Development, Testing, Parameterization, and Calibration of a Human Physiologically Based Pharmacokinetic Model for the Plasticizer, Hexamoll® Diisononyl-Cyclohexane-1, 2-Dicarboxylate Using In Silico, In Vitro, and Human Biomonitoring Data

Abstract

A physiologically based pharmacokinetic model for Hexamoll^® diisononyl-cyclohexane-1, 2-dicarboxylate was developed to interpret the biokinetics in humans after single oral doses. The model was parameterized with in vitro and in silico derived parameters and uncertainty and sensitivity analysis was used during the model development process to assess structure, biological plausibility and behavior prior to simulation and analysis of human biological monitoring data. The model provided good simulations of the urinary excretion (Curine) of two metabolites; cyclohexane-1,2-dicarboxylic acid mono hydroxyisononyl ester (OH-MINCH) and cyclohexane-1, 2-dicarboxylic acid mono carboxyisononyl ester (cx-MINCH) from the biotransformation of mono-isononyl-cyclohexane-1, 2-dicarboxylate (MINCH), the monoester metabolite of di-isononyl-cyclohexane-1,2-dicarboxylate. However, good simulations could be obtained, with and without, a lymphatic compartment. Selection of an appropriate model structure was informed by sensitivity analysis which could identify and quantify the contribution to variability in Curine by parameters, such as, the fraction of oral dose that directly entered the lymphatic compartment and therefore by-passed the liver and the fraction of MINCH bio-transformed to cx-MINCH and OH-MINCH. By constraining these parameters within biologically plausible limits the presence of a lymphatic compartment was deemed an important component of model structure. Furthermore, the use of sensitivity analysis is important in the evaluation of uncertainty around in silico derived parameters. By quantifying their impact on model output sufficient confidence in the use of a model should be afforded. This type of approach could expand the use of physiologically based pharmacokinetic models since parameterization with in silico techniques allows for rapid model development. This in turn could assist in reducing the use of animals in toxicological evaluations by enhancing the utility of "read across" techniques.

Keywords: Hexamoll®; PBPK; biomonitoring; human; in silico; in vitro; plasticizer.

Figure 1

Postulated DINCH metabolism in humans showing only those metabolites measured in human biological monitoring and described in the PBPK model.

Figure 2

Determination of half-life for the estimation of in vitro intrinsic clearance of MINCH in human liver microsomes where, (×) represents non-specific binding, (○) total liver metabolism, and (□) specific liver metabolism.

Figure 3

A schematic of the model for DINCH and sub-model for MINCH. The DINCH model contained a lymphatic compartment (– – -) which received a portion of the DINCH oral dose from the stomach and gut, which by-passes the liver and enters the venous blood via the lymph node represented by the blue arrow. The biotransformation of DINCH to MINCH occurs in both the liver and gut. In the MINCH sub-model the biotransformation of MINCH to cx- and OH-MINCH is ascribed only to the liver and urinary excretion of metabolites were described with first-order elimination rate constants and a bladder compartment.

Figure 4

Lowry plots of the 11 most influential parameters governing C_urine variance at (A) 2 h and (B) 10 h. FracMetab, FracDose, Creat, RUrine, BW, MPY, and K1 were the most important parameters throughout the entire 50 h simulation and accounted for almost 100% of variance in C_urine. In order to account for 100% variance the red broken line was drawn from 1 (100% variance point) on the y axis to the ribbon and then vertically down to the x axis. All parameters to the left of the vertical line account for all the variance in C_urine.

Figure 5

Simultaneous fit to both cx- (red line and symbols) and OH-MINCH (blue line and symbols) using a common set of calibrated parameters showing the 90% credible interval for volunteers A and B The upper panels show the fit when estimating PORALDOSE and lower panels when PORALDOSE is fixed.

Figure 6

Simultaneous fit to both cx-(red line and symbols) and OH-MINCH (blue line and symbols) using a common set of calibrated parameters showing the 90% credible interval for volunteer C with fixed PORALDOSE.

Figure 7

Comparison of the fit of two draws from the joint posterior distribution to demonstrate the wide range of PORALDOSE consistent with data of volunteer A [0.22 mg kg^-1 (solid line) and 0.4 mg kg^-1 (broken line)], [cx-(red lines and symbols) and OH-MINCH (blue lines and symbols)].

Figure 8

Sensitivity of plasma fraction bound for DINCH. The amount (milligram) of DINCH in plasma over a 1,000 h period is shown for the posterior mode parameter set (volunteer A) and default binding. A similar simulation is shown for a 10 fold reduction (corresponding to decreasing from 99.99 to 99.9% binding) in fraction bound: the small change in absolute value (10-fold multiplicative difference) results in an important change in the duration of clearance from plasma following ingestion [99.99% (solid line), 99.9% (broken line)].