Ultra-sensitive simultaneous LC-MS/MS quantification of ...bcn2013.europeanbioanalysisforum.eu/slides/PDF/P1-Pep-6 Erin... · ©2013 Waters Corporation 1 Ultra-sensitive simultaneous

©2013 Waters Corporation 1

Ultra-sensitive simultaneous LC-MS/MS

quantification of human insulin, glargine,

lispro, aspart, detemir and glulisine in human

plasma using 2D-LC and a novel high

efficiency column

Erin E. Chambers

Principal Applications Chemist


Outline

Background and Goals

Mass spectrometry development

Chromatography development

Sample Preparation Development

Validation Data

Conclusions


Insulin and Analogs

Insulin A Chain

Insulin B Chain

Human Insulin MW 5808

Insulin A Chain

Insulin B Chain

Insulin aspart (Novalog®) Avg MW 5826

Insulin A Chain

Insulin B Chain

Insulin A Chain

Insulin B Chain

Insulin A Chain

Insulin B Chain

Insulin detemir (Levemir®) Avg MW 5917

Insulin glargine (Lantus®) Avg MW 6063

Insulin glulisine (Apidra®) Avg MW 5823

Insulin A Chain

Insulin B Chain

Humalog (insulin lispro)


Background

Why bioanalysis for insulin analogs?

1. Many coming off patent between 2013 and 2015

o Bioequivalence studies

o Development of new versions

• bioanalysis

2. Methods needed to identify/differentiate specific insulins

• Need simultaneous quantification as combination therapies

common

o Forensic toxicology, cases of wrongful death

o Anti-doping

o Understanding/monitoring of patient dosing?

Current analytical methods

1. ELISA- based assays (lack of standardization)

2. Nano-flow or low flow LC-MS/MS assays

3. SPE-immuno affinity LC-MS/MS assays

4. Assays where insulin has been digested or disulfide bonds reduced


Specific Challenges in Developing an LC-MS/MS Assay for Insulin Analogs

Key challenge: distinguish human

insulin and Humalog (lispro) whilst

obtaining adequate specificity for low

level detection

Obtain sensitivity similar to LBAs

Specificity in matrix

High level of non-specific binding

(NSB)

Low MS sensitivity

– Poor fragmentation

– Multiple precursors

Chromatographic peak shape

Protein binding


Outline


– Challenges in Developing Ultra-Sensitive LC/MS Methods for

Peptides

o Specific insulin challenges




Validation Data

Conclusions


MS scan of Insulin Glargine

m/z 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

%

0

100

Scan ES+ 3.47e7 867.27

759.06

197.03

181.02 256.09

279.23

304.11 527.34

387.32 637.36

764.32

1011.73

874.98 1213.76

1020.66

7+

8+ 6+

5+

Lantus infused at 10 µL/min teed into LC effluent containing 40% ACN


MS Specificity: Avoiding Immonium Ion Fragments

02-Mar-2012

Time 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40

%

0

100

0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40

%

0

100

1: MRM of 4 Channels ES+ 867 > 136 (Lantus)

9.40e5 0.99

0.56

1.68

1.58

1.20 1.15

1.30 1.49 1.40

1: MRM of 4 Channels ES+ 867 > 984 (Lantus)

8.97e5 0.99

867 -> 136 (tyrosine immonium ion)

867 -> 984

Lack of Specificity


MSMS spectra for insulin glulisine, aspart, detemir, and glargine

m/z100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

%

0

100

m/z100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

%

0

100

m/z100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

%

0

100

m/z100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

%

0

1001370

13691161

1161

346

346

328301

1161

1160347447 961512 1092

1046

1162

11621369

11651344

1166

1370

1370

1426

1371 142716891470 1739

971

969

968136661 925

972

972

973

113911231005 1301

12631403

1180

454

13661184

1362

1011

1008

98314311791164

Insulin glulisine MSMS of 5+ 1165

Insulin aspart MSMS of 6+ 972

Insulin detemir MSMS of 5+ 1184

Insulin glargine MSMS of 5+ 1011

*

*

*

*


MSMS spectra from 5+ precursors of human insulin and insulin lispro

m/z100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

%

0

100

m/z100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

%

0

100226

219

201

345

345

292227

248327 652

652346446

405 609

11291065927653

877808758753711

960 13311130 12171159

1241 13561400 1517 1571

217

217

213

1162219

1159230

1159652292270

345361 446

11141065

948

1163

14101391

13311185

1298

1421

1497

Human insulin MSMS of 5+ 1163

Insulin lispro MSMS of 5+ 1163

*

*


MSMS of different Humalog (lispro) Precursors

MSMS of 5+ Precursor 1163 -> 217.3


Lower m/z precursor yields higher intensity but lower signal to noise

Time 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50

%

0

100

0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50

%

0

100

S/N:RMS=633.04

S/N:RMS=202.52

S/N 633

S/N 202


Humalog Sample Analysis: Effect of Higher m/z Precursor

Time 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50

%

0

100

0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50

%

0

100 3.69

2.74

3.42 3.28

3.13

4.33 3.78

3.82

3.86

3.91

4.11

4.34

5.82

5.79

5.32 5.26

5.19 4.92 4.69

5.68

5.50

6.81

) 5.78 4.69

4.14

3.53 3.26 2.75

3.20

4.08

3.59

4.03

4.33

4.34

5.55

5.02

5.03

5.13

5.17

5.19

6.82




Xevo TQ-S Triple Quadrupole MS conditions

Specific Insulin MRM Transition

Cone Voltage

(V)

Collision

Energy (eV)

Glargine 1011->1179 60 25

867->984 60 18

Lispro 1162-> 217 50 40

968.5->217 50 40

Detemir 1184-> 454.4 60 20

1184-> 1366.3 60 20

Aspart 971.8 -> 660.8 60 18

971.8 -> 1139.4 12 18

Glulisine 1165 -> 1370 14 22

1165 -> 346.2 14 22

Bovine (IS) 956.6 -> 1121.2 60 18

Human insulin 1162 -> 226 50 40

968.5->217 50 40

Note: highlighting indicates the primary transitions used for quantification


Outline



Peptides





Validation Data

Conclusions


Evolution of Insulin Method: Traditional C18 to Charged Surface Hybrid (CSH™) C18

Bovine Insulin MW 5734

ACQUITY UPLC BEH C18 1.7 µm 2.1 X 50mm

ACQUITY UPLC CSH C18 1.7 µm 2.1 X 50mm

Peak Width 3.6 sec

Peak Width 11 sec

Time 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40

%

0

100

0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40

%

0

100 1.46

1.29

1147.5 > 315.2


Improvement for insulin using solid core charged surface columns

Time 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

%

0

100

1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

%

0

100 1162 > 217 (Humalog)

1.44e5

4.26

5.20

5.24

5.28 5.81

1162 > 217 (Humalog) 1.44e5

4.32

5.75

4.92 5.02

5.90

6.63 6.80

Time 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

%

0

100

1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00

%

0

100

1165.2 > 1370 (Apidra)

3.87e5

4.27

5.22

5.15 5.31

1165.2 > 1370 (Apidra)

3.87e5

4.35

4.38

4.44 4.92

5.75

Solid core

Fully porous

Humalog Apidra

26% area increase

52% area increase


LC Method: At-column-dilution and Trap and Back Elute

Analytical Column: CORTECS C18+ 2.1 X 50mm, 1.7 µm

Trap column: XBridge C18 IS, 3.5 µm, 2.1 X 20mm

Mobile phase A= 0.1% formic acid in water

Mobile phase B= 0.1% formic acid in ACN

Loading time: 2 minutes

At Column Dilution

Elution

– 15 to 40% B over 4 minutes

Analytical Column Temp: 60°C

Sample Temp: 15°C

Injection Volume: 30 µL (can inject 45 µL without breakthrough)

SNW: 50/25/24/1 ACN/IPA/H20/FA


ACQUITY IClass with 2D Technology: Valve Diagram

Pump 3

MS

Pump 1 (Injector)

WastePump 3

MS

Pump 1

Waste

TC

AC

Pump 2 Pump 2

T T

AC

w Guard w Guard

Pump 1: Loading pumpPump 2: Dilution pumpPump 3: Elution pump

TC

TC= trapping columnAC= analytical column

POSITION 1 POSITION 2


Insulin detemir RT 5.52 min

Time 2.00 4.00 6.00 8.00

%

0

100

2.00 4.00 6.00 8.00

%

0

100

2.00 4.00 6.00 8.00

%

0

100 5.52

4.29

5.26

4.30

3.60

5.26 5.46

UPLC-MS/MS Chromatograms of human insulin, insulin analogs, and bovine insulin (IS)

Time 2.00 4.00 6.00 8.00

%

0

100

2.00 4.00 6.00 8.00 %

0

100

2.00 4.00 6.00 8.00

%

0

100

2.00 4.00 6.00 8.00

%

0

100 4.28

3.60

5.81

4.13

5.28 5.73

4.27

4.04 5.78

4.23

3.21 5.19

Insulin glulisine RT 4.29 min

Insulin lispro RT 4.28 min

Human insulin RT 4.30 min

Insulin aspart RT 4.27 min

Insulin glargine RT 4.13 min

Bovine insulin (IS) RT 4.23 min


Outline



Peptides





Validation Data

Conclusions


PPT followed by Oasis® MAX µElution 96-well plate

PPT: 250 µL human plasma sample precipitated 1:1 with 50/50 ACN/MeOH

+ 1% FA, vortex spin 10 min at 13K rcf,

dilute supernatant with 900 µL 5% NH4OH in water

SPE: Oasis® MAX µElution 96-well plate

Condition: 200 µL methanol

Equilibrate: 200 µL water

Load Sample: entire diluted supernatant in 2 steps of ~ 700 µL each

Wash: 200 µL 5% NH4OH in water

Wash: 200 µL 5% methanol, 1% acetic acid in water

Elute: 2X 25 µL 60% methanol, 10% acetic acid in water

Dilute: 50 µL water

Inject 30 µL

Extraction Conditions

Plasma detection limit: 50 pg/mL


Time 3.90 4.00 4.10 4.20 4.30 4.40 4.50 4.60 4.70 4.80 4.90 5.00 5.10 5.20

%

-1

99

3.90 4.00 4.10 4.20 4.30 4.40 4.50 4.60 4.70 4.80 4.90 5.00 5.10 5.20

%

-1

99

3.90 4.00 4.10 4.20 4.30 4.40 4.50 4.60 4.70 4.80 4.90 5.00 5.10 5.20

%

-1

99 1165.2 > 1370 (Apidra)

4.58e4 Area

4.34 1156

1165.2 > 1370 (Apidra)

4.58e4 Area

4.34 630

1165.2 > 1370 (Apidra)

4.58e4 Area

Current LOD and LLOQ for Insulin Glulisine in Human Plasma

0.05 ng/mL Apidra

0.1 ng/mL Apidra

Blank human plasma


Current Method: Insulin glargine (Lantus) at the LOD and the low QC in Human Plasma

Low QC 150 pg/mL (25 fmol/mL)

LOD 50 pg/mL (8.25 fmol/mL)

Blank human plasma

~248 amol on column at LLOQ

Time 2.60 2.80 3.00 3.20 3.40 3.60 3.80 4.00 4.20 4.40 4.60 4.80 5.00 5.20 5.40 5.60 5.80 6.00

%

-1

99

2.60 2.80 3.00 3.20 3.40 3.60 3.80 4.00 4.20 4.40 4.60 4.80 5.00 5.20 5.40 5.60 5.80 6.00

%

-1

99

2.60 2.80 3.00 3.20 3.40 3.60 3.80 4.00 4.20 4.40 4.60 4.80 5.00 5.20 5.40 5.60 5.80 6.00

%

-1

99

MRM of 8 Channels ES+ 1011 > 1179 (Lantus)

4.13 1129


4.12 391


5.27

5.11

2.76 3.34

3.22

5.06 4.35 4.25

3.98 3.83 3.57 4.85 4.72

5.74

5.57 5.49


Time 4.00 4.20 4.40 4.60 4.80 5.00 5.20 5.40 5.60 5.80 6.00

%

-0

100

4.00 4.20 4.40 4.60 4.80 5.00 5.20 5.40 5.60 5.80 6.00

%

-0

100

4.00 4.20 4.40 4.60 4.80 5.00 5.20 5.40 5.60 5.80 6.00

%

-0

100

MRM of 8 Channels ES+

1162 > 217 (Humalog)

4.28 1565


1162 > 217 (Humalog)

4.27 627


1162 > 217 (Humalog)

5.24

4.57

4.04 4.85

4.68

5.11 5.03

5.81 5.41 5.46

5.73 5.61

Insulin lispro (Humalog) at the LOD and the low QC

Low QC 150 pg/mL (26 fmol/mL)

LOD 50 pg/mL (8.6 fmol/mL)

Blank human plasma


Outline



Peptides





Validation Data

Conclusions


Standard Curve Statistics

Analyte Std. Curve Range pg/mL

Std. Curve

Range

fmol/mL

r2, linear fit,

1/x weighting

Mean %

accuracy

of all

points

Insulin lispro 50-10,000 8.6-1720 0.998 99.99

Insulin glargine 50-10,000 8.3-1650 0.996 99.98

Human insulin 50-10,000 8.6-1720 0.996 100

Insulin detemir 200-10,000 33.8-1690 0.998 96.4

Insulin glulisine 50-10,000 8.6-1720 0.995 100

Insulin Aspart 100-10,000 17.2-1716 0.995 100

For reference: 1µU/mL = 35 pg/mL = 6 fmol/mL Volund et al 1991


Human insulin QC Statistics

Human Insulin Avg basal level was 1937 pg/mL

Inter-day n=9

QC conc.

(pg/mL)

Mean Calc.

Conc. Std Dev % CV

Mean

Accuracy

150 1915.1 125.4 6.5 92.0

750 2542.5 141.0 5.5 94.8

2500 4326.0 146.7 3.4 97.6

7500 9819.0 960.3 9.8 104.0

Intra-day n=3 Basal level was 1872 pg/mL

QC conc.

(pg/mL)

Mean Calc.

Conc. Std Dev % CV

Mean

Accuracy

150 2056.5 16.7 0.8 90.2

750 2506.3 46.6 1.9 99.3

2500 4269.8 206.4 4.8 101.3

7500 10233.2 265.2 2.6 100.3


Insulin lispro QC Statistics

Insulin Lispro

Inter-day n=9

QC conc.

(pg/mL)

Mean Calc.

Conc. Std Dev % CV

Mean

Accuracy

150 144.0 17.5 12.2 96.0

750 721.8 32.3 4.5 96.2

2500 2447.1 202.9 8.3 97.9

7500 7697.5 634.8 8.2 102.6

Intra-day n=3

QC conc.

(pg/mL)

Mean Calc.

Conc. Std Dev % CV

Mean

Accuracy

150 164.6 14.9 9.1 109.8

750 748.2 19.8 2.6 99.8

2500 2417.6 230.4 9.5 96.7

7500 8215.4 243.1 3.0 109.5


Insulin glargine QC Statistics

Insulin Glargine

Inter-day n=9

QC conc.

(pg/mL)

Mean Calc.

Conc. Std Dev % CV

Mean

Accuracy

150 150.1 18.7 12.4 102.7

750 718.4 47.3 6.6 95.8

2500 2369.3 131.2 5.5 94.8

7500 7648.5 511.3 6.7 102.0

Intra-day n=3

QC conc.

(pg/mL)

Mean Calc.

Conc. Std Dev % CV

Mean

Accuracy

150 167.4 16.6 9.9 111.6

750 757.7 62.4 8.2 101.1

2500 2378.0 184.9 7.8 95.1

7500 7949.5 257.9 3.2 106.0


Further Method Assessment and Implementation

Pilot Study with Patient Samples*

– 22 type I and type II diabetic volunteers

o Received one or several insulins

– Dosage regime blind to analytical site

– Results concur with multi-dosing therapies

Human insulin over-spike

– Samples spiked with human insulin at 200X the ULOQ

o Represent possible high levels expected in diabetics

– No interference with quantification of any analogs including lispro

*manuscript submitted


Outline



Peptides





Validation Data

Conclusions


Conclusions/Key Points

Detection limits approx. 4X lower (than previous method) for most analogs

– Only other LC/MS method that reaches these detection limits uses nano-flow and 3-step sample prep involving affinity purification followed by 2 SPE extractions

The use of the CORTECS C18+ column provided significantly improved sensitivity and peak shape for insulin analogs versus charged-surface fully porous columns and traditional C18 columns

– Excellent batch-to-batch reproducibility

2D LC enables higher loading and further cleanup

Selective PPT/mixed-mode SPE cleanup significantly reduces endogenous interferences

Test performed to verify absence of interference when human insulin present at >200X higher concentrations than other analogs

– For example type II diabetics

– No cross-talk or impact on quantification observed

All FDA criteria for accuracy and precision met

– Average accuracies for standard curve points and QC samples were >92%, with most being close to 99%

– Inter- and intra-day precision for all QC samples better than 7.5%

– CV’s of matrix factors, for all analogs, across 6 lots of human plasma were <15%


Waters Corporation

Kenneth Fountain

Stephan Koza

Thomas Wheat

Acknowledgements

Kings College London, UK

Norman Smith

Cristina Legido-Quigley

Janaka Karalliedde

Prof. David Cowan

Documents

Ultra-sensitive simultaneous LC-MS/MS quantification of ...bcn2013.europeanbioanalysisforum.eu/slides/PDF/P1-Pep-6 Erin... · ©2013 Waters Corporation 1 Ultra-sensitive simultaneous