WEBINAR: Absolute quantification of mAbs and ADCs: forget amino acid analysis and shift to the new gold standard

1Arnaud Delobel, PhD – Webinar Quality Assistance – May 23rd 2019
Absolute quantification of
mAbs and ADCs: forget
amino acid analysis and
shift to the new gold
standard
Arnaud Delobel, PhD
Director, R&D

Agenda
• Presentation of Quality Assistance
• Absolute quantification: why and how?
• Development of an ICP/MS method for protein quantification
• Method validation
• How to apply the method to ADCs?

Quality Assistance sa
100% analytical services
100% (bio)pharmaceutical industry
189 highly-qualified employees
> 60% university graduates
77 worldwide R&D companies
& 200 projects (2018)
Product dedicated support
Customised project management
Compliance with EMA / FDA regulations
From discovery to market place
All laboratories on one site
5700 m²
9600 hours
5 clients
8 projects
ADC expertise
(2018)
+35 years experience
~1.1 M € in machinery
& equipment (2018)
22200 hours
16 clients
35 projects
mAbs expertise
(2018)

• Quantity of the active substance in a drug product is a critical
quality attribute, directly linked to the efficacy and safety of the drug
• A precise and accurate method is required to evaluate batch-to-
batch consistency
• Determination of protein concentration is a prerequisite to measure
the extinction coefficient ε of a protein
▪ ε is used as part of the fine characterisation of a protein
▪ ε is used for comparability or biosimilarity studies
Quantification of proteins: why?

• Absolute vs relative quantification: What does absolute mean ?
▪ Quantification without the reference substance
▪ No need for a well characterised standard of the protein to be quantified!
• Usual « absolute » methods
▪ Colorimetric (Lowry, Bradford, BCA, …)
• Usually with BSA as standard: large variability, important bias (up to 50 %)
▪ Amino Acid Analysis
▪ UV spectroscopy: requires knowledge of the molar extinction coefficient, in other
words prior availability of the reference protein
• Classical relative methods
▪ ELISA, LC-MS, LC-UV, …
▪ All require calibration with a standard solution of the protein to be assayed. Such a
standard is not readily available in most cases
Quantification of proteins: how?

• Complex workflow:
• Low throughput method
• High variability (%RSD: 5 – 10 %)
• Low accuracy (90 – 110 %)
Amino Acid Analysis
Vacuum acidic
hydrolysis
Derivatisation
UPLC/UV
analysis

Inductively-Coupled Plasma / Mass Spectrometry
ICP/MS

• Metallo-proteins
▪ ICP-MS obvious for a protein containing a metallic hetero-element with a
known stoechiometry.
• Fe for Hemoglobin
• Se for a Seleno-protein
▪ But applicable for a few proteins only !
• Common proteins
▪ Most proteins (> 95 %) contain Sulphur (Methionine and Cysteine residues)
▪ If elemental composition known:
Sulphur determination = protein quantification
▪ e.g. BSA : C2932H4614N780O898S39 : 1.883 % (w/w) S
Why ICP/MS for protein analysis?

• High ionization potential (10.4 eV) → low sensitivity
• For 32S+, 34S+, 32SO+, 34SO+ : many polyatomic interferences
Sulphur determination by ICP/MS: a challenge!

Sulphur determination by ICP/MS
• New instrument technology available: ICP-QQQ
(Agilent 8800/8900 Triple Quadrupole)
• Allows operation in MRM-like mode
• Elimination of virtually all interferences
• High sensitivity
• LOQ governed by residual sulphur in reagents and solvents

• Sample preparation: digestion is needed
• High sensitivity requires low contamination
• Sulphur contained in the formulation must be taken into account
• To increase accuracy and precision: isotope dilution with 34SO4
(IDMS)
Method development
Undigested sample solutions Sulphur recovery
1 ppm S as H2SO4 100 %
1 ppm S as Methionine > 95 %
1 ppm S as BSA < 90 %

Quantification using Isotope Dilution

• Isotope ratio: Q1 & Q2 are set alternatively to 32/48 or 34/50 to measure 32S or
34S
• Double IDMS
▪ Black marbles = 34SO4 solution added to samples
▪ 34SO4 spiking solution is prepared by oxidation of 34S Sulfur: its exact concentration is not
accurately known.
▪ Concentration of the 34SO4 solution is determined by reverse IDMS with a certified solution of
SO4 (natural Sulfur isotopic abundance), NIST traceable.
▪ Reverse isotope dilution allows also to compensate for sentivity differences observed in
ICP-MS between 32S and 34S (mass discrimination)

Workflow of the method
34S
Oxidation
(HNO3 / H2O2, 30’)
Standard stock
solution
Addition of sulphur
ICP standard

34S
Mineralisation
(HNO3 / H2O2, 30’)
Standard stock
solution
Addition of sulphur
ICP standard

Standard stock
solution
Sample solution
(50 µg S in 1-5 mL)
3 kDa MWCO-filtered
sample solution
Mineralisation
(HNO3 / H2O2 / HCl, 30’)
ICP-MS/MS analysis
Sulphur (32S/34S) quantification by IDMS
Absolute protein
quantification

ICP-MS parameters
• RF power: 1550 W
• Ar carrier gas: 0.25 L/min
• Dilution gas (Ar): 0.85 L/min
• O2 in reaction cell: 0.35 ml/min
• First quadrupole at m/z = 32 (or 34 for 34S)
• Last quadrupole at m/z = 48 (or 50 for 34SO+)
Total: 1.1 L/min

• According to ICH Q2(R1) guideline
(Validation of analytical procedures – Text and methodology)
• Model compound: BSA (Bovine Serum Albumin)
NIST certified solution: 67.38  1.38 g/L ( 2.0 %)
• Additional compounds tested
▪ Trastuzumab solution (label concentration: 21 mg/mL)
▪ Human IgG (Immunoglobulins mixture from Ph. Eur.)
▪ USP BSA for Protein Quantitation RS (2.00 mg/mL)
Method validation

• Linearity of the method:
▪ R = 0.99998
▪ Precision: RSD < 0.4 %
▪ Accuracy: 101.2 – 101.3 %
Method validation - Linearity
BSA (SRM 927e)
(mg)
Sulphur
(µg)
Sulphur
(µg/g)
Average
recovery (%)
RSD
(%)
n
1.6 30 0.6 101.29 0.34 3
2.1 40 0.8 101.30 0.21 3
2.6 50 1.0 101.26 0.22 6
3.2 60 1.2 101.17 0.13 3
3.7 70 1.4 101.17 0.12 3
y = 19.02520x + 0.09521
R² = 0.99996
20
30
40
50
60
70
80
1 1.5 2 2.5 3 3.5 4
Sulfurfound,µg
Sample mass (BSA dose, mg)

Method validation - Precision
(repeatability and intermediate precision)
• Two different analysts carried out independently 6 determinations of
the same solution of BSA and human IgG .
• Samples
▪ BSA: NIST, SRM 927e: 67.38 mg/mL
▪ IgG: human immunoglobulin BRP batch 1
• Ph. Eur. Reference standard for molecular size: unknown content
Sample Analyst 1 Analyst 2 Total RSD (%)
Average % RSD (n = 6) Average % RSD (n = 6) n = 12
NIST BSA (mg/mL) 67.28 0.41 67.50 0.52 0.6
IgG (mg S/g) 8.39 0.13 8.33 0.19 0.7

• Evaluation of matrix effect
▪ Same amount of BSA spiked with different formulation ingredients
Method validation – Matrix effect
Matrix Mean recovery (%) % RSD (n = 3)
0.9 % NaCl 100.27 0.05
PBS 100.24 0.36
10 % Sucrose 100.16 0.61
0.1 % PS-80 100.26 0.57
15 mg/mL Glycine 99.61 0.37
15 mg/mL Histidine 100.76 0.23
Worst case 100.33 0.21

• Quantification in the presence of non-proteinaceous sulphur:
▪ Source of non-protein sulphur: Sulphate
▪ A BSA solution is spiked with SO4 ( 5 % of BSA Sulphur)
▪ Part of this solution is filtered through 3 kDa
• Corrected BSA recovery in spiked solution: 99.6 %
Sample Sulphur (µg/g)
A: unspiked BSA solution 36.50
B: spiked BSA solution 38.81
C: filtered solution 2.44
B-C 36.37

• Quantification in the presence of non-proteinaceous sulphur:
▪ Non-protein sulphur: Methionine
▪ Methionine was added to BSA and IgG solution
▪ Solutions were filtered and Methionine recovery measured in filtrate
25
BSA sol IgG sol
Methionine added (mM) 2.05 2.05
Methionine in filtrate (mM) 2.03 2.04
% recovery 99.1 99.5

Method validation - Stability of solutions
• Stability of standard, sample and blank solutions were studied after
digestion and dilution
• Storage in metal-free polypropylene tubes
• 34S/32S ratios measured on different days:
no change > 1 % in 7 days
→ All solutions are stable for at least 7 days

Method validation – Impact of sample dilution
• What if we have a sample with low protein concentration?
→ Need for a large volume of sample to get 50 µg of S
→ Dilution of acids used for digestion
→ Will the digestion be complete?
• Test:
▪ Constant amount of BSA solution (40 mg) and of 34SO4
▪ Constant volumes of digestion reagents
(2 mL HNO3, 0.5 mL HCl, 1 mL H2O2)
▪ Variable volumes of water added before digestion (1 to 5 mL)

Method validation – Impact of sample dilution
BSA solution
(mg)
H2O
(mL)
Mean recovery
(%)
RSD
(%)
n
40 0.9 100.51 0.32 3
40 1.4 100.92 0.34 3
40 2.0 100.59 0.18 3
40 2.5 100.64 0.07 3
40 3 100.32 0.14 3
40 4 100.89 0.15 3
40 5 100.77 0.18 3

Study on Trastuzumab and USP BSA
Trastuzumab solution
• Expected concentration: 21 mg/mL
• Unknown formulation
• Non proteinaceous Sulphur: 1.7 ppb ( 0.2 % of Trastuzumab Sulphur)
• Found: 20.53 mg/g (RSD = 0.02 %, n = 3)
• Recovery: 97.8 % (vs. the nominal concentration)
• Result should be corrected for density (unknown)
BSA for Protein Quantitation (USP Reference Standard)
• Label concentration: 2.00 mg/mL (unknown accuracy, determined on basis of UV
280nm absorption)
• Found: 1.96 mg/mL (RSD = 0.4 %, n=3)
• Recovery: 98.0 % (vs. label concentration)

Conclusions
• The method characteristics are:
▪ Good sensitivity: 1 or 2 mg of protein in 5 mL is enough
▪ High reproducibility: RSD < 1 %
▪ Great accuracy: < 2 % bias
▪ Insensitivity to matrix effect
▪ Same method for any protein or peptide
▪ Uses only very common reagents and standards
▪ Fast: results obtained within half a day of lab work in routine use
• It is suitable for:
▪ Assay of a single protein or mAb in solution
▪ Assay of a protein adsorbed on an adjuvant (vaccines)
▪ Accurate determination of a protein or peptide standard solution concentration
▪ Determination of the molar extinction coefficient
• Excellent orthogonal method to colorimetry, AAA, theoretical molar extinction
coefficient

Comparison of protein quantification methods not requiring
the reference protein
Colorimetry
Extinct. coeff.
(calculated)
Amino Acids
Analysis
Sulfur by
ICP-MS
Single protein Yes Yes Yes Yes
Need for a standard
protein
Yes (BSA) No
No
(Amino Acids)
No
(Sulfate sol.)
Bias Significant
Can be
significant
5 - 10 % < 2 %
Precision (% RSD) > 5 % N/A > 5 % < 1 %
Interferences Yes
Yes
(ex: ADC)
No No

How to apply the method to ADCs?
• If the drug/linker contains no sulphur: easy!
• If the drug/linker contains sulphur:
Determination
of DAR by LC/MS Calculation of mean number
of sulphur atoms per ADC
ICP-MS/MS analysis and
determination of ADC
concentration

Take-Home Messages
AAA is no longer the
reference method for
absolute quantification of
proteins
You no longer have to
accept > 10%
uncertainty in your
extinction coefficient
determination
Our ICP/MS method was
validated and showed
excellent precision and
accuracy
(RSD < 1%, bias < 2%)
The method can be
applied on a wide range
of formulations without
further optimisation The method can be run
in a full GMP
environment

Acknowledgements
• Juliusz Bianga
• Damien Mouvet
• Philippe De Raeve

arnaud.delobel@quality-assistance.be
+32 71 53 47 81
www.quality-assistance.com
Technoparc de Thudinie, 2
B-6536 Donstiennes (Belgium)
Thank you for
your attention
Any question?

WEBINAR: Absolute quantification of mAbs and ADCs: forget amino acid analysis and shift to the new gold standard

Recommended

Recommended

More Related Content

What's hot

What's hot (7)

Similar to WEBINAR: Absolute quantification of mAbs and ADCs: forget amino acid analysis and shift to the new gold standard

Similar to WEBINAR: Absolute quantification of mAbs and ADCs: forget amino acid analysis and shift to the new gold standard (20)

More from Quality Assistance s.a.

More from Quality Assistance s.a. (15)

Recently uploaded

Recently uploaded (20)

WEBINAR: Absolute quantification of mAbs and ADCs: forget amino acid analysis and shift to the new gold standard