (RT = 2.13 min) standards and IS (RT = 2.30 min) in artificial CSF at 5000 pM. (C) Representative chromatogram of native 4-BBS-derivatized HA (RT = 2.05 min) and
t-MHA (RT = 2.13 min) in human CSF samples with their 4-BBS IS (RT = 2.30 min). The boxed numbers indicate ion intensity for each peak (in arbitrary units).
Table 2
Linearity of the method.
Theoretical concentration (pM)
n
HA
t-MHA
Calculated concentration (pM)
CV (%)
Mean bias (%)
Calculated concentration (pM)
CV (%)
Mean bias (%)
50
100
500
1000
2500
5000
6
6
6
6
6
6
53
97
482
998
2537
4984
5.4
7.2
4.7
4.4
2.5
2.4
5.7
ꢀ3.0
ꢀ3.6
ꢀ0.2
1.5
50
105
476
991
2550
4979
9.4
5.1
6.5
3.8
0.6
1.2
ꢀ0.6
4.7
ꢀ4.8
ꢀ0.9
2.0
ꢀ0.3
ꢀ0.4
Table 3
Precision and accuracy of the method.
Theoretical concentration (pM)
n
HA
t-MHA
Mean bias (%) Intraassay CV (%) Interassay CV (3 days) (%) Mean bias (%) Intraassay CV (%) Interassay CV (3 days) (%)
50
500
5000
15 ꢀ7.1
15 ꢀ1.4
15 ꢀ2.2
10.6
3.7
2.2
10.9
3.7
2.2
ꢀ6.4
ꢀ1.2
ꢀ1.1
8.1
2.4
3.7
11.0
3.0
3.7
Table 4
HA and t-MHA levels in human CSF before and after spiking with known concentration (n = 5).
Analyte
Mean value before
spiking (pM)
Spiking concentration
level (pM)
Mean theoretical spiking
value (pM)
Mean experimental spiking
value (pM)
CV (%)
390
1007
590
2713
3293
944
50
500
5000
50
500
5000
440
1507
5590
2763
3793
5944
459
1522
5407
2798
3980
5606
2.9
0.7
2.4
0.9
3.4
4.1
HA
t-MHA

34
Histamine and tele-methylhistamine quantification / M. Croyal et al. / Anal. Biochem. 409 (2011) 28–36
A
B
S/N:PtP=11.44
S/N:PtP=10.67
99
99
-1
-1
0.50
1.00
1.50
2.00
2.12
2.50
3.00
0.50
1.00
1.50
2.00
2.50
3.00
3.12
97
1.10
2.87
3.14
3.23
3.43
2.05
2.62
2.55
2.37
0.70
0.78
0.94
2.74
2.22
1.87
1.55
3.47
1.26
0.44
0.66
3.07
2.89
0.10
2.25
0.35
0.25
1.57
1.66
0.91
3.01
3.00
2.09
1.06
3.23
0.55 0.66
1.13
1.24
0.08
0.23
2.51
2.66
3.41
0.80
1.71
0.35
1.44
1.99
-3
4
Time (min)
Time (min)
0.50
1.00
1.50
2.00
2.50
0.50
1.00
1.50
2.00
2.50
3.00
Fig.3. Representative chromatogram of 4-BBS-derivatized HA (A) and t-MHA (B) at the LOD (12.5 pM) compared with an artificial blank CSF sample.
Table 5
Stability of the biogenic amines.
Analyte
n
Concentration level (pM)
4 h at room temperature
Mean recovery SEM (%)
Three freeze–thaw cycles
Mean recovery SEM (%)
1 year at ꢀ80 10 °C
Mean recovery SEM (%)
HA
5
5
5
50
500
5000
86.4 4.9
96.0 4.5
100.8 2.0
85.8 7.9
103.9 0.8
102.8 0.9
92.6 6.1
100.0 2.0
98.2 1.6
t-MHA
5
5
5
50
500
5000
107.8 8.3
97.1 3.7
101.1 2.0
99.3 6.2
103.8 0.6
100.3 1.1
105.1 3.5
96.3 3.4
100.6 0.7
Table 6
curves and QC samples. The CV values were calculated comparing
the newly determined sample concentration and the initially re-
ported one for each sample. For HA, 97.4% of repeat values were
found within 20% (0.1–19.8%); for t-MHA, 97.4% of repeat values
also were found (0.2–18.9%). Thus, the reproducibility of the meth-
od developed here was established over 1 year, according to the in-
curred sample reproducibility recommendations by the European
Bioanalysis Forum [20].
HA and t-MHA levels in the CSF of patients with narcolepsy–cataplexy and
neurological controls.
Narcolepsy–cataplexy Neurological controls
(n = 7)
(n = 32)
Age (years, median [range]) 28 [8–87]
HA (pM, mean SEM)
t-MHA (pM, mean SEM)
52 [20–86]
402 72
392 64
2431 461
2209 463
Matrix effect
patients with narcolepsy–cataplexy and a group of neurological
control subjects.
A matrix effect can be observed as ion suppression, in cases
where the values were positive, and as ion enhancement, in cases
where negative values were found [21]. The mean matrix effect
SEM values calculated here were of ꢀ1.4 2.4% for HA,
4.6 1.7% for t-MHA, and ꢀ11.7 1.3% for IS in artificial CSF sam-
ples and were 3.0 6.5% for HA, ꢀ1.6 4.7% for t-MHA, and
0.3 0.9% for IS in native human CSF samples. The slight differ-
ences in compound areas ratios demonstrate that the matrix effect
of artificial and native human CSF is negligible on the quantifica-
tion of both amines.
Discussion
We have developed and validated a new method for the com-
bined assay of HA and t-MHA in human CSF based on precolumn
derivatization of both amines [21–25] as para-bromobenzenesulf-
onamides followed by UPLC™–MS/MS analysis and positive ESI.
The interest of the process lies on easier isolation of the hydropho-
bic derivatives on a UPLC™ column of the C18 type and the sensi-
tive mass spectrometric detection of bromine-containing
compounds, leading to stronger isotopic signals with duplicate
peaks corresponding to the two isotopes of the halogen. Indeed,
our LOD (12.5 pM for each amine) was 200 times lower than the
recently published LC–MS/MS method [16] in which the LOD of
HA was reported to be 2727 pM in cell culture medium.
HA and t-MHA levels in human samples
As shown in Table 6, we found no significant difference in the
mean SEM levels of CSF HA and t-MHA between a group of

Histamine and tele-methylhistamine quantification / M. Croyal et al. / Anal. Biochem. 409 (2011) 28–36
35
The UPLC™ conditions selected allowed us to differentiate the
two derivatized biogenic amines in terms of RTs with distinct
peaks, and reproducible values were nevertheless generated in
the various experiments; furthermore, the two derivatives were
clearly distinguished by their different mass fragments (Scheme 2).
In contrast to HPLC coupled to electrochemical or fluorescent
detection modes, in which analytes can be identified only by their
RTs, UPLC™–MS/MS can provide high specificity due to the addi-
tional structural information depicted using MS/MS. This indicates
that the combination of the two analytical processes used here is
able to solve the inherently difficult problem of analysis of two
amines with close physicochemical properties.
The conditions of the derivatization reaction of both amines and
their chemically synthesized IS were selected for the reaction to
occur selectively on the primary amine group (temperature and ki-
netic optimization), leaving intact the imidazole nucleus in the
case of HA derivatization, thereby avoiding the formation of multi-
ple derivatives with variable stability and lesser reproducibility.
The chemical derivatization reaction is simple, uses a commercially
available reagent, and is fast (<10 min reaction time), specific, and
quantitative under mild reaction conditions. As a result, a nearly
quantitative derivatization was found with reaction yields of
103% and 99% for HA and t-MHA, respectively.
A successful validation of the assay was indicated by the high
linearity of calibration curves and the low inter- and intraassay
variation coefficients even at the LOQ.
The LOQ was targeted at 50 pM for each derivatized amine, at
which concentration the signal-to-noise ratios were 34.2 and
31.1 for HA and t-MHA, respectively, thereby indicating a high reli-
ability of the assays even at a low concentration. In agreement, the
signal-to-noise ratios at the lowest levels of the amines in the
various CSFs analyzed were approximately 62 and 95 for HA and
t-MHA, respectively.
levels found here did not significantly differ in CSF from narcolep-
tic and neurological controls, a discrepancy attributable to the
already discussed methodological problems [6] and/or to the
rather limited size of the two populations studied here. A larger
size clinical study using the currently developed methodology is
in progress and should allow us to better settle this question.
Acknowledgments
The contribution of Stéphanie Le Meur and Anthony Chéné in
chemical synthesis is gratefully acknowledged. This work was sup-
ported by Bioprojet Biotech and Institut National de la Santé et de
la Recherche Médicale (INSERM U888).
References
[1] H.L. Haas, O.A. Sergeeva, O. Selbach, Histamine in the nervous system, Physiol.
Rev. 88 (2008) 1183–1241.
[2] J.S. Lin, Y. Dauvilliers, I. Arnulf, H. Bastuji, C. Anaclet, R. Parmentier, L. Kocher,
M. Yanagisawa, P. Lehert, X. Ligneau, D. Perrin, P. Robert, M. Roux, J.M.
Lecomte, J.C. Schwartz, An inverse agonist of the histamine H3 receptor
improves wakefulness in narcolepsy: studies in orexinꢀ/ꢀ mice and patients,
Neurobiol. Dis. 30 (2008) 74–83.
[3] E. Mignot, G.J. Lammers, B. Ripley, M. Okun, S. Nevsimalova, S. Overeem, J.
Vankova, J. Black, J. Harsh, C. Bassetti, H. Schrader, S. Nishino, The role of
cerebrospinal fluid hypocretin measurement in the diagnosis of narcolepsy
and other hypersomnias, Arch. Neurol. 59 (2002) 1553–1562.
[4] K.M. Verburg, R.R. Bowsher, D.P. Henry, A new radioenzymatic assay for
histamine using purified histamine N-methyltransferase, Life Sci. 32 (1983)
2855–2867.
[5] M. Garbarg, H. Pollard, M.D. Trung Tuong, J.C. Schwartz, C. Gros, Sensitive
radioimmunoassays for histamine and tele-methylhistamine in the brain, J.
Neurochem. 53 (1989) 1724–1730.
[6] T. Toyo’oka, Separation assay of histamine and its metabolites in biological
specimens, Biomed. Chromatogr. 22 (2008) 919–930.
[7] R.D. Velasquez, G. Brunner, M. Varrentrapp, D. Tsikas, J.C. Frölich, Helicobacter
pylori produces histamine and spermidine, Z. Gastroenterol. 34 (1996) 116–
122.
The CSF levels of t-MHA found here (ꢁ2000 pM) were in excel-
lent agreement with those reported by Khandelwal and coworkers
[16] and Prell and coworkers [26] and were in the same range as
those of Swahn and Sedvall [10], who also used a combination of
the GC and MS methods.
[8] L. Autio, P. Ermala, Determination of histamine in cerebrospinal fluid in some
neuro-psychiatric cases, Ann. Med. Exp. Biol. Fenn. 27 (1949) 87–101.
[9] I.J. Jackson, B. Rose, Observations on the histamine content of the cerebrospinal
fluid in man, J. Lab. Clin. Med. 34 (1949) 250–254.
[10] C.G. Swahn, G. Sedvall, Identification and determination of tele-
methylhistamine in cerebrospinal fluid by gas chromatography–mass
spectrometry, J. Neurochem. 37 (1981) 461–466.
In contrast, HA levels found here (ꢁ400 pM) were markedly
lower than those reported previously in studies using different
methods. In Khandelwal and coworkers [16] and Jackson and Rose
[9], values of 400,000 and 90,000 pM were reported using a radio-
enzymatic assay and a spectrophotometric assay, respectively,
with the huge difference being tentatively attributable to the lack
of previous chromatographic isolation of HA in the latter study.
Nishino and coworkers [12] and Kanbayashi and coworkers [13]
found HA levels in CSF from neurological controls of approximately
3000 pM. In both cases, the same HPLC method with fluorometric
detection, initially developed for the assay of HA in human serum
and mouse tissues [11], was used.
Nevertheless, Kiviranta and coworkers [14], who also used this
HPLC method with fluorometric detection, reported HA levels
(ꢁ400–800 pM) close to those found here. Using another tech-
nique, Borg and coworkers [15] reported HA levels (ꢁ800 pM) sim-
ilar to ours.
Because no detail was provided by Nishino and coworkers [12]
and Kanbayashi and coworkers [13] regarding the validation of
their method in human CSF, and because the fluorometric detec-
tion is based on nonspecific derivatization of any amine (or amino
acid) [11], it could be tentatively proposed that the nearly 10-fold
higher value as compared with ours is related to interference by
other amines with chromatographic behavior resembling that of
HA, as is the case of t-MHA, for which we found levels of approxi-
mately 3000 pM (i.e., close to those reported by these authors).
In contrast to the reports of Nishino and coworkers [12] and
Kanbayashi and coworkers [13], the values for HA and t-MHA
[11] A. Yamatodani, H. Fukuda, H. Wada, T. Iwaeda, T. Watanabe, High-performance
liquid chromatographic determination of plasma and brain histamine without
previous purification of biological samples: cation-exchange chromatography
coupled with post-column derivatization fluorometry, J. Chromatogr. 344
(1985) 115–123.
[12] S. Nishino, E. Sakurai, S. Nevsimalova, Y. Yoshida, T. Watanabe, K. Yanai, E.
Mignot, Decreased CSF histamine in narcolepsy with and without low CSF
hypocretin-1 in comparison to healthy controls, Sleep 32 (2009) 175–180.
[13] T. Kanbayashi, T. Kodama, H. Kondo, S. Satoh, Y. Inoue, S. Chiba, T. Shimizu, S.
Nishino, CSF histamine contents in narcolepsy, idiopathic hypersomnia, and
obstructive sleep apnea syndrome, Sleep 32 (2009) 181–187.
[14] T. Kiviranta, L. Tuomisto, E.M. Airaksinen, Histamine in cerebrospinal fluid of
children with febrile convulsions, Epilepsia 36 (1995) 276–280.
[15] J. Borg, J.M. Warter, J.L. Schlienger, M. Imler, C. Marescaux, G. Mack,
Neurotransmitter modifications in human cerebrospinal fluid and serum
during hepatic encephalopathy, J. Neurol. Sci. 57 (1982) 343–356.
[16] J.K. Khandelwal, L.B. Hough, A.M. Morrishow, J.P. Green, Measurement of tele-
methylhistamine and histamine in human cerebrospinal fluid, urine, and
plasma, Agents Actions 12 (1982) 583–590.
[17] J. Koyama, A. Takeuchi, C. Tode, M. Shimizu, I. Morita, M. Nobukawa, M.
Nobukawa, N. Kobayashi, Development of an LC–ESI–MS/MS method for the
determination of histamine: application to the quantitative measurement of
histamine degranulation by KU812 cells, J. Chromatogr. B 877 (2009) 207–212.
*
[18] E. Mignot, R. Hayduk, J. Black, F.C. Grumet, C. Guilleminault, HLA DQB1 0602 is
associated with cataplexy in 509 narcoleptic patients, Sleep 20 (1997) 1012–
1020.
[19] G.J. Durant, J.C. Emmett, C. Robin Ganellin, A.M. Roe, R.A. Slater, Potential
histamine H2-receptor antagonists: 3. Methylhistamines, J. Med. Chem. 19
(1976) 923–928.
[20] P. Timmerman, S. Luedtke, P. Van Amsterdam, M. Brudny-Kloeppel, B.
Lausecker, S. Fischmann, S. Globig, C.J. Sennbro, J.M. Jansat, H. Mulder, E.
Thomas, M. Knutsson, D. Kasel, S.A. White, M.A. Kall, N. Mokrzycki-Issartel, A.
Freisleben, F. Romero, M.P. Andersen, N. Knebel, M. de Zwart, S. Laakso, R.S.
Hucker, D. Schmidt, B. Gordon, R. Abbott, P. Boulanger, Incurred sample
reproducibility: views and recommendations by the European Bioanalysis
Forum, Bioanalysis 1 (2009) 1049–1056.

36
Histamine and tele-methylhistamine quantification / M. Croyal et al. / Anal. Biochem. 409 (2011) 28–36
[21] H.L. Cai, R.H. Zhu, H.-D. Li, Determination of dansylated monoamine and amino
acid neurotransmitters and their metabolites in human plasma by liquid
chromatography–electrospray ionization tandem mass spectrometry, Anal.
Biochem. 396 (2010) 103–111.
[22] O. Ozdestan, A. Uren, A method for benzoyl chloride derivatization of biogenic
amines for high performance liquid chromatography, Talanta 78 (2009) 1321–
1326.
[23] H. Kawanishi, T. Toyo’oka, K. Ito, M. Maeda, T. Hamada, T. Fukushima,
M. Kato, S. Inagaki, Hair analysis of histamine and several metabolites
in C3H/HeNCrj mice by ultra performance liquid chromatography with
electrospray ionization time-of-flight mass spectrometry (UPLC–ESI–TOF–
MS): influence of hair cycle and age, Clin. Chim. Acta 378 (2007) 122–
127.
[24] M.K. Handley, W.W. Hirth, J.G. Phillips, S.M. Ali, A. Khan, L. Fadnis, C.E. Tedford,
Development of a sensitive and quantitative analytical method for 1H-4-
substituted imidazole histamine H3-receptor antagonists utilizing high-
performance liquid chromatography and dabsyl derivatization, J. Chromatogr.
B 716 (1998) 239–249.
[25] C. Ji, W. Li, X.D. Ren, A.F. El-Kattan, R. Kozak, S. Fountain, C. Lepsy, Diethylation
labeling combined with UPLC/MS/MS for simultaneous determination of a
panel of monoamine neurotransmitters in rat prefrontal cortex
microdialysates, Anal. Chem. 80 (2008) 9195–9203.
[26] G.D. Prell, J.P. Green, C.A. Kaufmann, J.K. Khandelwal, A.M. Morrishow, D.G. Kirch,
M. Linnoila, R.J. Wyatt, Histamine metabolites in cerebrospinal fluid of patients
with chronic schizophrenia: their relationships to levels of other aminergic
transmitters and ratings of symptoms, Schizophr. Res. 14 (1995) 93–104.