Kinetically-derived maximal dose (KMD) confirms lack of human relevance for high-dose effects of octamethylcyclotetrasiloxane (D4)

Abstract

The kinetically-derived maximal dose (KMD) is defined as the maximum external dose at which kinetics are unchanged relative to lower doses, e.g., doses at which kinetic processes are not saturated. Toxicity produced at doses above the KMD can be qualitatively different from toxicity produced at lower doses. Here, we test the hypothesis that high-dose-dependent toxicological effects of octamethylcyclotetrasiloxane (D4) occur secondary to kinetic overload. Octamethylcyclotetrasiloxane (D4) is a volatile, highly lipophilic monomer used to produce silicone polymers, which are ingredients in many consumer products and used widely in industrial applications and processes. Chronic inhalation at D4 concentrations 10⁴ times greater than human exposures produces mild effects in rat respiratory tract, liver weight increase and pigment accumulation, nephropathy, uterine endometrial epithelial hyperplasia, non-significant increased uterine endometrial adenomas, and reduced fertility secondary to inhibition of rat-specific luteinizing hormone (LH) surge. Mechanistic studies indicate a lack of human relevance for most of these effects. Respiratory tract effects arise in rats due to direct epithelial contact with mixed vapor/aerosols and increased liver weight is a rodent-specific adaptative induction of drug-metabolizing hepatic enzymes. D4 is not mutagenic or genotoxic, does not interact with dopamine receptors, and interacts at ERα with potency insufficient to cause uterine effects or to alter the LH surge in rats. These mechanistic findings suggest high-dose-dependence of the toxicological effects secondary to kinetic overload, a hypothesis that can be tested when appropriate kinetic data are available that can be probed for the existence of a KMD. We applied Bayesian analysis with differential equations to information from kinetic studies on D4 to build statistical distributions of plausible values of the K_m and V_max for D4 elimination. From those distributions of likely K_m and V_max values, a set of Michaelis-Menten equations were generated that are likely to represent the slope function for the relationship between D4 exposure and blood concentration. The resulting Michaelis-Menten functions were then investigated using a change-point methodology known as the "kneedle" algorithm to identify the probable KMD range. We validated our K_m and V_max using out of sample data. Analysis of the Michaelis-Menten elimination curve generated from those V_max and K_m values indicates a KMD with an interquartile range of 230.0-488.0 ppm [2790-5920 mg/m³; 9.41-19.96 µM]. The KMD determined here for D4 is consistent with prior work indicating saturation of D4 metabolism at approximately 300 ppm [3640 mg/m³; 12.27 µM] and supports the hypothesis that many adverse effects of D4 arise secondary to high-dose-dependent events, likely due to mechanisms of action that cannot occur at concentrations below the KMD. Regulatory methods to evaluate D4 for human health protection should avoid endpoint data from rodents exposed to D4 above the KMD range and future toxicological testing should focus on doses below the KMD range.

Keywords: Human relevance; Kinetically-derived maximum dose (KMD); Maximum tolerated dose (MTD); Pharmacokinetics/toxicokinetics; Regulatory policy; Threshold.

Fig. 1

Verification of mean Km and mean Vmax values. We used the mean Km and mean Vmax values that we estimated based on the 75 ppm D4 blood levels over time to see how well the Michaelis–Menten equation fits the data. We can see that the model does well in estimating the blood D4 levels following the 75 ppm exposure

Fig. 2

Verification of the Uncertainty from the Km and Vmax estimates. Here we performed 100 individual simulations using our distribution of Km and Vmax estimates to predict the D4 blood levels following a 75 ppm exposure. We have plotted the spaghetti plot (each simulation is 1 line in the plot). We can see that most of the ground-truth D4 blood levels are within the simulated values. The only exception is the 24 h measure which is only slightly outside of our predicted values

Fig. 3

Holdout Validation of the Approach. We trained our model to estimate Km and Vmax based on data for the 75 ppm exposure. Here we are showing how well the Km and Vmax estimate the D4 blood concentrations following the 150 ppm exposure. Overall, the fit is very good with a root mean square error of 0.16 units