Comparability of a human IgG1 after cell line switching: A retrospective view Peter Lloyd Head of PK-PD, Novartis Biologics

Disclaimer The views and opinions expressed in the following PowerPoint slides are those of the individual presenter and should not be attributed to Drug Information Association, Inc. (“DIA”), its directors, officers, employees, volunteers, members, chapters, councils, Special Interest Area Communities or affiliates, or any organization with which the presenter is employed or affiliated. These PowerPoint slides are the intellectual property of the individual presenter and are protected under the copyright laws of the United States of America and other countries. Used by permission. All rights reserved. Drug Information Association, DIA and DIA logo are registered trademarks or trademarks of Drug Information Association Inc. All other trademarks are the property of their respective owners.

Outline of the presentation: Introduction to case study: IgG PK and mAb-ligand binding models: - consequences for characterisation of extent and duration of effect Retrospective analysis: - can rodent replace NHP in PK studies to explore inherent IgG behaviour? - can immunogenic potential be detected with appropriate PK and PD analytical strategies? - can appropriate clinical strategies during clinical development provide more definitive information on comparability? Summary: - “yes they can”

Case study:  IgG 1 monoclonal Ab  binds to and neutralises a soluble cytokine target - target not detected at baseline in systemic circulation  cross-reactivity established with marmoset and human target (similar affinity)  PK assay for total drug in serum (marmoset / human)  PD assay for total ligand in serum (human)

Product AProduct BProduct CProduct D NS0-derived ≥52 mg/mL Sp2/0-derived ≥65 mg/mL Sp2/0-derived ≥180 mg/mL Sp2/0-derived ≥175 mg/mL - human and marmoset cross- reactivity Comparability exercise: - phys - chem - PK (NHP) Comparability exercise: - phys - chem - PK (NHP) Comparability exercise: - phys - chem -marmoset: 4 and 26 wk iv, 13 wk s.c. toxicology and EFD -- Clinical studies NS0 = expression cell line (mouse myeloma) Sp2/0 = expression cell line (mouse myeloma) HSA = Human serum albumin Case study:

Case study: Phase IV Phase III Phase II Phase I Pre-clin Research cell line switch

Case study: PK comparability non-human primate – NSO  – SP2/0 Time (days) Serum mAb concentration (  g/mL) single sc dose x-over study n=16 marmosets PK comparability established: - AUC - C max and t max

Outline of the presentation: Introduction to case study: IgG PK and mAb-ligand binding models: - consequences for characterisation of extent and duration of effect Retrospective analysis: - can rodent replace NHP in PK studies to explore inherent IgG behaviour? - can immunogenic potential be detected with appropriate PK and PD analytical strategies? - can appropriate clinical strategies during clinical development provide more definitive information on comparability? Summary:

slow clearance V ~ 7L t ½ ~ 300 h (human) iv dose elimination mAb mAb FcRn protects IgG from degradation & explains long serum half-life Roopenian and Akilesh. Nature Reviews Immunology 2007; 7: 715 A simple mAb PK model:

mouse IgG kinetics scale reasonably well using an allometric approach NB only when target mediated disposition is absent rat non-human primate Mouse, rat and cyno FcRn recognize human IgG man

mAb – ligand complex elimination mAb – ligand complex ligand input ligand elimination ligand + slow clearance V ~ 7L t ½ ~ 300 h dose elimination mAb mAb A simple mAb PK-PD model:

 Inherent pharmacokinetics of the mAb (IgG behaviour): - species differences often well understood and easily characterised; good prediction to man  Binding affinity to the target ligand: - species differences understood during characterisation of the mAb  Turnover of the ligand and clearance of the mAb-ligand complex: - species differences and behaviour of mAb-ligand complex sometimes not well understood - if cell surface ligand is not saturated with mAb then the mAb-ligand complex may be cleared very quickly (target mediated clearance) - for a soluble ligand the mAb-ligand complex will tend to follow IgG clearance mechanisms (eg case study) Components of the PK-PD model:

Single dose: 0.01 mg/kg 0.1 mg/kg 1.0 mg/kg 10.0 mg/kg Assumptions mAb with typical IgG kinetics; Kd = 0.37nM; soluble ligand, turnover (half-life 3h) soluble ligand Case study: PK-PD model

anti-CD11a mAb – Raptiva (efalizumab) Joshi et al An overview of the pharmacokinetics and pharmacodynamics of efalizumab: a monoclonal antibody approved for use in psoriasis J Clin Pharmacol 2006; 46: cell surface ligand Literature example:

–PK of monoclonal antibodies will generally follow “typical IgG behaviour” and scale reasonably well to man and/or exhibit Target Mediated Drug Disposition (TMDD) and be dependent on the amount of target present and its rate of turnover –Once maximum ligand binding is achieved then increasing the dose will primarily increase the duration of response –The shape of the exposure response curve can be simulated from pre- clinical data by adjusting parameters in the model (eg binding affinity, target expression) –The amount of ligand present and the rate at which it can be replaced (turnover) will be key drivers of the extent and duration of response Summary: Ab-ligand PK-PD binding models

Case study: consequences for comparability testing For an antagonistic mAb cleared primarily by IgG mediated clearance pathways, a study with PK endpoints will explore only inherent IgG characteristics this data could be generated equally well in rodents, as rodents clear human IgG by similar pathways to man Differences in conventional non-compartmental PK parameters (Cmax, tmax) at high saturating doses may not impact on efficacy Binding to target is often not addressed in a PK-only study design

Literature example: glycosylation pattern (rodent) Milward et al Effect of constant and variable domain glycosylation on pharmacokinetics of therapeutic antibodies in mice Biologicals 2008; 36: 41-47

Case study: impact of immunogenicity drug anti-drug antibody mAb “typical IgG kinetics” increase in IgG clearance increase in IgG clearance no effect on target binding decrease in target binding (total ligand, receptor occupancy) (decrease in total ligand) Target Mediated Disposition increase IgG clearance increase/decrease in clearance? (change in inflection point?) (change in inflection point?) no effect on target binding decrease in target binding non-neutralizing neutralizing

PKPD (soluble target) * Immunogenicity Analytical strategy: impact of neutralizing immunogenicity residual bioactive drug anti-human IgG residual bioactive drug total drug (non-human only) mAb (drug) ligand capture Ab NB: capture Ab has different ligand binding epitope compared with the drug and no steric interference * - for cell surface target use receptor occupancy and/or PK profile a) b) c) a) b) c) nIG may interfere in the PK and PD assays; detected as short PK t½. Immune complexes may also be cleared more rapidly. Decrease in total target may be apparent nIG may intefere in the PK and PD assays; detected as increase / decrease in exposure. Decrease in total target may be apparent nIG does not interfere in the PK assay; although immune complexes may be cleared more rapidly. Decrease in total target may be apparent drug-ligand complex / total target or LC-MS

Case study: population analysis of clinical data Phase III clinical study - incorporate comparability testing into clinical design; product A (NS0) and B (SP2/0) as covariates (n=35 patients per treatment) - serum total mAb and total ligand measured with sparse sampling strategy predict free ligand from PK-PD binding model - ligand capture (and therefore ligand suppression) will also detect neutralizing immunogenicity

Case study: population analysis of clinical data NSO SP2/0 Time (days) Serum mAb concentration (  g/mL) Serum ligand concentration (pg/mL) mAb (PK profile) Total ligand (PD profile) Conclusions for materials derived from both cell lines: - similar sc bav - similar in-vivo K D derived from clinical PK-PD binding model - no loss total ligand over time (ie no evidence of neutralizing immunogenicity) Typical PK and PD profile from the PK-PD model

Case study: summary Can rodent replace NHP in PK studies to explore inherent IgG behaviour? Industry perception and concern that NHP primate PK studies are necessary is driving acceptance that these will be the norm; however, for this case study (mAb targeted against a soluble ligand), pre- clinical studies in NHP with PK-only endpoint could probably be replaced by a PK study in rodent, even when the mAb does not bind the target in rodent (only the inherent IgG behaviour is characterised)

Case study: summary Can immunogenic potential be detected with appropriate PK and PD analytical strategies? If total ligand capture can be easily measured, then neutralizing immunogenicity should be apparent as a decrease in total ligand; this can easily be monitored in Phase III clinical trials; providing a robust assessment of immunogenic potential; additional neutralizing immunogenicity assays are probably not required

Case study: summary Can appropriate clinical strategies during clinical development provide more definitive information on comparability? doses are often given at saturation of target ligand binding and differences in dose therefore affect duration of action rather than extent; conventional PK parameters to assess comparability (eg Cmax and tmax) may not be appropriate clinical studies with target capture and appropriate analytical strategy are more informative than pre-clinical studies regarding comparability of efficacy and immunogenic potential during drug development comparability testing can be built into clinical study design

THANK YOU for your attention Acknowledgements: Jennifer Sims (Head Translational Sciences and Safety, Novartis Biologics) Phil Lowe (Modelling and Simulation, Novartis) Annette Zaar (Head of Immunogencity, Novartis Biologics) Fabienne Deckert (Head of PK-PD Bioanalytics, Novartis Biologics)