Development and validation of a liquid chromatography and ion spray tandem mass spectrometry method for the quantification of artesunate, artemether and their major metabolites dihydroartemisinin and dihydroartemisinin-glucuronide in sheep plasma

Article abstract of DOI:10.1002/jms.1883

Recently, promising fasciocidal activities of artesunate and artemether were described in rats and sheep. Therefore, a liquid chromatography-tandem mass spectrometry (LC-MS/MS) method was developed to quantify artesunate, artemether and their metabolites dihydroartemisinin and dihydroartemisinin-glucuronide in sheep plasma. Protein precipitation with methanol was used for sample workup. Reversed-phase high-performance liquid chromatography (HPLC) was performed using an Atlantis C18 analytical column with a mobile phase gradient system of ammonium formate and acetonitrile. The analytes were detected by MS/MS using selected reaction monitoring (SRM) with electrospray ionisation in the positive mode (transition m/z 267.4 → 163.0). The analytical range for dihydroartemisinin, dihydroartemisinin-glucuronide and artesunate was 10-1000 ng/ml and for artemether 90-3000 ng/ml with a lower limit of quantification of 10 and 90 ng/ml, respectively. Inter- and intra-day accuracy and precision deviations were < 10%. Consistent relative recoveries (60-80%) were observed over the investigated calibration range for all analytes. All analytes were stable in the autosampler for at least 30 h (6 °C) and after three freeze and thaw cycles. The validation results demonstrated that the LC-MS/MS method is precise, accurate and selective and can be used for the determination of the artemisinins in sheep plasma. The method was applied successfully to determine the pharmacokinetic parameters of artesunate and its metabolites in plasma of intramuscularly treated sheep.

Full text of DOI:10.1002/jms.1883

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Research Article
Received: 6 October 2010
Accepted: 11 December 2010
Published online in Wiley Online Library: 24 January 2011
(wileyonlinelibrary.com) DOI 10.1002/jms.1883
Development and validation of a liquid
chromatography and ion spray tandem mass
spectrometry method for the quantification
of artesunate, artemether and their major
metabolites dihydroartemisinin
and dihydroartemisinin-glucuronide in sheep
plasma
c
Urs Duthaler,a,b Jennifer Keisera,b∗ and Jorg Huwyler
¨
Recently, promising fasciocidal activities of artesunate and artemether were described in rats and sheep. Therefore, a liquid
chromatography–tandem massspectrometry(LC–MS/MS)methodwasdevelopedtoquantifyartesunate, artemetherandtheir
metabolites dihydroartemisinin and dihydroartemisinin-glucuronide in sheep plasma. Protein precipitation with methanol was
used for sample workup. Reversed-phase high-performance liquid chromatography (HPLC) was performed using an Atlantis
C18 analytical column with a mobile phase gradient system of ammonium formate and acetonitrile. The analytes were detected
by MS/MS using selected reaction monitoring (SRM) with electrospray ionisation in the positive mode (transition m/z 267.4 →
163.0). The analytical range for dihydroartemisinin, dihydroartemisinin-glucuronide and artesunate was 10–1000 ng/ml and
for artemether 90–3000 ng/ml with a lower limit of quantification of 10 and 90 ng/ml, respectively. Inter- and intra-day accuracy
and precision deviations were <10%. Consistent relative recoveries (60–80%) were observed over the investigated calibration
range for all analytes. All analytes were stable in the autosampler for at least 30 h (6 ◦C) and after three freeze and thaw cycles.
The validation results demonstrated that the LC–MS/MS method is precise, accurate and selective and can be used for the
determination of the artemisinins in sheep plasma. The method was applied successfully to determine the pharmacokinetic
c
parameters of artesunate and its metabolites in plasma of intramuscularly treated sheep. Copyright ꢀ 2011 John Wiley & Sons,
Ltd.
Keywords: artemisinins; metabolites; glucuronide; fascioliasis; liquid chromatography–tandem mass spectrometry
Introduction
liver and bile parasitising food-borne trematodes Fasciolahepatica
and F. gigantica are the causative agents of fascioliasis (fasciolosis).
As many as 17 million people might be infected with Fasciola
species. Inaddition, theveterinaryimpactisconsiderableassheep,
cattleandbovinearegloballyaffectedbythedisease.[6,7] Currently,
triclabendazole is the sole drug for the treatment of human
fascioliasis. Additional drugs are commonly used in the treatment
of veterinary infections with Fasciola spp. including albendazole,
closantel, hexachlorophene, nitroxynil and rafoxanide. Resistance
The endoperoxide artemisinin (qinghaosu), a secondary plant
compound of the herb Artemisia annua (Chinese wormwood),
is an efficacious and fast acting antimalarial drug.[1] Because
artemisinin itself has biopharmaceutical shortcomings such as a
poor bioavailability that limits its effectiveness, the sesquiterpene
lactone scaffold of artemisinin was modified and the semi-
synthetic derivatives have been developed. For example, the
methyl ether derivative artemether (AM, Fig. 1) is characterised by
a higher antiplasmodial activity than artemisinin.[1] The succinate
derivative, artesunate (AS, Fig. 1), is water soluble and can be
applied intravenously and is therefore indispensable for the
treatment of cerebral malaria.[2] The semi-synthetic artemisinins
are converted rapidly to dihydroartemisinin (DHA, Fig. 1) and are
eliminated primary over bile by glucuronidation in vivo (Fig. 1).[1,3]
Importantly, DHA and partly also the glucuronide adducts possess
high-to-moderate antimalarial activity.[4]
∗
Correspondence to: Jennifer Keiser, Department of Medical Parasitology and
Infection Biology, Swiss Tropical and Public Health Institute, P.O. Box, CH-4002
Basel, Switzerland. E-mail: jennifer.keiser@unibas.ch
a
Department of Medical Parasitology and Infection Biology, Swiss Tropical and
Public Health Institute, Basel, Switzerland
b
c
University of Basel, P.O. Box, CH-4003 Basel, Switzerland
The artemisinins do not only exhibit antiplasmodial but also
trematocidalactivities, sincehaemoglobinmetabolismiscommon
in Plasmodia and several trematodes including Fasciola spp.[5] The
Department of Pharmaceutical Sciences, Division of Pharmaceutical Technol-
ogy, University of Basel, Klingelbergstrasse 50, CH-4056 Basel, Switzerland
c
J. Mass. Spectrom. 2011, 46, 172–181
Copyright ꢀ 2011 John Wiley & Sons, Ltd.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Quantification of artesunate, artemether and their major metabolites
Figure 1. Hypothetical metabolic pathway of artemether and artesunate. (A) Parent drug: artemether (R1, AM, molecular weight (MW): 298.37 g/mol) or
artesunate (R2, AS, MW: 384.42 g/mol). (B) Primary metabolite: dihydroartemisinin (DHA, MW: 284.35 g/mol). (C) Phase II metabolite: dihydroartemisinin-
glucuronide (DHA-G, MW: 460.47 g/mol).
of triclabendazole and other trematocidals is widely spreading in
livestock, but fortunately it has not been demonstrated in humans
to date.[8] Promisingly, AM was effective against a triclabendazole-
resistant F. hepatica strain in rats.[9]
land). Dihydroartemisinin-12-α-o-β-d-glucuronide (DHA-G) was
purchased from Clearsynth (Mumbai, India). The chemical struc-
tures of AS, AM, DHA and DHA-G are depicted in Fig. 1. HPLC-grade
acetonitrileandmethanolwereproductsofBisolve(Valkenswaard,
The Netherlands) and J.T. Baker (Deventer, The Netherlands), re-
spectively. Ammonium formate (LC-MS grade), sodium nitrite,
formic acid (98%, LC–MS grade) and acetic acid (glacial) were
purchased from Fluka (Buchs, Switzerland). Potassium phosphate
(KH2PO4), potassium hydroxide (reagent grade, >90%) and β-
glucuronidasetypeIX-AfromEscherichiacoli(1 000 000–5 000 000
U/g protein) were obtained from Sigma–Aldrich (Buchs, Switzer-
land). Ultrapure water was produced with an Arium 61 215 water
purification system (Sartorius Stedim Biotech, Go¨ttingen, Ger-
many) and applied for the preparation of mobile phase. Plasma
from untreated sheep was collected in the framework of our PK
studies in the Campania region of Italy.[21,22]
Various analytical methods have been established in order to
quantify the artemisinin derivatives in biological fluids and tissues.
In the recent past, liquid chromatography–mass spectrometry
(LC–MS) and tandem mass spectrometry (LC–MS/MS) methods
were developed for experimental pharmacokinetic (PK) studies
of α,β-arteether and DHA in rat plasma and serum.[10,11] In ad-
dition, different high performance liquid chromatography (HPLC)
methods with mass spectrometric or electrochemical detection
(HPLC-ECD) were developed for the measurement of AS or AM and
DHA plasma levels in humans including a range of different sam-
ple procession methods (e.g. solid-phase extraction, liquid–liquid
extractionandproteinprecipitation).[12–16] Lately,analyticalmeth-
ods were also developed for the simultaneous determination of up
to14antimalarialdrugsinhumanplasmasincetheartemisininsare
nowadaysroutinelyappliedincombinationwithotherantimalarial
drugs to prevent resistance development.[17–20] However, to date
no analytical method exists for the concurrent analysis of the semi-
synthetic artemisinins, DHA and dihydroartemisinin-glucuronide
(DHA-G).
Recently, we studied the efficacy and safety of single oral,
intramuscularandintravenousdosagesof ASandAMin F.hepatica
infected sheep.[21,22] Plasma samples were collected at selected
time points to support the drug efficacy study with PK data. We
were interested to monitor the disposition of AS, AM as well as the
two metabolites DHA and DHA-G, since experimental studies of
humansandratsexposedtotheartemisininsshowedthepresence
of glucuronide metabolites in urine, bile and plasma.[23,24]
The aim of the present study was to develop and validate a
LC–MS/MS method to quantify AS, AM and their metabolites DHA
and DHA-G in sheep plasma for the prospective application to PK
studies.
LC–MS/MS equipment and conditions
The high-performance liquid chromatography (HPLC) system
(Shimadzu, Kyoto, Japan) consisted of two LC-20AD pumps, a
CTC HTS PAL autosampler (CTC Analytics, Zwingen, Switzerland),
a thermostated Jones chromatography column oven (Omnilab,
Mettmenstetten, Switzerland) and an online DG-3310 degasser
(Sanwa Tsusho, Tokyo, Japan). Chromatographic separation was
performed at 25 ◦C using a 2.1 mm × 20 mm Atlantis T3 3-µm
analytical column (Waters, Milford, MA, USA). A 2.1 mm × 10 mm
Atlantis 3-µm guard cartridge (Waters) was connected to the
analytical column.
The mobile phase consisted of a mixture of 5 mM ammonium
formate plus 0.15% (v/v) formic acid in ultra pure water (mobile
A) and 0.15% (v/v) formic acid in acetonitrile (mobile B). The
following stepwise gradient elution program was used: 0–1 min,
B 2%; 1–10 min, B 2–45%; 10–13 min, B 45–50%; 13–15 min, B
50–95%; 15–20 min, B 95%; 20–21 min, B 95–2%; 21–22 min,
B 2%. The gradient program allowed for a baseline separation
of all analytes. Carryover (<0.1% carryover between subsequent
analytical cycles) was managed by adding a wash cycle (95%
mobile B for 5 min) at the end of each run. The flow rate of
the mobile phases was set at 0.3 ml/min. Conditioning of the
column was necessary to avoid shifts in retention time of certain
analytes, i.e. AS, after a prolonged standby time of the instrument.
Experimental
Chemicals and reagents
AM and α,β-DHA were kindly provided from Dafra Pharma (Turn-
hout, Belgium). AS was obtained from Mepha AG (Aesch, Switzer-
c
J. Mass. Spectrom. 2011, 46, 172–181
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
U. Duthaler, J. Keiser and J. Huwyler
In more detail, the analytical column was conditioned with a
replicate injection (n = 6) of the artemisinins standards (10 µg/ml
in 1 : 1 methanol/mobile A) at the beginning of each series of
experiments. The HPLC system was coupled to an API 365 triple-
quadrupole mass spectrometer (PE Biosystems, Foster City, CA,
USA) equipped with a turbo ion spray interface, operated in
positive ionisation mode. A six-port switching valve (VICI Valco
Instruments,Schenkon,Switzerland)wasusedtodiverttheeffluent
from the analytical column to the ion-spray interface of the mass
spectrometer during 0–8 and 18–22 min of each run to avoid
contamination of the mass spectrometer. The mass spectrometer
was tuned by direct infusion (Harvard apparatus infusion pump
11, Massachusetts, USA) of 10 µg/ml of each artemisinin derivative
solved in acetonitrile plus 0.1% (v/v) formic acid. All analytes were
detected by selected reaction monitoring (SRM) with a transition
of m/z 267.4 → 163.0. DHA-G, DHA and AS were analysed with
a scan width and scan time of 3 amu and 2 s, respectively. AM
was measured with a scan width and scan time of 5 amu and
3 sec, respectively. The major MS instrumentation settings are
summarised in Table 1. Instrumentation control and data analyses
were performed with Analyst 1.4.2. software (PE Biosystems) and
Microsoft Excel 2003.
Plasma sample extraction procedure
An aliquot of 90 µl sodium nitrite solution (3 M) containing 1%
(v/v) acetic acid was added to 300 µl of the sheep plasma samples
and stirred (75 rpm) for 30 min at 37 ◦C in a TH/KS 15 incubator
(Edmund Bu¨hler, Hechingen, Germany). The addition of sodium
nitrite was shown to prevent the degradation of the artemisinins
in the presence of haemoglobin (Fe2+-haeme) in haemolysed
plasma samples.[19,25] The plasma–sodium nitrite mixture was
vortex-mixed (Vortex Genius 3, IKA, Staufen, Germany) with 10 µl
of IS working solution. For protein precipitation, 1000 µl of ice-
cooled methanol was added to each sample and mixed for
1 min. The precipitated samples were cooled on ice for 10 min
and subsequently centrifuged for 15 min at 16 100g (Eppendorf
centrifuge 5415 R, Hamburg, Germany) at 4 ◦C. The supernatant
was transferred to a 2-ml microtube and stored on ice. The
pellet was vortex mixed with 500 µl of methanol for another
minuteandcentrifugedasdescribedabove.Themethanolextracts
were combined, mixed and evaporated to dryness under a
stream of nitrogen. The residue was reconstituted in 250 µl of
methanol–mobile phase A (1 : 1, v/v), vortex mixed, centrifuged
(15 min, 16 100g, 4 ◦C) and transferred to an autosampler vial.
The autosampler rack was cooled at 6 ◦C. A 30-µl aliquot of each
sample was injected into the LC–MS/MS system for analysis.
Standard, quality control and internal standard preparation
Calibration curves
Calibration curves were established using the internal standard
method for the quantification of artemisinins plasma concentra-
tions. Concentrations were plotted versus the peak-area ratios of
the analytes to the IS. For α,β-DHA, the sum of the area under the
curve of both anomers was used. Each calibration set consisted
of one blank plasma sample (matrix sample processed without in-
ternal standard), one zero sample (matrix sample processed with
internal standard) and six calibration samples including the lower
limit of quantification (LLOQ). Concentrations chosen for the cali-
bration curves ranged from 10.2 to 1000.0 ng/ml (DHA-G, DHA and
AS) and 93.8–3000.0 ng/ml (AM). The six-point calibration stan-
dard curves were calculated and fitted either by linear (DHA-G and
DHA) or quadratic regression (AS and AM). The Akaike information
criterion (AIC) was used to select the most favourable type of
regression for each calibration.[26] The quadratic regression model
was applied because an increased signal strength was observed at
higher concentration (fit by y = ax2 + bx + c) and not because of
saturation of the signal (i.e. fit by y = −ax2 +bx +c). To determine
the best weighting factor (none, 1/x or 1/x2), concentrations were
back-calculated and the model with the lowest total bias across
the concentration range was selected.
Standard stock solutions (1 mg/ml) were prepared in methanol.
Appropriate volumes of stock solutions were serially diluted
in a 1 : 1 mixture of methanol and mobile phase A to obtain
working solutions of 0.3–30.0 µg/ml for DHA-G, DHA and AS and
2.8–90.0 µg/ml for AM. Plasma calibration samples were freshly
prepared by diluting working solutions with blank sheep plasma
(1 : 30, total volume of 300 µl), resulting in final concentrations of
1000.0, 400.0, 160.0, 64.0, 25.6 and 10.2 ng/ml for DHA-G, DHA
and AS and 3000.0, 1500.0, 750.0, 375.0, 187.5 and 93.8 ng/ml for
AM. Quality control (QC) samples were prepared with blank sheep
plasma at low, medium and high concentrations (DHA-G, DHA, AS:
10.2,160.0and1000.0 ng/ml;AM:93.8,750.0and3000.0 ng/ml).AS
or alternatively AM was used as the internal standard (IS). Internal
standard working solutions of 7.5 µg/ml AS or 150.0 µg/ml AM
were diluted with blank sheep plasma (1 : 30, final concentration:
250.0 ng/ml or 5.0 µg/ml, respectively).
Table 1. LC–MS/MS instrumentation settings
Parameter
Value
Glucuronide identification
Source temperature
400 ◦C
15 l/min
15 l/min
26 V
A mass spectrometric approach followed by verification with
an enzyme assay was conducted to assign and identify the
glucuronide metabolite (DHA-G) in our sheep samples.
Nebuliser gas (NEB)
Curtain gas (CUR)
Declustering potential (DP)
Focusing potential (FP)
Entrance potential (EP)
Collision cell exit potential (CXP)
Ion spray voltage (IS)
Plasma samples (10 × 300 µl) of AS or AM-treated sheep[21,22]
were subjected to protein precipitation by the addition of
1000 µl ice-cooled methanol followed by centrifugation for
15 min at 16 100g at 4 ◦C. The supernatants were collected,
stored at 4 ◦C and DHA-G was isolated by HPLC using the
same chromatography program as described above (100 µl
injection volume). Chromatography peaks representing DHA-G
were collected on dry ice (retention time: 8.0–10.5 min). Collected
samples were taken to dryness under nitrogen, resuspended in
230 V
6 V
6 V
5500 V
Collision gas N2 (CAD)
Collision energy (CE)
3 l/min
13 eV
Polarity of analysis
Positive
267.4 → 163.0 m/z
Mass transition for the artemisinins
c
wileyonlinelibrary.com/journal/jms
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2011, 46, 172–181

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Quantification of artesunate, artemether and their major metabolites
acetonitrile/formic acid 0.1% (v/v) and introduced into the MS/MS
by infusion (10 µl/min). First, a Q1 single MS scan from m/z 100
to 1000 was carried out. MS conditions are described in Table 1
(CAD and CE were 0 l/min and 0 eV, respectively). Second, each
signal with a relative intensity >30% was further examined using
the artemisinin-specific transition m/z 267.4 → 163.0 in the SRM
mode(CADandCEwere3 l/minand10–20 eV, respectively). Third,
resultswerecomparedwithreferencespectraofDHA-G(10 µg/ml)
obtained under identical conditions.
investigated at 93.8, 750.0 and 3000.0 ng/ml and the recovery of
DHA-G, DHA and AS at 10.2, 160.0 and 1000.0 ng/ml.
Matrix effects of the artemisinins were assessed as the ratio of
the absolute peak areas of blank plasma samples spiked after the
extraction to the absolute peak areas of the analytes solved in a
mixture of methanol–mobile A (1 : 1, v/v).
Stability: autosampler and freeze–thaw stability
Autosampler stability and freeze–thaw stability studies were
included in our method validation. QC samples at low, medium
and high concentration (as described above) were used to
test stabilities under different conditions. Autosampler stability
was evaluated by the analysis of QC samples (n = 3 per
concentration) over a period of 30 h. For the freeze–thaw stability,
QC samples were frozen at −80 ◦C (1 h) and thawed at room
temperature (1 h). This cycle was repeated three times before
analysis. The concentrations of these samples were compared
with concentrations of freshly prepared QC samples. The drug
solutions were considered as stable with a deviation of not more
than 15% and 20% at the LLOQ.
For the enzyme assay, the collected eluent was taken to
dryness under nitrogen and resuspended in 1000 µl of 75 mM
phosphate buffer, pH of 6.8, containing 1 mg/ml β-glucuronidase.
This mixture was incubated for 120 min at 37 ◦C and agitated at
600 rpm using a Thermomixer (Eppendorf Thermomixer Comfort,
Hamburg, Germany). The β-glucuronidase was inactivated by the
addition of 1000 µl acetonitrile and the incubation mixture was
analysed by LC–MS/MS. A 30-µl aliquot was measured before
(t = 0) and after 120 min exposure to β-glucuronidase.
Method validation
Our method was validated for selectivity, precision, accuracy,
recovery and stability according to the bioanalytical method
validation guidance for industry 2001 by the Food and Drug
Administration (FDA).[27]
Sample dilution
PK samples with drug concentrations exceeding the ULOQ
were diluted using blank sheep plasma. The dilution effect was
evaluatedasdescribedbyHodelet al.[19] toassurethattheaccuracy
oftheanalysiswasnotaffected.Inbrief,blanksheepplasma(n = 3)
was spiked with a concentration exceeding two times the ULOQ.
The samples were then fourfold diluted with blank plasma to
achieve a concentration within the calibration range. The accuracy
of drug quantification was defined as the percentage ratio of the
measuredconcentrationtothenominalconcentration.Adeviation
of 15% of the measured concentration from the nominal value
was accepted in our study.
Selectivity
Blank plasma obtained from sheep of different origin (n = 6) was
examined for interferences with endogenous substances using
the above-described extraction procedure but without adding the
IS working solution.
Lower limit of quantification
The LLOQ was selected as the minimal concentration in plasma
samples, which could be analysed with a precision of ≤20% and
accuracybetween80%and120%.Inaddition,theanalyteresponse
was at least five times above the noise level of the blank response.
Pharmacokinetic study
The PK studies were carried out in the framework of two
efficacy studies published recently.[21,22] Here, we present the
PK profile of one sheep treated with 60 mg/kg AS intramuscularly
to demonstrate the usefulness of the developed and validated
LC–MS/MS method. Blood samples were withdrawn from the
jugular vein into lithium-heparin-coated vacutainer tubes (BD,
Franklin Lakes, NJ, USA) at 0, 0.25, 0.5, 0.75, 1.0, 1.5, 2, 3, 4, 6, 8 and
24 h posttreatment. Plasma was produced by centrifugation and
stored at −80 ◦C. AM (5.0 µg/ml) was used as IS for the analysis.
PK samples with drug concentrations exceeding the ULOQ were
diluted with blank sheep plasma as described above.
Accuracy and precision
Accuracy and precision of the method was evaluated by analysing
QC samples (n = 5) at the LLOQ, a medium concentration (middle
QC) and at the upper limit of quantification (ULOQ). The intra- and
inter-day accuracy/precision was determined within a single run
and between different assays (n = 3), respectively. The following
concentrations were selected for QC samples, covering the entire
range of the calibration curve: 93.8, 750.0 and 3000.0 ng/ml for
AM and 10.2, 160.0 and 1000.0 ng/ml for DHA-G, DHA and AS.
Freshly prepared calibration standards were used for analyses.
The precision was calculated using the coefficient of variation
(CV [%]). The accuracy was determined as the percentage ratio
of the measured concentration to the nominal concentration. A
precision of 15% (LLOQ: 20%) and accuracy between 85% and
115% (LLOQ: 80–120%) was accepted in our study.
Results and Discussion
Method development
Todate, novalidatedanalyticalmethodexistsforthesimultaneous
quantification of the antimalarials AS, AM and their metabolites
DHA and DHA-G for the analysis of plasma samples from PK stud-
ies. We, therefore, decided to develop a LC–MS/MS method for
the quantification of these drugs and metabolites in plasma sam-
ples obtained from F.hepatica-infected sheep. Importantly, DHA-G
seems to be the major glucuronide metabolite of DHA and the pri-
mary elimination occurs over the bile.[3] Furthermore, since adult
Recovery and matrix effect
Relative recoveries (RRE) of the artemisinin derivatives were
determinedbycomparing theabsolutepeakareasofblankplasma
samples spiked before and after the extraction at three different
concentrations (n = 3 per concentration). The recovery of AM was
c
J. Mass. Spectrom. 2011, 46, 172–181
Copyright ꢀ 2011 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms