PBPK An Old Hat with New Tricks Amin Rostami-Hodjegan, PharmD, PhD, - PowerPoint PPT Presentation

PBPK An “Old Hat” with “New Tricks” Amin Rostami-Hodjegan, PharmD, PhD, FCP Professor of Systems Pharmacology University of Manchester, Manchester, UK & Vice President R&D Simcyp , Sheffield, UK amin.rostami@manchester.ac.uk

PBPK/IVIVE Linked Models http://www.natu re.com/clpt/jour nal/v92/n1/cov ers/index.html

Systems Approach: e.g. Inter-Individual Variability in PK Americans/Europeans Japanese/Chinese Age / Genetics / Environment / Disease

In Vitr In itro o DMP DMPK K Too ools: ls: e.g. e.g. Meta Metabo boli lism sm HLM HIM HKM rhCYP rhUGT hepatocytes Homogenisation CYP450 S9 FMO HOMOGENATE MAO Aldehyde oxidase Centrifugation @ Aldehyde dehydrogenase 9,000g Epoxide hydrolase Xanthine oxidase S9 Esterases UGT Nuclear / Mitochondrial pellet SULT Centrifugation @ hepatocytes Aldehyde oxidase 100,000g Glutathione S transferase cytosol Alcohol dehydrogenase Cytosol Xanthine oxidase S9 Microsomes

Dr Drug ug-Foc ocuse used Mode d Modell lling ing

System-Foc System ocuse used d Mode Modell lling ing Same Same Type ype of of In Init itial ial Data (CL of of Caf affeine eine and and Theo eophyll hylline ine) BUT UT Afte fter Dec econ onvolution olution to to Acc ccou ount nt for or Oth Other er Age Rela elated ted Compo mponen ents ts of of Clear learan ance (Siz Size, e, Blood lood Flo low, Pr Prote otein in Binding inding, mg mg Micr icrosoma osomal Pr Prote otein in per per Gram am Liv iver er etc) and etc and Sep Separ aration tion of of Ren enal al Pathway thway Relative CYP1A2 Activity 6 4 (Paediatric:Adult) 3 2 1 0 25 30 35 20 40 60 (weeks) (years) Age (PMA)

In In Vitr itro o vs vs In In Viv ivo o Ont Ontog ogen eny y CYP CYP2D6 2D6 an and d 3A4 3A4 Clin Pharmacol Ther 2007 CYP2D6 activity was detectable and concordant with genotype by 2 weeks of age, showed no relationship with gestational age, and did not change with post natal age up to 1 year. However: we know that: Thus, the development of renal function from birth may change in parallel with the development of the enzyme such that the drug/metabolite ratio may be relatively constant !!!!

Ma Matu turation tion of of Ren enal al Clear Clearan ance ce 150 250 GFR (ml/min/1.73m2) Schwartz Rowland 200 GFR (ml/min) Rubin data 100 Simcyp 150 100 50 y = 87.674x - 14.497 R 2 = 0.9988 50 0 0 0 0.5 1 1.5 2 0 50 100 150 200 250 BSA (m2 ) Age (months) Simcyp vs Rhodin Model

Bott Bottom om-Up Up App pproa oach h Meet Meets s Top op-Do Down wn Clin Pharmacol Ther 2008 1 1 CYP3A4 activity (DX/3HM CYP2D6 activity (DM/DX ratio) relative to adult ratio relative to adult (A) 0.8 0.8 (B) 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 4 8 12 0 4 8 12 Age (Months) Age (Months) Figure 1. Changes in CYP2D6 (a) and CYP3A4 (b) activity relative to adult values. The data of Blake et al , corrected for the development of renal function, are indicated by the diamonds. The simulated change in in the activity of each enzyme (solid line) was derived from in vitro data on hepatic enzyme expression and increase in liver weight with age.

Not Not Just ust Cl Clea earan ance ce: Physiolog : Physiology of y of Abso Absorpt ption ion 100 Stomach contents remaining vs time Mean % remaining in studies (all literature reports) 75 50 25 0 0 15 30 45 60 75 90 105 120 135 150 165 180 Time (minutes) β β 1i 2i t t ij ij γ γ 1i 2i y (D PR ) e PR e ij i i i No significant effect by postnatal or gestational age, weight or volume of intake but Food Type a significant COVAR: Aqueous < Breast Milk < Formula Milk < Semi-Solid < Solid (44.8 min) < (56.6 min) < (64.1 min) < (87.0 min) < (97.7 min)

Well ell Rec ecog ognised nised by Le by Lead ading ing Regu gula lato tory Agen y Agencies cies B. PBPK Model components A. Intrinsic/extrinsic Factors System component Drug-dependent EXTRINSIC (drug-independent) component Smoking Environment Lung ADME, PK, PD and INTRINSIC Rapidly perfused MOA Pregnancy Race organs Diet Organ Dysfunction DDI Metabolism Blood Slowly perfused Blood Obesity Gender Active transport organs Age Disease Passive diffusion Genetics Protein binding Alcohol Kidney Regulatory Drug-drug interactions Medical Practice Receptor binding Liver Intestines Huang and Temple, 2008 Individual or combined effects PBPK Model Dosing on human physiology Elimination Predict, Learn, Confirm Zhao P, et al Clin Pharmacol Ther 2011

Just ust Made Made It! It! ASC ASCPT PT 20 2012 12 YES NO

DDI in DDI in Neo Neona nate tes s an and d Inf Infan ants ts Relationship between age and enzyme maturation 1.000 Whole liver relative expression CYP2C9 0.100 CYP1A2 0.010 0.001 0 5 10 15 20 Age (y) Fraction of drug metabolized via a pathway of interest with age 120 100 80 50% 65% % fm 80% 60 95% Pathway 2 40 Pathway 1 50% 20 35% 20% 5% 0 neonate infant child adult 1 3 5 7 9 11 Age

Pr Pregn gnan ancy y as Ano as Anoth ther er E Exa xample mple Time – varying system parameters (anatomical, physiological and biological; CYPs abundance, etc.) Clinical Pharmacokinetics 2012 A PBPK Model to Predict Disposition of CYP3A- metabolized Drugs in Pregnant Women: Verification and Discerning the Site of CYP3A Induction A.B. Ke, S.C. Nallani, P. Zhao, A. Rostami- Hodjegan, J. D. Unadkat Br J Clinical Pharmacol 2012 CPT: PSP 2012

The Blood-Brain Barrier Intestine The Blood-CSF Barrier MDR1 Blood OATP1A2 BCRP Lumen MRP4 BCRP (P-gp) Enterocyte MRP4 OATP2B1 basolateral luminal Blood MDR1 (P-gp) MRP Choroid epithelium Endothelial cells apical BCRP abluminal MDR1 (P-gp) PepT1, Astrocyte feet PepT2, Cerebrospinal fluid (CSF) Brain parenchyma OATP1A2, OATP2B1, OCT3, OCTN1, OCTN2, IBAT, CNT1, CNT2, MCT1, MCT4, MCT5

Systems Systems App pproa oach: Absor h: Absorpt ption ion ADAM Model Duodenum Jejunum II Jejunum I Ileum IV Ileum III Ileum II Ileum I Colon Luminal Transit Stomach Emptying Segregated Blood Flows Enzymes (CYP3A4) vs Transporters (Pgp)

Seco Second nd Gues Guessing Bioa sing Bioavaila vailabili bility: ty: Baria Bariatric tric Sur Surge gery SG RYGB BPD-DS JIB Gastric resection Small intestinal bypass Invasiveness (Elder and Wolfe 2007; Padwal et al 2009; Tucker et al 2008)

Mi Mimic micking king Rou oux-en en-Y Y Gas Gastric tric Bypa Bypass ss - Using Using AD ADAM AM Dissolution / Precipitation / Super-Saturation Pgp (Darwich et al 2012; JPP)

Cy Cyclospo losporine rine – Post ost JI JI-Bypa Bypass ss

Systems App Systems pproa oach: Br h: Brain ain Brain mass Spinal CSF Q Ssink Volume pH Q Sout Q Sin CL met Volume pH Q bulk Volume Q Csink pH PSE Cranial CSF CL Cout CL Bout CL Bin CL Cin PSC PSB Volume Qbrain Qbrain pH Brain blood • CSF: circulation, pH, volume, … (Eyal et al., 2009) • Brain: anatomy, physiology, … • Passive permeability at three interfaces • Active transporters at BBB/BCSFB

Systems Systems App pproa oach h : Kidn : Kidney ey Transporters are available in all MechKiM: Filtration; Secretion (passive + active) Reabsorption (passive + active), Metabolism three proximal tubule cell compartments on the apical and Urinal tubule Cell (renal mass) Renal blood basal membrane. The model can handle: • Regional distribution/activity of transporters • Nephrotoxicity as well as changes in systemic exposure • Interplay between uptake, efflux and passive permeation • Interplay between metabolism and transporters

Assessme Assessment nt of of Proa oarrh rhyth ythmic mic Pot oten ency y : V : Var aria iabili bility ty Toxicology Mechanisms and Methods, 2012 J. of Cardiovasc. Trans. Res. (2012) integration step via the apropriate model for cell (extended to the ionic IONIC MODULE channels attributes of the whole heart wall) ION MEASURED CHANNELS CELL/TISSUE MODULE OUTPUT hERG ionic necessary – channels HUMAN HEART ACTION POTENTIAL ESTIMATED not enough VENTRICULAR CELL (APD90) MODEL SYSTEM POPULATION demography physiology MODULE genetics ATTRIBUTES EPICARDIUM covariates for HUMAN HEART the dynamic VENTRICULAR CELL effects MODEL MIDMIOCARDIUM HUMAN HEART pseudoECG VENTRICULAR CELL (QTc) MODEL LADME IVIVE ENDOCARDIUM HUMAN HEART VENTRICULAR CELL MODEL

Rema emaining ining Que Question stions P s PBPK BPK - Par art t of of MBDD MBDD: : Why? hy? WE ALL KNOW that: (1) All Models Are Wrong, but Some Models Are Useful ! George EP Box 1987 (1) Moving Away from ‘Drug Focused’ to ‘System Focused’ Modelling (2) Requires Different Type of Data (3) Requires Huge Integration Task (4) Appropriate Tools Are Essential PWC - Kate Moss June 2008