Society Awards 2021
Researches on the evaluation of pesticide safety in humans using a pharmacokinetic approach
2021 Volume 46 Issue 3 Pages 290-296
Details
2021 Volume 46 Issue 3 Pages 290-296
Similar to the pharmaceutical compounds, pesticides require human safety assessment for their registration and distribution; however, it is absolutely impossible to assess human safety by dosing humans with pesticides. Thus, how to appropriately evaluate the safety of pesticides in humans remains a great subject of debate. In this article, we present some examples of pesticide toxicity studies that identify species differences in toxicity and evaluate human safety by applying combinations of novel in vivo, in vitro, and in silico techniques to separately assess the key toxicodynamic (i.e., sensitivity) and/or toxicokinetic (i.e., exposure) factors. Because it is scientifically sound, the safety assessment strategy illustrated for three compounds in this article is expected to play an important role in the human safety assessment of agricultural compounds.
It is generally agreed that pesticides can play a significant role in stabilizing food production, and thus they are regarded as efficient and indispensable buffers against the consequences of forthcoming population increases and resulting food shortages. Similar to pharmaceutical compounds, pesticides require human safety assessment for their registration and distribution; however, clinical trials of pesticides in humans are obviously not allowed, thus the safety of pesticides in humans has been evaluated by dividing the safe dose levels obtained in laboratory animals by a safety factor. Recently, many in vitro and in silico techniques for human safety evaluation have been developed, with great improvement in assay variability and accuracy. With the development of the techniques, the importance and necessity of using them for the indirect human safety assessment of non-pharmaceutical chemicals including pesticides has gained greater recognition. In this article, we present some researches carried out to identify the species differences in pesticide toxicity and, by applying a combination of these techniques, evaluate the human safety of some pesticides with known toxicity profiles that differ between species. More specifically, scientifically-sound human safety assessment could be achieved by combining several in vitro (e.g., enzymes or cells), in vivo (e.g., chimeric mice with humanized liver), and in silico (e.g., PBPK modeling) techniques, which are described in this review.
Procymidone (Sumilex, its structure presented in Fig. 1) is a commercially available fungicide with both protective and curative properties. It is used to control plant diseases such as fruit rot (i.e., gray mold on fruits, vines, and vegetables) and Sclerotinia rot of kidney beans and vegetable crops.1) The metabolic pathways of procymidone in mammals are known.2) Procymidone is first hydroxylated at the methyl group of the imide ring. This hydroxylated metabolite is metabolized further in one of two ways: the hydroxymethyl group is conjugated to form a glucuronide or oxidized to form the carboxylic acid metabolite. Both the glucuronide of hydroxylated procymidone and the carboxylic acid metabolite are more hydrophilic and thus more readily excreted than the hydroxylated procymidone. As a major toxicity of procymidone, external genital abnormality (a reproductive toxicity) has been observed in male rats, but not in rabbits and monkeys.3) The key mechanism of the species differences in the toxicity has already been identified. In vivo experiments in rabbits and monkeys showed that the hydroxylated metabolite, which has antiandrogenic activity, is conjugated in liver, transferred to blood, and then rapidly excreted in urine (Fig. 2). On the other hand, the kinetics of the hydroxylated metabolite was found to be different in rats, where the glucuronide was shown to be excreted in bile and deconjugated in the gastrointestinal tract and reabsorbed, resulting in recycling through the enterohepatic circulation and increased exposure in rats (Fig. 2).3) The species difference in the maternal exposure to the hydroxylated metabolite results in a significant difference in fetal exposure.4) From these results, it was suggested that the species difference in the developmental toxicity of procymidone is mainly due to variation in the level of exposure to the causal substance, hydroxylated metabolite, and this variation stems from interspecies differences in the biliary excretion route of the glucuronide. To assess the teratogenic risk associated with procymidone in humans, the excretion route and rate of formation of glucuronides, which are key factors involved in the mechanism of the toxicity must be clarified. To compare the metabolic profiles of procymidone in humans and rats, we planned to conduct in vivo experiments using chimeric mice.
Chimeric mice with human or mouse hepatocytes were provided by PhoenixBio Co., Ltd. (Hiroshima, Japan). 14C-Labeled procymidone was orally administered to the chimeric mice and other laboratory animals, and the plasma concentrations of procymidone and its metabolites were measured.5,6) The pharmacokinetic parameters of chimeric mice and other laboratory animals are summarized in Table 1. When the parameters are compared among species, higher exposure (i.e., AUC and Cmax) was observed in rats and rat chimeric mice whereas the exposure was rather low in rabbits, monkeys, and human chimeric mice for both procymidone and its hydroxylated metabolite. In the previous articles on procymidone metabolism in intact rats,5) delayed Tmax and higher exposure to the hydroxylated metabolite were caused by the entero-hepatic circulation of its glucuronide. In the present research, the delayed Tmax for the hydroxylated metabolite was well reproduced in rat chimeric mice and reported to result in higher exposure to this metabolite, suggesting this animal model with rat or human hepatocytes is a useful model to investigate species differences in the pharmacokinetic, metabolic, and excretion profiles. On the other hand, the plasma concentration of the hydroxylated metabolite in human chimeric mice shifted synchronously to that of procymidone. As a result of species differences in the transition from the plasma concentration of the hydroxylated metabolite to that of procymidone, the AUC of the hydroxylated metabolite was two to five times larger in rat chimeric mice than in human chimeric mice.
| (A) Procymidone | |||||
|---|---|---|---|---|---|
| Parameter | Rat5) | Chimeric mice with rat hepatocytes6) | Rabbit5) | Monkey5) | Chimeric mice with human hepatocytes6) |
| Cmax [µg equiv./mL] | 6.80 | 4.80 | N.D. | 2.90 | 3.01 |
| Tmax [hr] | 4 | 1 | N.D. | 10 | 2 |
| AUC0–72 hr [µg equiv.hr/mL] | 89.87 | 48.3 | 0 | 73.59 | 44.3 |
| (B) Hydroxylated procymidone | |||||
| Parameter | Rat5) | Chimeric mice with rat hepatocytes6) | Rabbit5) | Monkey5) | Chimeric mice with human hepatocytes6) |
| Cmax [µg equiv./mL] | 10.42 | 9.57 | 0.33 | 1.35 | 1.77 |
| Tmax [hr] | 12 | 12 | 1 | 6 | 1 |
| AUC0–72 hr [µg equiv.hr/mL] | 332.06 | 120.8 | 0.875 | 40.2 | 36.1 |
To further elucidate the excretion route of the glucuronide, bile duct-cannulation studies were performed with chimeric mice and other laboratory animals.6) Metabolite profiles were analyzed in each species, and the excretion of the glucuronide was evaluated in urine and bile. The excretion ratios of the glucuronide are summarized in Fig. 3. The biliary excretion ratio of glucuronide was high in rat chimeric mice (4.52), like rats (3.70), but low in human chimeric mice (0.12).
In conclusion, the pharmacokinetics of the hydroxylated metabolite in urine and bile clearly show that the main excretion route of the glucuronide differs between rat chimeric mice and human chimeric mice, which results in lower exposure to the hydroxylated metabolite in humans (and as previously observed in rabbits and monkeys) than in rats. Because of the lower exposure to the hydroxylated metabolite in humans, which is a key teratogenicity trigger in rats, it is concluded that the teratogenicity of procymidone may be much lower in humans than in rats.
Flumioxazin (Sumisoya, its structure presented in Fig. 4) is an N-phenylimide herbicide that is widely used for controlling annual broadleaf weeds in soybean crops.7) The herbicidal activity of flumioxazin was shown to be due to its inhibition of protoporphyrinogen oxidase (PPO), one of the key enzymes in porphyrin biosynthesis. PPO is an essential enzyme involved in the synthesis of chlorophyll in plants. Inhibition of the enzyme in plant cells causes accumulation of intermediate tetrapyrroles, including protoporphyrin IX (PPIX), and then photoreactions that produce oxygen free radicals to destroy the cell membranes.8,9)
It is well known that PPO exists also in mammals, and works as a key enzyme in heme biosynthesis.10) Regarding the toxicity of PPO inhibitors in mammals, similar key events of PPO inhibition and further accumulation of PPIX leading to cell damage may also occur. Thus, it is believed that PPO inhibitors can cause toxicity especially in PPO-generating organs such as the liver and can occasionally induce hepatocellular tumors as a consequence of hepatotoxicity.11,12) Also, PPO catalyzes the oxidation of protoporphyrinogen IX to PPIX during heme biosynthesis in both plants and mammals. In mammals, the inhibition of PPO blocks the production of hemoglobin and erythrocytes, resulting in anemia.13–15) Since the molecular mechanism of heme biosynthesis is common to all mammals, the possibility that PPO inhibitors are hepatotoxic and/or cause anemia in humans must be considered, especially when the toxicity of PPO inhibitors is observed in rodents and other mammals. Flumioxazin has developmental toxicity in rats, causing embryonic lethality, teratogenicity (mainly ventricular septal defects, VSD, and wavy ribs), and growth retardation.16,17) The mechanism of this developmental toxicity in rats has already been shown to be embryonic anemia resulting from PPO inhibition and following the elimination of heme biosynthesis.18) Even though all mammals have PPO for heme biosynthesis, which is essential for vital activity, a remarkable species difference in the toxicity caused by flumioxazin-induced PPO inhibition has been reported. Oral administration of flumioxazin at 3,000 mg/kg, which was two orders of magnitude larger than the teratogenic dose of 30 mg/kg in rats, was shown not to cause developmental toxicity in rabbits.16) Thus, further mechanistic studies were warranted to assess the human safety of flumioxazin and evaluate the possibility of its developmental toxicity in humans.
First, we focused on identification of the toxic agent. Metabolism study with pregnant/non-pregnant rats and rabbits revealed that the major metabolites in the fetus are 3-OH flumioxazin, 4-OH flumioxazin, and 6-amino-7-fluoro-4-(2-propynyl)-2H-1,4-benzoxazin-3(4H)-one (APF).19,20) The inhibitory activity of these three metabolites as well as the parent flumioxazin were tested in a PPO inhibition assay with mitochondrial fractions of rat liver, the results of which demonstrated the significant high inhibitory activity of flumioxazin followed by 4-OH flumioxazin (14.0-times weaker), 3-OH flumioxazin (150-times weaker), and non-active APF20) as shown in Fig. 5A. From the quantitative evaluation in dynamics and kinetics of metabolites, we concluded that the agent causing embryonic anemia is the parent compound flumioxazin.
Next, we arranged to assess the human safety. The PPO inhibition assay was also conducted with liver mitochondrial fractions to compare the sensitivity between rats and humans. The results showed rats were more sensitive than humans, and the IC50 of flumioxazin was 4.56 times larger in humans than in rats, with lower inhibitory activity of the metabolites20) as shown in Fig. 5B. The details underlying species differences have been revealed. A molecular dynamics study (i.e., docking simulation study) of flumioxazin in several species clearly demonstrated that the affinity of substrate binding to PPO varied among species21) and was inconsistent with the species differences in PPO inhibitory activity observed in the in vitro assay.
Since PPO exists in complex and highly organized structures in mitochondria and thus accessibility of flumioxazin to PPO in cells can be different from that in the in vitro mitochondrial extract model, further confirmation of the species difference in a cell-based test system was sought and obtained. Rat, rabbit, monkey, and human hepatocytes were exposed to flumioxazin, and accumulation of PPIX resulting from PPO inhibition was monitored. The results (Fig. 6) showed similar tendencies to those shown in the mitochondrial assays, that the PPO inhibitory activity of flumioxazin was stronger in rats than in humans. The results suggested that the higher-order structure of PPO in cells does not affect its inhibition, and both the cell-based hepatocyte assay and the isolated mitochondria assay can adequately evaluate the inhibitory activity of chemicals, thus indicating that the inhibition is three to five times more sensitive in rats than in humans. For further evaluation of the effect of species difference in dynamics on the downstream outcome of heme biosynthesis inhibition, we investigated the impact of flumioxazin-induced PPO inhibition on heme biosynthesis in human and rat erythroid cells.22) Treatment with flumioxazin led to dose-dependent accumulation of PPIX in both human erythroleukemia (K562) cells and rat erythroleukemia (REL) cells; however, differences in heme biosynthesis were observed between species. As presented in Fig. 7, heme biosynthesis was not affected in human K562 cells by flumioxazin up to 5.0 μM, whereas a statistically-significant decrease in heme production was observed in rat REL cells. To complete the human risk assessment, a physiologically-based pharmacokinetic (PBPK) model was developed for pregnant women, and human fetal exposure to flumioxazin was estimated.23) By the in silico technique, the maximum concentration of flumioxazin in human embryos was estimated to be 1.92 μM at an exposure level equivalent to a maternal dose of 1,000 mg/kg/day. This concentration is lower than the no-effect level of 5.0 μM for heme production in human erythroid cells. From all these results, we concluded that human exposure could not result in levels that would produce toxicological effects and thus flumioxazin was not a developmental toxicity issue in humans.
Epyrifenacil (S-3100, with structure shown in Fig. 8) is a new herbicide under development. The herbicidal mechanism of epyrifenacil is similar to that of flumioxazin, which is the inhibition of PPO. As mentioned above, PPO inhibitors may cause accumulation of PPIX leading to cell damage in PPO-generating organs such as the liver11,12) and/or cause anemia resulting from inhibition of heme biosynthesis.13–15)
First, the results of a subchronic 90-day toxicity study in laboratory animals24) demonstrated that adverse outcomes caused by epyrifenacil were limited to hepatotoxicity and anemia in rats and mice, which were thought to be consequences of PPO inhibition. In these toxicological observations, hepatotoxicity was considered to be a more important risk and its exploration in detail was considered necessary since it may lead to neoplastic transformation when exposure is prolonged. As presented in Fig. 8, epyrifenacil has an ester bond in its structure, which is expected to be easily cleaved and thereby produce the carboxylic acid metabolite (CA). An in vitro metabolism study proved that ester cleavage in epyrifenacil was very rapid in mice, rats, and humans,24) which suggests that epyrifenacil, when dosed orally, is mostly metabolized to CA by the first-pass effect, and CA is the major metabolite in mammals. In order to conclude that CA is the main cause of epyrifenacil-induced hepatotoxicity in rodents, we also conducted an in vitro PPO inhibition assay. The results showed that CA inhibited PPO in rodents to a similar extent as the parent epyrifenacil.24) From these results, we concluded that the principal cause of epyrifenacil-induced hepatotoxicity in rodents is CA.
Since the molecular mechanism of heme biosynthesis is common to all mammals, the possibility of epyrifenacil-induced hepatotoxicity in humans must be considered. To conduct a precise risk assessment, we planned to look for evidence of species differences in the dynamics (i.e., PPO inhibitory activity) and kinetics (i.e., uptake by the liver) of PPO inhibition separately. Mitochondrial PPO inhibition assay showed a significant species difference in the inhibitory activity by CA. The IC50 in humans was 8.3 times larger than in rats and 13.9 times larger than in mice, demonstrating that the sensitivity to PPO inhibition by CA is less in humans than in rodents (Fig. 9). The details underlying species differences have not been revealed; however as mentioned above, a molecular dynamics study of flumioxazin clearly demonstrated the species difference in the binding affinity of the substrate to PPO,21) so the binding affinity of CA for PPO was deemed to vary similarly.
In view of the kinetics, we have obtained some preliminary data which suggest the contribution of active uptake and accumulation of CA in mouse livers. Since CA, the ester-cleaved metabolite of epyrifenacil, contains a carboxylic acid moiety, it is highly likely to be in its ionized form, a COO− anion, under physiological conditions. It is known that uptake of such anions into organs via organic anion transporters (OATs) or organic anion transporting peptides (OATPs) leads to significant exposure to the chemicals.25,26) It is also reported that there are species differences in the active transport of chemicals and thereby species differences in exposure and toxicity.27–29) In the in vitro metabolism study, no quantitative and qualitative species differences in metabolite production and metabolic degradation were present, thus we focused on evaluating species differences in the hepatic uptake of CA as one of the predominant toxicity factors. As shown in Fig. 10, the results of in vitro assays using hepatocytes from rodents and humans demonstrated that the uptake of CA is 6 to 13 times larger in mice and two times larger in rats than in humans.24) This result clearly indicates that exposure to CA in the liver is smaller in humans than in rodents, especially mice. Since species differences were proved in both dynamics and kinetics, it was concluded that the risk of CA-induced hepatotoxicity was lower in humans than in rodents. Our research provides only phenomenological results at this time, thus further mechanistic researches to explain the significant differences among species are highly desired.
It is always debatable whether it is appropriate to evaluate the safety and toxicity of pesticides directly in humans; however, it is absolutely impossible to assess the human safety of pesticides by direct dosing of humans. In order to address the dilemma, many in vivo, in vitro, and in silico techniques for evaluating safety in humans have been recently developed with great improvement in variability and accuracy. For procymidone, as explained above, the biliary excretion and consequent high exposure to the causal metabolite in plasma were identified as the key toxicity factors, and the low risk to humans of exposure was confirmed using a novel animal model of humanized chimeric mice. For flumioxazin, lower PPO inhibitory activity in human mitochondria and consequent higher no-effect level of no less than 5.0 μM in human erythroid cells were determined in in vitro assays, and then, using the in silico PBPK technique, it was confirmed that the concentration of flumioxazin in human fetus does not reach the no-effect level even at the extremely high dose level. For epyrifenacil, the levels of hepatic exposure to the causal metabolite and inhibition of the target enzyme were evaluated in vitro and were significantly different between mice and humans. The safety assessment strategy illustrated for the three compounds in this article to separately assess the key toxicodynamic factor (i.e., sensitivity) and/or toxicokinetic factor (i.e., exposure) is scientifically-sound approach, thus it is expected to play an important role in the human safety assessment of agricultural compounds.
These researches were performed at Sumitomo Chemical Co., Ltd. I would like to express my deepest gratitude to Dr. Yoshitaka Tomigahara and Dr. Satoshi Kawamura, who designed the overall strategy of human risk assessment for procymidone and flumioxazin. I also express my appreciation to all my colleagues in the Environmental Health Science Laboratory.
