Practical Application of PBPK in Neonates and Infants, Including Case Studies

Neil Parrott - 5 May 2014

Presented at the conference : Innovative Approaches to Pediatric Drug Development and Pediatric Medical Countermeasures: A Role for Physiologically-Based PK?

Physiologically based pharmacokinetic model use at Roche

Post-doc. research Pre-clinical group in discovery

M&S expert group + DMPK + formulation

M&S expert group + DMPK +Formulation + Clinical Pharmacology

Training

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

• PBPK modelling now applied routinely during small molecule development

• When PBPK sufficiently verified in adults, pediatric-PBPK is method of choice for PK prediction & dose setting in children

• Pediatric-PBPK is applied in many development projects Ph 2 to aid internal decisions as well as for PIP and PSP

• Allows us to leverage PK knowledge gained in adults and growing physiological knowledge on ontogeny of PK processes

Physiologically based pharmacokinetic models for neonates and infants

• Profound changes in PK processes in infants and neonates make PBPK particularly beneficial for this age group

• However, sparse data and many gaps in our knowledge of this population, especially oral drug absorption including intestinal UGT’s & transporters

Kearns N Engl J Med. 2003

Salem, Johnson, et al. Clin Pharmacokinet. 2014

Mooij, de Wildt et al Expert Opin. Drug Metab. Toxicol. 2012

birth 1 yr

4

Case Study : Oseltamivir use in infants

• Approved for oral treatment of influenza and for prophylaxis in adults and children ≥1 year of age

• In the light of the pandemic, several Health Authorities (e.g. FDA, EMEA, Australia, Canada) issued compassionate use authorization in children <1 of age for oral and > 1 year of age for IV route

• In communications with EMA the agency indicated – For IV registration in the EU-label and IV compassionate use in children <1 year of age

additional pre-clinical data are needed

EMEA requested a repeated dose IV toxicology study in juvenile marmosets

Background

5

Toxicity studies to support IV use in children <1 year

Juvenile studies (n=8) – Single oral dose in rats (PND 7, 14, 24 and 42) – 14-day oral repeat dose in rats (PND 7-21 and 21-49) – Single SC dose in rats (active metabolite; PND 7)

General toxicity studies i.v. – 14-day rat and marmoset (pro-drug & active metabolite adult animals) – Safety Pharmacology (IRWIN mouse and CV dog)

General toxicity studies oral – Single and repeat-dose toxicity in rats (27 wk) and marmoset (39 wk) – Special studies (incl. Safety Pharm.) in mice, rats, guinea pig, rabbit and dog – Reproductive toxicity studies in rats and rabbits – Carcinogenicity studies in mice, rats and transgenic mice

What was available ?

Clinical: Ongoing NIH trial in children <1 year (oral) and IV Compassionate Use in EU

6

Data to support IV use in children

• No pre-clinical juvenile studies by the intravenous route

What is missing?

What do we have to add?

Hardly feasible ?

Ethically justified ?

EMEA request:

repeated dose IV toxicity

in 3 day old marmosets

An alternative?

Building the bridge with M&S • Oral to IV

• Adult to juvenile

• Animal to human

SDPK IV in adult and newborn marmosets

PBPK is optimal method for species scaling of PK

7 PREDICTION ACCURACY ~ 75%, n=34

0.1

1

10

100

1000

10000

100000

1000000

0.1 1

10

10

0

10

00

10

00

0

10

00

00

10

00

00

0

Observed AUC (hr*ug/L)

Pre

dic

ted

AU

C (

hr*

ug

/L)

8

Compound specific

physicochemical and in vitro data

1 Adult Animal Model

Refinement

Simulation Adult animal PK

Age dependencies in

Enzyme/Transporters

Protein binding

Renal function

Distribution/Absorption

3 Juvenile Animal Model

Refinement

Simulation Juvenile animal PK

Compound specific

physicochemical and in vitro data

2 Adult Human Model

Refinement

Simulation Adult human PK

Age dependencies in

Enzyme/Transporters

Protein binding

Renal function

Distribution/Absorption

4 Juvenile Human Model

Refinement

Simulation Juvenile human PK

A strategy for pediatric-PBPK

Verify in Animal Predict in Man

9

oseltamivir carboxylate (OC)

Log D = -2.25

Base 8.2, Acid 3.6

very low permeability

oseltamivir (Os)

Log D = 0.36

Base 7.7

good permeability

Physicochemical & in vitro properties

carboxylesterase

Conversion of pro-drug in liver in both monkey and human

active metabolite pro-drug

10

PBPK models for pro-drug and metabolite run in parallel

Renal elimination

of OC

Renal elimination

of Os

HA = Hepatic Artery

LU = Lung

ART = Arterial Supply

VEN = Venous Return

AD = Adipose

MU = Muscle

LI = Liver

ACAT = Gut

SP = Spleen

HE = Heart

BR = Brain

KI = Kidney

SK = Skin

REP = Repro Organs

REDM = Red Marrow YELM = Yellow Marrow

ROB = Rest Of Body

oseltamivir oseltamivir carboxylate

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

Model for OModel for OC

Conversion of

O to OC in

liver

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

LU

SP ART

LI

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

Model for OModel for OC

Conversion of

O to OC in

liver

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

Model for OModel for OC

Conversion of

O to OC in

liver

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

LU

SP ART

LI

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

LU

SP ART

LI

ACAT

AD

HE

KI

MU

BR

SKREP

YELMREDM

ROB

VEN

HA

Model for OModel for OC

Conversion of

O to OC in

liver

11

PBPK model refinement in the monkey

• Metabolism scaling from in vitro verified in monkey

• Liver disposition for poorly permeable OC verified in monkey

OC

O

PBPK Liver model for Os PBPK Liver model for OC

Intracellular space Intracellular space

Extracellular space Extracellular space

Conversion

scaled from

in vitro

O O O O O

O O O O O O crosses cellular

membrane easily but once

converted to OC exit is

permeability limited

Permeability barrier

no barrier

to diffusion

Model verification in the animal is an essential step in PBPK

12

PBPK model refinement in the monkey

Model verification in the animal is an essential step in PBPK

1. Model verification for OC dosed IV refinement of renal CL

2. Addition of conversion of Os->OC & simulation

3. Refinement of Os->OC conversion

4. Refinement of OC release from liver

1.

2. 3 & 4

13

Simulations in adult human

15 mg infused over 1 hour

1

10

100

1000

0 6 12 18 24

Plasma Conc. (ng/mL)

Tim

e (

hrs

)

OC

Os 1

10

100

1000

10000

0 2 4 6 8

Time (hrs)

Pla

sm

a C

on

c. (n

g/m

L)

1

10

100

1000

10000

0 2 4 6 8

Time (hrs)

Pla

sm

a C

on

c. (n

g/m

L)

1

10

100

1000

0 6 12 18 24

Plasma Conc. (ng/mL)

Tim

e (

hrs

)

100 mg oral dose

OC

Os

1

10

100

1000

10000

0 2 4 6 8

Time (hrs)

Pla

sm

a C

on

c. (n

g/m

L)

1

10

100

1000

10000

0 2 4 6 8

Time (hrs)

Pla

sm

a C

on

c. (n

g/m

L)

Good simulations in adult human using model refinements scaled from monkey

14

0 1 2 3 4 5

1 day

3 weeks

6 weeks

3 months

4 years

Ag

e

Intrinsic clearance (uL/min/mg protein)

Development of metabolism with age

Integration of age dependencies

0

100

200

300

400

500

600

700

1-2

days

2-4

wks

2 mo

6 mo

1 yr

2 yr

6 yr

12 yr

PA

H C

leara

nce (

mL

/min

/1.7

m2)

0

20

40

60

80

100

120

140

160

GF

R (

mL

/min

/1.7

m2)

Maturation of renal function

1

10

100

1000

10000

100000

0 2 4 6 8

Time (hrs)

Pla

sma

Con

c. (n

g/m

L)

Simulation/Verification in newborn marmoset

Development of body and organ size with age

15

Simulated tissue concentration verification

Liver/plasma ratio difference with age is driven by balance of metabolism and release from liver

Terminal liver concentrations of Os and OC provided further verification of the hepatic disposition model

16

Development of esterases in human

Yang, D. et al, Human carboxylesterases HCE1 and HCE2: Ontogenic expression, inter-individual variability and differential hydrolysis of oseltamivir, aspirin, deltamethrin and permethrin. Biochemical pharmacology, 2009. 7(7): p. 238 – 247.

Shi, D., et al. Surged expression of carboxylesterase-1 (CES1) during the neonatal stage and significance in the activation of the anti-influenza pro-drug oseltamivir. J Infect Dis. 2011 Apr 1;203(7):937-42

Tamiflu cleavage is approx. 10-fold reduced in human newborns – similar to marmosets

17

Simulation of oral dosing in infants

A dose of 3 mg/kg b.i.d. given to infants of < 2 years

0

1

10

100

1000

36 41 46 51 56

Time (hrs)

Pla

sm

a C

on

c. (n

g/m

L)

.

AUC12 for OC

(ng.hr/mL)

Reported 4326 +/- 1878

Simulated 5700

Acosta, Edward P., P. Jester, P. Gal, J. Wimmer, J. Wade, Richard J. Whitley, and David W. Kimberlin, Oseltamivir Dosing for Influenza Infection in Premature Neonates. The Journal of Infectious Diseases, 2010. 202(4): p. 563-566.

The simulated steady state AUC is within the reported range

18

Simulation of oral dosing in newborns

A dose of 1.73 mg/kg b.i.d. given to premature neonates

1

10

100

1000

10000

96 108 120

Time (hrs)

Pla

sm

a C

on

c. (n

g/m

L) AUC12 for OC

(ng.hr/mL)

Reported 9250

Simulated 23000

The simulated steady state AUC approx 2x the reported value

Acosta, Edward P., P. Jester, P. Gal, J. Wimmer, J. Wade, Richard J. Whitley, and David W. Kimberlin, Oseltamivir Dosing for Influenza Infection in Premature Neonates. The Journal of Infectious Diseases, 2010. 202(4): p. 563-566.

19

• PBPK model simulations were verified sequentially in adult marmoset, adult human and newborn marmosets

• Simulations in human newborns were in line with the limited available clinical data

• Slow IV infusion produces plasma profiles of OC close to those of oral dosing while levels of Os are approx. 3-fold increased

Conclusions

In view of safety margins IV doses are not expected to have higher safety risk than the oral dose

20

• The modeling and simulation work was documented in a comprehensive report including an overview of all available non-clinical and clinical safety data

• The comprehensive report also stressed the technical feasibility & ethical concerns of a repeated dose toxicological study in newborn marmosets

• The report was submitted to the EMEA rapporteur

Impact

EMA withdrew request for the IV tox. study and accepted existing data and M&S as sufficient to support IV use in infants < 1 year

21

Acknowledgements

Doing now what patients need next