Yes this is an important concern. See my review of the previous items 7.1 and 7.4 where I also discuss about that. An additional protective (uncertainty) factor for such inter-individual differences could be applied. In the current toxicological risk assessment only a repetitive dosing scenario has been considered. Thus, the severe glutathione depletion would significantly decrease the elimination. Inversely, during intermittent dosing scenario (probably the real life condition in humans) by contrast to the current repetitive dosing scenario (in mice) would result to a greater recuperation of GSH at depletion for a greater elimination of the compound, and, hence, the KMD value should become lower for the same dose in mg/kg. This is why I suggest to use the more conservative and lower KMD of 10 ppm instead of 20 or 30 ppm for KMD, which is also based on the figure 5. In the white paper, the composite UF (uncertainty factor) is only 30. I believe the authors need to be more explicit to demonstrate that inter-individual differences in glutathione depletion and dosing scenario (intermittent versus repetitive) would not exceed the factor of 10 considered for the inter-species and inter-individual differences. I'm pretty sure that levels of glutathione in human population might be available in the literature to derive a mechanistic uncertainty factor in that domain while calculating the margin of exposure (MOE). A PBPK model can also be used to verify the effect of the intermittent versus repetitive dosing scenario. Furthermore, I still have another major concern in that domain: it is well assumed that only the free moiety in plasma (blood) would be active in tissue. Significant differences in proteins and lipids binding in plasma between the animal model of toxicity and human would result to significant interspecies difference in the toxic response at the same dose. Relatively lipophilic compounds similar to the current pesticide are know to significantly bind to lipoproteins and proteins in plasma from hydrophobic interactions even if they are not charged like a drug. Such a difference in the plasma (blood) binding between the animal model and human is still unknown, and, hence, may influence the interspecies and interindividuals extrapolations as well as the low-to-high dose extrapolations if the binding in blood become saturated as well or a lower binding is observed in human compared to mouse (i..e, more active free drug in human for potentially more toxicity). These data are not available unfortunately. Therefore, the current uncertainty factor of only 3 used for the inter-species differences by these authors might be limited. Again, using a lower KMD of 10 ppm would be conservative to cover the remaining uncertainty and variability effect. However, as mentioned in the item 7,1, the current MOEs are very large using a KMD of 30 ppm, which means that the human exposure is safe. Changing the KMD from 10 to 30 ppm would probably not change the conclusion of the present study.
A physiologically based TK model containing information on human basal GSH levels, rates of metabolism and rates of GSH replenishements could help in perfoming such extrapolations. Information on state of disease and impact on GSH basal levels and GSH replenishment would be necessary.
As already mentioned previously 1,3-Dichloropropene toxicokinetics in humans appear to be similar to those observed in rodents in terms of uptake, detoxification and formations of major metabolites. Since the glutathione conjugation is the primary pathway of detoxification of 1,3-D, severe depletion of glutathione can certainly result in non-dose proportional increase in 1,3-D blood levels in mice after single/multiple dose exposures. This was demonstrated by treatment with Diethylmaleate (DEM), which is known to decrease GSH levels in various organs by enzymatic conjugation with reduced GSH catalyzed by GSH transferase. Depletion of GSH by DEM would spare1,3-D conjugation and elevate 1,3-D levels and could potentially increase toxicity.