Toxicokinetics of β-Amanitin in Mice and In Vitro Drug–Drug Interaction Potential

The toxicokinetics of β-amanitin, a toxic bicyclic octapeptide present abundantly in Amanitaceae mushrooms, was evaluated in mice after intravenous (iv) and oral administration. The area under plasma concentration curves (AUC) following iv injection increased in proportion to doses of 0.2, 0.4, and 0.8 mg/kg. β-amanitin disappeared rapidly from plasma with a half-life of 18.3–33.6 min, and 52.3% of the iv dose was recovered as a parent form. After oral administration, the AUC again increased in proportion with doses of 2, 5, and 10 mg/kg. Absolute bioavailability was 7.3–9.4%, which resulted in 72.4% of fecal recovery from orally administered β-amanitin. Tissue-to-plasma AUC ratios of orally administered β-amanitin were the highest in the intestine and stomach. It also readily distributed to kidney > spleen > lung > liver ≈ heart. Distribution to intestines, kidneys, and the liver is in agreement with previously reported target organs after acute amatoxin poisoning. In addition, β-amanitin weakly or negligibly inhibited major cytochrome P450 and 5′-diphospho-glucuronosyltransferase activities in human liver microsomes and suppressed drug transport functions in mammalian cells that overexpress transporters, suggesting the remote drug interaction potentials caused by β-amanitin exposure.

Amanita phalloides poisoning can cause acute hepatitis, leading to the rapid development of liver insufficiency and ultimately coma and death [2,13]. Nephrotoxicity is reported less frequently [17]. Yilmaz et al. [7] reported that an oral intake of approximately 50 g of fresh Amanita phalloides, representing a dose of 0.32 mg/kg of amatoxins, can be lethal. A few reports are available that describe the toxicokinetics of αand β-amanitin after the oral administration of toadstool extract to beagle dogs and rats [9,18,19]. In addition, the toxicokinetics of α-amanitin were investigated after intravenous (iv), intraperitoneal, and oral administration of amatoxin in rats and mice [14,18,20,21]. Low absolute bioavailability (3.5-4.8%) and substantial distribution to the intestines, kidneys, and liver were observed after the oral administration of α-amanitin at doses of 2, 5, or 10 mg/kg to mice [14]. Bang et al. [22] developed a sensitive analytical method for the simultaneous determination of αand β-amanitin in mouse plasma using high-resolution mass spectrometry (LC-HRMS) and measured plasma concentrations of αand β-amanitin after the oral administration of αand β-amanitin at 1.5 mg/kg dose in male ICR mice. αand β-amanitin showed similar toxicokinetic profiles at low dose exposure [22].
In addition, bile drainage caused more than a 70% reduction in plasma concentration of αand β-amanitin and reduced hepatic toxicity following the oral administration of dried powder of Amanita exitialis (60 mg/kg) [19]. The results suggested that biliary excretion of αand β-amanitin plays a crucial role in the toxicokinetics of these toxins, and the contribution of biliary excretions of αand β-amanitin to their absorption may be crucial to understanding their toxicokinetics as well as hepatic toxicity. However, the toxicokinetics and tissue distribution of β-amanitin ( Figure 1) upon exposure to different doses and different routes of administration remain unclear. Thus, this study aimed to investigate the toxicokinetic profile of β-amanitin after its oral and intravenous administration with various doses. To investigate the toxicokinetically susceptible tissues, we aimed to investigate the tissue distribution of β-amanitin in mice. Moreover, the involvement of hepatic and renal transporters, which are critical for excretion and tissue distribution of drugs [27][28][29], for αand β-amanitin need to be investigated. Organic anion transporting polypeptides (OATP) 1B1 and OATP1B3 were selected as hepatic transporters that are exclusively expressed in the liver and are critical for liver distribution and biliary excretion of substrates [29]. Organic cation transporter (OCT)2, multidrug and toxin extrusion protein (MATE)2-K, organic anion transporter (OAT)1 and OAT3 were selected as renal transporters that are exclusively expressed in the kidney and regulate kidney distribution and urinary excretion of substrates [29]. In addition, the interactions of β-amanitin with hepatic and renal transporters could increase susceptibility to β-amanitin and cause toxindrug interactions, similar to the interaction between OATP1B3 and α-amanitin [30]. We also aimed to evaluate the in vitro inhibition of major cytochrome P450 (CYP) and 5 -diphosphoglucuronosyltransferase (UGT) enzymes by β-amanitin using human liver microsomes (HLMs) and on hepatic and renal transporters using a mammalian cell overexpression system [14,31,32].

Tissue Distribution Experiments
Male mice fasted for 12 h with free access to water. Blood, brain, liver, kidney, heart, lung, stomach, intestine, and spleen were collected at 0.5, 1, 2, and 4 h after iv (0.8 mg/kg) injection or at 0.5, 1, 3, and 6 h after oral (10 mg/kg) administration of β-amanitin (n = 4 at each time point). Plasma samples (5 µL each) were harvested by centrifugation at 13,000× g for 3 min at 4 • C. Tissue samples were washed with cold normal saline, weighed, and stored at −80 • C until analysis.

Excretion Experiments
β-amanitin in water was administered by iv bolus injection into the tail vein at 0.8 mg/kg dose (n = 4) and by oral gavage at 10 mg/kg dose (n = 3) to male mice. Mice were returned to metabolic cages, and urine and feces samples were collected individually at 0-6, 6-12, 12-24, 24-36, and 36-48 h. Urine and feces samples were stored at −80 • C until the analysis. Concentrations of β-amanitin in biological samples were evaluated following our previous methodology [22]. The standard calibration curves for β-amanitin in mouse plasma ranged from 0.5 to 500 ng/mL and were linear, with a correlation coefficient of 0.9952. The inter-and intra-day accuracy and precision fell within the acceptance criteria (3.1-14.6% of coefficients of variation (CVs) and 92.5-115.0% of accuracy). The extraction recovery and matrix effect of this analytical method were 82.8-88.9% and 93.0-98.6%. Three freeze-thaw cycles (88.7-97.2% accuracy with CVs of 3.6-14.3%), short-term storage for 2 h on ice (94.7-97.9% accuracy with CVs of 3.6-7.4%), long-term storage for 2 weeks at −80 • C (90.4-105.0% accuracy with CVs of 3.7-9.1%), and post-preparation stability for 24 h in 4 • C autosampler (90.7-94.1% accuracy with CVs of 6.8-7.5%) showed negligible effect on the stability of β-amanitin [22]. The standard calibration curves for β-amanitin in various tissue homogenates showed fairly good linearity and the intra-day accuracy and precision results from the quality control samples prepared in various tissue homogenates were all in acceptance criteria (Table S1, Supplementary Materials).
An aliquot (5 µL) of plasma sample was mixed with 15 µL of 4 -hydroxydiclofenac (internal standard (IS), 5 ng/mL) in methanol and vortexed for 2 min. The mixture was centrifuged at 13,000× g at 4 • C for 5 min, and the supernatant was then transferred to an autosampler vial. An aliquot (5 µL) was injected into the LC-HRMS system for analysis.
Each tissue and feces sample was homogenized in water (1:3, w/v). In total, 50 µL aliquots of urine, tissue homogenates, or feces homogenates were mixed with 150 µL of 4 -hydroxydiclofenac solution (IS, 5 ng/mL in methanol). The other steps were as described for plasma sample preparation.

Toxicokinetic Parameters and Statistical Analysis
The toxicokinetic parameters of β-amanitin were analyzed by non-compartmental analyses (WinNonlin, Pharsight; Mountain View, CA, USA) ( Table 1). Data were expressed as mean ± standard deviations (SD). Comparisons of data between groups were performed using one-way ANOVA or Student t-test. Values of p < 0.05 were considered statistically significant.

Transport Studies of α-and β-Amanitin
HEK293 cells overexpressing OCT2, MATE2-K, OAT1, OAT3, OATP1B1, and OATP1B3 transporters and HEK293-control cells were maintained in a humidified atmosphere of 5% CO 2 at 37 • C in DMEM supplemented with 10% FBS and 5 mM non-essential amino acids. In the case of HEK293-OATP1B1, -OATP1B3, and MATE2-K cells, 2 mM sodium butyrate was added to the culture medium to enhance transport activity [28,33]. Cells were seeded at 2 × 10 5 cells/well in poly-D-lysine-coated 24-well plates. After 24 h, the growth medium was discarded, and attached cells were washed with HBSS and preincubated for 20 min in HBSS at 37 • C. Stock solutions of αand β-amanitin and representative inhibitors of transporters were diluted in HBSS to make the final concentration: cimetidine 1 mM (for OCT2), cimetidine 100 µM (for MATE2-K), probenecid 100 µM (for OAT1 and OAT3), and rifampin 100 µM (for OATP1B1 and OATP1B3). Uptake of 50 µM αand β-amanitin was measured in the absence and presence of representative inhibitors for 5 min at 37 • C. Plates were placed immediately on ice, and cells were then washed twice with 1 mL of ice-cold HBSS. Residual HBSS was removed thoroughly from the plates. Subsequently, 150 µL of 80% methanol containing IS was added to each sample well, and the cell plates were shaken gently for 20 min at 4 • C. The remaining steps were the same as described for plasma sample preparation.

Toxicokinetics of β-Amanitin in Mice
First of all, intravenous bolus administration of 0.2, 0.4, and 0.8 mg/kg via tail vein and oral administration of 2, 5, and 10 mg/kg via oral gavage did not cause death or serious toxicity in mice in this study.
Mean plasma concentration-time curves of β-amanitin after the iv administration of β-amanitin at doses of 0.2, 0.4, and 0.8 mg/kg demonstrated rapid elimination from plasma; t 1/2 was in the range of 18.3-33.6 min (Figure 2A and Table 2). High systemic clearance (CL) after iv injection was observed for all three doses (Figure 2A and Table 2). Additionally, iv-injected β-amanitin showed linear kinetics in the dose range of 0.2-0.8 mg/kg, evidenced by a dose-proportional increase in AUC ( Figure 3A) and dose-independent in CL and V ss (Table 2 and Figure 3C,D). Consistent with the dose linearity of these parameters, the initial plasma concentration (C 0 ) was also increased dose proportionally in the intravenous dose range of 0.2~-0.8 mg/kg.
More than 60% of iv-injected β-amanitin was recovered as a parent form, and it suggested the limited in vivo metabolism of β-amanitin.
Dose linearity was also observed after oral administration at doses of 2-10 mg/kg (Figures 2B and 3B and Table 3). However, t 1/2 values of orally administered β-amanitin (105.0~132.0 min) were much longer than iv-injected β-amanitin (18.3~33.6 min) (Tables 2 and 3). Moreover, T max showed a broad range from 10~120 min (Table 3). Thus, the absorption of β-amanitin likely occurred throughout the intestine, and delayed absorption may decrease the elimination rate of orally administered β-amanitin. Cumulative urinary and fecal excretion of β-amanitin for 48 h following oral administration was 2.4 ± 1.2% and 72.4 ± 24.7%, respectively ( Figure 4B). These values also contrast the values measured after iv injection. Greater fecal recovery of β-amanitin could be attributed to unabsorbed compounds following oral administration, supporting the low estimate for oral bioavailability (i.e., 7.3-9.4%).    Table 3. Toxicokinetic parameters of β-amanitin in male ICR mice after oral administration of three doses (mean ± SD).

Tissue Distribution of β-Amanitin in Mice
The tissue distribution of β-amanitin differs depending on the route of administration ( Figure 5 and Table 4). Intravenously administered β-amanitin was distributed to the kidneys, spleen, heart, lungs, and liver and eliminated from these tissues at a slower rate compared to the elimination from plasma. Levels of β-amanitin in the brain, stomach, and intestines were below detection limits ( Figure 5A), suggesting that the distribution of β-amanitin was restricted to some tissues. Tissue-to-plasma AUC ratios of β-amanitin in the heart, lungs, liver, spleen, and kidneys were 0.2, 0.7, 0.5, 2.6, and 2.5, respectively (Table 4). Conversely, tissue concentration of β-amanitin in the stomach, intestines, kidneys, lungs, and spleen was greater than that in plasma ( Figure 5B), with high tissue-to-plasma AUC ratios of 28.8, 87.3, 6.7, 1.6, and 1.8, respectively (Table 4). These values in po administration were greater than those in iv injection. Tissue distribution to heart, lung, liver, and kidney was 0.5, 1.6, 0.5, and 1.8, respectively, which was similar to distribution after iv injection (Table 4). Thus, the stomach, intestine, and kidney may be target organs after oral β-amanitin poisoning. The heart, lung, liver, and spleen may also be target organs after β-amanitin exposure intravenously and orally in mice. In the case of the kidneys, liver, and spleen, tissue concentrations of orally administered β-amanitin tend to increase after 6 h, and it may cause delayed toxicity ( Figure 5B). Table 4. The area under the plasma and tissue concentration-time curves (AUC last ) for β-amanitin in male ICR mice after intravenous (iv) and oral (po) administration of β-amanitin at doses of 0.8 or 10 mg/kg, respectively, (mean ± SD) (n = 4).

Discussion
The Amanitaceae family is responsible for nearly 95% of all fatal mushroom poisonings, and amatoxins are the primary threats to human life [7]. The content of amatoxins varies by Amanita species, but αand β-amanitin are the most abundant substances. For example, an amatoxin content of 9.3 mg/g was observed in dried mushrooms and αand β-amanitin accounted for 56% of toxins found in dried powder from Amanita phalloides [5]. αand β-amanitin account for 82% of toxins (2.87 mg of amanitin/3.49 mg peptide toxins/g dried powder) in Amanita exitialis [9,19]. Among amatoxins, α-amanitin is of interest for toxicological investigation. The lethal dose of α-amanitin is reported to be 0.3-0.6 mg/kg in mice and 4.0 mg/kg in rats following intraperitoneal injection, and 0.1 mg/kg in humans after oral administration [2,20,38]. However, little information regarding toxicokinetics, tissue distribution, and toxic doses of β-amanitin is available in the literature. Bolus administration of 0.8 mg/kg via tail vein and 10 mg/kg via oral gavage did not cause toxicity in the present study.
β-amanitin showed a short t 1/2 and high CL for all iv doses ( Table 2). The rapid elimination of β-amanitin in the urine in parent form is likely the explanation ( Figure 3B). Sun et al. reported that α-amanitin, excreted through glomerular filtration, was reabsorbed in renal tubules and caused kidney toxicity [20,39]. In addition, α-amanitin is actively taken up into the liver via hepatic transporters, OATP1B1 and OATP1B3 (Figure 3.3), which could lead to hepatotoxicity [10,26]. Similar to the case of α-amanitin, accumulation of β-amanitin in the liver and kidneys might also correlate with liver and kidney toxicity. OATP1B1-and OATP1B3-mediated hepatic uptake and OAT3-mediated kidney uptake of β-amanitin (Figure 3.3), along with reabsorption into renal tubules, may be involved. The potency of α-amanitin was 10-fold greater than β-amanitin in MCF-7 cells [37]. Thus, in spite of approximately two-fold higher cellular uptake mediated by hepatic or renal transporters and higher liver or kidney accumulation of β-amanitin, the contribution of β-amanitin to in vivo amatoxin toxicity may be lower than α-amanitin. αand β-amanitin share common features in the elimination process, including major elimination in the urine without significant metabolism [11,40]. This similarity likely reflects their structural similarity. The only difference in structure is a -NH 2 group in α-amanitin vs. an -OH group in β-amanitin. The hydroxyl group confers greater affinity for OAT3, higher cellular penetration, and higher accumulation in the heart, kidney, and spleen.
Interestingly, orally administered αand β-amanitin showed distinctive accumulation in the stomach and intestine ( [14] and Table 4 in this study), leading to prolonged absorption of toxicity and, consequently, a delay in elimination and potentiation of toxicity. Sun et al. [19] showed that the interruption of enterohepatic cycling of amatoxins by biliary drainage in dogs induced a 70% reduction in intestinal amatoxin absorption and lessened signs of severe toxicity. Taken together, intestinal absorption of and exposure to amatoxin may correlate with toxicity and could underlie different amatoxin toxicities depending on the route of poisoning. In this study, concentrations of orally administered β-amanitin in the kidney, liver, and spleen tend to increase after 6 h. This could lead to delayed toxicity in these tissues, which is consistent with the latent period of hepatic and kidney toxicity [41].
The intestinal tract, liver, and kidney are vulnerable to amanitin toxicity [2,13] and also express drug-metabolizing enzymes and transporters that regulate the absorption, distribution, metabolism, and excretion of endogenous substrates and xenobiotics [30,42]. In vitro inhibitory effects of β-amanitin on major human drug-metabolizing enzymes and transporters were thus evaluated. β-amanitin showed weak inhibition of CYP2A6, CYP2B6, and CYP2D6 with IC 50 values of 93.9, 38.0, and 76.2 µM, respectively. Inhibition of CYP1A2, CYP2C8, CYP2C9, CYP2C19, and CYP3A4 at 100 µM in HLMs was almost negligible (Figure 7). IC 50 values of β-amanitin toward CYP2A6, CYP2B6, and CYP2D6 were much higher than concentrations observed (15.9 to 162 ng/mL) in a person suffering from poisoning [26]. Thus, a β-amanitin-induced drug interaction is not likely to be caused by CYP inhibition. β-amanitin inhibited MATE2-K and OAT3 with IC 50 values of 2.1 and 37.4 µM, respectively. Again, these concentrations are much higher than the plasma concentrations found in a case of poisoning [26]. β-amanitin did not significantly inhibit other transport activities (i.e., OCT1, OCT2, OAT1, OATP1B1, OATP1B3, NTCP, P-gp, or BCRP) ( Figure 9) and UGT enzyme activities (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, or UGT2B7) ( Figure 8). Thus, the possibility of drug interaction caused by βamanitin poisoning is remotely related to drug-metabolizing enzymes and transporters, and these results also seem to be reassuring that αand β-amanitin weakly or negligibly interacts with CYPs and UGTs, and drug transporters. In this study, β-amanitin is a substrate for OAT3, OATP1B1, and OATP1B3, and it inhibits the transport activity of MATE2-K and OAT3. The interaction between these transporters and β-amanitin could be caused by the direct interaction between β-amanitin and substrate probe drugs such as metformin, estrone-3-sulfate, and estradiol-17-β-D-glucuronide. After the 1 h incubation of β-amanitin and substrate probe drugs, the mass signal of β-amanitin was not affected by the presence of substrate drugs and vice versa ( Figure S1, Supplementary Materials). The absence of interference between β-amanitin and substrate drugs indicated no significant chemical interaction between them.
In addition, the involvement of MATE2-K, OAT3, OATP1B1, and OATP1B3 could be used as the detoxification mechanism of β-amanitin, as the clinical therapy in amatoxin poisoning implies. Single or combined treatment of benzylpenicillin, silibinin, and Nacetylcysteine has been used [2]. The mechanism of these treatments may be attributed to the hepatoprotective and anti-oxidative effect of silibinin and N-acetylcysteine and the reduced hepatic distribution of α-amanitin by inhibiting OATP1B3 using benzyl penicillin and silibinin [2,30,43,44]. In this case, the detoxification mechanism of α-amanitin poisoning can be explained by inhibiting the interaction between α-amanitin and OATP1B3 rather than the direct interaction between a-amanitin and the treated drugs, such as benzylpenicillin, silibinin, and N-acetylcysteine.
In addition to the interaction between β-amanitin and drug transporters, further toxicokinetic features of β-amanitin after iv and oral dose escalation, including tissue distribution and excretion characteristics, as well as in vitro assessment of drug interactions with drug-metabolizing enzymes and transporters, would provide additional useful information for a safety assessment using an in vivo prediction model.

Conclusions
Intravenous and oral administration of β-amanitin suggests low bioavailability, high CL, and low V ss : The absorption of β-amanitin is limited, and excretion is rapid, mainly via the renal route. Tissue distribution studies showed that β-amanitin was accumulated in gastrointestinal tissues based on stomach-and intestine-to-plasma ratios and unabsorbed β-amanitin that was recovered in feces after oral administration. β-amanitin showed weak or negligible inhibition of major human CYP and UGT enzymes and drug transporters, suggesting the likelihood that β-amanitin causes drug interactions via inhibition of CYP and UGT enzymes and drug transporters is remote.