b c d e f g h i j k l ma

0

0

350

10340

20330

30

320

40

310

50

300

60

290

70

280

80

270

90

260

100

250 110

240 12

0

230

130

220

140

150

200

160

190

170

180

210

180

170

190 160

200150

210

140

220

130230

120

240

110

250

100

260

90

270

8028070290

60300

50

310

40320

30

330

20

340

10

350

0

0

30

330

60

300

90270

120

240 150

210

180

180

150

150

120

120

90 90

60

6030

30

0

0

0

350

350

10

340

340

20

330

330

30

320

320

40

310

310

50

300

300

60

290

290

70

28028

0

80

270

270

90

260

260

100

250

250 110

240

240 120 230

230 13

0

220

220

140

210

210150

200

200160

190

190

170

180

180

180

170

170

190

160

160

200

150

150

210

140

140

220

130

130230

120

120240

110

110250

100 100

260

90 90

270

80

80280

70

70290

60

6030050

50310

40

40320

30

30330

20

20

340

10

10

350

a

b

c

d

e

f

a

b

c

d

e

f

Immunoglobulin protein | ca. 150,000 Daltons | participates in the immune reaction as the antibody for a specific antigen | There are five main types: IgA, IgD, IgE, IgG, and IgM

Humanized IgG antibody fragment (Fab) | 50,000 Daltons | VH, CH1 and VL, CL regions, linked by an intramolecular disulfide bond.

Conjugation Site (ADC)

Adduct formation

N-terminal pyroE formation

Site of Aggregation

Characterization Result

Charge Variant Profile Determined

Protein Aggregation Determined

Intact and Sub-unit Mass Analysis Determined

Native Mass Analysis Determined

STRUCTURAL INSIGHTS

AREA OF DETAIL

Sources of Charge Related Heterogeneity Effect Species formed

Sialylation COOH addition Acidic

Deamidation COOH formation Acidic

C-terminal lysine cleavage Loss of NH2 Acidic

Adduct formation COOH formation or loss of NH2 Acidic

Succinimide formation Loss of COOH Basic

Methionine, cysteine, lysine, histidine, tryptophan oxidation Conformational change Basic

Disulfide-mediated Conformational change Basic

Asialylation (terminal Galactose) Loss of COOH Basic

C-terminal lysine and glycine amidation NH2 formation or loss of COOH Basic

Deamidation

Promoting the cooperation between science and business, researchers of the CDL develop workflows for biosimilar characterization with the goal of translating analytical methods into routine, transferable, and compliant solutions. Prof. Christian Huber and co-workers recently published a study entitled “A Generic HPLC Method for Absolute Quantification of Oxidation in Monoclonal Antibodies and Fc-Fusion Proteins Using UV and MS Detection” describing a middle-up approach for the characterization of oxidation variants in therapeutic monoclonal antibodies (mAbs) and Fc-fusion proteins1.

Accelerating innovations in biopharmaceutical characterization

CASE STUDY The Christian Doppler Laboratory

for Biosimilar Characterization

A Platform Method for Oxidation Monitoring in Monoclonal Antibodies and Fc-Fusion Proteins—a Biopharmaceutical Method Development Case Study

© S

chei

nast

The Christian Doppler Laboratory (CDL) for Innovative Tools for Biosimilar Characterization at the University of Salzburg, Austria, aims at accelerating innovations in biopharmaceutical characterization.

“ Biosimilars are the fastest growing sector within the pharmaceutical industry. To ensure the effectiveness and safety of these drugs, an in-depth characterization at the molecular level is obligatory. Through the cooperation of academic research groups and industrial partners, the University of Salzburg is contributing to the development of safer and more affordable pharmaceuticals.”

—Prof. Dr. Christian Huber, Chemist, Christian Doppler Laboratory Head

© Luigi Caputo

0

0

30

330

60

300

90270

120

240 150

210

180

180

150

150

120

120

90 90

60

6030

30

0

0

0

350

350

10

340

340

20

330

330

30

320

320

40

310

310

50

300

300

60

290

290

70

28028

0

80

270

270

90

260

260

100

250

250 11

0

240

240 12

0 230

230 13

0

220

220140

210

210150

200

200160

190

190

170

180

180

180

170

170

190

160

160

200

150

150

210

140

140

220

130

130230

120

120240

110

110250

100 100

260

90 90

270

80

80280

70

70290

60

6030050

50310

40

40320

30

30330

20

20

340

10

10

350

a

b

c

d

e

f

The CDL for Biosimilar Characterization comprises a multidisciplinary team of scientists focusing on method development and workflow optimization for the analysis of biosimilar drugs. Their aim is to accelerate innovations in biopharmaceutical characterization and to ensure that the developed methods can be translated into routine, compliant solutions for pharmaceutical companies. Collaboration with industrial partners within the CDL provides insights and alignment with regulatory requirements for biosimilars, whilst mentoring and developing the next generation of biopharmaceutical analytical scientists in an academic environment.

With this research focus in mind, the team embarked on the development of a fast and generic method to quantitate protein oxidation in biotherapeutics. Oxidation of methionine is a post-translational modification (PTM) which is considered a potential critical quality attribute (CQA) in biopharmaceutical production—this implies stringent monitoring all the way from drug development and manufacturing to lot release/quality control (QC). Methionine oxidation has been shown to affect the

structure of the Fc region of mAbs2 which may alter binding to the neonatal Fc-receptor (FcRn)3, impacting pharmacokinetic properties, potentially leading to faster plasma clearance and altered clinical efficacy of the protein drug. Monitoring protein oxidation is typically performed using a conventional peptide-based bottom-up approach. One issue with bottom-up methods in this context is the risk of introducing artificial oxidation during tryptic digestion.

Characterization at the intact protein or subunit level, on the other hand, provides valuable information on a protein’s structure, as the context of modifications is maintained. The CDL for Biosimilar Characterization therefore prioritizes the application of intact protein and subunit analysis over peptide mapping approaches, which commonly involve lengthy, variable sample preparation and a loss of information on the context of the detected modifications.

© KBFilm Salzburg

© KBFilm Salzburg

0

0

30

330

60

300

90270

120

240 150

210

180

180

150

150

120

120

90 90

60

6030

30

0

0

0

350

350

10

340

340

20

330

330

30

320

320

40

310

310

50

300

300

60

290

290

70

28028

0

80

270

270

90

260

260

100

250

250 11

0

240

240 12

0 230

230 13

0

220

220140

210

210150

200

200160

190

190

170

180

180

180

170

170

190

160

160

200

150

150

210

140

140

220

130

130230

120

120240

110

110250

100 100

260

90 90

270

80

80280

70

70290

60

6030050

50310

40

40320

30

30330

20

20

340

10

10

350

a

b

c

d

e

f

“ Protein oxidation is a PTM which is considered a potential CQA in biopharmaceutical production—this implies stringent monitoring all the way from drug development and manufacturing to lot release/QC.”

—Dr. Therese Wohlschlager, Biotechnologist, Christian Doppler Laboratory

The authors of the study initially determined three main requirements to be fulfilled by the method in development:

1. Rather than being tailored to one specific molecule, the method should be applicable for a range of molecules sharing one common property— the Fc domain of IgG1 antibodies.

2. The method should ideally be UV-based, with the possibility of coupling chromatographic separation directly to mass spectrometry (MS). This analytical setup facilitates easy method transfer from the discovery and development laboratory into a QC environment.

3. The approach should provide insights into protein oxidation while maintaining the context of other modifications such as glycosylation, applying a so called “middle-up” approach. Additionally, this strategy minimizes artificial oxidations which may be introduced during proteolytic digestion.

Selecting the parts for biosimilar method development success

The middle-up workflow for protein oxidation quantitation developed within the CDL comprises:

• Thermo Scientific™ MAbPac™ RP column

• Thermo Scientific™ UltiMate™ 3000 BioRS UHPLC system

• Thermo Scientific™ Q Exactive™ Hybrid Quadrupole-Orbitrap™ mass spectrometer

• Thermo Scientific™ Chromeleon™ 7.2 Chromatography Data System (CDS) software

Other studies have shown protein oxidation analysis by hydrophobic interaction chromatography (HIC) or ion exchange chromatography (IEC)4, 5. HIC and IEC methods tend to be HPLC-UV-based only, and usually do not offer the opportunity to hyphenate to MS for further characterization. In order to overcome this limitation, the method developed by Regl et al. involved an ion-pair reversed phase (IP-RP) separation, which could readily be coupled to MS in order to deliver further structural information1.

Initial attempts provided evidence that oxidized and non-oxidized mAb variants could not be baseline resolved by HPLC and MS at the intact protein level: On the one hand, mass spectra revealed overlap of isotopic envelopes, making separation by IP-RP-HPLC increasingly important. On the other hand, baseline chromatographic separation at the intact protein level— a prerequisite for quantitation—could not be achieved. Because of these limitations at the intact protein level, a middle-up approach involving the analysis of Fc/2 subunits was pursued as described in Figure 1.

© Sergey Gusenkov

© KBFilm Salzburg

Figure 1. Typical workflow for quantification of oxidation in the Fc/2 subunits of monoclonal antibodies and fusion proteins.

Biopharmaceuticals containing oxidized variantsA generic workflow developed for characterization of oxidation in biopharmaceuticals involving:

25 kDa150 kDa

IdeS

25 kDa25 kDa

25 kDa

23 kDa23 kDa

TCEP

Proteolysis with IdeS & reductionDigestion of the mAb with IdeS (FabRICATOR®, Genovis AB) and reduction of disulfide bonds with tris(2-carboxyethyl)phosphine (TCEP).

1

2

4

3

© KBFilm Salzburg

Separation & detectionThis middle-up approach enables mass spectrometric resolution of the resulting Fc/2 subunits for accurate quantitation both of oxidized and non-oxidized forms of the mAb. Since the IdeS digestion step is very short and is performed at moderate temperatures, the risk of artificial oxidation is minimized. The obtained mAb subunits are subsequently separated by IP-RP-HPLC on a MAbPac RP column before detection in the Thermo Scientific™ Dionex™ UltiMate™ 3400RS VWD Variable Wavelength Detector and the Q Exactive MS.

Quantification of oxidationOxidation is identified based on the chromatographically separated peaks as well as extracted ion chromatograms revealing mono- or double oxidation, as indicated by +16 and +32 Da mass shifts with respect to the non-oxidized variant. Calibration and quantification of UV- and MS data were all performed using Chromeleon 7.2 CDS.

Figure 2. Overlay of ten representative chromatograms of MabThera® (Rituximab, Roche) originator at different concentrations, as used for external calibration both of the UV-spectroscopic and mass spectrometric detection systems. Conditions: Sample: 6.0–1200 ng Fc/2 of MabThera on column. Proteolysis with IdeS for 1 h at 22 °C; Reduction: 5 mM TCEP for 30 min at 60 °C; HPLC: MAbPac RP, 4 µm, 1500 Å, 150.0 × 2.1 mm i.d., 200 µL/min; 80 °C; 28.9% acetonitrile in 0.1% trifluoroactic acid for 5 min, 28.9–29% acetonitrile in 5 min 29–30% acetonitrile in 9 min, 30–45% acetonitrile in 5 min; Detection: UV @ 214 nm, Q Exactive MS; HESI (200 °C) at +4.0 kV; Full Scan m/z 1000–2500; R = 140,000 @ m/z 200.

Figure 3. Chromatograms illustrating the different detection modes evaluated using UV- and MS detection, specifically UV detection at 214 nm (UV), full-scan total ion current detection (TICC), extracted ion current chromatogram detection of the seven most abundant charge state signals (XICC), and extracted ion current chromatograms of two fragments (b34 and y45) obtained upon all ion fragmentation (AIF-MS). Experimental conditions as in Figure 2.

b c d e f g h i j k l ma

0

0

350

10340

20330

30

320

40

310

50

300

60

290

70

280

80

270

90

260

100

250 110

240 12

0

230

130

220

140

150

200

160

190

170

180

210

180

170

190 160

200150

210

140

220

130230

120

240

110

250

100

260

90

270

8028070290

60300

50

310

40320

30

330

20

340

10

350

0

0

30

330

60

300

90270

120

240 150

210

180

180

150

150

120

120

90 90

60

6030

30

0

0

0

350

350

10

340

340

20

330

330

30

320

320

40

310

310

50

300

300

60

290

290

70

28028

0

80

270

270

90

260

260

100

250

250 110

240

240 120 230

230 13

0

220

220

140

210

210150

200

200160

190

190

170

180

180

180

170

170

190

160

160

200

150

150

210

140

140

220

130

130230

120

120240

110

110250

100 100

260

90 90

270

80

80280

70

70290

60

6030050

50310

40

40320

30

30330

20

20

340

10

10

350

a

b

c

d

e

f

a

b

c

d

e

f

Immunoglobulin protein | ca. 150,000 Daltons | participates in the immune reaction as the antibody for a specific antigen | There are five main types: IgA, IgD, IgE, IgG, and IgM

Humanized IgG antibody fragment (Fab) | 50,000 Daltons | VH, CH1 and VL, CL regions, linked by an intramolecular disulfide bond.

Conjugation Site (ADC)

Adduct formation

N-terminal pyroE formation

Site of Aggregation

Characterization Result

Charge Variant Profile Determined

Protein Aggregation Determined

Intact and Sub-unit Mass Analysis Determined

Native Mass Analysis Determined

STRUCTURAL INSIGHTS

AREA OF DETAIL

Sources of Charge Related Heterogeneity Effect Species formed

Sialylation COOH addition Acidic

Deamidation COOH formation Acidic

C-terminal lysine cleavage Loss of NH2 Acidic

Adduct formation COOH formation or loss of NH2 Acidic

Succinimide formation Loss of COOH Basic

Methionine, cysteine, lysine, histidine, tryptophan oxidation Conformational change Basic

Disulfide-mediated Conformational change Basic

Asialylation (terminal Galactose) Loss of COOH Basic

C-terminal lysine and glycine amidation NH2 formation or loss of COOH Basic

Deamidation

0 5 10 15 20 25 30 35

- 12.5

0.0

12.5

25.0

37.5

50.0

mA

U

Oxidation in calibration standard

(0.6-120.0 ng/µL)Fc/2)

min

0 5 10 15 20 25 30 35

0

100

200

300

400

500

600

700

mA

U XICC

TICC

AIF-MS

UVOxidized

Fc/2 Non-oxidized Fc/2

Fd‘

LC

min

“This method is unique in its ability to monitor oxidation; it is generic for any IgG 1 mAb, and it can be used in research and QC environments—where UV is preferred for detection and cGMP compliance is required—it even allows the site of oxidation to be determined via top- or middle-down MS!”

—Prof. Dr. Christian Huber, Chemist, Christian Doppler Laboratory Head

© KBFilm Salzburg

Christof Regl and his colleagues believe that this middle-up workflow of Fc/2 subunits could be developed into a ‘multi-attribute monitoring’ (MAM) method. Moreover, the approach enables absolute quantification of any IgG1 mAb at the Fc/2 level. This provides a broader picture of the influence of oxidation and other PTMs (e.g., lysine clipping, pyroglutamate formation, degradation, glycosylation) on the mAb. The method is unique in its ability to monitor oxidation; many other studies offer structural analysis and assignment of modifications, but absolute quantification of the oxidation on these samples is rarely performed. The method developed within the CDL can be applied in both research (top- or middle-down MS for site-specific determination of oxidation) and QC environments (UV-only for detection and quantitation in a cGMP environment).

Data analysis using Chromeleon 7.2 CDS gave calibration curves with correlation coefficients close to 1.0 and very narrow confidence intervals, as indicated by the orange lines in Figure 4. Calculation of the relative process standard deviations for the different quantification approaches yielded values in the range of 7.2–13.6%, which clearly indicates that the quantification method is robust and reliable. Compared to quantification via UV detection, the XICC method proved to give the most accurate quantitative results, showing a relative process standard deviation of 10.3% and a mean bias of only 7%. Raw data as well as data evaluations are freely available for download (DOI: 10.1021/acs.analchem.7b01755, and www.proteomexchange.org accession No. PXD006873) and may be utilized as benchmarking data.

Additionally, the study demonstrates that non-oxidized mAb can be used as a calibration standard, alleviating the need for isotopically-labelled standards or other molecules to be introduced into the analysis. This makes the entire protocol a ‘platform approach’ because you do not need to synthesize isotopically-labelled proteins or peptides. Even better, the standard sample does not need to be completely free of oxidized products, as this can be corrected for based on prior HPLC-UV analysis.

A generic ‘platform method’—delivering specific results Oxidation is a major tell-tale modification observed in mAb products that have passed their expiration date6. As oxidized variants of a therapeutic antibody may show altered clinical efficacy, reliable and accurate quantitation of oxidation is critical to ensure patient safety. With this middle-up technique, the CDL for Biosimilar Characterization has developed a fast, generic method for monitoring oxidation in Fc/2 subunits of any IgG1 mAb or Fc fusion proteins.

The benefits to this approach are multiple:

• Maintained context of the modification at the subunit level

• Detection by UV only, or coupled to MS for further characterization

• Easy transfer from an R&D environment to a QC laboratory

• Absolute quantitation of oxidized versus non-oxidized methionine for any IgG1 mAb or Fc fusion protein

b c d e f g h i j k l ma

0

0

350

10340

20330

30

320

40

310

50

300

60

290

70

280

80

270

90

260

100

250 110

240 12

0

230

130

220

140

150

200

160

190

170

180

210

180

170

190 160

200150

210

140

220

130230

120

240

110

250

100

260

90

270

8028070290

60300

50

310

40320

30

330

20

340

10

350

0

0

30

330

60

300

90270

120

240 150

210

180

180

150

150

120

120

90 90

60

6030

30

0

0

0

350

350

10

340

340

20

330

330

30

320

320

40

310

310

50

300

300

60

290

290

70

28028

0

80

270

270

90

260

260

100

250

250 110

240

240 120 230

230 13

0

220

220

140

210

210150

200

200160

190

190

170

180

180

180

170

170

190

160

160

200

150

150

210

140

140

220

130

130230

120

120240

110

110250

100 100

260

90 90

270

80

80280

70

70290

60

6030050

50310

40

40320

30

30330

20

20

340

10

10

350

a

b

c

d

e

f

a

b

c

d

e

f

Immunoglobulin protein | ca. 150,000 Daltons | participates in the immune reaction as the antibody for a specific antigen | There are five main types: IgA, IgD, IgE, IgG, and IgM

Humanized IgG antibody fragment (Fab) | 50,000 Daltons | VH, CH1 and VL, CL regions, linked by an intramolecular disulfide bond.

Conjugation Site (ADC)

Adduct formation

N-terminal pyroE formation

Site of Aggregation

Characterization Result

Charge Variant Profile Determined

Protein Aggregation Determined

Intact and Sub-unit Mass Analysis Determined

Native Mass Analysis Determined

STRUCTURAL INSIGHTS

AREA OF DETAIL

Sources of Charge Related Heterogeneity Effect Species formed

Sialylation COOH addition Acidic

Deamidation COOH formation Acidic

C-terminal lysine cleavage Loss of NH2 Acidic

Adduct formation COOH formation or loss of NH2 Acidic

Succinimide formation Loss of COOH Basic

Methionine, cysteine, lysine, histidine, tryptophan oxidation Conformational change Basic

Disulfide-mediated Conformational change Basic

Asialylation (terminal Galactose) Loss of COOH Basic

C-terminal lysine and glycine amidation NH2 formation or loss of COOH Basic

Deamidation

“ We found the MAbPac RP column to be very robust and well-resolving. Although we used a very shallow gradient for the separation, we observed variability below 1% when we evaluated the intra- and inter-day standard deviations for retention times. Additional peak parameters were very robust, as corroborated by the RSD values for peak area, peak height, and peak width at half height.”

—Christof Regl, MSc, Biologist, Christian Doppler Laboratory

© KBFilm Salzburg

Find out more at thermofisher.com/IntactMS

© 2018 Thermo Fisher Scientific Inc. All rights reserved. FabRICATOR is a registered trademark of Genovis AB. MabThera is a registered trademark of Roche in Europe. All other trademarks are the property of Thermo Fisher Scientific and its subsidiaries. This information is presented as an example of the capabilities of Thermo Fisher Scientific products. It is not intended to encourage use of these products in any manners that might infringe the intellectual property rights of others. Specifications, terms and pricing are subject to change. Not all products are available in all countries. Please consult your local sales representatives for details. CS72702-EN 0518M

Figure 4. Calibration curves for the non-oxidized Fc/2 fragment of MabThera (Rituximab) using UV-spectroscopic and mass spectrometric

detection. R2 = correlation coefficient, Vx0 = relative process standard deviation.

The method is very robust; digestion with IdeS minimizes the risk of inducing artificial oxidation and the MAbPac RP column offers outstanding separations, excellent resolution of oxidized species and demonstrated reproducibility over more than 600 injections. Moreover, protein carry-over between injections is minimal, which represents a basic requirement for reliable quantification. Prof. Huber and his co-workers believe that this workflow represents an appropriate standard method for any biopharmaceutical analytical lab which is challenged with the routine monitoring of methionine oxidation of an Fc-containing protein.

For more information on the Christian Doppler Laboratory for Innovative Tools for Biosimilar Characterization, please visit cdl-biosimilars.sbg.ac.at and www.thermofisher.com/CDL.

References1. Regl, C.; Wohlschlager, T.; Holzmann, J.; and Huber, C. G. A Generic HPLC Method for

Absolute Quantification of Oxidation in Monoclonal Antibodies and Fc-Fusion Proteins Using UV and MS Detection. Analytical Chemistry, 2017, 89 (16), 8391–8398.

2. Mo, J.; Yan, Q.; So, C. K.; Soden, T.; Lewis, M. J.; Hu, P. Understanding the Impact of Methionine Oxidation on the Biological Functions of IgG1 Antibodies Using Hydrogen/Deuterium Exchange Mass Spectrometry. Analytical Chemistry, 2016, 88 (19), 9495–9502.

3. Tsuchida, D.; Yamazaki, K.; Akashi, S. Comprehensive Characterization of Relationship Between Higher-Order Structure and FcRn Binding Affinity of Stress-Exposed Monoclonal Antibodies. Pharmaceutical Research, 2016, 33 (4), 994–1002.

4. Stracke, J.; Emrich, T.; Rueger, P.; Schlothauer, T.; Kling, L.; Knaupp, A.; Hertenberger, H.; Wolfert, A.; Spick, C.; Lau, W.; Drabner, G.; Reiff, U.; Koll, H.; Papadimitriou, A. A novel approach to investigate the effect of methionine oxidation on pharmacokinetic properties of therapeutic antibodies. MAbs, 2014, 6 (5), 1229–1242.

5. Haverick, M.; Mengisen, S.; Shameem, M.; Ambrogelly, A. Separation of mAbs molecular variants by analytical hydrophobic interaction chromatography HPLC: overview and applications. MAbs, 2014, 6 (4), 852–858.

6. Houde, D.; Kauppinen, P.; Mhatre, R.; Lyubarskaya, Y. J. Determination of protein oxidation by mass spectrometry and method transfer to quality control. Journal of Chromatography A, 2006, 1123 (2), 189–198.

“ Even when I massively overloaded the column, it regenerated completely. It really is very robust. The MAbPac RP column did not have any problem with carryover and was stable for over 600 injections.”

—Christof Regl, MSc, Biologist, Christian Doppler Laboratory

© KBFilm Salzburg

0

10

20

30

0 10 20 30 40 50 600

1

2

3

0 20 40 600

2

4

6

0 10 20 30 40 50 60

Cou

nts

. min

, . 1

0-5

Cou

nts

. min

, . 1

0-6

MA

U .

min

Cou

nts

. min

, . 1

0-4

ng µL-1 ng µL-1ng µL-1ng µL-1

0

1

2

3

0 10 20 30 40 50 60

TICC, R2 = 0.9956 AIF-MS, R2 = 0.9878 XICC, R2 = 0.9950UV, 214 mm, R2 = 0.9975

Vx0=7.2% Vx0=9.4% Vx0=13.6%Vx0=10.3%