Challenges in early clinical development of biologics Peter Lloyd March 2011 Acknowledgements: Jennifer Sims, Phil Lowe, Sebastian Spindeldreher, Andrew Warren
Protein therapeutics biologics Current generation recombinant proteins, binding proteins, cytokines; pegylated proteins: antibodies (mabs), antibody fragments (e.g. Fabs), antibody drug conjugates (ADCs) Next generation? above proteins with non-natural amino-acids / post-translational modifications (eg altered glycosylation pattern), nanobodies, bi-specific antibodies, new scaffolds...... 2 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Monoclonal antibodies vs NCEs: mab New Chemical Entity 150,000 dalton 200-500 dalton Biological production process heterogeneous (post-translational modifications) High species selectivity (affinity / potency) Multi-functional target binding, Fc effector function, FcRn binding Toxicity largely on target mediated exaggerated pharmacology Target can affect PK behaviour (Target Mediated Drug Disposition) Drug -Drug Interaction few examples and mostly PD related Immunogenicity sometimes observed Chemical production process - homogeneous Generally less selective Single target Toxicity often off target mediated Mostly linear PK; non-linearity from saturation of metabolic pathways DDI many examples and metabolic and/or PD related Immunogenicity rarely observed 3 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Outline of the presentation: PK-PD behaviour of mabs - inherent IgG characteristics - ligand binding models and Target Mediated Drug Disposition Developing an integrated bio-analytical strategy - PharmacoKinetics - PharmacoDynamics (target ligand) - Immunogenicity Selection of appropriate starting doses for early clinical testing - FDA guidance 2005 - lessons learnt from TGN1412 incident 2006 and NOAEL / MABEL Summary 4 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Key questions prior to clinical testing: Optimal characterisation of preclinical PK-PD relationships in pharmacologically relevant 1 species to enable an understanding of the clinical therapeutic dose range How much drug do we need to get to the target and for how long? pharmacokinetics / delivery feasible dose and route of administration feasible dosing frequency How will we know if it is safe what is the potential safety liability? How will we know if it works - what are the required PD characteristics? markers of efficacy (biomarkers / surrogate endpoints) 1 what is relevant? - affinity / potency - target expression / turnover - downstream markers 5 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
A simple mab PK model mab iv dose elimination mab limited volume of distribution ~ 7L slow clearance t ½ ~ 20 days (human) FcRn (tight binding at ph6) protects IgG from degradation & explains long serum half-life Roopenian and Akilesh. Nature Reviews Immunology 2007; 7: 715 6 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Mouse, rat and monkey FcRn recognise human IgG NB - binding affinity of human IgG to mouse FcRn is 10x higher than to human FcRn rat cynomolgus monkey mouse - inherent human IgG kinetics scales reasonably well across species - NB when target mediated disposition is absent 7 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
lower FcRn binding affinity Engineered proteins may have altered FcRn binding fusion proteins chimeric mabs human mabs - cept - iximab - umab shorter half-life Suzuki T et al: Importance of neonatal FcR in regulating the serum half-life of therapeutic proteins containing the Fc domain of human IgG1: a comparative study of the affinity of monoclonal antibodies and Fc-fusion proteins to human neonatal FcR J Immunol 2010; 184: 1969-1976 8 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
A simple mab PK-PD model input ligand mab + ligand mab ligand complex dose elimination mab slow clearance elimination ligand rapid clearance Soluble target: mab-ligand complex tends to take on the elimination characteristics of the mab elimination mab ligand complex Cellular target: mab-ligand complex tends to take on the elimination characteristics of the ligand (TMDD apparent) 9 PK/PD of Biologics Mark Milton PJM Offsite Nov 2010
Typical mab PK-PD model(s) Lowe PJ et al: On setting the first dose in man: Quantitating biotherapeutic drugtarget binding through PK and PD models Basic & Clin Pharmacology & Toxicology 2009; 106: 195-209
Example 1: mab against cell surface ligand anti-cd11a mab Raptiva (efalizumab) - rapid clearance of mab-receptor complex - high doses saturate target binding capacity with a return to normal IgG kinetics mab complex ligand increasing dose: change in IgG kinetics (TMDD) Target Mediated Drug Disposition increasing dose: increases duration of effect 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: 10-20
Concentration (linear scale) nm Concentration (log scale) nm Example 2: mab against soluble ligand (MCP-1) mab complex ligand - slow elimination of mab-ligand complex - accumulation of total ligand - rapid turnover of ligand - rapid saturation of binding capacity Accumulation inactive complex 1 x 0.1 mg/kg 2 x 0.3 mg/kg total ligand total mab total mab free ligand (not measured) total ligand free mab (not measured) Time (days) Time (days)
Example 3: mab against soluble ligand (IL1- ) - slow elimination of mab-ligand complex - saturable accumulation of total ligand mab complex ligand Accumulation inactive complex Measured ACZ885 (2 x 10mg/kg) Measured total IL-1β 13 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Example 3: mab against soluble ligand (IL1- ) mab complex ligand Accumulation inactive complex Measured total ACZ885 increasing dose: no change in IgG kinetics Simulated free IL1- increasing dose: increases duration of effect 10 mg/kg 3 mg/kg 1 mg/kg 0.3 mg/kg 0.1 mg/kg 0.1 mg/kg 0.3 mg/kg 1 mg/kg 3 mg/kg 10 mg/kg 14 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Scalable components of the PK-PD model: Inherent pharmacokinetics of the mab and clearance of the mab-ligand complex: - PK of monoclonal antibodies will generally follow typical IgG behaviour and scale reasonably well to man and/or exhibit Target Mediated Disposition and be dependent on the amount of target present and its rate of turnover Binding affinity and potency against the target ligand: - species differences understood during characterisation of the mab Expression and turnover of the ligand: - key drivers of the extent and duration of response; can affect PK behaviour - once maximum ligand binding is achieved then increasing the dose will primarily increase the duration of response: are SAD/MAD designs appropriate for early clinical development? - species differences often not well understood - healthy individuals vs disease often not well understood 15 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Outline of the presentation: PK-PD behaviour of mabs - inherent IgG characteristics - ligand binding models and Target Mediated Drug Disposition Developing an integrated bio-analytical strategy - PharmacoKinetics - PharmacoDynamics (target ligand) - Immunogenicity Selection of appropriate starting doses for early clinical testing - FDA guidance 2005 - lessons learnt from TGN1412 incident 2006 and NOAEL / MABEL Summary 16 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Bioanalytical strategy - PK Free- bioactive mab Total mab sandwich ELISA anti-human IgG mab (drug) ligand anti-human IgG mab (drug) anti-human IgG sandwich ELISA (pre-clinical only) mab (drug) competitive ELISA ligand protein digestion bridging ELISA ligand mab (drug) ligand protein LCMS unique signature peptide 17 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Bioanalytical strategy PD (soluble ligand) total ligand (target capture) anti ligand Ab2 mab (drug) anti ligand Ab1 + + sandwich ELISA ligand endogenous binding protein free ligand anti ligand Ab4 anti ligand Ab3 sandwich ELISA mab (drug) ligand endogenous binding protein Technically challenging: detection of total ligand (increasing) in the presence of drug and endogenous binding protein (if relevant) careful selection of capture and detection reagents Technically very challenging: detection of free ligand (decreasing) in the presence of increasing concentrations of total ligand specificity and sensitivity (LLOQ) 18 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Bioanalytical strategy - immunogenicity Biacore Bridging ELISA mab (drug) labelled ADA ADA mab (drug) mab (drug) non-neutralising neutralising non-neutralising neutralising Consequences of immunogenicity: - increase in clearance of mab (immune complexes) - decrease in capacity for target binding (neutralising immunogenicity) - influence on PK and/or PD assays integrated assessment? 19 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
anti-human interleukin mab [µg/ml] Serum concentration rel. Response Unit [RU] Example 4: mab against cell surface ligand (receptor) Single dose 10 mg/kg i.v. n=3 animals 300 PK profile 1400 1200 Immunogenicity 200 1000 800 600 100 400 0 0 10 20 30 40 50 Time [day] 2001 2002 2003 competitive ELISA mab (drug) ligand 200 0-200 0 10 20 30 40 50 Time [day] 2001 2002 2003 Biacore ADA mab (drug) 20 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Serum concentration [ug/ml] Rel Response Untis [RU] Example 4: mab against cell surface ligand (receptor) Three doses 100 mg/kg i.v. days 0, 7, 14 (n=3 animals) 10000 1000 100 400 350 300 250 PK assay ligand mab (drug) ligand 10 200 150 bridging ELISA 1 0 100 50 0 0 20 40 60 80 100 120 Days Immunogenicity ADA - bridging ELISA overcomes limitations of competitive ELISA to measure exposure to drug in presence of ADA response Biacore mab (drug) 21 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Outline of the presentation: PK-PD behaviour of mabs - inherent IgG characteristics - ligand binding models and Target Mediated Drug Disposition Developing an integrated bio-analytical strategy - Pharmacokinetics - Pharmacodynamics (target ligand) - Immunogenicity Selection of appropriate starting doses for early clinical testing - FDA guidance 2005 - lessons learnt from TGN1412 incident 2006 and NOAEL / MABEL Summary 22 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
FDA guidance: 2005 Step 1 Determine No Observable Adverse Effect Level (NOAEL) Step 2 Step 3 Convert NOAEL to a Human Equivalent Dose (HED) - generally normalised to body surface area (low MW NCEs) - mg/kg normalisation recommended for proteins >100K daltons Select HED from the most appropriate species - additional factors: metabolism, receptors, binding epitopes - default: most sensitive species (lowest HED) Step 4 Apply a safety factor (>10-fold) to give a: Maximum Recommended Starting Dose (MRSD) Step 5 Adjust MRSD based on the pharmacologically active dose (PAD) 23 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
FDA guidance: 2005 Pro simple to use supported by historical evidence (mainly conventional NCEs) Con implicit in guideline primary focus: NOAEL secondary focus: pharmacologically active dose (PAD) over simplified scaling to man focus on dose not exposure one algorithm fits all step 5 often ignored 24 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
TGN1412 incident Northwick Park: 2006 Expert Scientific Group on Phase I clinical trials Sir Gordon Duff November 2006 245 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Duff Report recommendations: 2006 NB reliant on robust measures of pharmacology (PD biomarkers): - target (eg receptor occupancy or ligand binding) - mechanism (eg downstream signalling) - outcome (eg clinical response) 246 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
EMEA guidance: 2007 Risk mitigation strategies 27 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
FDA guidance: 2005 Step 1 Determine No Observable Adverse Effect Level (NOAEL) Step 2 Step 3 Convert NOAEL to a Human Equivalent Dose (HED) - generally normalised to body surface area (low MW NCEs) - mg/kg normalisation recommended for proteins >100K daltons Select HED from the most appropriate species - additional factors: metabolism, receptors, binding epitopes - default: most sensitive species (lowest HED) Step 4 Apply a safety factor (>10-fold) to give a: Maximum Recommended Starting Dose (MRSD) Step 5 Adjust MRSD based on the pharmacologically active dose (PAD) ie MABEL 28 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
NOAEL and MABEL NOAEL No Observable Adverse Effect Level FDA highest dose level that does not produce a significant increase in adverse effects an effect that would be unacceptable if produced by the initial dose of a therapeutic in a phase I clinical trial conducted in adult healthy volunteers MABEL - Minimal Anticipated (Acceptable) Biological Effect Level minimal exposure / dose level that is anticipated to produce an acceptable biological effect an effect that would be considered acceptable if produced by the initial dose of a therapeutic in a phase I clinical trial 29 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Example 5: target suppression in safety assessment mab complex ligand humanised mab; high affinity against soluble target Accumulation inactive complex K d man < cynomolgus monkey (~10-fold) typical IgG kinetics target ligand can be measured in the systemic circulation at baseline and is increased in disease mab acts as a capture system : mab-ligand complex (detected in serum) is a biomarker for suppression of free ligand via a PK/PD model previous experience with a similar molecule (lower affinity) against the same target 30 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Concentration (µg/ml) Concentration (µg/ml) Concentration (ng/ml) Concentration (ng/ml) Concentration (ng/ml) Concentration (µg/ml) Concentration Concentration (µg/ml) Concentration (ng/ml) Example 5: 4wk GLP toxicology study mab complex ligand 1000 100 10 PK (exposure) DOSE: 0 DOSE: 03 10000 1000 100 10 1 PD (total ligand) 1 0.1 DOSE: 0 DOSE: 10 3 0 DOSE: 10 15 3 1000 100 Accumulation inactive complex 10 1 0.1 1 0.1 DOSE: 10 DOSE: 20 15 10 DOSE: 120 15 PK-PD model: 1000 exposure and 100 total ligand 10 pre-clin exposure data conform 10000 10000 1000 1000 1000 100 100 10 10 10 1 1 10000 10000 1000 1000 1000 100 100 10 10 10 1 1 1 0.1 1 0.1 to a 2-compartment model DOSE: 20 0 28 DOSE: 120 56 84 DOSE: 120 0 28 56 84 0 10000 Time (days) Time (days 1000 1000 increase in total ligand (measured) 1000 100 100 fitted to ligand binding model 10 10 10 1 PK-PD model allows estimation of 1 0.1 1 free ligand (target 0 suppression) 28 56 84 0 28 56 84 0 28 56 84 Time (days) Time (days) NB: free ligand (green) not measured Time (days) 31 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Serum concentration (ug/ml) Example 5: predicted exposure (PK) in man mab complex ligand 10000 Accumulation inactive complex Cynomolgus monkey PK-PD model scaled to human: 1000 100 10 Cynomolgus monkey 120mg/kg NOAEL (observed mean data) >100-fold exposure safety margin (C max ) - IgG PK behaviour 1 10.0 mg/kg - increased target binding affinity (10-fold) 0.1 0.01 0.1 mg/kg 0.3 mg/kg 1.0 mg/kg 3.0 mg/kg 0.001 0 20 40 60 80 100 120 140 160 Time (days) 32 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Concentration (ng/ml) Concentration (ng/ml) Example 5: predicted PD (effect) in man mab complex ligand Simulation: predicted target suppression in man - pre-clin exposure data adjusted to man - binding affinity in PK-PD model adjusted to man Accumulation inactive complex 1000 MABEL predicted free ligand 100 10 0.1 mg/kg 0.3 mg/kg 1000 Monte-Carlo simulation 100 1 1 mg/kg 3 mg/kg 10 mg/kg 10-28 -14 0 14 28 42 56 70 84 Time (days) 98 1 0.1 112 126 140 154 168 182-28 -14 0 14 28 42 56 70 84 98 Time (days) 112 126 140 154 168 182 33 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Example 5: Justification of starting dose in man Toxicology (exposure) [1] NOAEL 120 mg/kg [2/3] HED 120 mg/kg - adjust for anticipated exposure in man? [4] Apply >10-fold safety factor 12 mg/kg Pharmacology (response) [5] PAD / MABEL - justify based on pharmacology - max pharmacology achieved at 20mg/kg (6-fold lower than NOAEL) - adjust for anticipated exposure in man (0.1 mg/kg projected to give >100-fold margin) - include anticipated duration of effect - adjust for inter-species differences in affinity / potency (>50% ligand suppression for <48h) 0.1 mg/kg Maximum Recommended Starting Dose - define anticipated safety window based on NOAEL and MABEL * 0.1 mg/kg * - NB an additional factor may be added based on uncertainty of data / prediction and relative risk 34 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Benefits of an Integrated Approach Simple mathematical representation of known biology also represents components of the model which cannot be measured (e.g. low circulating level of free ligand) Hypothesis testing the ability to test assumptions prior to experimental design, leading to better pre-clinical and clinical studies Sensitivity analysis elements of the model which are key drivers of the desired outcome (e.g. target expression / target turnover / affinity / potency) 35 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
Summary: Challenges in early clinical development of biologics Inherent IgG characteristics make mabs attractive drug targets; however: target biology can affect PK behaviour of mabs / duration of action - target expression (how much and where?) and target turnover (how quickly is it replaced?) - normal vs disease, gender, species differences should be understood - other sinks (e.g. shed or decoy receptors), off target binding, endogenous binding proteins A bioanalytical strategy and appropriate tools needs to be developed - integrated interpretation of PK, target binding and immunogenicity An integrated PK/PD-(IG) approach can quickly identify appropriate dose and dosage regimen to achieve desired target suppression - influence and impact on traditional study design 36 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011
37 Early Cin Dev of Biologics Peter Lloyd BeAPP March 2011 Thank you for your attention