TABLE 7

Case examples illustrating the predictive utility of modeling and simulation approaches

	Development Stage & Question	Data Available	Modeling Approach	Outcome and impact	comments
Compound-1	Clinical development.	• In vitro metabolic phenotyping	• PBPK model	• Refined mechanistic understanding of disposition	Shardlow et al., 2013
	Can PBPK modeling inform the optimal ketoconazole^a drug interaction study design for a drug with an extended t1/2?	• Human mass balance	• CL based on in vitro CL_int	• Supported improved clinical study design for subsequent studies
		• Phase I human PK	• F_a, k_a, and V_dss based on in vivo data
		• DDI study with ketoconazole	• Retrospective simulation of Comp-1 –ketoconazole DDI was used to build confidence in the model
			• Simulations were used to compare alternative dosing regimens
Compound -2	Clinical development.	• Metabolism in recombinant system	• PBPK	• Provided high confidence in ability of model to accurately predict DDI with potent CYP3A4 inducers	In-house example.
	What is the DDI risk for compound that is predominantly metabolized by CYP3A4 in vitro?	• In vitro metabolic phenotyping	• CL based on IVIVE and in vitro phenotyping	• Supported additional simulations of DDI with moderate CYP3A4 inhibitors	J & J (Janssen R&D in-house example)
		• Metabolism in human hepatocytes	• Clinical PK data in healthy subjects and in ketoconazole DDI trial used to verify model relevance
		• Human mass balance	• Sensitivity analysis conducted on f_u,gut
		• DDI study with ketoconazole
Compound -3	Discovery to early development	• Human mass balance	• Static model	• Allowed DDI prediction when PBPK model and extensive in vitro and in vivo data were not available	Lu et al., 2007, 2010
	Can DDI of CYP3A victim drug with strong and moderate CYP3A inhibitor be predicted using in vitro data and static model?	• Reaction phenotyping	• Measured in vitro f_ACYP and fm_CYP linked to represent the factor of (1/(1+ I/K_i)
		• P450 activity remaining (fA_CYP) in the presence of ketoconazole or fluconazole (human hepatocyte suspended in human plasma)	• f_A,CYP corrected by comparing extracellular inhibitor concentration (determined in vitro) with in vivo C_max
			• Calculated steady-state DDIs compared with clinical observations
Compound -4	Clinical development	• Metabolic phenotyping (HLM)	• PBPK model	• Improved mechanistic understanding of the observed DDIs	In-house example.
	What is the risk of DDI for a compound primarily metabolized by CYP3A4 in vitro? Can PBPK model explain the observed clinical data and make predictions of the outcome of novel scenarios (DDI and pediatrics)?	• Rat QWBA	• CL based on in vivo CL_IV and retrograde extrapolation of in vivo CL_int	• Suggested previously unexpected role of efflux transport in fraction absorbed—which was subsequently verified by an in vitro study	J & J (Janssen R&D in-house example)
		• Human mass balance study data available	• Predicted V_ss consistent with intravenous dose data and rat QWBA	• Supported design of clinical pediatric study
		• Phase I human PK	• Model was verified using observed clinical PK data from single, multiple-dose and three clinical DDI studies
		• DDI studies (with three inhibitors)
Compound -5	Clinical Development	• In vitro metabolism (HLM and hepatocytes)	• PBPK model	• Guided selection of single-dose over proposed multiple-dose study to maximize DDI potential	Novartis in-house example
	What is the DDI effect of ketoconazole and rifampin on comp-4 after a single dose and at steady state, respectively, in clinical trials?	• In vitro phenotyping (HLM and rhCYP)	• CL based on clinical data
		• Rat ADME data	• V_dss predicted from physiochemical data
		• Clinical DDI with rifampin or ketoconazole	• Sensitivity analysis for f_u,gut conducted
			• Model verified with clinical DDI data
Compound- 6	Before FIH	• In vitro metabolism (HLM and hepatocytes—across species)	• PBPK model	• Exclusion of CYP2C9*3 genotype in FIH trials due to safety risk for the compound in this population	Novartis in-house example
	Should PMs of CYP2C9 be excluded from FIH trials?	• In vitro phenotyping (HLM and rhCYP)	• CL based on rhCYP2C91, 2, *3 kinetic data
		• rhCYP2C91, 2, *3 kinetics
		• Rat ADME data

↵^a Historically, ketoconazole was preferred as a clinically administered potent inhibitor of CYP3A4. However, a recent FDA memo FDA Drug Safety Communication: FDA limits usage of Nizoral(ketoconazole) oral tablets because of potentially fatal liver injury and risk of drug interactions and adrenal gland problems. http://www.fda.gov/Drugs/DrugSafety/ucm362415.htm and industry white paper (PMID: 26044116) have proposed that alternative inhibitors be administered in clinical DDI studies
PBPK, physiologically based pharmacokinetic; PK, pharmacokinetics; P450, cytochrome P450; F_A, fraction absorbed; k_a, absorption rate constant; CL, clearance; V_dss, steady-state volume of distribution; CL_int, intrinsic clearance; IVIVE, in vitro-in vivo extrapolation; f_m, fraction metabolized; f_A, fraction activity remaining ; HLM, human liver microsomes; f_u,gut, fraction unbound enterocytes; rh, recombinant human; QWBA, quantitative whole body autoradiography; FIH, first in human; ADME, absorption distribution metabolism excretion; PM, poor metabolizer.