Challenges in peptide mapping mass spectrometry of biopharmaceuticals

07.MAR.2018

Will Burkitt, Ben Holmes, Johanna Paris, Nisha Patel, John O’Hara

Characterisation within UCB2

Context to the type of analysis performed

UCB is a global biopharma focused on severe diseases in two therapeutic areas: neurology and immunology

Characterisation is on the biology side of the business:

• Biopharmaceuticals rather than small molecules

• Antibody based molecules produced in cell culture (mammalian and e-coli)

Discovery

• Target identification• V-region discovery• Candidate selection

Development

• Toxicology• Phase I• Phase II• Phase III

Commercial

• Phase IV

Characterisation

Characterisation within development3

Who do we work with and what do we do?

Activities with in development:

Process scale-up and optimisation

• Upstream (cell-culture)

• Down stream purification

Formulation

Product understanding

Analytical

• Method development

• Stability

• Bioassay

Types of study:

• Comparability

• Reference standard characterisation

• Process understanding

• Force degradation study

• Stability

• Investigations

What types of analysis do we perform?4

Wide range of physical/biophysical

Gel separations

• SDS-PAGE

• CE

HPLC separations

• Cation exchange

• Size exclusion

Higher Order

• CD, FTIR, DSC, AUC, Biacore, DLS

Mass spectrometry

• Peptide mapping reduced and non-reduced

• Intact mass denatured and native

Gives information on the primary structure

What people want to know…

• Expected sequence confirmation

• PTM levels

• Sequence variants

Peptide mapping5

Method

Sample preparation (digestion)• Denature (guanidine hydrochloride)

• Reduce (dithiothreitol)

• Alkylate (iodoacetamide)

• Buffer exchange

• Digest (trypsin for 3 hours)

Separation• Reversed phase C18

• Water:acetonitrile with 0.1% formic acid

• 45 minute run time

Detection• Q-Exactive Plus Orbitrap

Data processing• Biopharama Finder

Ideal method:

• Reproducible

• Not induce PTMs during analysis

• Scalable

• Fast

• Specific cleavage

• Sensitive

Manufacture overview of a biopharmaceutical6

Modification levels

Peptide: DTLMISR

Non-modified selected ion chromatogram:

Oxidised selected ion chromatogram:

7

Example of how the level at which a PTM is present i s calculated (oxidation)

How are these calculated?

Easy example:

• One modification site

• Good chromatographic separation

• 16 Da mass shift

To calculate oxidation level:

Integrate peak area of non-modified and modified

Not quantitative

Can be used to compare levels across batches within the same sample set

0

10

20

30

40

50

60

70

80

90

100

Rel

ativ

e In

tens

ity

SIC: E:\Data\UCB\Innovate\A33\06JUL2017\11JUL2017_02_IgG4_Demo.raw15.81

16.2812.68 16.9513.89 17.36 18.06 18.5214.95 15.65 19.4013.5910.95

10 11 12 13 14 15 16 17 18 19RT (min)

NL: 3.04E8

0

10

20

30

40

50

60

70

80

90

100

Rel

ativ

e In

tens

ity

SIC: E:\Data\UCB\Innovate\A33\06JUL2017\11JUL2017_02_IgG4_Demo.raw13.92

15.8114.1313.1712.33 14.55 15.2313.52 16.36 19.5818.2316.7011.26

10 11 12 13 14 15 16 17 18 19RT (min)

NL: 4.35E7

Modification Level =���������������

���������(�����������������)

Example of measurement of modification level

Repeatability

Method can give linear and repeatable results

Linearity

8

Easy example norleucine misincorporation for methioninine

y = 0.04536x + 0.00189R² = 0.99996

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

5.0%

0 0.2 0.4 0.6 0.8 1

Mea

sure

d ra

tio o

f nor

leuc

ine

to

met

hion

ine

Ratio of low level batch to high level batch

Lower

concentration

at about 1.3%

Repeat

injections

at about 2.3%

Higher

concentrations

at about 4.4%

Std Dev CV Std Dev CV Std Dev CV

Norleucine

level0.06% 3.6% 0.05% 2.0% 0.13% 2.6%

Modifications searched for

Standard modifications:

• Glycosylation (glycoforms)

• C-terminal lysine clipping

• N-terminal pyroglutamic acid formation

• Methionine oxidation

• Deamidation

• Glycation

9

How many should be included?

Additional modifications:

• Tryptophan oxidation

• Histidine oxidation

• O-linked glycosylation

• Aspartic acid isomerisation

• Glycation degradation products (AGEs)

• Leader sequence

• Sequence variants

• …

• Non-hypothesis modifications (wild card search)

What are the issues with this approach?10

Levels reported are not quantitative

• Comparison between runs and instruments is difficult

• Comparison between occasions can be difficult

• How should levels be calculated?

When is a sample different

• Analytical difference Vs biochemically relevant difference

Detection of repeated sequences

• Would the presence of additional light chain be detected?

Too much data

• Large numbers of samples

• How to spot relevant differences quickly

• Which differences relate to the various parameters?

What improvements are we trying to make?

Isotopically labelled products

• For each product make a small amount that has been isotopically labelled so that this can be used as an internal standard.

• Focus on the amount of unmodified peptide present

• Any modification of a peptide will reduce the amount of unmodified peptide

• Doesn’t matter if you know what modifications to look for

11

Two separate approaches

Statistical methods

• Focus only on the PTMs and the levels that have been measured for each

• Takes some account of loading and digestion differences

• Include all modifications: hypothesis/non-hypothesis and sequence variants

Mass Spectrometry Analysis

Labelled internal standard methodology12

Principle

Trypsin Digestion

mAb X mAb XReferenceStandard

LabelledmAb X

LabelledmAb X

Equal amounts

Labelled internal standard methodology (2)13

Labelled peptide Labelled peptide

Peptide from mAb X Peptide from mAb X Ref Std

C DBA

IA = Intensity of peptide from mAb X IC = Intensity of peptide from Ref StdIB = Intensity of labelled peptide in mAb X sample ID = Intensity of labelled peptide in Ref Std sample

Measure intensity of peptide signals:

Calculate ratio of peptide and labelled peptide in each sample:

��

��

��

��

Calculate ratio of ratios to give the amount of the mAb X peptide relative to the mAb X Ref. Std. peptide:

��������

=��

��

��

��= ��� !!"#�$%&'�!#��()��)

Labelled internal standard methodology (3)14

Differences

Labelled peptide Labelled peptide

Peptide from mAb X Peptide from mAb X Ref. Std.

C DBA

Decrease in signal from peptide results in decrease measured level of peptide

��������

=��

��

��

��= ��� !!"#�$%&'�!#��()��)

Decrease in signal results from modification of the peptide e.g. deamidation, oxidation etc……even modifications that we don’t know to look for

Labelled internal standard results15

Com

parison of similar batches

0.9

0

0.9

5

1.0

0

LC01

LC02

LC04

LC05

LC06

LC07

LC08

LC10

LC11

LC13

LC14

LC15

LC16

LC18

LC19

HC01

HC02

HC03

HC04

HC06

HC08

HC09

HC10

HC11

HC12

HC13

Tw

o b

atch

es sa

me

pro

cess

LC02: 1.3+

0.9%H

C02:+

1.4+0.7%

0.4% less W

ox

0.9

0

0.9

5

1.0

0

LC01

LC02

LC04

LC05

LC06

LC07

LC08

LC10

LC11

LC13

LC14

LC15

LC16

LC18

LC19

HC01

HC02

HC03

HC04

HC06

HC08

HC09

HC10

HC11

HC12

HC13

Tw

ore

fere

nce

stan

da

rds d

iffere

nt

pro

cesse

s

0.9

5

1.0

0

LC01

LC02

LC04

LC05

LC06

LC07

LC08

LC10

LC11

LC13

LC14

LC15

LC16

LC18

LC19

HC01

HC02

HC03

HC04

HC06

HC08

HC09

HC10

HC11

HC12

HC13

Labelled internal standard results

Change of cell line and m

edia

Cell line 1 +

Media A

Cell line 2 +

Media A

Cell line 1 +

Media B

0.9

5

1.0

0

LC01

LC02

LC04

LC05

LC06

LC07

LC08

LC10

LC11

LC13

LC14

LC15

LC16

LC18

LC19

HC01

HC02

HC03

HC04

HC06

HC08

HC09

HC10

HC11

HC12

HC13

0.9

5

1.0

0

LC01

LC02

LC04

LC05

LC06

LC07

LC08

LC10

LC11

LC13

LC14

LC15

LC16

LC18

LC19

HC01

HC02

HC03

HC04

HC06

HC08

HC09

HC10

HC11

HC12

HC13

Labelled internal standard results17

Acidic species

Statistical methods18

Heatmaps

• Only data used is the levels of PTMs that have been measured for each sample

• Aids the data analysis as using the PTM levels provides some normalisation between the samples

• All modifications are included: hypothesis/non-hypothesis and sequence variants

• For each modification the levels are normalised to the mean and the number of standard deviations from the mean plotted

Values Std Dev from mean

Sample 1 0.59 0.71

Sample 2 0.36 -0.11

Sample 3 0.18 -0.75

Sample 4 0.31 -0.29

Sample 5 0.66 0.96

Sample 6 0.19 -0.71

Sample 7 0.06 -1.18

Sample 8 0.87 1.71

Sample 9 0.59 0.71

Sample 10 0.06 -1.18

Mean 0.39

Std Dev 0.28

Heatmaps19

As originally envisaged…

Heat maps20

What I got

Heatmaps21

Bioreactor conditions

ן Too much data

ן Combining datasets form multiple occasions

ן Processing data does not scale well

Issues still to be solved22

ן Automation

o Current demand requires improvements in automation of sample preparation and data processing

ן Host Cell Proteins (HCPs)

o Current levels of sensitivity allow detection some of these with our existing methodology

ן Higher Order Structure

o Can peptide mapping methods be extended to give information on higher-order structure?

Issues still to be solved23

ן Peptide mapping with mass spectrometry can provide a very detailed analysis of biopharmaceutical samples.

ן Large amounts of data are produced.

ן Plotting this data as a heat map can aid interpretat ion of this data and focus attention on potential issues.

ן Using isotopically labelled internal standards of the product of interest has the potential to give a rapid comparison of the similarity of batches.

ן There is still lots more work to do…

Conclusions24

Acknowledgements25

Characterisation Upstream

Questions?26

Thanks!