for MA and Beyond John Gabrielson CASSS 2 nd International Symposium on Higher Order Structure of Protein Therapeutics

Transcription

1 Higher Order Structure Characterization for MA and Beyond John Gabrielson CASSS 2 nd International Symposium on Higher Order Structure of Protein Therapeutics February 12, 2013

2 Proteins form multiple levels of structure Primary Structure Amino Acid Sequence Peptide map Secondary structure Sub-structure (α-helix, β-sheet, etc.) FT-IR Tertiary Structure Three dimensional structure Mean nresidueellipticity(degcm 2 decimole -1 ) Near UV CD Wavelength (nm) Quaternary Structure Self-associated structure c(s) Monomer Aggregate SV-AUC sedimentation coefficient (S) 2

3 Higher order structure characterization methods support two components of product characterization Higher order structure of proteins in bulk solution Elucidation of structure for MA Characterization of reference standard Evaluation of analytical comparability Higher order structure of protein product variants Linking primary structural modifications to biological function Quality by design 3

4 Analytical comparability evaluations comprise several components Process C Comparability bilit Batch Analysis Evaluate Process changes In process controls Evaluate Historical ranges Post-change trends Product Ch Characterization t i ti Stability and Forced Degradation Evaluate 1, 1 2, 2 3 structure t t Orthogonal methods Evaluate Product P d t stability t bilit Deg rates / pathways Product characterization enables additional understanding of product structure & profile and potential impacts of a process change to drug safety and efficacy Some slide elements courtesy of Gustavo Grampp 4

5 Analytical comparability throughout the product lifecycle duct Compa arability Rigo or An nalytical Pro First in human Report results Fewer methods Scale or site change Protocol More methods Wider AC, less data Final commercial process Protocol All relevant methods Tighter AC, more data Quantitative AC Qualified methods MA enabling comparability Marketing Authorization Major change Degree of rigor should be the same as MA enabling study Minor change Degree of rigor should be the same as MA enabling study, smaller subset of methods may be justified Phase 1 Phase 2 Phase 3 Post-approval Commercialization Timeline Magnitude of Manufacturing Change 5

6 Case Study 1: Setting quantitative acceptance criteria for comparability d 2 A/dλ 2 FT-IR Red = structurally altered MRE (deg cm 2 dmol -1 re esidue -1 ) Near UV CD Red = native Wavenumber (cm -1 ) Visual assessment: PASS Numerical assessment: FAIL Difference: > 10 std deviations Safety/efficacy risk! Wavelength (nm) Visual assessment: FAIL Numerical assessment: PASS Difference: < 3 std deviations Manufacturer risk! Manufacturer and patient both incur greater risk with a visual evaluation ation of comparability because of the subjective nature of the assessment Teska, et al., Anal Biochem, 434 (2013)

7 Case Study 1: Setting quantitative acceptance criteria for comparability Precision Method error Lot-to-lot variation Sensitivity Ability to discriminate between results which are truly different Blending studies λ 2 d 2 A/d Wavenumber, λ (cm -1 ) % β-sheet h tprotein 90% β-sheet (10% α-helix) 80% β-sheet (20% α-helix) 70% β-sheet (30% α-helix) 50% β-sheet h t (50% α-helix) h ) 30% β-sheet (70% α-helix) 20% β-sheet (80% α-helix) 10% β-sheet (90% α-helix) 100% α-helix protein Difference (W WSD) Weigh hted Spectral β-sheet h tprotein α-helix protein Percent Impurity Changes detectable at Method qualification allows meaningful acceptance criteria to be established 7

8 Quality by Design: Product understanding & Process control Cell culture Purification UF/DF Drug Substance Process Control Media Feed Strategy Set Points Metals Enzyme Inhibitors Column Pooling Deamidation V L Isomerization Product Attributes C L V H C H 2 Oxidation C H 3 Pro Amidation Cyclization C H 1 Cleavage Disulfide variants Glycosylation Oxidation C-Terminal Lys Biological relevance of product attributes (criticality) Process design for control of CQAs (understanding process impact) Some slide elements courtesy of Bob Bailey 8

9 Applying HOS methods to QbD: Linking structural modifications to biological function Characterization defined by phase Ph 1/2 Limited HOS characterization Ph 3 Advanced MA - Comprehensive Selection of variants for characterization Criticality Abundance Tiered analysis based on method sensitivity E.g., DSC, CD, FT-IR Cascading approach Primary structural modifications Folding, HOS, stability MOA, potency, safety Deamidation V L Isomerization Product Attributes C L V H C H 2 Oxidation C H 3 Pro Amidation Cyclization C H 1 Cleavage Disulfide variants Glycosylation Oxidation C-Terminal Lys 9

10 Case Study 2: Oxidation, conformational stability Cytokine 1 Cytokine 2 Cytokine 3 Increasing oxidation Offset Cp Offset Cp Offset Cp Temperature Temperature Temperature IgG A IgG B IgG C Increasing oxidation Offset Cp Offset Cp Offset Cp Temperature Temperature Temperature 10

11 Case Study 2: Oxidation, conformational stability, biological activity Potency (% %) Comparable Tm1 Comparable Potency Tm1 (CH2 domain) Tm shift ( ΔTm, C) Potenc cy (%) Comparable Tm2 Comparable Potency Tm2 (Fab domain) Tm shift ( ΔTm, C) Increasing oxidation from left to right in both graphs 11

12 Summary & Conclusions Higher order structure characterization should be phase appropriate, reflecting comprehensive product understanding for MA submission Elucidation of structure, reference standard characterization, analytical comparability, and characterization of relevant product variants Analytical comparability for MA and post-approval process changes should reflect product understanding Characterization methods selected to fit the context of the process change Quantitative acceptance criteria based on scientific understanding and product history HOS characterization of product variants Consider criticality of primary structure modifications Consider abundance and types of variants Goal is to strengthen link between structure and function 12

13 Acknowledgements Amgen Analytical Sciences Amgen R&D Amgen Quality Kelly Arthur Yijia Jiang Leslie Sidor Nikita Dinh Linda Narhi Art Chirino Brad Winn Cynthia Li Darrin Cowley Brent Kendrick Bob Bailey Drew Kelner University of Colorado Brandon Teska John Carpenter 13

14 Thank you 14

15 Why study higher order structure? For proteins, structure defines function The secondary and tertiary structures of protein products are critical product quality attributes Loss of higher-order structure will impact efficacy Loss of higher-order structure may impact safety Regulatory expectations for analytical comparability are increasing Manufacturers should confirm that higher order structure is maintained in the product. (2005: ICH issued international a guidance document Q5E) Image courtesy of Article: Therapeutic antibodies 15

16 There are many approaches to assess the similarity between two spectra Correlation coefficient Area of spectral overlap Derivative correlation algorithm Multivariate curve resolution -50 Euclidean distance Weighted spectral difference (WSD) WSD = n i= 1 ( y y ) 1i 2i n y 1i y j= 1 1 j y 1 and y 2 are vectors that represent the reference and sample spectra to be compared Prestrelski et al., Biophys. J. 1993, 65, Kendrick et al., J. Pharm. Sci. 1996, 85, Jiang et al., J. Pharm. Sci. 2011, 100, Li et al., J. Pharm. Sci. 2011, 100, Shashilov and Lednev, Chem. Rev. 2010, 110, dmol -1 resid due -1 ) RE (deg cm 2 M Wavelength (nm) What differences are important? Absolute difference Shape Local details

17 Regulatory expectations for analytical comparability are increasing 1996: FDA issued a guidance document on comparability Provide extensive chemical, physical and bioactivity comparisons with side-by- side analyses of the old product and new product. Tests should include those routinely used for release and others used to fully characterize product structure. 2005: ICH issued international guidance document Q5E Characterize the product when a manufacturing change has been made that has the potential to impact quality attributes. Manufacturers should confirm that higher order structure is maintained in the product. Recent feedback from FDA Objective and pre-defined acceptance criteria should be developed for characterization ti assays used to evaluate product comparability. Protein higher order structure data should be provided from the most discriminating techniques available. Higher order structural ral characterization ation of protein products is consistently addressed in regulatory expectations of analytical comparability Some slide elements courtesy of Brent Kendrick 17