1. 0.2 0.4 0.6 0.8
0
20
40
60
80
100
120
140
Time, min
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
20
40
60
80
100
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
200
400
600
800
1000
1200
1400
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
200
400
600
800
1000
1200
1400
1600
Intensity,cps
• 96-Well protein precipitation extraction of 50 µL
of K2EDTA human plasma
• The supernatant (50 µL) was diluted with 100 µL
of 2:8 acetonitrile/water
• Step Gradient: 10% to 70% MP B
• Retention Time: 0.43 minutes
• Total Cycle Time: 2 minutes
• HPLC Column 2: Thermo Hypercarb
• Calibration Range: 1 to 1000 ng/mL
Chromatographic System 1
HPLC Column 1 HPLC Column 2
Programmed
Mobile Phase
Method
1
Method
2
Method
3
Method
4
Method
5
Switching Valve
Chromatographic System 2
Indicates the location of the stable-isotope label(s).
Getting the Most Out of Your Mass Spectrometer – Multiplexing LC/MS/MS Assays for High Throughput
Todd Lusk, Megan Benson, James Lauzun, and Daniel Mulvana
Quintiles, 19 Brown Road, Ithaca, New York
Overview
Purpose
• To demonstrate multiplexed analysis of five separate high-throughput method
using liquid chromatography with tandem mass spectrometry (LC/MS/MS) assays
autonomously on a single mass spectrometer (AB SCIEX API 4000) for discovery
applications.
Method
• Five separate methods were analyzed in a high-throughput format using two
interchangeable chromatography systems and two high-performance liquid
chromatography (HPLC) columns on a single mass spectrometer with autonomous
software control.
• Common LC mobile phase (MP) solutions, chromatographic conditions, and
mass spectrometry conditions were used; however, the gradient conditions were
individualized to each analyte.
• Method 5 for niacin and its metabolites used extraction and LC conditions required for
separation of the analytes. The other method conditions were selected based on a
“best-of-fit” model for the LC system required for Method 5.
Results
• Five analytical methods were run in sequence employing switching valves for two
chromatographic systems, as well as LC column switching and diversion of hte flow to
either waste or the mass spectrometer.
• This system ran five separate LC/MS/MS methods each with 2 minute average cycle
time totaling 950 injections in 32 hour analysis time under autonomous control.
Introduction
Conclusions
• By multiplexing LC/MS/MS methods, five separate high-throughput assays
were sequentially analyzed autonomously on a single mass spectrometer.
• Multiplexing interchangeable parallel systems increased adaptability because
independence of the chromatographic systems allowed for alterations while the
LC/MS/MS was running on the alternate system.
• Starting conditions can be quickly altered for optimization and the
system components and solvent systems can be updated to “best-of-fit”
chromatographic conditions to multiplex several assays on interchangeable
systems.
• Method development can be expedited because short cycle time allowed in situ
changes to be made quickly. Changes in MP and NW solution and switches
between solvent systems can be quickly managed manually or automated to
test a spectrum of chromatographic conditions.
• Software control of the instruments and components allowed for complete
control of the system for automated method changes and remote operation of
LC/MS/MS system components.
• Method 5 was developed to ensure the elution of niacin prior to nicotinamide
due to high concentrations of nicotinamide that prevented accurate integration
of niacin peaks at the high flow rate employed for increased sample throughput.
• The multiplexed system has successfully analyzed up to three GLP acceptance
criteria assays (≤15% or ≤20% at the LLOQ); however, the cycle time reflected
the increased cleanup time of the LC system that prevented true high-
throughput analysis.
• In the future, the multiplex components should translate easily into setup for
two-dimensional chromatography and in-line sample extraction.
Accuracy and Precision in Quality Control SamplesHPLC System Schematic and Common Conditions
Challenges in Method Development
HPLC Methods Sequence Overview Method 3/Chromatography System 2 ‒ Loperamide
Method 4/Chromatography System 2 ‒ Tramadol
Method 5/Chromatography System 2 ‒ Niacin,
Nicotinamide, and Nicotinuric Acid
• 96-Well protein precipitation extraction of 50 µL
of K2EDTA human plasma
• The supernatant was evaporated and reconstituted
with 75 µL of 2:8 acetonitrile/water
• Step Gradient: 10% to 60% MP B
• Retention Time: 0.37 minutes
• Total Cycle Time: 2 minutes
• HPLC Column 1: Waters XBridge BEH C18
• Calibration Range: 0.05 to 50 ng/mL
• 96-Well protein precipitation extraction of 50 µL of K2EDTA
human plasma.
• Standard working solutions were prepared in
2:8 acetonitrile/water
• QC samples were prepared with charcoal-stripped (twice)
plasma to remove the endogenous levels of analytes in the
matrix.
• The organic layer was evaporated and reconstituted with
75 µL of 5:95 1% acetic acid in acetonitrile/water
• Linear Gradient: 8% to 32% MP B
• Retention Times: 0.44 minutes for niacin, 0.51 minutes for
nicotinamide, and 0.64 minutes for nicotinuric acid
• Total Cycle Time: 2 minutes
• HPLC Column 2: Thermo Hypercarb
• Calibration Ranges: 50 to 25000 ng/mL for niacin and 10 to
5000 ng/mL for nicotinamide and nicotinuric acid
0.2 0.4 0.6 0.8
Time, min
0.0
2.0e4
6.0e4
1.0e5
1.4e5
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
5
10
15
20
25
30
35
40
Intensity,cps
Diclofenac
Selectivity Blank
m/z 296.1 → 214.1
Diclofenac
LLOQ QC
m/z 296.1 → 214.1
[13C6]-Diclofenac
m/z 302.1 → 220.1
0.2 0.4 0.6 0.8
Time, min
0.0
2.0e4
4.0e4
6.0e4
8.0e4
1.0e5
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
500
1000
1500
2000
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
5
10
15
20
25
30
Intensity,cps
Loperamide
Selectivity Blank
m/z 477.1 → 266.1
Loperamide
LLOQ QC
m/z 477.1 → 266.1
Loperamide-d6
m/z 483.1 → 272.1
0.2 0.4 0.6 0.8
Time, min
0.0
2.0e4
4.0e4
6.0e4
8.0e4
1.0e5
1.2e5
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
500
1000
1500
2000
2500
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
50
100
150
200
Intensity,cps
Tramadol
Selectivity Blank
m/z 264.1 → 58.1
Tramadol
LLOQ
m/z 264.1 → 58.1
Tramadol-d6
m/z 270.1 → 64.1
Common Chromatographic Conditions
Autosampler: LEAP CTC PAL
Pumps: Shimadzu LC-20AD
Controller: SCL-10A
Switching Valves: Valco 2-position and Shimadzu FC-20
Flow Rate: 1 mL/min
Injection Volume: 5 µL
Common Mass Spectrometry Conditions
Detection: AB SCIEX API 4000 with Turbo Ion
Spray ionization in positive ion mode
Software: Analyst 1.4.2
Acquisition Time: 1 minute
Chromatography System 1
MP A: 0.1% Ammonium hydroxide in water
MP B: 0.1% Ammonium hydroxide
in acetonitrile
NW 1: 1% Ammonium hydroxide in water
NW 2: 1% Ammonium hydroxide in acetonitrile
Chromatography System 2
MP A: 25 mM Ammonium formate
with 0.1% formic acid in water
MP B: 0.1% Formic acid in acetonitrile
NW 1: 1% formic acid in water
NW 2: 1% Formic acid in acetonitrile
HPLC Column 1: Waters XBridge BEH C18
(20 x 2.1 mm, 5-µm particle size)
HPLC Column 2: Thermo Hypercarb
(50 x 2.1 mm, 5-µm particle size)
N
O
O
CH3
CH3
CH3
H3C CH3
NO
O
H3C
H3C
N
H
O
O
CH3
CH3
H3C CH3
NO
O
H3C
H3C
Method 2/Chromatography System 1 ‒ Diclofenac
• 96-Well protein precipitation extraction of 50 µL
of K2EDTA human plasma
• The supernatant (50 µL) was diluted with 100 µL
of 2:8 acetonitrile/water
• Linear Gradient: 10% to 25% MP B
• Retention Time: 0.35 minutes
• Chromatography System 1
• HPLC Column 1: Waters XBridge BEH C18
• Calibration Range: 1 to 1000 ng/mL
H
N
Cl
Cl
O
HO
N
OH
O
N
NH2
O
N
H
O
N
O
OH
Method 1/Chromatography System 1 ‒ Verapamil and Norverapamil
• 96-Well protein precipitation extraction of 50 µL
of dipotassium ethylenediaminetetraacetic acid
(K2EDTA) human plasma
• The supernatant (50 µL) was diluted with 100 µL
of 2:8 acetonitrile/water
• Step Gradient: 40% to 70% MP B
• Retention Times: 0.36 minutes for norverapamil
and 0.41 minutes for verapamil
• Total Cycle Time: 2 minutes
• HPLC Column 1: Waters XBridge BEH C18
• Calibration Range: 0.4 ng/mL (LLOQ; lower limit
of quantitation) to 200 ng/mL (ULOQ; upper limit
of quantitation)
Norverapamil
C26H36N2O4
Exact Mass: 440.27 Da
Verapamil
C27H38N2O4
Exact Mass: 454.28 Da
Diclofenac
C14H11Cl2NO2
Exact Mass: 295.02 Da
Loperamide
C29H33ClN2O2
Exact Mass: 476.22 Da
Tramadol
C16H25NO2
Exact Mass: 263.19 Da
Niacin
C6H5NO2
Exact Mass: 123.03 Da
Nicotinamide
C6H6N2O
Exact Mass: 122.05 Da
Nicotinuric Acid
C8H8N2O3
Exact Mass: 180.05 Da
0.2 0.4 0.6 0.8
Time, min
0
100
200
300
400
500
600
700
800
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
500
1000
1500
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0.0
1.0e4
2.0e4
3.0e4
4.0e4
5.0e4
6.0e4
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
100
200
300
400
500
600
700
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
1000
2000
3000
4000
5000
6000
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0.0
5.0e4
1.0e5
1.5e5
2.0e5
2.5e5
3.0e5
3.5e5
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
200
400
600
800
1000
1200
1400
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0
1000
2000
3000
4000
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0.0
5.0e4
1.0e5
1.5e5
2.0e5
Intensity,cps
Niacin
Selectivity Blank
m/z 124.0 → 80.0
Niacin
LLOQ QC
m/z 124.0 → 80.0
Niacin-d3
m/z 127.0 → 83.0
Nicotinamide
Selectivity Blank
m/z 123.0 → 80.0
Nicotinamide
LLOQ QC
m/z 123.0 → 80.0
Nicotinamide-d4
m/z 127.0 → 84.0
Nicotinuric Acid
Selectivity Blank
m/z 181.0 → 135.0
Nicotinuric Acid
LLOQ QC
m/z 181.0 → 135.0
Nicotinuric Acid-d4
m/z 189.0 → 139.0
Niacin Nicotinamide
CV% (coefficient of variation) = standard deviation/mean x 100
RE% (relative error) = (calculated mean – nominal value)/nominal value x 100
Acknowledgments: S. Spencer, Ken Ruterbories, and Matt Chapple
Verapamil
Selectivity Blank
m/z 455.3 → 165.1
Verapamil
LLOQ QC
m/z 455.3 → 165.1
Verapamil-d6
m/z 461.3 → 165.1
Norverapamil
Selectivity Blank
m/z 441.3 → 165.1
Norverapamil
LLOQ QC
m/z 441.3 → 165.1
Norverapamil-d6
m/z 447.3 → 165.1
0.2 0.4 0.6 0.8
Time, min
0.0
5.0e4
1.0e5
1.5e5
2.0e5
2.5e5
Intensity,cps
0.2 0.4 0.6 0.8
Time, min
0.0
5.0e4
1.0e5
1.5e5
Intensity,cps
Method/
Analyte
Expected
Concentration
(ng/mL)
Sample
Name Mean
Precision
(CV%)
Accuracy
(RE%)
Method 1 0.400 LLOQ QC 0.374 10.5 -6.5
Verapamil 1.20 QC1 1.21 2.4 0.5
60.0 QC2 57.4 3.2 -4.4
150 QC3 144 3.5 -4.2
200 ULOQ QC 195 3.0 -2.6
Method 1 0.400 LLOQ QC 0.338 6.5 -15.4
Norverapamil 1.20 QC1 1.12 5.6 -6.9
60.0 QC2 57.0 11.1 -5.0
150 QC3 145 5.1 -3.3
200 ULOQ QC 195 3.2 -2.7
Method 2 1.00 LLOQ QC 1.03 6.9 3.1
Diclofenac 3.00 QC1 3.00 7.0 0.0
300 QC2 285 3.3 -4.9
750 QC3 744 7.1 -0.9
1000 ULOQ QC 938 5.5 -6.2
Method 3 0.0500 LLOQ QC 0.0531 3.0 6.1
Loperamide 0.150 QC1 0.148 10.5 -1.2
15.0 QC2 15.2 1.8 1.6
37.5 QC3 37.8 2.3 0.7
50.0 ULOQ QC 50.0 1.1 0.1
Method 4 1.00 LLOQ QC 0.990 6.3 -1.0
Tramadol 3.00 QC1 2.99 6.2 -0.3
300 QC2 294 4.2 -1.8
750 QC3 702 1.0 -6.4
1000 ULOQ QC 920 0.9 -8.0
Method 5 50.0 LLOQ QC 55.4 10.5 10.8
Niacin 150 QC1 161 6.4 7.1
1000 QC2 1080 3.4 7.7
18800 QC3 19800 7.1 5.8
25000 ULOQ QC 25600 8.0 2.5
Method 5 10.0 LLOQ QC 12.4 5.7 24.4
Nicotinamide 30.0 QC1 36.6 4.8 21.8
200 QC2 235 3.0 17.7
3750 QC3 4090 4.9 9.1
5000 ULOQ QC 5200 3.7 4.1
Method 5 10.0 LLOQ QC 13.7 12.9 36.8
Nicotinuric 30.0 QC1 31.8 9.6 6.1
Acid 200 QC2 205 2.5 2.6
3750 QC3 3740 3.7 -0.3
5000 ULOQ QC 5120 4.7 2.4
Analysis Sequence
Method 5 showed a positive bias at the LLOQ across all three analytes. The
bias was not related to the multiplexed analysis as reinjection of the run in a
separate analysis showed similar bias (data not shown).
Test
injections
and tuning
for
optimization
for all
analytes
System
suitability of
all methods
for selectivity
and signal-to-
noise ratio
Method 1
Analysis
Method 2
Analysis
Valve
switching
program to
change MP
and NW
solutions
Method 3
Analysis
Method 4
Analysis
Method 5
Analysis
Instrument
Shutdown
Reset to
initial
conditions
Mini-cycle of
methods in
sequence
MP solution
equilibration
solution purge
and NW
• Development of high-throughput LC/MS/MS assays has facilitated increased workload
and decreased cycle time (<2 minutes); however, this may result in instrument idle time.
• Multiplexing reduces the need for redundancy in expensive mass spectrometry systems.
An autosampler setup of twelve 96-well plates gives a capacity of ~1000 samples per
mass spectrometer.
• Clinical applications require multiple methods for therapeutic screening that are run on a
single system to minimize analysis time to provide quick results.
• Efficiency of compound screening, where several different shorter assays are required,
has increased using high-throughput multiplexing.
• Operating multiplexed chromatographic systems in parallel enabled changes to be
made to one part of the system while the instrument was running, virtually eliminating
instrument downtime for routine tasks.
• The use of LC switching valves under Analyst 1.4.2 software control enabled analysis of
multiple runs in sequence autonomously maximizing instrument time. Software control
allows for autonomous system actions including programed changes of MP systems and
system equilibration.
• Implementing Analyst 1.4.2 software control allowed access to the system to make
changes remotely (i.e., repeating analysis or preparing the instrument for the next analysis).
• Method development was approached from a “best-of-fit” model, in which the best
choices of MP and needle wash (NW) solutions for all five methods were employed as
the two interchangeable chromatographic systems.
• Determining the conditions acceptable for all analytes was conducted the first day
of method development. Methods for the analysis of loperamide and tramadol used
a standard system, which was adapted for the analysis of diclofenac, verapamil/
norverapamil, and niacin/metabolites. On the following day, preparation and extraction
of all analytical runs was performed concurrently with LC/MS/MS system performance
evaluation and analysis of the five methods began in the afternoon for the 32 hour duration.
• A standard starting gradient was used for initial retention, peak shape considerations, and
LC column determinations, and then was amended to achieve the best chromatography
using the existing MP systems.
• The high-throughput nature of the compound screening methods allowed for fast
in situ method development. New LC conditions were tested per each injection and LC
gradients were altered in real time to achieve an optimized method in minimal time.
• For problematic separations or analytes requiring more sensitivity or increasing signal-
to-noise ratio, MP solutions and HPLC columns were changed and additional analytes
re-tested to confirm the viability of new LC conditions.
• An examination of the acceptability of running the multiple systems in tandem was
conducted to determine the best sequence of injection for the runs to eliminate possible
sources of suppression from previous HPLC MP or NW solutions.
• Increasing the flow rate during the post-elution column wash phase reduced column
carryover and minimized the autosampler wash time to shorten the overall cycle time.
• Calibration of the MS by infusion was streamlined by the use of mixed analyte infusion
solutions.
N
OH
O
CH3
CH3
lCN
O N
OH
CH3
CH3