Applied Protein Services

Transcription

1 Company / City, Country Applied Protein Services First Last / Lonza Inc. / DD Month YYYY

2 Agenda Overview Capabilities Project Example & Highlights Case Studies slide 2

3 Applied Protein Services: Core Technologies Algonomics Epibase & Epibase IV Computer-based and cellular immunoprofiling for lead selection, lead comparison, optimization and surveillance Tripole Protein Engineering Technology platform for protein structure analysis, modeling, engineering, deimmunization and antibody optimization AggreSolve Applied Protein Services is a platform of immunogenicity, stability and protein engineering services An integrated in vitro-in silico technology platform to help resolve protein aggregation and stability challenges slide 3

4 Applied Protein Services: Risk Assessment & Optimization Discovery Development Manufacture Distribution basic research disease discovery drug discovery drug development clinical trials production packaging marketing sales distribution Early Risk Assessment Improve Lead Properties Quality / Comparability lead discovery lead selection In silico Assessment Immunogenicity Stability lead optimization process development & formulation In vitro Assessment Immunogenicity Stability Mitigate risks Improve quality & safety Reduce attrition slide 4

5 Therapeutic Protein Development Challenges / Risks Immunogenicity Antibody response against protein/antibody therapeutics Loss of efficacy, PD, PK changes, danger to patients. (EPO) Immunogenicity must be addressed for each biologic Aggregation Increase in production costs Low productivity Additional purification steps QC controls, longer timelines Stability, Regulatory Compliance Reduced half-life and reduced compliance (i.e. particles) Activity & Pharmacology Altered activity and bioavailability, PK / PD Enhanced immunogenicity slide 5

6 Agenda Overview Capabilities Project Example & Highlights Case Studies slide 6

7 Applied Protein Services: Capabilities Immunogenicity Screening Aggregation Removal Tripole Epibase Epibase IV AggreSolve Deimmunization Humanization Deaggregation Rapid MS In vitro Validation Immunogenicity Aggregation Activity Expression Cell Line / Strain Construction slide 7

8 Protein Services & Solutions Lead Selection Stage Early risk assessment, high throughput Select optimal leads early for lower immunogenicity risk Lead Optimization Stage Optimize lead candidates for improved safety and developability Removal of T-cell epitopes (reduce immunogenicity risk) Removal of aggregation hot-spots (improved stability) Quality / Comparability Assessment / Filing & Data Support Quality monitoring, batch comparability, formulations, biosimilars In vitro immunogenicity, stability assessment slide 8

9 Predict and Reduce Immunogenicity B cells need T-cell help to produce high affinity antibodies (i.e. IgG against protein / antibody drugs) B-cell epitopes: The surface patches required for B-cell response Eliminate T-helper epitopes will potentially reduce / diminish immunogenicity T-cell epitopes: Peptide sequence of the protein needed for effective T-cell help slide 9

10 Epibase In Silico Immunoprofiling Epibase Epibase IV lead discovery lead selection lead optimization process development & formulation preclinical clinical trials Epibase Epibase IV In silico T-Epitope Identification In vitro T-cell Assays Anti-Drug Antibody Screening Patented T-cell epitope screening platform to predict potential immunogenicity risk of protein leads. Supports ranking of antibody and protein leads and the selection of leads with minimum risk Crucial information to reduce immunogenicity risk by engineering Epibase patent granted QI 2007; FASTER patent granted QIII 2006 slide 10

11 Epibase Explained Model Building Template identification: Retrieve HLA subtypes of known 3D structure that are at least 50% identical to a given HLA subtype Build a 3D structure Run the Proprietary FASTER Algorithm Select relevant part of the receptor Include the flexibility of side chains Determine Binding affinity Promiscuity EP , Proteins, 2002; Proteins, 2005 slide 11

12 Epibase In Vitro Screening Strategy HLA typing SSP-PCR or InnoLiPa Check by independent reviewer Donor 50 PBMC preparation Check quality PBMC by polyclonal T-cell activation assay Peptides/proteins Assay development Screening 50 donor cell preps Technology platform Flow cytometry ELISpot Readout & data management Following projectspecific flow charts QC/QA of raw data and manipulations Data analysis Using statistical program «R» Final report slide 12

13 Tripole A proprietary platform for structure analysis & protein modeling Builds / analyzes antibody and protein 3-D structure Safeguards antibody structure integrity during in silico engineering Identifies desirable substitutions that de-immunize a protein Humanizes antibodies Explores and enhances protein / antibody structure characteristics VH, VL Structure Modeling CDR Conformation Modeling Exploration of Best VH, VL and CDR Combinations Generate Feasible Antibody Structures slide 13

14 AggreSolve Protein Aggregation Solutions AggreSolve is a technology platform introducing Quality by Design (QbD) to resolve aggregation and stability issues Combined In vitro-in silico Service For proteins/antibodies with documented aggregation issues Addresses aggregation hotspots and stability weak points Evaluate effect of amino acid substitutions on stability and aggregation Protein Re-design Prediction, modelling, protein engineering and subsequent in house in vitro testing of potential lead molecules Aggregation Hotspot slide 14

15 Epibase In Vitro Immunoprofiling Epibase Epibase IV lead discovery lead selection lead optimization process development & formulation preclinical clinical trials Epibase Epibase IV In silico T-Epitope Identification In vitro T-cell Assays Anti-Drug Antibody Screening Measuring and analyzing T-cell response using donor peripheral blood mononuclear cells (PBMCs) on population level Immunoprofiling for lead comparison of novel proteins and biosimilars, cell line profiles slide 15

16 CDR Grafting and Humanization A humanised AB contains sequences from 6 complementary determining region (CDRs) from a non-human MAB into a human framework. This process is known an CDR grafting Apart from the antigen binding loops of the CDRs the presence of key residues in the framework is also very important to retain antibody affinity The key expertise is to keep non-human sequences to a minimum without affecting the antibody affinity Apart from affinity framework sequences are also selected for stability and other desirable properties slide 16

17 Humanization of Murine Antibody Mouse V H and V L Mouse CDRs* Human constant domains Human constant domains & framework regions Chimeric Humanized *CDR: Complementary Determination Region, where antigen binds slide 17

18 Lonza s Technology Build 3D antibody model of the non human antibody Select most suitable frameworks based on the structure and sequence homology Graft CDRs into these suitable frameworks Deimmunization: Epibase analysis and avoid/remove T-Cell epitopes wherever possible Candidate mabs are carefully designed to safeguard affinity and stability slide 18

19 Our Value Proposition Extensive scientific and technological know-how on antibody structure analysis 3-D modelling capabilities (Tripole) in house for CDR grafting Peer reviewed publications for AB modelling related work Our combined approach of humanization and deimmunization will result in a less immunogenic molecule than performing classic humanization Epibase is the leading in silico tool in identifying T cell epitopes Track record in deimmunization Integrated in silico and in vitro approach we make variants, deimmunize followed by expression and functional analysis In vitro screening include a rapid productivity and aggregation screening to choose best candidate for future development (manufacturability) slide 19

20 Humanization Work Flow CDR Grafting Epibase Input: Customer Sequence Ab model buiding (Tripole) In silico CDR grafting (Humanization) Analyze/deimmunize using Epibase In silico modelling (4 weeks) Transient Expressing and purification Activity Analysis Output: Optimized Sequence(s) Expression and screening (binding affinity) of humanised variants in-house Supply lead candidates to Customer Output: Humanized Leads Sequences In vitro screening (16 weeks) Output Option: Humanized Material slide 20

21 OD 450nm Log EC50 (pm) Relative Yield Real Life Example: Anti Hen Egg Lysozyme (D1.3) Binding affinity retained Good productivity Parental Variant 1 Variant 2 0 Parental Variant 1 Variant 2 In-house control Parental 1.0 Variant Variant pm slide 21

22 OD 450nm Relative Yield Log EC50 (pm) Real Life Example: Anti-IFNγ Ab Binding affinity retained EC50 (pm) Parental Control Variant Variant Parental Humanised Control Variant 2 Variant 3 Chimeric Humanised control Variant Variant log pm Productivity can be improved Parental Humanised control Variant 2 Variant 3 In-house control slide 22

23 Agenda Overview Capabilities Project Example & Highlights Case Studies slide 23

24 Example: Protein Screening & Optimization Projects Development Candidate In Silico Epibase In Vitro Epibase IV Epitope map Whole protein Protein Engineering Tripole Deimmunization Humanization AggreSolve Engineering Preferred / Optimized Candidate slide 24

25 Highlight: Epibase Light Server Model Customer-specific Epibase Application Server: Multi-user, password-protected system, work from different sites Near unlimited number of sequence submissions Automated epitope prediction and report generation Uses the latest Epibase release Analyse sequence groups as a batch notification of results Input data, predictions, reports stored in the internal database Web-based, customized user interface Secure, encrypted data transfer, using SSL certificate slide 25

26 Customers widely accept our Applied Protein Services Novel Products 4-Antibody Genmab Octopharma Ablynx GSK Pierre Fabre Alligator Bioscience IAVI Pieris Arana/Cephalon J&J ProtAffin Bayer Healthcare Kyowa Hakko Kirin TcL Pharma BioWa Lundbeck Tibotec Crucell MAT Biopharma Transgene GeNeuro Morphosys UCB slide 26

27 Agenda Overview Capabilities Project Example & Highlights Case Studies slide 27

28 Viventia Case Study: De-immunization of VB6-845 Background Viventia s anti-epcam recombinant immunotoxin V I V E N T I A Biotechnologies Inc. Humanized Fab fragment fused to a deimmunized toxin (bouganin) Targets and mediates cell death in EpCAM-positive solid tumors First-in-man phase 1 trial assessed the safety of VB6-845 in 13 patients with various EpCAM-positive cancers Objective Low or no antibody responses to deimmunized bouganin portion Observed immune response to Fab moiety Minimize the potential immunogenicity risk of the fusion protein by de-immunizing the Fab portion slide 28

29 Viventia Case Study: De-immunization of VB6-845 In silico De-immunization Screening for T-cell epitopes using Epibase Antibody structure modeling Substitutions to eliminate T-cell epitopes based on structure integrity In Vitro Verification & Testing of De-immunized Protein Screening for T-helper cell responses using PBMCs from healthy donors Individual and population responses V I V E N T I A Biotechnologies Inc. slide 29

30 Viventia Case Study: De-immunized VB6-845 Fab Epibase Screening Epitope identification on Fab sequence Filtering of epitopes present in human germlines Critical epitopes are identified as strong binders to DRB1 and DRB 3/4/5 Identification of Mutations that Remove Epitopes Prevention of novel epitopes For other allotypes In overlapping frames Respecting structural integrity of the protein Stability Function (e.g. affinity for ligand) V I V E N T I A Biotechnologies Inc. slide 30

31 Viventia Case Study: De-immunized VB6-845 Fab Proposed Changes 19 mutations (11 in VH, 8 in VL) removed critical epitopes or decreased the affinity of remaining epitopes 14 of 19 proposed mutations (10 in VH, 4 in VL) retained expression and affinity for EpCAM: 74% success rate De-immunized Fab has a similar binding affinity to wild type Binding Affinity De-Fab: KD = 1.31x10-9 M Wild Type: KD = 1.56x10-9 M V I V E N T I A Biotechnologies Inc. slide 31

32 Viventia Case Study: In vitro Testing of De-immunized VB6-845 Fab V I V E N T I A Biotechnologies Inc. Compare the immunogenic potential of de-immunized Fab to wild type based on: Number of responsive donors Mean SI over the population Relative response slide 32

33 # responsive donors SI>2 Viventia Case Study: In vitro Testing Single Donor and Population Level V I V E N T I A Biotechnologies Single Donor Population Inc. In vitro testing: # positive donors (53 tested) Relative Response within the test population WT Fab De-immunized Fab WT Fab De-immunized Fab De-immunized Fab shows a substantial and significant reduction in its ability to raise T-cell responses slide 33

34 Viventia Case Study: Conclusions V I V E N T I A Biotechnologies Inc. De-immunized anti-epcam Fab showed a significantly reduced T-cell activation potential in vitro, compared to wild type A 2 nd -generation VB6-845 molecule has been engineered and is now ready for testing in phase 1 trials In silico de-immunization provides a cost-effective and rapid solution to further reducing or avoiding potential immunogenicity risk of therapeutic proteins slide 34

35 Affibody Case Study: Background Half-life extension by Albumin binding Serum albumin an ideal carrier Fusion Partner Albumin Binding Domain (ABD) Serum Albumin Long half-life (17-19 days) Favorable dosing Convenience Safety Wide distribution Present in plasma (40%) and tissues (60%) Albumin Binding Domain (ABD) high affinity for HSA Small size (5 kda) slide 35

36 Affibody Case Study: Objective Originally naturally occurring bacterial protein: ABD wt with known T-cell epitope (Goetsch 2003) Affinity maturation to femtomolar affinity (Jonsson 2008) Structure analysis and modelling of ABD variants Protein engineering for stability, expression yield and reduced immunogenicity Validation by Algonomics/Lonza T-cell epitope mapping in silico T-cell proliferation assay of selected deimmunized mutants ABD wt ABD 035 Bacterial Affinity matured Johansson02 Jonsson08 ABD 094 Deimmunized unpublished slide 36

37 Affibody Case Study: Epibase In Silico Profile of ABD and Variants Epibase screening In silico T cell epitope mapping and ranking of 131 variants Selected based on their stability, affinity and predicted antigenicity / immunogenicity Rational selection of best candidate for in vitro testing slide 37

38 Affibody Case Study: Epibase IV Testing of ABD and Variants Compare the immunogenic potential of wild type ABD and variants based on: Number of responsive donors Mean SI over the population Relative response slide 38

39 Number of responding individuals Affibody Case Study: Epibase IV Testing Results of ABD and Variants 52 * * ABD001 ABD094 rhsa KLH No significant immunogenicity was detected with ABD094 * Significantly different compared to buffer alone slide 39

40 Affibody Case Study: Conclusions In contrast to wildtype ABD, no significant immunogenicity was detected with ABD094 Combined in silico and in vitro approach used for testing of mutants allowed for discrimination of molecules differing in only one amino acid In silico mapping provides a cost-effective and rapid solution to further reducing or avoiding potential immunogenicity risk of therapeutic proteins slide 40

41 GSK Case Study: Deimmunization Humanized Antibody SKFH11L9 Binds beta-amyloid 1-40 Epibase analysis showed relatively high immunogenicity Courtesy of Volker Germaschewski, GSK Biopharm R&D, Stevenage, UK slide 41

42 GSK Case Study: Deimmunization Objectives Identify positions and residues for amino acid replacement to remove predicted Th epitopes Construct deimmunized variants and assess for binding function Combine successful mutations Re-assess the deimmunized molecule by Epibase Courtesy of Volker Germaschewski, GSK Biopharm R&D, Stevenage, UK slide 42

43 GSK Case Study: Tripole In Silico Deimmunization A model of SKFH11L9 was build and compared to murine parent VH/VL 6F6 Amino-acid changes were suggested: 7 in VH, 4 in VL All proposed changes were tested individually and in combinations in ELISA Combinations of VH/VL chains were tested for binding functionality Combinations were re-tested with Epibase to ensure low number of epitopes Courtesy of Volker Germaschewski, GSK Biopharm R&D, Stevenage, UK slide 43

44 GSK Case Study: Comparison of Mutant Affinities Courtesy of Volker Germaschewski, GSK Biopharm R&D, Stevenage, UK slide 44

45 GSK Case Study: Conclusion In silico epitope prediction provides a quick tool to Assess T-cell epitope content based on the major HLA subtypes Provide provide very specific information for amino acid replacements (in conjunction with structural data and modelling) Reduce immunogenicity yet preserve structural integrity By testing the constructs experimentally for binding function and combining only functional replacements in a sequential fashion the best compromise between deimmunization and function can be achieved slide 45

46 Back Up Slides slide 46

47 Examples: Immunogenic Proteins Protein Category Protein Type & Producer Cells Clinical Consequences Non-Human Proteins Insulin Natural Loss of efficacy uncommon Human Proteins Factor VIII Natural Loss of efficacy IFN-α2a rdna Loss of efficacy GM-CSF rdna rdna Loss of efficacy No loss of efficacy Homologous to Native Proteins G-CSF rdna No loss of efficacy IFN-β rdna Loss of efficacy Epo IL-2 rdna rdna Cross reacted with endogenous protein and caused adverse effects. Loss of efficacy associated with both types of antibodies. Sequence Variants IFN-β rdna Loss of efficacy Chemically Modified Pegylated MGDF rdna Cross reacted with endogenous protein and caused adverse effects. Hybrid Molecules GM-CSF/IL-3 hybrid (PIXY 321) rdna Clinical efficacy abrogated TNFR2-Ig rdna No correlation with clinical responses or adverse effects. slide 47

48 Preclinical Immunogenicity Assessment GUIDELINE ON IMMUNOGENICITY ASSESSMENT OF BIOTECHNOLOGY-DERIVED THERAPEUTIC PROTEINS Doc. Ref. EMEA/CHMP/BMWP/14327/ Non-clinical assessment of immunogenicity and its consequences Therapeutic proteins show species differences in most cases. Thus, human proteins will be recognized as foreign proteins by animals. For this reason, the predictivity of non-clinical studies for evaluation of immunogenicity is considered low. Non-clinical studies aiming at predicting immunogenicity in humans are normally not required. However, ongoing consideration should be given to the use of emerging technologies (novel in vivo, in vitro and in silico models), which might be used as tools. slide 48

49 Technical References (1/4) Stas Ph, Gansemans Y, Lasters I. Immunogenicity Assessment of Antibody Therapeutics Current Trends in Monoclonal Antibody Development and Manufacturing Springer AAPS 2010, p Stas Ph, Lasters I. Immunogénicité de proteins d intérêt thérapeutique Medecine Sciences 2009; 25(12): Bancos S, Cao Q, Bowers W, Crispe N. Dysfunctional memory CD8+ T cells after priming in the absence of the cell cycle regulator E2F4 Cell Immunol 2009; 257(1-2): Stas Ph, Pletinckx J, Gansemans Y, Lasters I. Immunogenicity Assessment of Antibody Therapeutics Recombinant Antibodies for Immunotherapy, edt. M. Little Cambridge University Press 2009, p Stas Ph, Lasters I. Strategies for preclinical immunogenicity assessment of protein therapeutics. IDrugs 2009; 12(3): Throsby M, van den Brink E, Jongeneelen M, Poon L, Alard P, Cornelissen L, Bakker A, Cox F, van Deventer E, Guan Y, Cinatl J, ter Meulen J, Lasters I, Carsetti R, Peiris M, de Kruif J a, Goudsmit J. Heterosubtypic Neutralizing Monoclonal Antibodies Cross-Protective against H5N1 and H1N1 Recovered from Human IgM+ Memory B Cells. PlosOne 2008; 3(12):e3942,1-15. Barderas R, Desmet J, Timmerman P, Meloen R, and Casal I. Affinity maturation of antibodies assisted by in silico modeling. PNAS 2008; 105: Van Walle I, Gansemans Y, Parren P, Stas Ph, Lasters L. Immunogenicity screening in protein drug development. Expert Opin Biol Ther. 2007; 7: Fontayne A, Vanhoorelbeke K, Pareyn I, Van Rompaey I, Meiring M, Lamprecht S, Roodt J, Desmet J, Deckmyn H. Rational humanization of the powerful antithrombotic anti-gpibalpha antibody: 6B4. Thromb Haemost. 2006; 96: slide 49

50 Technical References (2/4) Staelens S, Desmet J, Ngo T, Vauterin S, Pareyn I, Barbeaux P, Van Rompaey I, Stassen J, Deckmyn H, Vanhorelbeke K. Humanization by variable domain resurfacing and grafting on a human IgG4, using a new approach for determination of nonhuman like surface accessible framework residues based on homology modelling of variable domains. Mol Immunol. 2006; 43: Desmet J, Meersseman G, Boutonnet N, Pletinckx J, De Clercq K, Debulpaep M, Braeckman T, Lasters I. Anchor profiles of HLA-specific peptides: analysis by a novel affinity scoring method and experimental validation. Proteins. 2005; 58: Van Walle I, Lasters I, Wyns L. Align-m a new algorithm for multiple alignment of highly divergent sequences. Bioinformatics. 2004, 20: Van Walle I, Lasters I, Wyns L. Consistency matrices: quantified structure alignments for sets of related proteins Proteins. 2003, 51:1 9. Boutonnet N, Janssens W, Boutton C, Verschelde JL, Heyndrickx L, Beirnaert E, van der Groen G, Lasters I. Comparison of predicted scaffold-compatible sequence variation in the triple-hairpin structure of human immunodeficiency virus type 1 gp41 with patient data. J Virol. 2002; 76: Desmet J, Spriet J, Lasters I. Fast and accurate side-chain topology and energy refinement (FASTER) as a new method for protein structure optimization. Proteins. 2002; 48: van den Beucken T, van Neer N, Sablon E, Desmet J, Celis L, Hoogenboom HR, Hufton SE. Building novel binding ligands to B7.1 and B7.2 based on human antibody single variable light chain domains. J Mol Biol. 2001; 310: Boutonnet N, Van Belle D, Wodak S. Infuence of ligand binding on the conformation of ligand of torpedo acetylcholinesterase. Theor Chem Acc. 2001; 106: Desmet J, De Maeyer M, Spriet J, Lasters I. Flexible docking of peptide ligands to proteins. Methods Mol Biol. 2000; 143: slide 50

51 Technical References (3/4) De Maeyer M, Desmet J, Lasters I. The dead-end elimination theorem: mathematical aspects, implementation, optimizations, evaluation, and performance. Methods Mol Biol. 2000; 143: Hufton SE, van Neer N, van den Beucken T, Desmet J, Sablon E, Hoogenboom HR. Development and application of cytotoxic T lymphocyteassociated antigen 4 as a protein scaffold for the generation of novel binding ligands. FEBS Lett. 2000; 475: Boutonnet N, Kajava A, Rooman M. Structural classification of a and b supersecondary structure units in proteins. Proteins. 1998; 30: Desmet J, Wilson IA, Joniau M, De Maeyer M, Lasters I. Computation of the binding of fully flexible peptides to proteins with flexible side-chains. FASEB J. 1997; 11: De Maeyer M, Desmet J, Lasters I. All in one: a highly detailed rotamer library improves both accuracy and speed in the modeling of side-chains by dead-end elimination. Folding Des. 1997; 2: Lasters I, Desmet J, De Maeyer M. Dead-end based modeling tools to explore the sequence space that is compatible with a given scaffold. J Protein Chem. 1997; 16: Desmet J, De Maeyer M, Lasters I. Theoretical and algorithmical optimization of the dead-end elimination theorem. in 'Proceedings of the Pacific Symposium on Biocomputing 1997' (1997) Altman RB, Dunker AK, Hunter L, Klein TE. (eds.) World Scientific, Singapore, pp Lasters I, De Maeyer M, Desmet, J. Enhanced dead-end elimination in the search for the global minimum energy conformation of a collection of proteins side chains. Protein Eng. 1995; 8: Boutonnet N, Rooman M, Wodak S. Automatic analysis of protein conformational changes of by multiple linkage clustering. J Mol Biol. 1995; 253: Boutonnet N, Rooman M, Ochagavia ME, Richelle J, Wodak S. Optimal protein structure alignments by multiple linkage clustering: application to distantly related proteins. Protein Eng. 1995; 8: Desmet J, De Maeyer M, Lasters I. The 'dead end elimination' theorem as a new approach to the sidechain packing problem. in 'The Protein Folding Problem and Tertiary Structure Prediction' (1994) Merz K, LeGrand S. (eds.) Birkhäuser Boston, Inc., pp slide 51

52 Technical References (4/4) Lasters I, Desmet J. The fuzzy-end elimination theorem: correctly implementing the side-chain placement algorithm based on the dead-end elimination theorem. Protein Eng. 1993; 6: Desmet J, De Maeyer M, Hazes B, Lasters I. The dead end elimination theorem and its use in protein sidechain positioning. Nature 1992; 356: slide 52