Applying physiologically-based pharmacokinetic (PBPK) modeling

for special populations

John DiBella, M.S.E.

Vice President, Marketing & Sales

john.dibella@simulations-plus.com

April 27th, 2017

Outline

• What is required with GastroPlus™ PBPK models?

• Pediatric populations – special considerations

• Disease state modeling – hepatic impairment

• Conclusions

Mechanistic Absorption Modeling (MAM)

Physiologically based Pharmacokinetics (PBPK)

What’s defined in a PBPK model?

• Each compartment represents a tissue:– Specific volume(s) *

– Blood perfusion rate *

– Enzyme/transporter expression levels *

– Volume fractions of lipids & proteins *

– Tissue:plasma partition coefficient (Kp)

• Estimated from drug properties:

– logD vs. pH

– pKa(s)

– Plasma protein binding

– Blood:plasma concentration ratio

* From literature sources

GastroPlus

Dissolution and absorption

in vitro constants:Vmax(s), Km(s), Ki(s), EC50, etc…

Scale to in vivo processes

Nonlinear kinetics (and DDI)

Physical properties -Peff, Sw, pKa, logP,

fup, Rbp

Formulation: dose, dosage form,

particle size,release profile

Structure in silico

in vitro experiments

Plasma/tissue concentration profiles

PKPlus- Vd, CL, K12, K21, K13, K31

PBPKPlus - CLint

The Big Picture – Drug Inputs

Therapeutic/Adverse Effect Data

PBPK/PD modeling

IV/Oral PK data

in vitrometabolism

Structure in silico

Not asking you to generate more data:Let’s just make better use of it!

NH

O

OH

OCH3

CH3

CH3

Discovery Preclinical Clinical

Discovery PK/PreclinicalVirtual candidate selection

First-in-human (FIH) dose selection

Safety research and risk assessment

Pharmaceutical DevelopmentFormulation optimization

Quality by Design (QbD) implementation

Understand food effects

Clinical PK/PharmacologyDrug-drug interactions (DDIs)

Special populations

Virtual bioequivalence trials

Pediatric Populations:Special Considerations

NHANES Average Body Weight:2003 - 2004

Children & Adults:Tissue Sizes

Compare 1975 to 1962 liver weight

0

500

1000

1500

2000

2500

0 5 10 15 20

Age (yr)

Liv

er

We

igh

t (g

)

ICRP 1975

Altman 1962

PEAR

Haddad S. - J. Tox. Envir. Health 64:453 (2001) for ages <= 12Houtkooper, LB, J. Appl. Physiol. 72:366 (1992)Segal, KR, Am. J. Clin. Nutrition 47(1):7 (1988)Separate eqn. for Males and Females, Aged <10, 10-85

Age Dependent Tissue CompositionEffect of age on tissue compositions is included. Example plots for two of the

tissues, Adipose and Brain, are shown below:

PMA – postmenstrual age (gestational + postnatal age)

Scaling Pediatric FupFup scaling is based on changes in total plasma protein (albumin and 1-acid

glycoprotein) using previously published equation (McNamara, AAPS PharmSci, 2002, E4)

adult

adult

adult

pedped

fu

fu

P

Pfu

)1(1

1

Pped and Padult is binding protein concentration in pediatric and adult

subject, respectively; fuped and fuadult is fraction unbound in plasma in pediatric

and adult subject, respectively.

0.0

0.2

0.4

0.6

0.8

1.0

0 0.2 0.4 0.6 0.8 1

Fu child - obs

Fu

ch

ild

- p

red

Pediatric fup observed and predicted from published equation using pediatric plasma protein level as implemented in GastroPlus.

Reported values were for ages 1 day to ~ 4 months.

Plasma Protein and Hematocrit

CYP Enzyme Ontogeny

Johnson T., Clin Pharmacokinet 45(9):931 (2006)

Glomerular Filtration – Neonates

Kearns – New Engl J Med 2003, 349:1157

Intestinal Physiology

• Limited information available for some parameters

• For some parameters the information is only qualitative (i.e., underdeveloped villi structure in infants < 3 years old, or differences in bile salt composition and site of reabsorption)

• Changes made in the GastroPlus™ ACAT™ model include:

– Stomach pH in neonates

– Stomach volume

– Intestinal length and radius (and compartment volumes)

– Transit time

– Enzyme/transporter expression levels (scaled by surface areas)

Transporter Ontogeny

Alcorn Clin Pharmacokinet 2002; Fakhoury et al. DMD 2005De Woskin et al. Reg Toxicol Pharmacol 2008

PStc – Difficult to Predict…Easier to Scale

SAltyPermeabiliPStc

What is the cell surface area of individual tissues?

21 TissueTissue CellVolume

SA

CellVolume

SA

If we assume that average cell sizes are similar across tissues, then:

2

121

Tissue

TissueTissueTissue

CellVolume

CellVolumePStcPStc

Specific PStcFor easier scaling of PStcs across tissues, we introduce:

Specific PStc (= PStc per mL of cell volume)

in vitro CLdiff measured in hepatocytes [uL/min/106 cells] or [uL/min/mg MP]

in vivo Liver PStc

Hepatocellularity or MPPGL and liver size

in vivo Specific PStc

Liver cell volume*

Individual tissue PStc values

Individual tissue cell volumes** Need to account for contribution of apical and basolateral cell surface in certain tissues

• Eliminated by biliary secretion

• Substrate for:

• OATP1B1 and OATP1B3 – uptake transporters expressed on liver sinusoidal membrane

• MRP2 – efflux transporter expressed on apical membranes in liver, kidney and gut

• in vitro data available from rat and human suspended hepatocytes and human sandwich-cultured hepatocytes

• in vivo data available in rat and human (adult and pediatric for 1- to 16-year-old)

Example: Valsartan

Valsartan Model Structure

MRP2

MRP2

MRP2

OATP1B1/1B3

MRP2

Nishimura et al. DMD 2005

IVIVE with Scaling Factors

Liver

IVIVE with All Permeability-Limited Tissues

Liver

all other tissues

Requires PStcestimates for all tissues

IVIVE with PStc Scaling for ALL Tissues

Lukacova – 17th North American ISSX meeting 2011, Atlanta, GA

(Poirier – J Pharmacokinet Pharmacodyn 2009, 36:585)1.Predicted rat IV using in vitro data measured in rat hepatocytes

2.Predicted human IV using in vitro data measured in human hepatocytes

Valsartan: Refine Adult Model

• Passive diffusion through tissue membranes in all tissues scaled from liver PStc predicted from in vitro

• Liver uptake predicted from in vitro

• Secretion into bile, urine, and gut lumen via MRP2

• MRP2 expression in different tissues estimated from reported relative mRNA levels in liver, kidney, small intestine, and colon

Predict pediatric populations

Valsartan: Refine Adult Model

Refine adult model:- fitted passive diffusion through tissue membranes and intestinal permeability

Valsartan: Predict Pediatric DispositionDosing 2 mg/kg, experimental profiles are averages of 6-7 individuals

Need to account for lower villi density in very young children?

(initial assumption – the same transporter density as in adults)

Valsartan: Predict Pediatric DispositionDosing 2 mg/kg, experimental profiles are averages of 6-7 individuals

(initial assumption – the same transporter density as in adults)

Decreased absorptive surface area

Lukacova et al. (2014) AAPS Annual Meeting, San Diego, CA

Examples: Pediatric Populations

Samany et al. (2015) ASCPT Annual Meeting, New Orleans, LA

Oseltamivir

Disease State Modeling:Hepatic Impairment

Liver cirrhosis • Replacement of normal liver tissue with non-functional scar

tissue caused by chronic conditions:

– Alcoholism

– Viral hepatitis

– Non-alcoholic fatty liver, non-alcoholic steatohepatitis

– Bile duct disease

• Pathophysiology includes systemic physiological changes in:liver function, plasma protein concentration, kidney function,hepatic and gut blood flow, cardiac output, enzymeexpression, and hematocrit.

• Child-Pugh score is used to classify the degree of diseaseseverity

Physiological changes in liver cirrhosis

Ref: Johnson – Clin Pharmacokinet 2010, 49:189-206

Ref: Edginton – Clin Pharmacokinet 2008, 47:743-752

Buspirone physicochemical and biopharmaceutical properties

S+LogP = 1.97 (AP 7.2)Exp Log P =2.63 (Gertz 2008)

S+pKa = 7.16 (basic 1) , 2.93 (basic 2)Exp pKa = 7.32 (basic 1), 4.12 (basic 2) (Temple 1982)

S+Sw = 0.17 mg/mL @ pH 8.9 (AP 7.2)Exp solubility = NA

S+Peff = Human: 2.16 x 10-4 cm/s (AP 7.2)Exp Caco 2 Papp = 2.5 x 10-5 cm/s (Gertz 2010)Exp Caco-2 Papp = 3.6 x 10-5 cm/s (Absorption Systems)Converted Effective Human Peff = 4.3 x 10-4 cm/s

S+PrUnbnd (human) = 22.66 % (AP 7.2)Exp unbound = 5% (Bullen 1985)

S+Rbp (human) = 1.29 (AP 7.2)Exp Blood/plasma ratio = 0.81 (Shibata 2002)

AP 7.2 predicted metabolism by CYP3A4 and CYP2D6rCYP model: S+Km = 13.4 µM (3A4) and 7.8 µM (2D6)

S+Vmax = 64.1 (3A4) and 1.8 (2D6) nmol/min/nmol Enz.Experimental : Km = 8.7 µM (3A4) Zhu, Drug Metab. Disp. 33(4):500 (2005)

AP 7.2 = ADMET Predictor v. 7.2S+ stands for properties predicted with Simulations Plus models

Predicted metabolism of Buspirone

Observed metabolic pathway of BuspironePrimary route of Buspirone elimination is metabolism by CYP3A4

Ref: Zhu et al., DMD 33; 500–507, 2005

Drug Log P Basic pKa Tmax (h)

4.69 10.0 27

4.7 10.1 16

3.81 8.52 15

4.46 9.65 7.8

4.39 9.82 7

5.11 9.86 6

Lysosomal Trapping of Lipophilic Cations

Ref: Kazmi F., Drug Metab. Disp. 41(3):897 (2013)

Lipophilic

Amines

LogP > 1

pKa > 6.5

Simulated Caco-2 cellular kineticswith MembranePlus™

This information is used to parameterize fraction unbound in enterocytes in the PBPK models.

PO IR tablets 7.5 and 20 mg BID on day 5 Healthy subjects

Ref: Dockens et al., J Clin Pharm 46; 1308-1312, 2006

Predicted (lines) and observed (points) Cp-time profiles of buspirone (red), 1-pyrimidinylpiperazine metabolite (blue) and 6-hydroxybuspirone metabolite (pink) in healthy adult volunteers after 9 doses of 7.5 mg and 20 mg buspirone hydrochloride administered once a day.

PO IR tablet 10 mg BID on days 1, 5, and 10 Healthy subjects: Buspirone

Ref: Barbhiya et al., Eur J Clin Pharmacol (1994) 46-41-47

PO IR tablet 10 mg BID on days 1, 5, and 10 Healthy subjects: 1-PP metabolite

Ref: Barbhiya et al., Eur J Clin Pharmacol (1994) 46-41-47

PO IR tablet 10 mg BID on days 1, 5, and 10 Decompensated subjects: Buspirone

(default physiology)

Ref: Barbhiya et al., Eur J Clin Pharmacol (1994) 46-41-47

PO IR tablet 10 mg BID on days 1, 5, and 10Decompensated subjects: 1-PP metabolite

(default physiology)

Ref: Barbhiya et al., Eur J Clin Pharmacol (1994) 46-41-47

PEAR Physiology:Default Hepatic Impairment Models

PO IR tablet 10 mg BID on days 1, 5, and 10 Decompensated subjects: Buspirone

Ref: Barbhiya et al., Eur J Clin Pharmacol (1994) 46-41-47

PO IR tablet 10 mg BID on days 1, 5, and 10Decompensated subjects: 1-PP metabolite

Ref: Barbhiya et al., Eur J Clin Pharmacol (1994) 46-41-47

Macwan et al. (2014) AAPS Annual Meeting, San Diego, CA

Liver Cirrhosis Impaired Renal Function

Li et al. (2012) Acta Pharmacologica Sinica 33:1359

Examples: Special Populations

Gastric Bypass

Almukainzi et al. (2014) J Diabetes Metab 5:3

Conclusions

Importance of PBPK modeling for special population groups

• Unwarranted studies, due to the general nature of regulatory guidelines, may be avoided

• Alleviation of the ethical problems and recruitment issues associated with clinical studies using children or subjects with more severe impairment of organs

• Modeling helps to plan and optimize study design

• Model simulations help to predict likely outcome in special populations

Acknowledgements

• Dr. Viera Lukacova

• Dr. Michael Bolger

• Dr. Joyce Macwan

• Ms. Grazyna Fraczkiewicz

• Dr Eva Huehn

• Mr. Walter Woltosz

• All teams at Simulations Plus, Inc.

Thank you for your kind attention…

Questions?