DDMODEL00000102: Meibohm_2013_oncology_EMD525797_TMDD

  public model
Short description:
Semi-mechanistic model-based drug development of EMD 525797 (DI17E6), a novel anti-?v integrin monoclonal antibody. B. Meibohm, B. Brockhaus, M. Zühlsdorf, A. Kovar PAGE 2013, 6/2013
PharmML (0.6.1)
  • Semi-mechanistic model-based drug development of EMD 525797 (DI17E6), a novel anti-?v integrin monoclonal antibody
  • B. Meibohm, B. Brockhaus, M. Zühlsdorf, A. Kovar
  • PAGE 22 (2013), 11/2015
  • The University of Tennessee Health Science Center, Memphis, TN, USA Merck Serono, Darmstadt, Germany Merck Serono, Darmstadt, Germany Merck Serono, Darmstadt, Germany
  • Objectives: The objective of this analysis was to develop a semi-mechanistic population PK model for EMD 525797 incorporating receptor occupancy that forms the basis of a model-guided dose rationale. Methods: A stepwise approach was used to develop and iteratively refine a population PK/PD model throughout clinical development of EMD 525797 using nonlinear mixed effect modeling with NONMEM. The model was initially derived for describing the PK data in a dose-escalation study in Cynomolgus monkeys, and was subsequently refined with human concentration-time data. At the current iteration of model refinement, 815 serum concentrations of 51 subjects have been included in the analysis: 37 healthy volunteers receiving EMD 525797 single doses of 35, 100, 250, 500, 1000, and 1500 mg, and 14 patients with metastatic castrate-resistant prostate cancer (mCRPC) receiving EMD 525797 250 or 500 mg every 2 weeks. EMD 525797 was administered as intravenous infusion over 1 hour. Results: The disposition of EMD 525797 was best described by a 2-compartment model with 2 parallel elimination pathways, an unspecific proteolytic pathway, and a saturable, target-mediated process with Michaelis-Menten-type kinetics, which is an approximation of the target-mediated drug disposition model (TMDD) [1]. A disease modifier function was included for Vmax based on disease status. The obtained parameter point estimates and their between-subject variability (% CV) were: Vmax=493 µg/hr (21.4%), with a 35.3% increase in mCRPC patients, Km=0.571 µg/mL, V1=4.41 L (22.0%), V2=3.44 L (40.4%), Q=0.0444 L/hr (56.9%), and CLproteolytic=0.00857 L/hr (25.8%). Saturation of the saturable elimination process was assumed to be indicative of receptor occupancy for the targeted ?v integrins, and IC90, IC95, and IC99 of ?v integrin inhibition could be derived from Km. Model-based stochastic simulations were performed to explore dosing regimens with regard to their likelihood to achieve steady-state trough concentration exceeding these landmarks. Conclusions: Modeling and simulation provides a rational basis for EMD 525797 dose selection. Next steps are to use the model to compare between different populations and to test whether other TMDD model approximations could fit better the data [2].
Nadia Terranova, Kheizurane_ElMekki
Context of model development: Mechanistic Understanding;
Long technical model description: Semi-mechanistic population PK model including an approximation of the target-mediated drug disposition model (TMDD) for EMD 525797 to understand the receptor occupancy. It is a 2-compartment model with two parallel elimination pathways: one is saturable and follows Michaelis-Menten kinetics, the other one with first order kinetics;
Model compliance with original publication: Yes;
Model implementation requiring submitter’s additional knowledge: No;
Modelling context description: Development of semi-mechanistic population PK model for incorporating receptor occupancy ;
Modelling task in scope: estimation;
Nature of research: Early clinical development (Phases I and II);
Therapeutic/disease area: Oncology;
Annotations are correct.
This model is not certified.
  • Model owner: Nadia Terranova
  • Submitted: Dec 11, 2015 5:01:15 PM
  • Last Modified: Jul 18, 2016 11:09:52 AM
Revisions
  • Version: 7 public model Download this version
    • Submitted on: Jul 18, 2016 11:09:52 AM
    • Submitted by: Nadia Terranova
    • With comment: Updated model annotations.
  • Version: 4 public model Download this version
    • Submitted on: Dec 11, 2015 5:01:15 PM
    • Submitted by: Kheizurane_ElMekki
    • With comment: Edited model metadata online.

Independent variable T

Function Definitions

additiveError(additive)=additive

Structural Model sm

Variable definitions

dCENTRALdT=((((-VM ×C1)(KM+C1)-(CL ×C1))-(Q ×C1))+(Q ×C2))
dPERIP1dT=((Q ×C1)-(Q ×C2))
C1=CENTRALV1
C2=PERIP1V2
LOG_C1={-5  if  (C1≤0)log(C1)  otherwise

Initial conditions

CENTRAL=0
PERIP1=0

Variability Model

Level Type

DV

 residualError

ID

 parameterVariability

Covariate Model

Continuous covariate DIS

Parameter Model

Parameters
POP_VM;  POP_KM;  POP_V1;  POP_V2;  POP_Q;  POP_CL;  BETA_VM_DIS;  RUV_ADD;  OMEGA_VM;  OMEGA_V1;  OMEGA_V2;  OMEGA_Q;  OMEGA_CL;  GRPVM=((1000 ×POP_VM) ×(1+(BETA_VM_DIS ×DIS))); 
ETA_VM∼N(0.0,OMEGA_VM) — ID
ETA_V1∼N(0.0,OMEGA_V1) — ID
ETA_V2∼N(0.0,OMEGA_V2) — ID
ETA_Q∼N(0.0,OMEGA_Q) — ID
ETA_CL∼N(0.0,OMEGA_CL) — ID
EPS_Y∼N(0.0,1.0) — DV
VM=(GRPVM+ETA_VM)
log(V1)=(log(POP_V1)+ETA_V1)
log(V2)=(log(POP_V2)+ETA_V2)
log(Q)=(log(POP_Q)+ETA_Q)
log(CL)=(log(POP_CL)+ETA_CL)
KM=(1000 ×POP_KM)

Observation Model

Observation Y
Continuous / Residual Data

Parameters
Y=(LOG_C1+(additiveError(RUV_ADD) ×EPS_Y))

Estimation Steps

Estimation Step estimStep_1

Estimation parameters

Initial estimates for non-fixed parameters

  • POP_VM=0.5001
  • POP_KM=0.533
  • POP_V1=4.38
  • POP_V2=3.5
  • POP_Q=0.559
  • POP_CL=0.0065
  • BETA_VM_DIS=0.01
  • RUV_ADD=0.094
  • OMEGA_VM=0.00436
  • OMEGA_V1=0.059
  • OMEGA_V2=0.132
  • OMEGA_Q=0.123
  • OMEGA_CL=0.508
Estimation operations
1) Estimate the population parameters
    Algorithm FOCE

    Step Dependencies

    • estimStep_1
     
    Help