Dressman Bath 2014

PK-UK 2014

Challenges and benefits of using PBPK

to evaluate an IVIVC for drugs with non-ideal solubility and/or permeability

Bath, November 2014

Prof. Dr. Jennifer Dressman

Why IVIVC and what are the challenges to establishing an IVIV relationship?

IVIVCs

Support bridging studies during clinical studies

Support formulation and manufacturing changes after registration

BUT,

So far, most IVIVCs have been for MR dosage forms

Even for these, the in vitro results are usually generated

with tests that resemble QC tests, so the IVIVC may not

be very robust

IVIV relationships for IR dosage forms have been difficult

to attain with classical methods

Dressman

Recommendation: pay attention to

the in vitro test design!

Part I: biorelevant dissolution tests –

what are these?

Dressman

Dressman

General approach to dissolution testing

Hypothesis:

the closer the

dissolution test

conditions to the

physiology, the

better the chances

of predicting product

performance

0

200

400

600

800

1000

1200

1400

0 12 24 36 48 60 72

Time (hour)

Co

nc (

ng

/mL

)

Simulation (fasted)

Simulation (fed)

Fasted

Fed

Finding the right dissolution test…..

THREE important considerations:

WHERE in the GI tract is drug

released from the dosage form?

HOW LONG does the dosage form

have to release the drug?

COMPOSITION of the fluids into

which the drug is released?

WHERE in the GI tract is drug released from the dosage

form? This will vary with the drug product e.g.

Immediate release dosage forms

Enteric coated dosage forms

Extended release dosage forms

Pulsatile delivery….

The site(s) of release and/or % released at each site of

release are often also dependent on whether the

dosage form is given before or after a meal, so the

dissolution test should also reflect the dosing

conditions

Finding the right dissolution test…..

HOW LONG does the dosage form have to release

the drug?

The drug must be released before or at its site(s)

of absorption, otherwise release will not result in

absorption. So it is important to understand the

permeability of the drug at various points in the

gut.

The passage of the dosage form through the

stomach depends on unit size and prandial state.

Finding the right dissolution test…..

COMPOSITION of the fluids into

which drug is released

The foods and drinks we

consume, gastric juices, bile,

pancreatic juices, bacterial

fermentation as well as water

re-uptake all combine to

influence the composition of

the GI fluids at various points

in the gut.

Finding the right dissolution test…..

GI-appropriate media composition and

volume: „biorelevant“ dissolution media

1. Fasted state

Stomach:

FaSSGF: simulates reduced

surface tension in the stomach

Small intestine:

FaSSIF to simulate basal

bile secretion in upper SI

Colon:

FaSSCOF to simulate conditions in a fasted state PK study

Vertzoni et al. EJPB 2005, Dressman et al. Pharm.Res. 1998, Vertzoni et al. Pharm. Res. 2010

2. Fed State

Stomach: FeSSGF: Milk/buffer pH 5 combination to

simulate gastric conditions after a standard breakfast

Small intestine: FeSSIF-V2 to simulate postprandial bile

secretion, lipolysis products, increased buffer capacity and osmolality in upper SI after food intake

Colon: FeSSCOF to simulate the ascending colon in the

fed state

Jantratid et al., Pharm. Res. 2008, Vertzoni et al., 2010

GI-appropriate media composition and

volume: „biorelevant“ dissolution media

Using „instant“ powders to make the biorelevant media:

„Study of a Standardized Taurocholate–Lecithin Powder for Preparing the Biorelevant Media FeSSIF and FaSSIF“

Dissolution Technologies

source: www.biorelevant.com

Biorelevant dissolution media

Case example 1: Relationship between

dissolution and PK of Danazol (a poorly

soluble but highly permeable drug)

Dressman

Case example 1. danazol

Aqueous solubility: 1µg/ml D:S 200 liters H2O

Dose: 200 mg 20 liters FaSSIF

pKa: neutral 6 liters FeSSIF

log P: 4.53

Danatrol dissolution profiles in

various media at 100 rpm

x

3

SIF

Danazol‘s food effect reflects its

dissolution characteristics

Plasma profiles of danazol after administration in the

fasted (●) and fed (○) state (from Charman et al.)

x

3

Part II: Combining biorelevant

dissolution testing with PBPK modeling

to achieve IVIV relationships for poorly

soluble drugs

Dressman

Biorelevant dissolution tests alone:

Qualitative forecast of food effects and formulation trends possible

But, how can we arrive at a more quantitative prediction??

Coupling with PBPK models (IV-IS-IV-R):

Quantitative forecast of food effects and formulation trends possible, since contributions of all steps affecting bioavailability can be addressed

Dressman

PBPK Modeling and PK Simulation

(Drug in Plasma) x BA Distribution

and elimination

Drug

Dissolved drug

Dissolution in gastric

media

Calculation of PK profile

Solid drug

Dissolution in intestinal

media

Gastric emptying

Stomach Small intestine

Solid drug

Dissolved drug

Precipitation water/meal

Permeation

Excretion

Gastric emptying

Initial Assumptions in the model

- Negligible absorption from the stomach

- Simultaneous solid and liquid emptying from the stomach (disintegrating dosage form)

- No intestinal permeability restrictions (for high permeable drugs)

IV-IS-IV Strategy to predict oral absorption

Set up in vitro dissolution testing Biorelevant media: fasted/fed, gastric/intestinal

Test parameters: apparatus, volume, hydrodynamics

Predict in vivo drug release Qualitative prediction based on the extent in dissolution,

comparing with the in vivo data (AUC, ranking

order)

Predict PK profile with in silico PBPK modeling Quantitative prediction based on the extent in dissolution,

comparing with the in vivo data (AUC, Cmax, Tmax)

Understand in vivo performance Further investigation to understand absorption mechanism and

indentify key factor which affect on the in vivo

performance

Linkage

Dis

so

lve

d %

Time

Conc (

ng/m

L)

Time

Development of a desired formulation

(no food effect, less variability, high BA)

Control strategy with a key attribute

(design space, quality by design)

Dressman

Case example 2: Coupling biorelevant

dissolution with PBPK to understand the effect

of particle size on PK of aprepitant (poorly

soluble, highly permeable)

Dressman

Case example: Aprepitant

pKa : 9.7 (basic)

Cs : 0.02 mg/mL in SGFsp (pH1.2), 0.0007 mg/mL in SIFsp (pH 6.8)

Log P : 4.8

Papp : 7.8 x 10-6 cm/sec (Caco-2)

BCS Class 2 (low solubility and high permeability)

Micronized, nanosized formulations

Indication: Emesis

Dressman

Dissolution of aprepitant 125mg

in biorelevant media

Micronized Nanoformulation

0

10

20

30

40

50

0 15 30 45 60

Time (minute)

Dis

so

lve

d %

: FaSSIF-V2

: FeSSIF-V2

: SIFsp

0

10

20

30

40

50

0 2 4 6 8 10

Time (min)

% d

isso

lve

d

FaSSIF-V2

FeSSIF-V2

SIFsp

Dressman

Simulated profiles of aprepitant

in the fasted and fed state

0

200

400

600

800

1000

1200

1400

0 12 24 36 48 60 72

Simulation (fasted)

Simulation (fed)

Fasted

Fed

0

200

400

600

800

1000

1200

1400

0 12 24 36 48 60 72

Time (hour)

Co

nc (

ng

/mL

)

Simulation (fasted)

Simulation (fed)

Fasted

Fed

Micronized Nanoformulation

Y. Shono et al. EJPB 76:95-104 (2010)

Value added: prediction of z value (=> particle size) at which food effect is eliminated

Value added of IVISIVC: mechanistic insight

1. Effects of dissolution rate (z) on PK profile

Fasted: highly dependent on dissolution rate in addition to solubility

Fed: no significant effect of dissolution rate ===> depends on gastric emptying

Fasted Fed

Time (h)

0 12 24 36 48

Pla

sm

a c

ele

co

xib

co

nce

ntr

atio

n (

ng/m

L)

0

200

400

600

800

1000

1200

2fastedz

5fastedz

fastedz

5.0fastedz

2.0fastedz

Time (h)

0 12 24 36 48

Pla

sm

a c

ele

co

xib

co

nce

ntr

atio

n (

ng

/mL

)0

200

400

600

800

1000

1200

1400

2~5.0fedz

2.0fedz

,fedz

2. Effects of gastric emptying rate on PK

profile in the fed state

Time (h)

0 12 24 36 48

Pla

sm

a c

ele

co

xib

co

nce

ntr

atio

n (

ng/m

L)

0

200

400

600

800

1000

1200

1400

1600

GER = 6 kcal/min

GER = 5 kcal/min

GER = 4 kcal/min

GER = 3 kcal/min

GER = 2 kcal/min

Dressman

Case example 3: prediction of PK of a Merck

& Co. development compound, which is

neither highly soluble nor highly permeable

and, being a weak base, might precipitate

upon entry into the small intestine

Dressman

Case example: Compound A

pKa : 4.2, 7.6, 9.1 (basic)

Cs : 0.02 mg/mL in SGFsp (pH1.2), 0.0007 mg/mL in SIFsp (pH 6.8)

Log D : 1.4 @ pH 1.5, 3.5 @ pH 7.4

Papp : 0.8 x 10-6 cm/sec (Caco-2)

BCS Class 4 (neither solubility or permeability is high)

N

NH

OH

NH

O2S

N

SCF

3

Compound A pharmacokinetic model

Dressman

A precipitation step for the fasted state and a limitation to

permeability were introduced into the Stella Model

Solid drug in

stomach

Solid drug in

small intestine

Dissolved drug

in stomach

Dissolved drug

in small intestine

Dissolution

in stomach

Dissolution in

small intestine

Distribution

and elimination

Gastric solid

emptying

Dose

Absorbed drug

in plasma

Plasma

concentration

Gastric liquid

emptying

Precipitaton

Absorption

through

SI mucosa

Solid drug in

stomach

Solid drug in

small intestine

Dissolved drug

in stomach

Dissolved drug

in small intestine

Dissolution

in stomach

Dissolution in

small intestine

Distribution

and elimination

Gastric solid

emptying

Dose

Absorbed drug

in plasma

Plasma

concentration

Gastric liquid

emptying

Precipitaton

Absorption

through

SI mucosa

Transfer model data revealed transfer rate depedent

precipitation in the fasted state, but no precipitation in the fed state

Dressman

Compound A transfer model data

fed

Compound A PBPK simulation

Only when biorelevant media are used and

precipitation in the fasted

state as well as a

permeability restriction is invoked, is it possible to

simulate the in vivo profiles

accurately

Dressman

0

10

20

30

40

50

60

70

0 10 20 30 40 50

Time [h]

Co

nc

en

tra

tio

n [

ng

/mL

]

in vivo, fasted

simulation, fasted

(FaSSGF/FaSSIF v2)

simulation, fasted

(FaSSGF/FaSSIF "old")

simulation, fasted

(FaSSGF/SIF)

simulation, fasted, no

permeability restriction

A

0

10

20

30

40

50

60

70

0 10 20 30 40 50

Time [h]

Co

nc

en

tra

tio

n [

ng

/mL

]

in vivo, fed

simulation, fed

(FeSSGF/FeSSIF v2)

simulation, fed

(FeSSGF/SIF)

simulation, fed, no

permeability restriction

B

Dressman

Case example 4: prediction of nifedipine

(poorly soluble, highly permeable, first pass

substrate) PK after administration of the

Adalat osmotic pump formulation

nifedipine pharmacokinetic model

Step 1: in silico description of nifedipine PK to account

for first pass (using PK-SIM)

Dressman

Dissolution

Peroral dose

First pass

metabolism in

SI and liver

Absorption

stomach

Small intestine

Distribution of

drug into

organs

Drug in

venous blood

pool

Drug in

arterial blood

pool

Metabolism and elimination

Intravenous dose

nifedipine release and dissolution in vitro

Step 2: release and dissolution in the Type 3 tester

(BioDis) using a biorelevant media profile

Dressman

0

20

40

60

80

100

120

0 10 20 30 40 50 60 70

Time [h]

% D

isso

lved

, %

Rele

ased

% released, Adalat OROS 60mg

% dissolved, Adalat OROS 60mg

Nifedipine PK

Step 3: prediction

of PK profile based on

PK-SIM model for first pass

AND biorelevant

dissolution testing

Conclusion: drug

dissolution after release

from the MR dosage form

is rate-limiting to

absorption

Dressman

0

10

20

30

40

50

0 10 20 30 40 50 60 70

Time [h]

Pla

sm

a c

on

cen

trati

on

[n

g/m

L]

0.01

0.1

1

10

100

0 20 40 60 80

Time [h]

Pla

sm

a c

on

ce

ntr

ati

on

[n

g/m

L]

0

10

20

30

40

50

0 10 20 30 40 50 60 70

Time [h]

Pla

sm

a c

on

cen

trati

on

[n

g/m

L]

0

0

0

1

10

100

0 20 40 60 80Time [h]

Pla

sm

a c

on

cen

trati

on

[n

g/m

L]

Dressman

Strategy to understand and forecast

oral drug absorption

Combine results from biorelevant

dissolution tests with physiologically

based pharmacokinetic modeling

to create

„in vitro – in silico – in vivo“

relationships

Dressman

IVIVC Challenges ahead

Integration of appropriately biorelevant dissolution

data (i.e. use the right level of biorelevant media)

Appropriate integration of dissolution data into the

commercial PBPK models (especially for MR dosage

forms that do not have robust, zero-order release

kinetics)

Appropriate integration of precipitation data into

commercial PBPK models

Dressman

Acknowledgments

Prof. Dr. Christos Reppas & Dr. Maria Vertzoni

(U-Athens)

Dr. Yasushi Shono (U-Frankfurt, now Takeda)

Dr.s Filippos Kesisoglou & Henry Wu (Merck & Co., Inc.)

Dr. Christian Wagner (U-Frankfurt, now FDA)

…and Many Thanks for your attention!