modernizing the platform characterization peptide map for ... · modernizing the platform...

Modernizing the Platform Characterization Peptide Map for

Accurate Assessment of Deamidation and

Isomerization by LC-MS/MS

Lisa A. Marzilli, Elaine Stephens, Michelle English,

Shibu Philip* and Jason C Rouse

Analytical R&D, Pharmaceutical Sciences, Andover, MA, *Pearl River, NY

Analytical Technology-CASSS March 16, 2017

Antibody Considerations as a Drug Candidate

Typical parameters in selecting lead antibody:

2

affinity

selectivity

activity manufacturability

stability

poly-reactivity

aggregation

propensity

viscosity

Antibodies have a propensity to degrade:

manufacturing, storage and in vivo administration

Determine Degradation

Hotspots

Sequence Liabilities or

CQAs?

Engineer to Eliminate Sequence Liability?

Early Product Development Challenge

3

Oxidation: Met & Trp

Deamidation: Asn and Gln

Isomerization: Asp

Glycation: Lys

Truncations: Asp-Pro

Molecular modeling

Peptide mapping

Does hotspot impact

antigen binding?

Focus on

complementarity

determining regions

(CDRs)

Constant region

hotspots may be

potential liabilities

Deamidation/Isomerization Scheme

4

Asn deamidation reactions proceed more

quickly at elevated pH and temperature

N+1 residue (NG, NS, NT, and NN)

Geiger T. and Clarke S., J of Biol. Chem., (1987), 262 (2), 785.

Asn, N

Asp, D

IsoAsp, isoD

Asp isomerization reaction rate most

impacted by lower pH and N+1 residue

N+1 residue (DG, DS, DT, and DD)

Li et al, J. Pharm. Sci., (2013), 102 (3), 947.

The Ideal Peptide Map

• Robust and easy to perform

• Complete sequence coverage

• High quality separations (in a MS-friendly solvent)

• High quality UV chromatograms

• Minimal method-induced artifacts (short digestion times, low pH)

• Accurate targeted quantitation (Extracted Ion Chromatograms)

5

Developed in-house, automated software to

highlight potential “hotspots”

6

Trastuzumab L Chain CDR Hotspots

1 DIQMTQSPSSLSASVGDRVTITCRASQDVN 30

31 TAVAWYQQKPGKAPKLLIYSASFLYSGVPS 60

61 RFSGSRSGTDFTLTISSLQPEDFATYYCQQ 90

91 HYTTPPTFGQGTKVEIKR

Trastuzumab H Chain CDR Hotspots

1 EVQLVESGGGLVQPGGSLRLSCAASGFNIK 30

31 DTYIHWVRQAPGKGLEWVARIYPTNGYTRY 60

61 ADSVKGRFTISADTSKNTAYLQMNSLRAED 90

91 TAVYYCSRWGGDGFYAMDYWGQGTLVTVSS 120

121

Automatically

underlines

CDR1, CDR2

and CDR3

Highlights

potential

hotspots in

blue

Case Study #1 Using Trastuzumab (Herceptin®) as

a Benchmark for Peptide Map Method Development

7

Trastuzumab L Chain CDR Hotspots Rel. Abundance

1 DIQMTQSPSSLSASVGDRVTITCRASQDVN 30 High level (>5%)

31 TAVAWYQQKPGKAPKLLIYSASFLYSGVPS 60 Low level (1-5%)

61 RFSGSRSGTDFTLTISSLQPEDFATYYCQQ 90 Potential (<1%)


Trastuzumab H Chain CDR Hotspots Rel. Abundance

1 EVQLVESGGGLVQPGGSLRLSCAASGFNIK 30 High level (>5%)

31 DTYIHWVRQAPGKGLEWVARIYPTNGYTRY 60 Low level (1-5%)

61 ADSVKGRFTISADTSKNTAYLQMNSLRAED 90 Potential (<1%)


Case Study #1 Using Trastuzumab (Herceptin®) as

a Benchmark for Peptide Map Method Development

8

Trastuzumab L Chain CDR Hotspots Rel. Abundance

1 DIQMTQSPSSLSASVGDRVTITCRASQDVN 30 High level (>5%)

31 TAVAWYQQKPGKAPKLLIYSASFLYSGVPS 60 Low level (1-5%)

61 RFSGSRSGTDFTLTISSLQPEDFATYYCQQ 90 Potential (<1%)


Trastuzumab H Chain CDR Hotspots Rel. Abundance

1 EVQLVESGGGLVQPGGSLRLSCAASGFNIK 30 High level (>5%)

31 DTYIHWVRQAPGKGLEWVARIYPTNGYTRY 60 Low level (1-5%)

61 ADSVKGRFTISADTSKNTAYLQMNSLRAED 90 Potential (<1%)


3

Digestion Time (h)

% I

so

me

riza

tio

n / S

ucc

inim

ide

Digestion Time (h)

D102G (H-CDR3)

% I

so

me

riza

tio

n / S

ucc

inim

ide

0

2

4

6

8

10

12

0 10 20

pH 7.5

pH 8.2

Platform

0

2

4

6

8

10

12

0 1 2

pH 7.5

pH 8.2

2

Digestion Time (h)

% D

ea

mid

ati

on

/ S

uc

cin

imid

e

Digestion Time (h)

N30T (L-CDR1)

% D

ea

mid

ati

on

/ S

uc

cin

imid

e

0

2

4

6

8

10

12

14

0 10 20

pH 7.5

pH 8.2

Platform

0

2

4

6

8

10

12

0 1 2

pH 7.5

pH 8.2

N55G (H-CDR2)

0

10

20

30

40

50

0 10 20

pH7.7

pH8.2

Platform

0

1

2

3

4

5

6

7

8

9

0 1 2

pH 7.5

pH 8.2

Digestion Time (h)

Asn Deamidation and Asp Isomerization vs. Trypsin

Digestion Time (Time Course) for Trastuzumab (Herceptin®)

Digestion Time Can Influence Level of Deamidation

9

%D

eam

idati

on

/ S

uccin

imid

e

%D

eam

idati

on

/ S

uccin

imid

e

Deamidation for Innovator Trastuzumab

10 10 10

Digestion Time (h)

% D

ea

mid

ati

on

/ S

ucc

inim

ide

Digestion Time (h)

Pfizer: N55G (H-CDR2) %

Dea

mid

ati

on

/ S

ucc

inim

ide

5

•Time-course digestion facilitates the

accurate measurement of deamidation at

T=0

•Digest for 30, 60 and 120 min

•Report accurate deamidation levels

Literature

% Deamidation Level at

N55G

(H-CDR2)

Genentech Harris

2001 ~1 (CEX-HPLC)

Roche Sydow

2014 4 (pH 6.0 map)

Pfizer Results

Innovator

Trastuzumab

44 to 2

New Trypsin Peptide Map

Shorter Digestion Times and Time-Course

11

•Time-course digestion facilitates the accurate

measurement of deamidation at T=0

•Digest for 30, 60 and 120 min

• Efficient reduction, alkylation and digestion with

Lys-C/trypsin (Promega) (pH 8.2)

• Waters XSelect XP CSH C18 column(formic acid)

• Reproducible UV chromatographic profiles

• Analyzed on a Thermo Orbitrap Fusion™ Tribrid™

Mass Spectrometer

• Unprecedented depth of analysis

• High quality MS/MS data (CID, ETD)

• Use Fusion for PTM discovery

Peptide Mapping

Goal: minimize

method-related

artifacts

Low-artifact Trypsin Digest Methods

12

12

Denaturation /

Reduction Alkylation Tryptic Digestion

Buffer exchange

Using Bio-Spin 6

7M GdHCl /

100mM Tris pH 8.2

DTT

30 min at 37 ˚C

IAA

30 min, dark,

RT

Buffer exchange

Bio-Spin 6 column

(Bio-Rad)

130 µL load volume

Digest at

pH 8.2

7M GdHCl /

20 mM MES (pH 6.0)

TCEP

60 min at 37 ˚C

Lys-C/trypsin (Promega)

18 hours at 37˚C in 0.5 mM TCEP

in 20 mM MES (pH 6.0)

1:10 enzyme: protein (w:w).

Digestion quenched using 1% TFA

Buffer exchange

Bio-Spin 6 column

(Bio-Rad)

130 µL load volume

Digest at

pH 6.0

Lys-C/trypsin (Promega)

30, 60, and 120 min

37˚C in 100 mM Tris (pH 8.2),

1:10; enzyme: protein (w:w).

Digestion quenched using 1% TFA

0

1

2

3

4

5

6

0 1 2

pH 7.5

pH 8.2

Time (h)

% d

ea

mid

atio

n

Elaine Stephens

Asn Deamidation and Asp Isomerization in

Trastuzumab (Herceptin®) Using Various Methods

13

Site Modification Method % Modification

N30T (L-CDR1) Deamidation

CEX-HPLC (UV)1 ~15

Trypsin pH 6.0 (overnight) map 2 11

PFE-Trypsin pH 7.5 (overnight) map 12

PFE-Lys-C/Trypsin pH 8.2 (T=0) map 10

N55G (H-CDR2) Deamidation

CEX-HPLC (UV)1 ~1


PFE-Trypsin pH 7.5 (overnight) map 44


D102G (H-CDR3) Isomerization

CEX-HPLC (UV)1 ~9


PFE-Trypsin pH 7.5 (overnight) map Not detected


1. R.J. Harris, B. Kabakoff, F.D. et al, J of Chromat. B, Biomedical sciences and applications, 752, 233-245 (2001). 2. J.F. Sydow, F. Lipsmeier, V. et al, PLoS One, 9, e100736 (2014).

Efficient Separation and ETD Spectra

Confirm ID of H12 (D102G) and H12 (iso-D102G)

14

400 600 800 1000 1200 1400 1600 1800 2000

m/z

697.3

929.3

1393.9

1155.4660.5

892.7 958.31328.4477.3

1466.7753.7 1234.7564.4390.3 1786.01506.9 1915.9

697.2

929.4

1394.01045.4

892.6660.6 1155.51466.7477.3 753.8

1785.01506.91916.9

z4

z5z6

[M+4H]4+

[M+3H]3+

[M+2H]2+

z17 z18z15

c12

C232+

z15

Z182+ C25

2+

H12 (D102G)

Peak 1

H12 (iso-D102G)Peak 2

C3+57 [Z23-57]2+

H12 peptide (WGGD102GFYAMDYWGQGTLVTVSSASTK)

* *

C242+C20

2+

Z172+

318.3

318.2

c4433.3

c3

490.3

490.3

c5

564.3

of in H-

Peak 1

H12 (D102G)

Peak 2

H12 (Iso-D102G)

Non-stressed Stressed

(14 days

@ pH8)

Rela

tive A

bun

dance

Time (min) Time (min)

Peak 1

H12 (D102G)

Peak 2

H12 (Iso-D102G)

• Waters XSelect XP CSH

C18 column with formic acid

can resolve isomerized peak

Extracted Ion

Chromatogram (EIC) for H12

Isomerization of Asp-102 led to lower potency (Harris et al, 2001)

Diagnostic ion (C3 + 57) Confirms Presence of

isoAsp in Trastuzumab

15

C N

H

C

O

C H

H

N

CH2

COOH

O

c4 c3

C N

H

C

O

C H

H

N CH2

COOH

O

c4 c3 c3 + 57 = 375

Iso-D102G

Peak 2

D102G

Peak 1

350 400 450 m/z

100

100

Rela

tive A

bundance

z5

477.3

z4

390.3 c4

433.3

477.3

c3+57 375.2

Rela

tive A

bundance

c3

318.3

c3

318.3

z4

390.3 c4

433.3

a5

a5

z3

z3

J.J. Cournoyer, J.L. Pittman, V.B. Ivleva, E. Fallows, L. Waskell, C.E. Costello, P.B. O'Connor, Prot. Sci., 14,

452-463 (2005).

of in H-

Peak 1

H12 (D102G)

Peak 2

H12 (Iso-D102G)

Non-stressed Stressed

(14 days

@ pH8)

Rela

tive A

bun

dance

Time (min) Time (min)

Peak 1

H12 (D102G)

Peak 2

H12 (Iso-D102G)

• Waters XSelect XP CSH

C18 column can resolve

isomerized peak

Extracted Ion

Chromatogram (EIC) for H12

Case Study #1

Peptide Map of Trastuzumab (Herceptin®)

16

Trastuzumab (Herceptin®) is useful as a benchmark for

peptide map method development

Artifactual deamidation is a real possibility

Some motifs are sensitive to digestion time and pH (ex.

N55G in H-CDR2)

Resolution of Asp isomerized peptides (D102G vs isoD102G)

is challenging

Peptides have the same mass

Can only quantitate Asp/isoAsp if separated

17

L Chain CDR Hotspots

1 XXXXXXXXXXXXXXXXXXXXXXXRSSQSLA 30

31 NSYGNTFLSXXXXXXXXXXXXXXXGISNRF 60

61 SXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 90

91 XXXLQGTHQPYTXXXXXXXX

H Chain CDR Hotspots

1 XXXXXXXXXXXXXXXXXXXXXXXXXGYRFT 30

31 NYWIHXXXXXXXXXXXXXXGINPGNNYATY 60

61 RRKFQGXXXXXXXXXXXXXXXXXXXXXXXX 90

91 XXXXXXXXEGYGNYGAWFAYXXXXXXXXXX 120

121

Case Study #2: Deamidation in an IgG4 mAb

Case Study #2: Deamidation in an IgG4 mAb

18

Three distinct acidic isoforms (~15%)

Acidic fractions collected and characterized for structure/function

Acidic

CEX-HPLC Profile

Bulk

B Ma

in

Should

er

A3 A2 A1 AA

AU

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

Minutes

120.00 121.00 122.00 123.00 124.00 125.00 126.00 127.00 128.00 129.00 130.00 131.00 132.00 133.00 134.00 135.00

Case Study #2: UV Chromatograms of Trypsin

Peptide Maps

H12 (N103Y)

A1 H12 (D/isoD103Y

A2 H12 (D/isoD103Y)

CEX-HPLC Fraction A1 and A2 show an increase in H12 deamidation

Antigen binding impacted A1 charge variant

CEX-HPLC Profile

Red = CEX Main Black = CEX A1 Blue = CEX A2 Green = CEX A3

19

20

Case Study #2: New Peptide Map Column for

Enhanced Chromatographic Separation

0

10

20

30

40

50

60

70

80

90

100

Rela

tive

Ab

un

da

nce

7.0% Deamidation

H12

LAM02OCT2012000002.d: EIC 967.4557 +All MS

LAM02OCT2012000002.d: EIC 967.7837 +All MS0.0

0.2

0.4

0.6

0.8

1.0

5x10

Intens.

0.0

0.5

1.0

1.5

5x10

120 121 122 123 124 125 126 127 128 Time [min]

H12

Deamidation ?

Previous Trypsin Map

Waters BEH C18, TFA

New Trypsin Map

Waters XSelect XP CSH C18, FA

H12 (N103Y) H12 (N103Y)

IgG4 mAb: H chain deamidation in CDR3

Localized deamidation site to: EGYGN103YGAWFAYWGQGTLVTVSSASTK

NY site not expected to deamidate (NG, NS, NT, NN)

Causes loss of target binding

104 106 108 110 112 114 116 118 120

Time (min)

7.0% Deamidation

Case Study #2: Deamidation of an IgG4 mAb

21

New HPLC column allows for better separation and accurate quantitation

of deamidation

Detected unexpected N103Y deamidation hot spot (beyond NG, NS, NT,

and NN)

In process of updating software reflect new findings

Site Modification Modification

(total Asp and

iso-Asp, %)

CEX Fraction

N33S (L-CDR1) Deamidation 3.2 A3

N55N (H-CDR2) Deamidation 0.1 - - -

N103Y (H-CDR3) Deamidation 7.0 A1 and A2

Case Study #3

Forced Deamidation Study of IgG1 mAb

22

Stress

Study

• Incubation at 25 ºC and 37 ºC in 100 mM sodium phosphate, pH 8.0

• T=0, 2 and 4 weeks

Peptide Map

Method

• New Lys-C/trypsin map, pH 8.2, 30, 60 and 120 min

• Waters XSelect XP CSH C18 column, FA

MS Analysis

• LC-MS/MS and EICs

• Collision induced dissociation (CID)

• Targeted electron transfer dissociation (ETD)

Case Study #3: Forced Deamidation Study of IgG1

mAb

23

IgG1 L Chain CDR Hotspots Rel. Abundance

1 DIVMTQTPLSLSVTPGQPASISCRSSQSLV 30 High level (>5%)

31 HSNGNTFLYWYLQKPGQSPQLLIYRVSNRF 60 Low level (1-5%)

61 SGVPDRFSGSGSGTDFTLKISRVEAEDVGV 90 Potential (<1%)

91 YYCFQATHVPWTFGGGTKVEIKR

IgG1 H Chain CDR Hotspots Rel. Abundance

1 QVQLVQSGAEVKKPGASVKVSCKASGYTFT 30 High level (>5%)

31 GYYIHWVRQAPGQGLEWMGWIYPGNFNTKY 60 Low level (1-5%)

61 NERFKGRVTMTTDTSTSTAYMELRSLRSDD 90 Potential (<1%)

91 TAVYYCAREDGSPYYAMDYWGQGTSVTVSS 120

Deamidation of Light Chain (L2 peptide with CDR1) 25SSQSLVHSN33GNTFLYWYLQKPGQSPQ50LLIYR55

24 99 100 101 102 103 104 105 106 107 108 109

37˚C

4 weeks

L2 (D33G)

3.3%

L2 (isoD33G)

9.5%

25˚C

4 weeks

T=0

L2(N33G)

120 min EICs shown

L2 (D33G)

13.0%

L2 (isoD33G)

30.0%

L2 (D33G)

0.7%

L2 (isoD33G)

2.1%

L2 (E50L)

1.5%

Deamidation of Heavy Chain (H5 with CDR2) 39QAPGQ43GLEWMGWIYPGN55FNTK59

25

114 115 116 117 118 119 120 121 122

Time (min)

0

H5(E43G) H5(D55F or isoD55F)

1.0%

37˚C

4 weeks

25˚C

4 weeks

T=0

H5(N55F)

120 min EICs shown

H5(D55F or isoD55F)

4.0%

H5(D55F or isoD55F)

12.9%

26

CID Spectra Showing Deamidation at Asn-55 39QAPGQGLEWMGWIYPGN55FNTK59

39QAPGQGLEWMGWIYPGNFNTK59

b14

b14

b13

b13

b12

b12

b11

b11

y9

y9

y8

y8

y7

y7

b132+

b132+

b122+

b122+

y82+

y82+

y72+

y72+

y4+

y4

39QAPGQGLEWMGWIYPGDFNTK59

b14

y7

300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000

m/z

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Rela

tive A

bundance

777.4454

940.5127

728.0817

671.4913

1341.7339 389.3110 1053.6013 470.8820 1155.6390

1454.8182 967.5503

1618.9397 620.4687 809.5502 1296.7688 357.1390 1582.9045 1767.1686 1887.0905

778.4114

941.5096

728.0490

671.4405

1054.6180 1341.7526 471.3714 1155.6852

389.7402 1455.8529

967.5021 1617.8817 621.0251 1297.8234 809.6602

244.1232 1492.6409 1745.0511 1845.8401

CID did not provide conclusive evidence for deamidation site

ETD Fragmentation Confirms Deamidation Site

at Asn-55 (N55F)

27

400 600 800 1000 1200 1400 1600 1800 2000

m/z

0

1

2

3

4

5

6

7

8

9

10

11

12

0

1

2

3

4

5

6

7

8

9

10

11

12

Re

lative

Ab

un

da

nce

940.4541z=1

1472.7207z=1

493.2530z=1

669.3685z=1346.1862

z=1 1727.7738z=1

607.2953z=1

1053.5317z=1

1903.9133z=?1359.6283

z=1 1591.7782z=?

393.6677z=?

903.3638z=? 1244.7306

z=?

1937.8198z=?

742.7191z=?

346.1848z=1

941.4358z=1

493.2527z=1

1471.7155z=1

669.3676z=?

984.4893z=1

1904.8741z=1

1618.7538z=1

1413.6531z=1

1728.7710z=1

1126.5248z=2

1297.6061z=?

427.2351z=?

843.7211z=?

1952.6699z=?

NL: 1.24E6

03Aug16_EFNA4_HR_ETD#36869-36989 RT: 116.56-116.94 AV: 15 F: FTMS + p ESI Full ms2 [email protected] [250.0000-2000.0000]

NL: 1.94E5

03Aug16_EFNA4_HR_ETD#37814-37974 RT: 120.12-120.67 AV: 20 F: FTMS + p ESI Full ms2 [email protected] [250.0000-2000.0000]

39QAPGQGLEWMGWIYPGDFNTK59

39QAPGQGLEWMGWIYPGNFNTK59

z3

z3

z4

z4

608.2818

z5

z5

c13

c13

c17

c17

z14

z14 z4

z4 z5

z5 z3

z3

Asn Deamidation and Asp Isomerization Results for

IgG1 mAb using New Low Artifact Trypsin Map

28

Site Modification Modification (total Asp and iso-Asp, %)

Unstressed Stressed 4

weeks at 25 C

Stressed 4

weeks at 37 C

N33G (L-CDR1) Deamidation 2.8 12.8 43.0

N55F (H-CDR2) Deamidation

1.0 4.0 12.9

D100G (H-CDR3) Isomerization 0.7 1.0 1.7

Important to analyze stressed material for hotspots (2.8 to 43.0%)

Detected unexpected N55F deamidation hot spot

D100G not a hotspot despite motif and location in H-CDR3

Summary

• New trypsin peptide map-shorter digestion times, efficient enzyme, time-course elements and updated C18 column with formic acid

– Detects method-induced artifacts

– Improved resolution of deamidated and isomerized peptides

– Accurate quantitation of deamidation/isomerization via time course

– Fragmentation (CID and ETD) determine sites of modification

• ETD also distinguishes Asp vs iso-Asp

• Deamidation and isomerization hotspots are sequence and location dependent and less predictable in proteins

– Detected unexpected deamidation hot spots (beyond NG, NS, NT, and NN)

• N103Y in heavy chain CDR3

• N55F in heavy chain CDR2

• Low and constant level of Gln deamidation (Q50L, Q40G)

29

Summary of Hotspots

30

Case Study 1

Trastuzumab

Originator

Case Study 2

IgG4 mAb

Ref Material

Case Study 3

IgG1 mAb-stressed

Ref Material -

Stressed

Rel. Abundance

N30T (L-CDR1)

N33S (L-CDR1) N33G (L-CDR1)

N33G (L-CDR1) High level (>5%)

N55G (H-CDR2)

N55N (H-CDR2) N55F (H-CDR2)

N55F (H-CDR2) Low level (1-5%)

D102G (H-CDR3)

N103Y (H-CDR3) D100G (H-CDR3)

D100G (H-CDR3) Potential (<1%)

Forced degradation studies help identify potential sequence liabilities

Acknowledgements

31

Tao He

Eric Sousa

Yan Zhang

Laura Lin

Bekah Ward

Neil Steinmeyer

Tom Porter

Meg Ruesch

Jason Rouse-co-author

Elaine Stephens-co-author

Michelle English-co-author

Shibu Philip-co-author

Robert Kutlik (software)

Kristin Boggio

Heather DeGruttola

Jennifer Ide

Natalia Kozlova

Mellisa Ly

Andrew Saati

Roger Theberge

Matt Thompson

Ying Zhang

modernizing the platform characterization peptide map for ... · modernizing the platform...

Documents