TO DOWNLOAD A COPY OF THIS POSTER, VISIT WWW.WATERS.COM/POSTERS ©2018 Waters Corporation

INTRODUCTION

Protein reversed phase chromatography, while preferred

for LC-MS, is heavily dependent on the conditions under

which it is performed. Methods employing polymeric

columns and trifluoroacetic acid (TFA) have been

preferred by chromatographers but are inherently

restricted to low pressure, low throughput analyses and

compromised MS detection.

Accordingly, a novel platform for LC-MS has been

developed that includes three critical breakthroughs: a

new particle technology to afford increased throughput, a

unique high coverage phenyl surface to lessen ion

pairing dependence, and a more MS-friendly mobile

phase system based on highly purified difluoroacetic

acid (DFA). Together, these advances achieve the

simultaneous optimization of both LC and MS

capabilities, making it possible to characterize mAbs and

ADCs to an unprecedented level of detail.

A NEW LC-MS APPROACH FOR ENHANCING SUBUNIT-LEVEL PROFILING OF mAbs AND ADCs Jennifer Nguyen1, Jacquelynn Smith2, Olga V. Friese2, Jason C. Rouse3, Daniel P. Walsh1, Ximo Zhang1, Nilini Ranbaduge1, and Matthew A. Lauber1 1Waters Corporation, Milford, MA, USA, 2Biotherapeutics Pharm. Sci., Pfizer WRD, St Louis, MO, USA, 3Biotherapeutics Pharm. Sci., Pfizer WRD, Andover, MA, USA

METHODS

Reduced, IdeS digested NIST mAb was acquired in the form of the Waters mAb

Subunit Standard (p/n 186008927). Therapeutic monoclonal antibodies (mAbs) and

antibody drug conjugates (ADCs) (manufactured by Pfizer) were subjected to IdeS

digestion and reduction according to standard procedures and performed at Waters

Corporation in Milford, MA. Reagent grade DFA was purified via distillation. ICP

quantitation of metals was performed by EAG Laboratories. Various ion pairing

conditions and concentrations using DFA, TFA, and formic acid (FA) were

investigated along with separation temperature, flow rate, and alternative eluents

such as isopropanol to demonstrate the optimization of methods.

Analyses were performed using an ACQUITY UPLC H-Class Bio, ACQUITY UPLC

TUV detector, and Xevo G2-XS QToF mass

spectrometer. Waters mAb Subunit Standard

separations were performed on a 2.7 µm, 2.1 x 50 mm

BioResolve RP mAb Polyphenyl column. ADC

separations were performed at 80 °C or 70 °C on a 2.7

µm, 2.1 x 150 mm BioResolve RP mAb Polyphenyl or a

1.7 µm, 2.1 x 150 mm ACQUITY BEH C4 300 Å

column. Samples were run using 0.1% or 0.15%

DFA, TFA, or FA in water (mobile phase A) and

0.1% or 0.15% of the same modifier in

acetonitrile or 90/10 (v/v) acetonitrile/isopropanol

(mobile phase B). The gradient was run from 15

-55% in 20 min at a flow rate of 0.6 mL/min for 150 mm columns and 0.2 mL/min for

50 mm columns. Analyses were performed with UV detection at 280 nm using

MassLynx 4.1 and UNIFI 1.8. LC/MS analyses were performed in sensitivity mode

and optimized for the reduction of adducts. MaxEnt was used for deconvolution.

References

1. Smith, J., Friese, O., Rouse, J., Nguyen, J., Lauber, M. , and Jayaraman, P. Characterization of Hydrophobic Monoclonal Antibodies and Antibody Drug

Conjugates. Presented at WCBP 2018, Washington D.C., United States, January 28-31, 2018.

2. Nguyen, J. M.; Rzewuski, S.; Walsh, D.; Cook, D.; Izzo, G.; DeLoffi, M.; Lauber, M. A. Designing a New Particle Technology for Reversed Phase Separations of

Proteins. (2018). Waters Application Note (PN: 720006168EN).

3. Fekete, S.; Veuthey, J.; Guillarme, D. New trends in reversed-phase liquid chromatographic separations of therapeutic peptides and proteins: Theory and

applications. J. Pharm. Biomed. Anal. 2012, 69, 9-27. of therapeutic peptides and proteins: Theory and applications. J. Pharm. Biomed. Anal. 2012, 69, 9–27.

With LC-MS quality DFA and the novel polyphenyl solid-core stationary phase of the BioResolve RP mAb column, a new platform method has been developed for both mAb and ADC subunit separations. Figure 8 shows how this new platform method can be implemented to improve ADC subunit profiling, starting with a previous employed method and working toward a newly optimized technique performed with DFA mobile phases and a 2.7 µm, 2.1 x 150 mm BioResolve RP mAb Polyphenyl column to give better MS sensitivity, better resolution and higher recovery of the Fd’(3) sepcies at both high (80 °C) or lowered temperatures (70 °C), such that on-column degradation can also be minimized.

CONCLUSION

A new platform technology has been developed to enhance LC-MS subunit profiling of mAbs and ADCs that is based on the use of highly purified DFA as a mobile phase ion pairing agent and the novel BioResolve RP mAb Polyphenyl stationary phase.

LC-MS quality DFA has been successfully prepared as verified through ICP metal quantitation and LC-MS application testing.

DFA confers notable gains in MS sensitivity versus TFA, even at higher concentrations, while providing comparable (and sometimes better) resolution.

An optimized concentration of DFA can also increase the recovery of challenging protein analytes, as has been exemplified with the increased recovery of a three payload Fd’(3) subunit encountered in the analysis of an ADC.

LC-MS quality DFA was then used to improve the subunit separation of an ADC sample. High resolution

separations with favorable baseline properties were achieved by using 0.15% DFA, as is highlighted with the

zoomed chromatograms of Figure 5. Moreover, it was found that a 0.15% DFA modified mobile phase also

gave the highest recovery of the challenging to recover three payload Fd’(3) subunit.

The raw MS spectra in Figure 6 indicated that, as predicted, a FA mobile phase gives the highest sensitivity

among the various ion-pairing agents. Interestingly, 0.15% DFA gives the next highest signal intensity. For

this comparison, there was, in fact, only a 15% decrease in sensitivity versus using 0.1% FA. Observations on

charge state distributions are equally noteworthy. That is, there is actually only a slight shift in average charge

state in comparing the spectra obtained with FA versus DFA versus TFA.

Deconvoluted MS spectra for the unmodified light chain [LC(0)] of the ADC are shown in Figure 7. The spectra

resulting from DFA modified mobile phase showed less sodium (Na) and potassium (K) adducts than FA which,

as a lower strength ion pairing agent, may more readily allow the undesirable creation of adducts. The TFA

separation also shows low metal content but gives a distinct signal for a gas phase TFA ion pair. With DFA

separations, there is little to no sign of a DFA gas phase ion pair. Use of a 10 eV collision energy ensures that

any DFA gas phase ion pairs that are formed upon ionization are eliminated upon low energy collisional

activation.

RESULTS AND DISCUSSION

It has been observed that the use of DFA in place of TFA can actually

afford higher chromatographic resolution, as exemplified in a separation

of NIST mAb subunits (Figure 2).

Moreover, versus TFA, DFA has been confirmed to yield higher

sensitivity MS detection of proteins due to its lower ion-pairing strength

and reduced ion suppression (Figure 3).

These preliminary data clearly indicated that there could be significant

promise in the use of DFA for protein LC-MS. However, two concerns

remained that had to be addressed if new methods based on DFA were

to be developed: 1) that there is no commercially available MS-quality

DFA and 2) that it might be necessary to optimize MS settings

specifically for DFA based mobile phases.

The issue of DFA purity and MS suitability was solved by performing

multiple distillations to produce a low metal content, LC-MS grade form

of the reagent. ICP quantitation of metals (Table 1) shows the

successful reduction of metals in DFA to make it comparable to LC-MS

quality FA and TFA reagents that are currently commercially available.

Figure 4 shows the benefit to using an LC-MS quality DFA reagent in

combination with optimized MS settings. A focus is made on the

deconvoluted mass spectrum of the NIST mAb light chain (LC) to

highlight the minimization of adduct levels. These spectra show that an

optimized MS system is also extremely important to minimizing metal

adducts.

A

0.1% DFA0.1% FA 0.1% TFAPc*=105Pc*=114Pc*=71

DFA0.1%

TFA FA

~2x

0.1% 0.1%

Figure 2. UV chromatograms of NIST mAb subunits separated using a

2.1 x 50 mm BioResolve RP mAb Polyphenyl column, 0.1% modified

mobile phases, a flow rate of 0.2 mL/min and temperature of 80°C.

Figure 3. MS sensitivity compari-

son of 0.1% modified mobile

phase additives. These results

were obtained from the analysis

of IdeS digested, reduced subu-

nits of NIST mAb using a QDa

mass detector.

Table 1. ICP quantitation of metals for DFA, distilled DFA, LC-MS grade

FA, and LC-MS grade TFA shows the effectiveness of using distillation

to produce an LC-MS quality DFA reagent.

Conc. (ppm)

Commercially Available

DFA

LC-MS Quality DFA*

LC-MS Grade FA

LC-MS Grade TFA

Na 41 <1 <1 3

Mg 2 <0.1 <0.1 0.3

K 2 <1 <1 <1

Ca 11 <1 <1 1

Fe 2 <1 <1 <1

Figure 4. Deconvoluted spectra of the NIST mAb LC subunit as obtained

with a 2.1 x 50 mm BioResolve RP mAb Polyphenyl column, 0.1% DFA

modified mobile phases, and Xevo G2-XS QTof mass spectrometer.

*Patent Pending

0.1% FA

Pc*=102.8

0.1% DFA

Pc*=158.2

0.15%

DFA

Pc*=158.4

0.1% TFA

Pc*=149.9

% Peak Area

Fd’(3) = 2.6

% Peak Area

Fd’(3) = 4.0

% Peak Area

Fd’(3) = 5.2

% Peak Area

Fd’(3) = 4.2

scF

c(0

)

LC

(0)

LC

(1)

Fd’(0

) Fd’(1

a)

Fd’(3

)

Fd’(1

a)

min

us H

2O

Fd’(2

b)

Fd’(2

a)

scF

c(0

) D

/P

scF

c(0

)

LC

(0)

LC

(1)

Fd’(0

) Fd’(1

a)

Fd’(3

)

Fd’(1

a)

min

us H

2O

Fd’(2

b)

Fd’(2

a)

scF

c(0

) D

/P

scF

c(0

)

LC

(0)

LC

(1)

Fd’(0

)

Fd’(1

a)

Fd’(3)Fd’(1

a)

min

us H

2O

Fd’(2

b)

Fd’(2

a)

scF

c(0

) D

/P

scF

c(0

)

LC

(0)

LC

(1)

Fd

’(0

) Fd’(1

a)

Fd’(3

)Fd’(2

b)

Fd’(2

a)

Fd’(1a) m

inus H

2O

Figure 5 . UV chromatograms of ADC subunits with zoomed views as

obtained with various acid modified mobile phases using a BioResolve

RP mAb Polyphenyl 2.1 x 150 mm column, flow rate of 0.6 mL/min, and

temperature of 80 °C.

Pc* = 138.1

% Peak Area

Fd’(3) = 4.3

B1.7 µm, 2.1 x 150 mm ACQUITY BEH C4 300 Å

0.1% TFA at 80 °C with 10% IPA in Mobile Phase B

2.7 µm, 2.1 x 150 mm BioResolve RP mAb Polyphenyl

0.1% TFA at 80 °C with 10% IPA in Mobile Phase B

A

Pc* = 152.5

% Peak Area

Fd’(3) = 4.3

2.7 µm, 2.1 x 150 mm BioResolve RP mAb Polyphenyl

0.1% TFA at 80 °C without IPA in Mobile Phase B

Pc* = 153.8

% Peak Area

Fd’(3) = 4.2

2.7 µm, 2.1 x 150 mm BioResolve RP mAb Polyphenyl

0.15% DFA at 80 °C without IPA in Mobile Phase B

Pc* = 158.4

% Peak Area

Fd’(3) = 5.2

Pc* = 134.3

% Peak Area

Fd’(3) = 4.2

Pc* = 142.1

% Peak Area

Fd’(3) = 5.4

DC

FE

Starting Point

New Platform Method Control1.7 µm, 2.1 x 150 mm ACQUITY BEH C4 300 Å

0.15% DFA at 70 °C without IPA in Mobile Phase B

2.7 µm, 2.1 x 150 mm BioResolve RP mAb Polyphenyl

0.15% DFA at 70 °C without IPA in Mobile Phase B

0.1% FA

1.29e6

0.1% DFA

5.23e5

0.15% DFA

9.85e5

0.1% TFA

6.61e4

3.24%

+Na2.3%

+K

3.07%

Glycation

+ 162 Da

5.47%

-H2O

1.74%

+Na 0.3%

+K

0.33%

Glycation

+ 162 Da

5.98%

-H2O

4.77%

-H2O 1.83%

+TFA1.16%

+Na 1.09%

+K

0.36%

Glycation

+ 162 Da

5.92%

-H2O

0.57%

+Na 0.2%

+K

0.1% FA

2.59e5

0.1% DFA

1.46e5

0.1% TFA

6.25e4

0.15% DFA

2.22e5

Figure 6. Raw mass spectra of the of the unmodified light chain [LC(0)] of the ADC using various modified mo-

bile phases and optimized MS settings using a BioResolve RP mAb Polyphenyl 2.1 x 150 mm column and a

Xevo G2-XS QTof mass spectrometer.

Figure 7 . Deconvoluted MS spectra of the unmodified light chain [LC(0)] of the ADC as obtained with various

acid modified mobile phases and optimized MS conditions using a BioResolve RP mAb Polyphenyl 2.1 x 150

mm column and Xevo G2-XS QTof mass spectrometer.

Figure 8. UV spectra of the ADC subunit separations showing the incremental improvements in performance

upon changing from the previous platform method (0.1% TFA modified mobile phase using a BEH C4 column)

to the new platform method (based on 0.15% DFA and a BioResolve RP mAb Polyphenyl column), in alphabeti-

cal order from A to F. Panel F provides a control comparison to show the enablement of the method via the use

of a BioResolve RP mAb Polyphenyl column.

Figure 1. A) Schematic representation of the BioResolve RP mAb Poly-

phenyl 450Å 2.7 µm stationary phase. B) Chemical structure of DFA.

MS

Initial Conditions

Desolvation Gas Flow (L/h): 600

Collision Energy: 6 eV

Final Conditions

Desolvation Gas Flow (L/h): 100

Collision Energy: 10 eV

Reag

en

t Q

uali

ty D

FA

LC

-MS

Qu

ali

ty D

FA

11.2%

+Na

10.0%

+K

7.02%

-H2O

1.98e6

2.89%

Glycation

6.71%

+Na

6.82%

+K

7.19%

-H2O

3.07%

Glycation

9.60e5

10.08%

+Na

7.7%

+K

6.78%

-H2O 2.39%

Glycation

2.59e6

2.46%

+Na

1.95%

+K

7.28%

-H2O

2.96%

Glycation

1.20e6

B

Fc/2

LC

Fd’