Clemens Stilling Characterisation, Analytical Sciences for Biologicals, UCB Pharma S.A. Alun, living with Parkinson s disease

Size: px

Start display at page:

Download "Clemens Stilling Characterisation, Analytical Sciences for Biologicals, UCB Pharma S.A. Alun, living with Parkinson s disease"

Britton May
5 years ago
Views:

1 Clemens Stilling Characterisation, Analytical Sciences for Biologicals, UCB Pharma S.A. Alun, living with Parkinson s disease

2 Outline Short intro Aggregate characterisation during candidate selection and development Case Study: Candidate selection Process flow Feedback loop to ensure a inherently developable candidate is chosen for development Case Study: Aggregation understanding during development Aggregation understanding by forced degradation studies Increase in aggregates in a manufacturing batch and linking back to FDS Not included: formulation or process development PEG mab Fab Fab-PEG 2

3 Introduction to aggregates and biopharms Aggregates are linked to immunogenicity Biopharma companies pro-actively aim to minimise the aggregate levels therefore minimising immunogenicity Selecting candidates early with low inherent aggregation propensities Developing a manufacturing process which reduces aggregates Developing formulations which are unfavourable towards aggregation Mapping out aggregation pathways and develop understanding of aggregates 3

4 Lifecycle management Discovery Candidate Selection Development Phase I-IV Commercial 4

5 Aggregation propensity minimisation during candidate selection Case study: Aggregation screening informs candidate selection Discovery Candidate Selection Development Phase I-IV Commercial 5

6 Candidate selection Many candidates against a target are evaluated by various developability criteria Biophysical properties are evaluated early in the process to test for inherent aggregation propensity This occurs in tandem with biochemical screening Aggregation, deamidation, oxidation, etc. This allows for re-engineering Number of samples Discovery Acquisition Informatics HTP T m Biacore Candidates Biochem /Biophys Screen OK Phase I-IV Commercial Re-engineer 6

7 Bioinformatics during candidate selection Bioinformatics plays an increasingly important role in discovery and candidate selection. Basic parameters MW, chemical formulae, pi, exctinction coefficient Homology modelling Deamidation prediction T m prediction Aggregation prediction Secondary and tertiary structure prediction Solvent accessibility Disulfide connectivity 7

8 Aggregation by agitation: inherent properties vs. formulation mM sodium acetate/90mm sodium chloride, ph 5 Absorbance at 595nm AA 2A B 3A C 4A D Vortexing 25 o C 1400rpm, Nephelometry at 595 nm Time(h) 25mM sodium citrate/125mm sodium chloride, ph 6 Stability at ph5: B<A<D<C Stability at ph6: A<B<D<C Absorbance at 595nm A B C D 1C 2C 3C 4C Distinct differences between molecules. the next generation biopharma leader 13 March 2013 Time(h) 8

9 Molecular charge and aggregation: pi and zeta potential Zeta potential = molecular charge in standard buffer (10mM NaPO 4 ). Zero lower ph than pi Can increase tendency to aggregate if molecular charge approaches zero Zeta potential (mv) 20 CDP850 D CDP571 E CDP7524 A 10 Emab C CDP6038 B 496_IgG F 0-10 Charge Size ph Most likely to aggregate when Zeta = 0 9

10 Effect of buffers on aggregation by vortexing Absorbance at 595nm mab A ZP=0 at ph Ac Absorbance at 595nm mab B ZP=0 at ph Time(h) Ac Absorbance at 595nm mab C ZP=0 at ph Ac Absorbance at 595nm mab D ZP=0 at ph Ac 10mM phosphate (zeta conditions) or acetate ph5 (Ac) In 10mM Na phosphate, fastest initial aggregation rate at ph closest to Zeta 0 Greater propensity to aggregate in 115mM (acetate) buffer, ph 5, than 10mM Na phosphate buffer ph

11 Melting point / start of melting as an indicator of stability DSC calorimetry of unfolding induced by heating IgG4 s melt at lower temp (more unstable) than IgG1, due to Fc (CH2) Overall stability includes Fab IgG A and B worse than D Tm1: IgG4 CH2 Tm2: Fab, CH3 B A D C 11

12 Formulation strategy to minimise aggregation Commonly aggregation is partially controlled by the addition of excipients to the final formulation Low concentrations of surfactant (common excipient) inhibit denaturation and aggregation at air-liquid interface Aggregation still has to be controlled in the manufacturing process Effect of vortexing 25 o C 1400rpm 1mg/ml +/- 0.02% tween 80 A acetate + tween A acetate - tween B acetate + tween B acetate tween A citrate + tween A citrate tween B citrate + tween B citrate - tween 12

13 Summary for candidate selection Molecular charge and ph: Choose buffer ph and type to avoid zero molecular charge Select/engineer molecule pi that allows required formulation ph Molecular stability: in given conditions, higher Tm = less aggregation tendency Select/engineer higher Tm Adopt more stable format, e.g. IgG1 or Fab -PEG Combinations of stresses may exacerbate aggregation Avoid combinations e.g. zero molecular charge and agitation Protect final DS formulation with surfactant 13

14 Aggregation characterisation in Development Case study: a change in aggregation profile during manufacturing in a mab Discovery Candidate Selection Development Phase I-IV Commercial 14

15 Incorporating Characterisation and FDS Studies into the Product Lifecycle Research Protein engineering Candidate selection Stress study Development Phase I Method development & Formulation GMP DS Process 1 Characterisation & FDS & investigations GMP DS Process 2 Characterisation & FDS & investigations Stability/ Accelerated Learnings Phase III Comparability: Characterisation FDS data 15

16 Observation of HMW species Initial observation from batch release data Increased level of HMW species Question 1: Are these new species? Question 2: Do we know the pathway? Investigation undertaken (purification & characterisation) Semi-prep SE-HPLC SDS-PAGE, native PAGE, DLS, SEC-MALLS, MALDI-MS Results Compare with learnings from FDS Question 3: could FDS data have prevented the investigation? Outcome

17 HMW 1 Ref Comparison of Batch SE-HPLC Profiles (Drug Substance) Monomer Batch 1 HMW 2 Batch Percent Peak Area (%) Aggregates HMW 2 HMW1/2 HMW 1 Total Monomer Batch 2 Ref Batch Batch Batch Batch 3 Formation of aggregates can potentially affect activity and immunogenicity profiles of biopharmaceuticals

Non-Reduced SDS-PAGE (3-8% Tris-Acetate) Denaturing conditions MW Markers Blank DS Batch 1 HMW 2 Batch 1 HMW 1 Batch 1 DS Batch 2 HMW 2 Batch 2 HMW 1 Batch 2 Blank MW Markers kda 500 290 240 ~ 1200

18 Non-Reduced SDS-PAGE (3-8% Tris-Acetate) Denaturing conditions MW Markers Blank DS Batch 1 HMW 2 Batch 1 HMW 1 Batch 1 DS Batch 2 HMW 2 Batch 2 HMW 1 Batch 2 Blank MW Markers kda ~ 1200 kda ~ 900 kda ~ 600 kda ~ 330 kda ~ 290 kda Extrapolated Dimer 160 ~ 160 kda Monomer Batch 1 Batch 2 Similar profiles Non-covalently bound species will dissociate under denaturing conditions Lane 5 & 8: ~ half of HMW 1 (dimer) are non-covalently bound Lanes 4 & 7: HMW 2 species are predominately (~80%) non-covalently bound

Clear Native Gel (3-8% Tris-Acetate) reserves integrity of non-covalent species MW Markers Blank Ref Std Blank DS Batch 1 HMW 2 Batch 1 HMW 1 Batch 1 DS Batch 2 HMW 2 Batch 2 HMW 1 Batch 2 Blank MW

19 Clear Native Gel (3-8% Tris-Acetate) reserves integrity of non-covalent species MW Markers Blank Ref Std Blank DS Batch 1 HMW 2 Batch 1 HMW 1 Batch 1 DS Batch 2 HMW 2 Batch 2 HMW 1 Batch 2 Blank MW Markers SEC profiles Mixture of oligomers Dimer Monomer Batch 1 Batch 2 Similar profiles Lane 7 & 10: Confirms HMW 1 are mainly a dimeric species Lanes 6 & 9: confirms HMW 2 species is a mixture of oligomers (di-,tetra-, hexamer )

20 SEC-MALLS Data SEC with multi-angle laser light-scattering (MALLS), viscometer and refractive index detectors Provides MW, hydrodynamic radii, intrinsic viscosity and % aggregates Ref Std Right angle Light Scattering (mv) HMW 2 HMW 1 DS from Batch 2 DS from Batch 1 Estimated average MW of HMW 2 > 1,500,000 Da (limit of working range of SEC column)

21 Dynamic Light Scattering Measures intensity of laser light that is scattered from particles Larger particles scatter light >> smaller particles Provides an estimation of diameter size of particles DS (14 nm) HMW 1 (17 nm - Dimer) HMW 2 (37 nm) DS HMW 1 (Dimer) HMW 2 Confirms that HMW 1 are a dimeric species; HMW 2 data suggest that average MW > decamer

22 Mass Spectrometry: Cross-linking aggregates Reference standard sample HMW purified sample

23 Stability and clinical data Batch 2 with increased HMW2 % monomer by SEC - stability at 5 o C within specification for aggregates at manufacture within specification for aggregates at end of shelf-life On stability and accelerated stability studies: Batch 1, Batch 2 Lower specification limit -70 o C, 5 o C and 25 o C batch 2 did not form aggregates at a faster rate than other batches. % monomer by SEC - accelerated stability at 25 o C Batch 1, Batch 2 23

24 HMW 1 and 2: Characterisation Summary HMW 1: Data consistent for a dimer (MS, AUC, SEC-MALLS,SDS and native PAGE, and DLS) About half of the dimer species is made from non-covalent bonds 97% of species was reducible to Heavy and Light chain species HMW 2: Analysed by SDS and native PAGE, SEC-MALLS, MS and DLS Predominantly (80%) non-covalently linked species Fully reducible to Heavy and Light chain species Mixture of oligomers with MW up to decamers Q1: Are they the same species? A1: Same species present in all batches but levels vary (0.02% to 0.33% for HMW 2)

25 Native PAGE (Silver Stain) of FDS Samples MW Markers Ref Std DS Batch 2 HMW 2 Batch 2 DS Batch 1 Blank DS Batch 1 oxidised 1 wk DS Batch 1 agitated 72 h DS Batch 1 photostressed 10h DS Ref ph wks DS Ref 50 o C 1 day DS Ref 37 o C 12 wks Mixture of oligomers Dimer Monomer

26 Development case study summary Increase of HMW2 at time of release HMW1 and HMW2 were purified and characterised by an array of thechniques. HMW1 represents dimer HMW2 represents oligomers up to decamer An initial FDS screen was performed to identify conditions which mimick HMW2 Identification of simple and informative methods - clear native PAGE (Silver/Sypro Ruby) Confirmation of identified stress conditions using orthogonal techniques e.g. native gels and SE-HPLC

27 Development case study summary Q2: Can we determine the degradation pathway? A2: Answer: photostability appears to mimic closely the observed aggregation pattern. However, aggregation are difficult pathways and to truly understand pathways considerably work has to be performed. Q3: Could we have used the FDS samples to prevent an investigation? A3: This depends on confidence levels The FDS gave 5 different options. Analysing the sample using the same methods allowed better understanding. Thus in this instance the investigation was still necessary Experience: number of studies, platform technology, sequence predictions etc Aggregation is a complex pathway often overlap between Photostability, agitation and oxidation. 27

28 Conclusions Candidate Selection Biophysics/biochemical screen early in project (including other factors: ph stability, chemical stability, etc) Select / re-engineer candidate to improve easier process development, more stable product Characterisation informs process development and formulation Don t diagnose problem, avoid it! Development Understand the process Have the tools to understand aggregation Perform stress studies early Investigations into abnormal events can make or break projects 28

29 Acknowledgements Characterisation John O Hara Xavier Perraud, Jenni Halley, Martin Hampel, Smita Thobhani, Jun Gan, Kieran Dawkins Protein Biophysics Alison Turner Bryan Smith, Kaushik Sarkar Structural Biology Alistair Henry James Heads Analytical Sciences for Biologicals Carine Dortu 29

Preview Slides for Introduction to Protein Formulation and Delivery

Preview Slides for Introduction to Protein Formulation and Delivery Tim Kelly, Ph.D. VP, Biopharmaceutical Development KBI Biopharma, Inc Durham, NC Pooja Arora, Ph.D. Senior Manufacturing Technical Specialist