High Resolution Mass Spectrometry of Antibody Drug...

Page 1

High Resolution Mass Spectrometry of Antibody Drug Conjugates Using

the Orbitrap Mass Analyzer

INTRODUCTION

The complexity of modern therapeutic proteins presents

a great analytical challenge. Most often a whole set of

Kai Scheffler1, Eugen Damoc2, Aaron Bailey3, and Jonathan Josephs3

Thermo Fisher Scientific, 1Dreieich, Germany, 2Bremen, Germany, 3San Jose, USA

the Orbitrap Mass Analyzer

RESULTS

There are many factors that play a key role in the analysis

of proteins, some of which relate to sample preparation

Figure 7: A) The Sliding Window feature in BioPharma

Finder provides a composite spectrum based on stepwise

deconvolution results across the full RT range. B) Low

abundant species are represented more accurately and are

less prone to be underrepresented which is in particular useful

Figure 4: LC-MS analysis of intact Brentuximab vedotin

under denaturing conditions: (A) Reversed phase

chromatogram and the resulting averaged MS spectrum

provides a complex mixture of charge state envelopes as well

as vcMMAE-specific reporter fragment ion at m/z 718. (B) Dataa great analytical challenge. Most often a whole set of

analytical techniques is used to characterize these

proteins. With mass spectrometry alone several

complementary methods are required to analyze protein

drugs on the intact protein and on the peptide levels.

Native mass spectrometry on the intact protein level

allows also for the analysis of molecules which rely on

non-covalent interactions to preserve critical structural

features, such as antibody-drug conjugates (ADC). In

addition the use of 100% aqueous buffers in native MS

analysis produces lower charge states detected at

higher m/z values compared to analysis under

denaturing condition and thus improves mass separation

of heterogeneous mixtures [1]. Recent technical

advance-ments on the benchtop Orbitrap mass

spectrometry platform offer now complete

of proteins, some of which relate to sample preparation

(buffers, solvents, additives) while others relate to the

mass spectrometer’s source conditions as well as the

physical environment inside the instrument [2,3]. The Q

Exactive Plus and Q Exactive HF mass spectrometers

(Figure 2A) have previously been introduced with the

Protein Mode option, which was one of many

advancements for intact protein analysis on the Orbitrap

platform. For these two instruments an automated HCD

gas control was introduced by using an electronically

controlled valve for nitrogen gas in the HCD cell for easier

optimization of experimental conditions required for

different types of analyses wished to run on a single

platform.

In Normal Mode pressure settings are factory-optimized,

[ [ [ [ [

0

20

40

60

80

100

Relativ

eIntens

ity

8.0 8.2 8.4 8.6 8.8 9.0 9.2 9.4 9.6 9.8RT(min)

100

mAb

less prone to be underrepresented which is in particular useful

for low abundant, low and high drug loads of ADCs.

A

Rela

tiv

e A

bu

nd

an

ce

6 7 8 9 10 11 12

Time (min)

0

20

40

60

80

100

Reversed phase LC-MS of Brentuximab vedotin yields ~9 masses of denatured species

500 1000 1500 2000 2500 3000 3500 4000 4500

m/z

0

20

40

60

80

100

Rela

tiv

e A

bu

nd

an

ce

Reversed phase LC-MS

50

60

70

80

90

100

25041.47

(32.5 ppm)

2 x G0F

125678.81

(16.0 ppm)

1 x G0F

76675.44

(1.0 ppm)

2 x G0F

103272.01

(5.6 ppm)1 x G0F

54269.74

2 x G0F

148086.47

Rela

tiv

e I

nte

nsit

y

as vcMMAE-specific reporter fragment ion at m/z 718. (B) Data

analysis with ReSpect deconvolution and Sliding Window

integration show roughly six covalently-structured forms of

unraveled cysteine-linked ADC.

A

B

Full MS spectrum

718.5121

762.5026

spectrometry platform offer now complete

characterization of the complex conjugates composed of

small molecule drugs attached to antibodies on a single

instrument platform. In this presentation relevant

workflows covering chromato-graphy, mass spectrometry

and data analysis for ADC characterization are laid out

and data obtained from two different types of ADCs are

presented.

In Normal Mode pressure settings are factory-optimized,

suitable for most analyses and ions are cooled in the C-

trap (Fig. 2B). In Protein Mode the default trapping gas

pressure setting is 0.2 and that corresponds to a ∆HV

which is 5x lower than in Normal Mode. Additionally, ions

are transferred and cooled in the HCD cell and thus have a

longer flight path (Fig. 2C). The new High Mass Range

(HMR) Mode that is especially required for the analysis of

proteins under native conditions when samples are kept in

aqueous buffers with no organic solvents involved at near

neutral pH. For HMR mode the default trapping gas

pressure setting is the same as in Normal Mode and can

even be slightly increased for even improved trapping of

certain species such as protein complexes and

heterogeneous large proteins (e.g. antibody drug

conjugates). The trapping path in HMR mode is the same CONCLUSIONS

Figure 1: Operating modes for the three major

BioPharma workflows: Normal Mode, Protein Mode and

HMR mode Normal Mode

MAbPac RP column

for denatured samples

Peptide Mapping

components

integrated 0

20

40

60

80

100

Relativ

eIntens

ity

8.0 8.2 8.4 8.6 8.8 9.0 9.2 9.4 9.6 9.8RT(min)

0

20

40

60

80

Rela

tiveIntens

ity

8.0 8.2 8.4 8.6 8.8 9.0 9.29.4 9.6 9.8RT(min)

mAb + 2 drugs

mAb + 4 drugs

-100

-80

-60

-40

-20

0

20

40

60

80

100147082.012 148038.597

148996.413

146125.059

149953.100

148262.672149219.126

147306.160145166.647

146347.724 150910.692150179.154

145392.644 148396.374147434.934 151867.722146481.051 151133.091149347.719 152828.100150299.831 153321.699 154341.000

147081.809 148039.009

146124.209 148996.509

147305.809 148262.209149954.309

145169.709 146345.509 149220.509

147438.509 150916.709150173.309148485.909146481.309 149381.309145391.209 152463.109150529.609 151508.509

Mass

0

1

2 34

5

6

78 9

151865

152824153773

152000 153000 154000

7 8 9

BFigure 5: LC-MS analysis of intact Brentuximab vedotin

under native conditions:

A) Size exclusion chromatogram and mass spectrum

obtained from averaging 2 min chromatographic time which

includes all DAR forms (DAR 0-8). B) ReSpect deconvolution

result using the Sliding Window feature. A pattern of lower

abundance species were detected corresponding to a low

abundance loss of 762 Da from each glycoform at each DAR.

Based on the individual deconvolved abundances of the

G0F/G0F glycoform, an average DAR value of 4.07 was

calculated which is consistent with previous reports [4].

0

10

20

30

40

50(5.6 ppm)54269.74

(4.6 ppm)

148086.47

(28.1 ppm)2 x G0F

105907.20

(23.5 ppm)

20000 30000 40000 50000 60000 70000 80000 90000 100000 110000 120000 130000 140000 150000

Mass

Rela

tiv

e I

nte

nsit

y

24279.51

(-761.96 Da)

MATERIALS AND METHODS

Samples:

Samples used in this study are Trastuzumab (Roche,

UK; tradename Herceptin),Trastuzumab emtansine

conjugates). The trapping path in HMR mode is the same

as in Protein Mode with ion cooling taking place in the HCD

cell. And also, mass detection is enabled ranging up to m/z

8000 compared to m/z 6000 in the two other modes.

CONCLUSIONS

• The BioPharma option extends the capabilities of the

Q Exactive Plus and Q Exactive HF mass spectro-

meters allowing for mass detection up to m/z 8000 in

the new High Mass Range (HMR) Mode

• This new operating mode extends the instrument’s

capabilities to cover all three major workflows for

BioPharma characterization and allows for intact

mass analysis of larger proteins such as antibodies

and antibody drug conjugates under native

conditions.

• Native LC/MS intact mass analysis of Brentuximab

SMART digest

MAbPac RP

column

Intact Analysis

native & denatured

Subunit Analysis

FabRICATOR®

(IdeS) digest

and/or

reduction

Acclaim VANQUISH C18 column

MAbPac SEC column

for native samples

Q Exactive Biopharma

• Protein Mode

• Enhanced Resolution Mode

• High Mass Range Mode ( m/z ≤8000)

Vanquish UHPLC

Protein Mode

HMR Mode

©2016 Genovis AB

DAR2

DAR4

DAR6

2 x G0F

153355.40

A

B

5200 5400 5600 5800 6000 6200 6400 6600 6800 7000 7200 7400 7600

m/z

0

10

20

30

40

50

60

70

80

90

100

Rela

tive A

bund

ance

G0F/G0FDAR

Mass

Accuracy

(ppm)

Relative

Abundance

DAR0 11.7 6.77

DAR2 23.1 69.23

DAR4 22.4 100.00

DAR6 40.5 69.75

DAR8 17.6 10.61

Average

Drug-to-Antibody

Ratio (DAR)

4.07

Brentuximab

vedotin

Buffer salts

50 mM

NH4OAc

isocratic

elution

0 2 4 6 8 10

Time (min)

0

20

40

60

80

100

Rela

tive A

bundance

Size Exclusion LC-MS

Figure 2: A) Schematic of the Q Exactive Plus/HF mass

spectrometers and differences in the trapping path in

the three different operating modes available: B) Normal

Mode, C) Protein Mode and D) HMR Mode. E) Illustration of

improvement in signal intensity for +17 charge state of a

mAb light chain comparing Protein Mode and Normal Mode.

Aregion where UK; tradename Herceptin),Trastuzumab emtansine

(Roche UK; tradename Kadcyla) and Brentuximab

vedotin (Seattle Genetics, US; tradename Adcetris).

All antifbody samples were prepared for intact mass

analysis by dissolving the dried substances according to

the manufacturer's instructions. For denaturing LC-MS

intact analysis 1 µg of protein samples were separated

using a 10 min gradient of 10-90% ACN in H2O and

0.1% formic acid (Thermo MAb-Pac RP; flow rate 250

µL/min). For native LC-MS intact analysis 10 µg of

sample was desalted online using size exclusion

chromatography with 50 mM NH4OAc using isocratic

elution at a flow rate of 300 µL/min and directly infused

into the mass spectrometer via electrospray ionization.Native MS intact protein analysis allows direct observation

of molecules which rely on non-covalent interactions to

preserve critical structural features, such as maintaining

• Native LC/MS intact mass analysis of Brentuximab

vedotin resulted in detection of intact ADC forms,

DAR0-8. ReSpect deconvolution and Sliding

Window integration showed an average DAR of

4.07, consistent with previous studies.

• Native LC/MS intact mass analysis of Trastuzumab

emtansine resulted in detection of intact ADC forms,

DAR0-8. ReSpect deconvolution and Sliding

Window integration showed an average DAR of

3.71, consistent with previous studies.

• The sliding window feature in BioPharma Finder

software is essential for accurate determination of

0

20

40

60

80

100

Re

lative

In

ten

sity

148000 149000 150000 151000 152000 153000 154000 155000 156000 157000 158000 159000 160000

Mass

147000

DAR0DAR8

2 x G0F

148086.53

2 x G0F

150720.06

153355.40

,

2 x G0F

155987.81

2 x G0F

158632.20

40

60

80

100

49+

53+52+

51+ 50+48+

60

80

100

Denatured

Figure 6: A) Mass spectra obtained for Trastuzumab

emtansine under denaturing and native conditions. B)

Calculation of DAR ratio based on C) the deconvoluted

spectra obtained using the Sliding Window feature in

BioPharma Finder software.

A

region where

source CID is

applied

Q Exactive Plus:

Standard Orbitrap mass

analyzer

Q Exactive HF:

Ultra High Field Orbitrap

mass analyzer

Trapping in Normal Mode:

trapping gas pressure setting 1

(fixed on setting 1)

Trapping in Protein Mode:

trapping gas pressure default setting 0.2

(range 0.2-1)

Trapping in HMR Mode:

trapping gas pressure default setting 1

(range 1.0-1.5)

1355.5 1356.0 1356.5 1357.0

m/z

0

10

20

30

40

50

60

70

80

90

100

Re

lative

Ab

un

da

nce

x 5.1

Protein Mode

Normal Mode

B

C

D

E

N2

N2

N2

Chromatography:

A Vanquish UHPLC system was used for all LC/MS

experiments. For native analysis, 50 mM ammonium

acetate buffer (99.99%, Sigma Aldrich) was used. On a

size exclusion column. Reversed phase chromatography

was performed with water/0.1% formic acid and

acetonitrile/0.1% formic acid on a MAbPac RP 2.1x50

mm column.

Mass Spectrometry:

Mass spectrometers used in this study were the

commercially available Thermo ScientificTM Q ExactiveTM

Plus and Q ExactiveTM HF equipped with BioPharma

Option. The instruments were operated under Tune 2.8

instrument control software in HMR mode which allows

preserve critical structural features, such as maintaining

interchain associations which hold together cysteine-linked

ADCs. The use of 100% aqueous physiological pH buffers

in native MS analysis produces a fewer number and lower

charge states (increased m/z) compared to denaturing

conditions (Figure 3) and improves mass separation which

is beneficial in particular for heterogeneous mixtures. We

demonstrate this phenomenon using the cysteine-linked

ADC Brentuximab vedotin (Figures 4 and 5) and the

Asparagine-linked ADC Trastuzumab emtansine (Figure 6).

Based on mass spectra acquired under native condition

DAR values for both ADCs were determined that are well in

agreement with previously reported numbers [4,5].

software is essential for accurate determination of

DAR ratios.

REFERENCES

[1] Rosati S. et al. In-depth qualitative and quantitative analysis

of composite glycosylation profiles and other micro-heterogeneity

on intact monoclonal antibodies by high-resolution native mass

spectrometry using a modified Orbitrap. mAbs. 2013;5(6):917-

924. doi:10.4161/mabs.26282.

[2] Fenn J.B. et al. Electrospray ionization for mass spectrometry

of large biomolecules. Science. 1989;246(4926):64-71.

[3] Fenn J..B, Electrospray wings for molecular elephants (Nobel

Lecture). Angew. Chem. Int. Ed. 2003. 42:3871–3894.

[4] Dabaene F. et al. Innovative native MS methodologies for

antibody drug conjugate characterization: High resolution native

100 ADC

Size Exclusion Chromatography

G0F/G1F

DAR

Mass

Accuracy

(ppm)

Relative

Abundance

DAR0 6.49 9.19

DAR1 21.69 34.26

DAR2 0.05 59.03

DAR3 6.81 100.00

DAR4 5.17 91.16

DAR5 6.69 67.42

Average

Drug-to-Antibody

Ratio (DAR)

3.71

2900 2950 3000 3050 31000

20

40

2500 3500 4500 5500 6500 7500m/z

0

20

40

60

80

100

0

20

40

Rela

tive

Ab

un

dan

ce

5950 6000 6050 6100 6150m/z

0

20

40

60

80

10026+

25+

Native

200 m/z

B

buffer salts

*

instrument control software in HMR mode which allows

for detection of a mass range up to 8,000 m/z. The S-

lens RF level was allowed to be increased to a setting of

200 in HMR mode and set to that level for all

experiments shown here. Resolution settings between

15,000 and 70,000 were applied.

Data Analysis:

Data analysis was performed with Thermo ScientificTM

BioPharma FinderTM 2.0 software using the ReSpectTM

algorithm and Sliding Window feature. DAR ratios were

calculated automatically based on deconvolution species

that were identified using the publicly-available FASTA

sequences for Brentuximab vedotin and Trastuzumab

emtansine, a mass tolerance of 50 ppm, and a static

antibody drug conjugate characterization: High resolution native

MS and IM-MS for average DAR and DAR distribution

assessment. Anal Chem. 2014. 86(21):10674-83.

[5] Lazar A.C. et al. Analysis of the composition of immunocon-

jugates using size-exclusion chromatography coupled to mass

spectrometry. Rapid Commun. Mass Spectrom 2005. 19:1806–

1814.

TRADEMARKS/LICENSING

© 2016 Thermo Fisher Scientific Inc. All rights reserved.

FabRICATOR is a trademark of Genovis AB. All other trademarks

are the property of Thermo Fisher Scientific and its subsidiaries.

This information is not intended to encourage use of these

Figure 3. Trastuzumab analyzed under native (A) and

denaturing (B) conditions resulting in highly similar

deconvolution results (C).

0

20

40

60

80

1005923.3

5695.6

6169.9

5484.6

6438.0

3080 3100m/z

3085.6

5900 5950m/z

5923.3

3000 4000 5000 6000 7000

m/z

3000 4000 5000 6000 70000

20

40

60

80

100

Rela

tiv

e A

bu

nd

an

ce

3085.6

2962.23291.2

2904.13366.0

2848.33444.2

3526.13612.1

3797.3

m/z

zoom

zoom

z=+25

z=+48

native

denatured

60 Th

30 Th

R=3462

R=4359

Deconvolution

4.7ppm

A

B

0 2 4 6 8 10

Time (min)

0

50

5600 5800 6000 6200 6400 66000

20

40

60

80

100

25+

zoom

5950 6000 6050 6100 6150 62000

20

40

60

80

100

R=70,000

DAR5 6.69 67.42

DAR6 15.20 40.46

DAR7 6.28 24.28

DAR8 3.78 3.84

C

buffer salts

modification of Gln>Pyro-Glu for the heavy chain. products in any manner that might infringe the intellectual

property rights of others.

-100

-80

-60

-40

-20

0

20

40

60

80

100148055.503

148217.603

148380.303

147909.303147848.203 148542.203

148056.888

148218.688

148379.888

148540.188147908.488147850.188

-2.9 ppm

5.0 ppm

1.2 ppm

-2 ppm

4.1 ppm

native

denatured

Rela

tiv

e I

nte

nsit

yC

5950 6000 6050 6100 6150 6200m/z

0

20

40

60

80

100

148000 149000 150000 151000 152000 153000 154000 155000 156000

Mass

D* *

**

** ** *