Toxicokinetic evaluation during intoxication of psychotropic drugs using brain microdialysis in mice

Abstract

In the event of an overdose, the pharmacokinetics of the drug may be altered, resulting in an unexpectedly rapid increase in blood and tissue drug concentrations. Because central nervous system (CNS)-acting drugs are the major cause of hospitalization for overdose, brain concentrations, which are closely related to the development of acute psychotropic symptoms, would be important. However, due to the lack of an appropriate model for overdose, it is difficult to predict the CNS symptoms of patients with acute poisoning. To clarify the toxicokinetics during intoxication with CNS-acting drugs, we investigated the relationship between the dose and concentrations in the blood and brain in mice. Therapeutic or toxic doses of phenobarbital, flunitrazepam, imipramine, and amoxapine were administered intraperitoneally to mice. Serum and extracellular fluid of the brain were collected up to 24 hr after administration and analyzed using LC-MS/MS to determine the pharmacokinetic parameters in the serum and brain. A comparison of the four psychotropic drugs showed that the toxicokinetics of amoxapine in the blood and brain are clearly different from others, with the brain concentrations being specifically highly susceptible to increase during dose escalation. These results are consistent with the CNS-related symptoms observed in amoxapine overdose. Therefore, the methodology of the current study could be useful for predicting CNS toxicity during psychotropic drug poisoning.

INTRODUCTION

The number of drug abuse cases has continued to increase (Matsumoto et al., 2020), and sleep medication and anxiolytics have become the main causative agents of drug-related psychiatric disorders. Overdose is caused by an unavoidable increase in the dosage or type of drug due to lack of improvement in symptoms or by a dependent situation due to long-term administration. In addition, some central nervous system (CNS)-acting drugs have a narrow safety margin, and there is a high risk of death due to overdose (Hikiji et al., 2016).

Standard treatments for acute drug intoxication include gastrointestinal lavage, forced diuresis, and hemodialysis (Müller and Desel, 2013). However, in overdose cases of CNS-acting drugs in suicide attempts, a certain length of time is required from drug intake to arrival at the hospital; therefore, gastrointestinal lavage is not effective in many cases. In addition, most CNS-acting drugs have high plasma protein binding rates or relatively large distribution volumes, which are not amenable to hemodialysis (King et al., 2019) and therefore do not often effectively reduce blood concentrations.

In general, when a therapeutic dose of a drug is administered, there is a correlation between the dose and blood concentration. A relationship between blood and brain concentrations and the mechanism of brain transport has been reported for some drugs, such as imipramine, a typical tricyclic antidepressant (O’Brien et al., 2012). However, there is a lack of information on tissue concentrations at toxic doses, and the relationship between doses and tissue concentrations is not clear. When intoxicating or lethal doses of drugs are taken, it is expected that protein binding sites, metabolism, and excretion may readily be saturated, resulting in unexpected increases in blood or tissue drug concentrations. Since drug concentrations in the brain are considered to be deeply involved in the development of acute poisoning symptoms of CNS-acting drugs, clarifying the relationship between brain and blood concentrations is important in promoting better emergency treatment by estimating brain concentrations from patients’ blood concentrations.

Although a compartmental model is widely known as a method for pharmacokinetic analysis, physiologically based pharmacokinetic (PBPK) models have recently been used to predict the blood and tissue concentrations of drugs. The PBPK model is a mathematics-based model and simulation method that is constructed using mechanism-based model structures and parameters to predict the absorption, distribution, metabolism, and excretion of drugs (Hartmanshenn et al., 2016). However, the analysis using the PBPK model has also only examined pharmacokinetic modeling at therapeutic doses, but not pharmacokinetics during intoxication (‘toxicokinetics’). Toxicity, including toxicokinetics, must be evaluated in animals treated with toxic doses during drug development; however, blood (but not tissue) drug concentrations have been measured in a series of toxicity tests (Ministry of Health, Labour and Welfare, 2014). Therefore, the purpose of this study was to clarify the changes in toxicokinetics and brain drug concentrations in mice treated with toxic doses.

The drugs examined in this study were phenobarbital, flunitrazepam, imipramine, and amoxapine. Phenobarbital was selected from the barbiturates, which are considered to have a narrow safety margin and therefore, may cause acute poisoning (López-Muñoz et al., 2005). Although benzodiazepines are considered to have a wider safety margin among CNS-acting drugs, they account for a large proportion of the total number of drug-induced acute poisonings due to their widespread use in the pharmaceutical market. Flunitrazepam was selected because it is a highly potent benzodiazepine and one of the leading causes of death among drug overdose fatalities (Hikiji et al., 2016). Imipramine was selected because it is an early-developed, first-generation tricyclic antidepressant and it can be used as a basis for other drugs. In addition, there is relatively more information regarding therapeutic dose administration. Amoxapine, a second-generation, tricyclic antidepressant, produces severe CNS toxicity in cases of intoxication. Amoxapine-induced seizures may also be resistant to benzodiazepines (Merigian et al., 1995), and treatment is often difficult. These clinical experiences suggest that the brain concentration of amoxapine increases rapidly during acute poisoning. In the present study, we attempted to compare the toxicokinetics and changes in brain concentration of amoxapine used as a model drug exhibiting CNS toxicity during acute intoxication with those of imipramine and other psychotropic drugs. To the best of our knowledge, no study has been reported on the brain concentration of drugs during the acute intoxication, and this study can serve as a basis for the development of therapies for acute poisoning.

MATERIALS AND METHODS

Materials

Imipramine hydrochloride, amoxapine and phenobarbital sodium salt were purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Desipramine hydrochloride, and flunitrazepam were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Imipramine-D3 maleate in methanol, amoxapine-D8 in methanol, phenobarbital-D5 in methanol and diazepam-D5 in methanol were purchased from Sigma-Aldrich Japan (Tokyo, Japan). All other reagents used were of the highest grade and commercially available.

Animals

ICR male mice (8 weeks old) were obtained from Sankyo Lab Service Corporation (Tokyo, Japan). Mice were housed in plastic cages in a temperature controlled room (22 ± 1°C) and maintained on a 12 hr light-dark cycle with free access to food and water. All procedures for animal care were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the Showa University (Approval No.23022). Every effort was made to minimize the number of animals used and their suffering.

Microdialysis surgery

The mice, anesthetized with a combination of medetomidine (0.75 mg/kg, i.p.), midazolam (4 mg/kg, i.p.) and butorphanol (5 mg/kg, i.p.), were placed in a stereotaxic apparatus. A microdialysis probe (D-I-6-02; 0.22 mm outer diameter, 2 mm membrane length; Eicom Co., Ltd., Kyoto, Japan) was implanted into the striatum at the following coordinates: AP: +0.5 mm, Ml: +1.7 mm relative to bregma and DV: −4.4 mm the skull (Kaizaki et al., 2014). The probes were secured onto the skull using dental acrylic. The mice were allowed to recover for at least 24 hr before the experiment was begun. After the experiments, the mice were decapitated and brain tissue was removed to confirm that the probe had been embedded in the striatum.

Drug treatment

The drugs used in the present study are taken orally in humans; however, most of their toxicological information comes from intraperitoneal administration to animals. In addition, in order to avoid the factor of intestinal absorption of the drug, intraperitoneal administration to mice was employed for all drugs in this study.

The therapeutic and toxic doses of each drug were determined as follows. The therapeutic dose of phenobarbital was set at 10 mg/kg (Ghosh et al., 2002; Ho and Ho, 1979), and the toxic dose was set at 150 mg/kg with reference to the LD₅₀ values (162–660 mg/kg) (Budavari et al., 1996). The therapeutic dose of flunitrazepam was set at 10 mg/kg (Depoortere et al., 1986). The LD₅₀ value of flunitrazepam is reported to be 1050 mg/kg (Miyagawa et al., 1985), but the toxic dose was set at 200 mg/kg because of its solubility. The therapeutic dose of imipramine was set at 20 mg/kg (Wróbel et al., 2015; Depoortere et al., 1986; Zhao et al., 2015), and the toxic dose was set at 100 mg/kg with reference to the LD₅₀ value (110 mg/kg) (Budavari et al., 1996). The therapeutic dose of amoxapine was set at 12 mg/kg (Malatynska et al., 2005), and the toxic dose was set at 60 mg/kg, because acute deaths were observed at 70 mg/kg in preliminary experiments.

In vivo microdialysis

The probes were perfused at 2 µL/min with artificial cerebrospinal fluid (Hataoka et al., 2017). One hour after the reflux, the dialysate sample was collected in 10-min fractions. Samples were collected up to 24 hr after drug administration. Samples at 0.5, 1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 22 and 24 hr after drug administration were subjected to liquid chromatography-mass spectrometry (LC-MS/MS) analysis directly.

Blood sampling

At 0.5, 1, 2, 4, 8, and 24 hr after drug administration, 30 µL of blood was collected from the tail vein of mice using microhematocrit tubes (Fisher Scientific, Tokyo, Japan). After standing for 30 min, the blood was centrifuged for 10 min at 5000 × g at 4°C and the serum was collected. The drug in the serum was extracted according to the QuEChERS method (Mathiaux et al., 2014). Ten microliters of serum was added to 290 µL of water, 100 mg of Agilent Bond Elut QuEChERS (Agilent, Tokyo, Japan), and 300 µL of acetonitrile containing the internal standard (IS) and mixed well using a vortex mixer. The mixture was centrifuged for 10 min at 15000 × g at room temperature. The supernatant was collected and subjected to LC-MS/MS analysis. Depending on the drug to be measured, the following IS were used: 3.3 µg/mL of phenobarbital-D5 for phenobarbital analysis, 0.067 µg/mL of diazepam-D5 for flunitrazepam analysis, 0.067 µg/mL of imipramine-D3 for imipramine and its metabolite desipramine analysis, and 0.067 µg/mL of amoxapine-D8 for amoxapine analysis.

LC-MS/MS analysis

LC-MS/MS was performed using a LC-40ADXR and LCMS-8045 (Shimadzu, Tokyo, Japan). Chromatographic separation was achieved on a Phenomenex Kinetex XB-C18 column (2.1 mm I.D. × 100 mm., 1.6 µm; Shimadzu) with an equivalent Phenomenex Security Ultra C18 guard column (2.1 mm ID; Shimadzu). The column temperature was set at 40°C. The injection volume was 1 µL or 5 µL, depending on the concentration of the sample. The mobile phases were (A) 10 mmol/L ammonium formate and 0.1% formic acid, (B) methanol containing 10 mmol/L ammonium formate and 0.1% formic acid. The initial elute condition was set at 95% A, the condition was changed linearly to 5% A in 7.5 min, held for 2.5 min, then immediately back to 95% A, and held for 5 min. The flow rate was set at 0.3 mL/min. After ionization with electrospray ionization, the samples were analyzed in multiple reaction monitoring mode. Phenobarbital was detected in negative mode. Flunitrazepam, imipramine, desipramine and amoxapine were detected in positive mode. The flow rates of the nebulizer gas, the drying gas and the heating gas are set at 3 L/min, 10 L/min, 10 L/min, respectively. Temperatures of the interface, the desolvation line and the heat block are set to 300°C, 250°C, and 400°C, respectively. The instrumental conditions for each drug are indicated in Table 1.

Table 1. Liquid Chromatography-Mass spectrometry (LC-MS/MS) conditions.

Compound	ESI mode	Precursor ion	Product ion	Collision energy	Retention time
	(+)/(-)	(m/z)	(m/z)	(V)	(min)
Phenobarbital	(-)	231.2	42.0	17	4.84
Phenobarbital-D5	(-)	236.2	42.0	17	4.83
Flunitrazepam	(+)	314.0	268.0	26	5.98
Diazepam-D5	(+)	290.1	198.1	31	6.81
Imipramine	(+)	281.2	86.1	17	6.08
Desipramine	(+)	267.2	72.1	15	6.15
Imipramine-D3	(+)	284.2	89.2	18	6.06
Amoxapine	(+)	314.1	271.1	15	5.99
Amoxapine-D8	(+)	322.1	276.0	25	5.98

Pharmacokinetic parameters

Pharmacokinetic parameters such as Cmax, Tmax, T1/2, and area under the drug concentration-time curve (AUC) were calculated using non-compartmental analyses. The AUC ratio was calculated as:

AUC ratio = (AUC in the brain / AUC in the blood) x 10^-3 (1)

The brain transfer unit (BTU) was calculated as:

The BTU is a value indicating the number of times the AUC ratio (brain/serum) increases when the dose is changed by one therapeutic unit; the larger the value, the easier the brain transfer.

Statistical analysis

All data were tested for equality of variance then analyzed by Student’s t-test or Welch’s t-test with JMP Pro 16.0 (SAS, Cary, NC, USA).

RESULTS

Toxicokinetics of phenobarbital

The time-dependent changes in phenobarbital concentrations in the serum and extracellular fluid of the brain of mice treated with the therapeutic (10 mg/kg) and toxic doses (150 mg/kg) are shown in Fig. 1. The Cmax in the serum and brain of the toxic dose were approximately 15 and 20 times the therapeutic dose, respectively, and were similar to the dose increase rate (15 times). Both serum and brain Tmax and serum half-life were not affected by the doses, but brain half-life was significantly prolonged as the dose increased (Table 2). The serum AUC_0-24 with the toxic dose was approximately 28 times higher than with the therapeutic dose, and the brain AUC_0-24 was approximately 52 times higher (Table 2), indicating that the increase rate of AUC in the brain in terms of dose increase was approximately two-fold higher than that in the serum.

Fig. 1

Time-dependent changes in phenobarbital concentration in the serum and brain of mice. Phenobarbital concentration in the serum (a, µg/mL) and the extracellular fluid of the brain (b, ng/mL) of mice treated with toxic (150 mg/kg, i.p., solid line) and therapeutic (10 mg/kg, i.p., dashed line) doses of phenobarbital were determined using LC-MS/MS. Concentrations were plotted in a semi-log plot against time after administration. Values represent the mean ± SEM (therapeutic dose group n = 5, toxic dose group n = 4).

Table 2. Pharmacokinetic parameters of four psychotropic drugs, phenobarbital, flunitrazepam, imipramine, and amoxapine, after a single therapeutic or toxic dose in mice.

Compound	Dose (mg/kg)	Cmax							Tmax							t1/2							AUC
		Serum (µg/mL)			Brain (ng/mL)				Serum (hr)			Brain (hr)				Serum (hr)			Brain (hr)				Serum (µg hr/mL)			Brain (ng hr/mL)			Ratio [Brain/ Serum] (×10−3)
Phenobarbital	10	18.5	±	2.1	207.4	±	17.9		1.40	±	0.37	0.70	±	0.12		8.99	±	5.79	15.6	±	2.5		161.3	±	10.5	1528	±	136##	9.50	±	0.76
	150	285.3	±	24.1**	4201	±	366**,##		1.25	±	0.25	0.75	±	0.14		8.52	±	1.68	32.7	±	2.3**,##		4587	±	370**	79842	±	8227**,##	16.7	±	2.1**
Flunitrazepam	5	1.02	±	0.15	15.6	±	1.5##		1.00	±	0.00	1.00	±	0.00		1.04	±	0.06	0.93	±	0.05		3.18	±	0.30	27.4	±	1.9##	8.83	±	0.53
	200	16.2	±	1.5**	371.1	±	43.8**,##		1.67	±	0.21*	2.18	±	0.17**		3.18	±	0.43**	2.78	±	0.52*		126.2	±	27.5**	2015	±	345**,##	16.6	±	1.3**
Imipramine	20	0.577	±	0.134	13.4	±	2.3##		0.92	±	0.08	0.67	±	0.11#		2.71	±	0.08	1.02	±	0.06##		2.95	±	0.35	32.0	±	4.7##	10.9	±	1.0
	100	2.09	±	0.12**	136.0	±	13.6**,##		0.92	±	0.83	0.83	±	0.11		2.75	±	0.20	1.30	±	0.15##		11.1	±	0.6**	300.3	±	24.0**,##	27.1	±	1.2**
Desipramine	(Imipramine 20)	0.13	±	0.01	0.73	±	0.05##		0.75	±	0.11	1.33	±	0.21#		2.29	±	0.13	N.A.				0.77	±	0.13	N.A.			N.A.
	(Imipramine 100)	0.64	±	0.08**	7.09	±	0.81**,##		2.83	±	0.54**	3.33	±	1.21**		4.08	±	0.28**	N.A.				7.57	±	0.95**	N.A.			N.A.
Amoxapine	12	2.16	±	0.20	0.98	±	0.08##		2.67	±	1.09	3.17	±	0.17#		5.28	±	0.95	6.01	±	1.05		23.7	±	2.0	3.76	±	1.45	0.17	±	0.06
	60	3.87	±	0.22**	27.3	±	4.8**,##		2.17	±	1.19	2.33	±	0.21		4.00	±	0.16	2.81	±	0.15		51.0	±	3.6**	116.3	±	13.4**,##	2.38	±	0.36**

Values represent the mean ± SEM from four to six animals; * and ** denote p < 0.05 and p < 0.01, respectively, in comparison to therapeutic dose; ^# and ^## denote p < 0.05 and p < 0.01, respectively, in comparison to serum levels; N.A., data not available because the brain concentration was below the detection limit after 8 hr or 18 hr of imipramine administration.

Toxicokinetics of flunitrazepam

The serum and brain concentrations of flunitrazepam with therapeutic (5 mg/kg) and toxic (200 mg/kg) doses are shown in Fig. 2. The serum and brain Cmax of the toxic dose were approximately 16 and 24 times the therapeutic dose, respectively, which were lower than the dose increase rate (40 times). Both serum and brain Tmax became significantly larger as the dose increased (Table 2). Both serum and brain half-lives were significantly prolonged at toxic doses (Table 2). The serum AUC_0-24 of the toxic dose was approximately 40 times higher than the therapeutic dose and the brain AUC_0-24 was approximately 74 times higher (Table 2), indicating that the increase rate of the AUC in the brain was approximately two-fold that in the serum.

Fig. 2

Time-dependent changes in flunitrazepam concentration in the serum and brain of mice. Flunitrazepam concentration in the serum (a, µg/mL) and the extracellular fluid of the brain (b, ng/mL) of mice treated with toxic (200 mg/kg, i.p., solid line) and therapeutic (5 mg/kg, i.p., dashed line) doses of flunitrazepam were determined using LC-MS/MS. Concentrations were plotted in a semi-log plot against time after administration. Values represent the mean ± SEM (therapeutic dose group n = 5, toxic dose group n = 6).

Toxicokinetics of imipramine and its metabolite desipramine

The serum and brain concentrations of imipramine with the therapeutic (20 mg/kg) and toxic (100 mg/kg) doses are shown in Fig. 3a and 3b, respectively, and those of desipramine are shown in Fig. 3c and 3d, respectively. The Cmax of imipramine in the serum and brain with the toxic dose was approximately 3.6 and 10 times the therapeutic dose, respectively. The ratio of the increase in Cmax in the brain with the toxic dose was approximately twice the dose increase. Both Tmax and half-life in the serum and brain were not affected by the dose. The serum AUC_0-24 with the toxic dose was approximately 3.8 times higher than the therapeutic dose, and the brain AUC_0-24 was about 9.4 times higher (Table 2), indicating a higher rate of increase in the AUC in the brain than in the serum. The AUC ratio of brain to serum increased 2.5 times by dose increase. The Cmax of desipramine in the serum and brain with the toxic dose of imipramine was approximately 5 and 10 times the therapeutic dose, respectively, which was higher than the dose increase. Both serum and brain Tmax were significantly higher at toxic doses than at therapeutic doses (Table 2). Serum half-life was significantly prolonged with toxic doses (Table 2). The serum AUC_0-24 with the toxic dose was approximately 9.8 times higher than the therapeutic dose. Since the brain concentration of desipramine was below the detection limit after 8 hr of administration for the therapeutic dose and after 18 hr for the toxic dose, the brain half-life, brain AUC_0-24, and AUC ratio were not determined.

Fig. 3

Time-dependent changes in imipramine and desipramine concentrations in the serum and brain of mice. Imipramine concentration in the serum (a, µg/mL) and the extracellular fluid of the brain (b, ng/mL), desipramine concentration in serum (c, µg/mL) and the extracellular fluid of the brain (d, ng/mL) of mice treated with toxic (100 mg/kg, i.p., solid line) and therapeutic (20 mg/kg, i.p., dashed line) doses of imipramine were determined using LC-MS/MS. Concentrations are plotted in a semilog plot against time after administration. Values represent the mean ± SEM (n = 6).

Toxicokinetics of amoxapine

The serum and brain concentrations of amoxapine with therapeutic (12 mg/kg) and toxic (60 mg/kg) doses are shown in Fig. 4. The serum and brain Cmax of the toxic dose were approximately 1.8 and 28 times, respectively, the therapeutic dose, indicating that the brain Cmax increased markedly with the toxic dose compared with the dose increase. Both Tmax and half-life in the serum and brain were not affected by the dose. The serum AUC_0-24 with the toxic dose was approximately 2.2 times higher than the therapeutic dose, and the brain AUC_0-24 was approximately 31 times higher (Table 2), indicating a higher rate of increase in the AUC in the brain than in the serum. The AUC ratio of brain to serum increased markedly with the toxic dose, which was approximately 14 times higher than the therapeutic dose.

Fig. 4

Time-dependent changes in amoxapine concentration in the serum and brain of mice. Amoxapine concentration in the serum (a, µg/mL) and the extracellular fluid of the brain (b, ng/mL) of mice treated with toxic (60 mg/kg, i.p., solid line) and therapeutic (12 mg/kg, i.p., dashed line) doses of amoxapine were determined using LC-MS/MS. Concentrations were plotted in a semilog plot against time after administration. Values represent the mean ± SEM (n = 6).

Comparison of BTU in psychotropic drugs

To compare the brain transferability between drugs, we calculated the BTU, which is a value indicating the number of times the brain/serum AUC ratio increases when the dose is doubled, using the therapeutic dose as a reference (Table 3). The BTU of the drugs used in this study was as follows: amoxapine, imipramine, phenobarbital, and flunitrazepam. Log P, an index of lipophilicity, followed the order imipramine > amoxapine > flunitrazepam > phenobarbital, indicating that the order of BTU was different from that of Log P.

Table 3. The brain transfer unit (BTU), protein binding rate, and log P of the four psychotropic drugs, phenobarbital, flunitrazepam, imipramine, and amoxapine, administered to mice.

Compound	BTU	Protein binding rate of human plasma (%)	Log-P[1]
Phenobarbital	0.234	20–45[1]	1.47
Flunitrazepam	0.0940	77–80[2]	2.06
Imipramine	0.993	60–96[1]	4.8
Amoxapine	5.46	90[1]	3.4

[1] DRUGBANK 5.1.8

[2] Suzuki et al., 2014

DISCUSSION

In the event of an overdose, the pharmacokinetics of the drug may be altered, resulting in an unexpectedly rapid increase in blood and tissue drug concentrations. Because CNS-acting drugs are the major cause of hospitalization for overdose, brain concentrations, which are closely related to the development of acute psychotropic symptoms, are important. However, due to the lack of a toxicokinetic model for overdose, it is difficult to predict the CNS symptoms of patients with acute poisoning. To clarify the toxicokinetics during intoxication with CNS-acting drugs, we investigated the relationship between the dose and concentrations in the blood and brain using a mouse model. Since the drugs used in this study are taken orally in humans, it would be desirable to examine oral administration as well. However, intraperitoneal injection was employed here to estimate the amount of intoxication in the animals. The individual PK parameters might be different between intraperitoneal and oral administrations; however, the feature of the toxicokinetic profiles is expected to be similar between intraperitoneal and oral administrations.

The half-life of phenobarbital, a drug known to be narrow safety margin, in the brain at the toxic dose was prolonged to about twice of that at the therapeutic dose (Table 2). An increase in the AUC ratio of brain to serum was observed with the toxic dose, indicating that the brain transfer rate of phenobarbital increased with dose escalation and the brain AUC was more likely to increase than that in serum (Fig. 1, Table 2). Phenobarbital has a plasma protein binding rate of 20–45% in humans (DRUGBANK 5.1.8) and binds predominantly to albumin (Breckenridge, 1974). Since albumin is abundant in plasma (Chuang and Otagiri, 2006; Shen et al., 2013), protein binding would not be easily saturated even at toxic doses. Nevertheless, the brain transfer increased with the toxic dose, which is probably due to the increase in the concentration of free drug in the plasma. These results indicate that brain concentrations of phenobarbital tended to increase but decreases during overdose were resisted. This indicates that toxic symptoms might be prolonged during intoxication, which is consistent with the symptoms observed in a clinical setting (Jana et al., 2014).

The relative increase in serum and brain Cmax at the toxic dose of flunitrazepam, a typical benzodiazepine with a wide safety margin, was lower than the dose increase (Fig. 2, Table 2). However, the serum and brain half-lives were prolonged and the AUC ratio of brain to serum was increased during intoxication, indicating that the brain AUC is more likely to be elevated compared to the serum. It has been suggested that brain concentrations are more likely to increase than the blood at toxic doses of psychotropic drugs, even in benzodiazepine. In addition, a similar degree of Tmax delay and half-life prolongation were observed in the serum and brain when flunitrazepam was administered at the toxic dose, suggesting that it is possible to estimate the brain transition from blood concentrations in this type of drug.

In this study, we examined tricyclic antidepressants, for which acute intoxication often leads to hospitalization and fatalities and the life prognosis is sometimes guarded (McKenzie and McFarland, 2007). Cardiotoxicity is a typical acute symptom of first-generation tricyclic antidepressants such as imipramine (Nguyen et al., 2021; Giwa and Oey, 2018; Gheshlaghi et al., 2012). CNS toxicity (causing, for example, convulsions) is more problematic than cardiotoxicity in amoxapine poisoning (Kulig et al., 1982). Litovitz et al. reported that seizures with amoxapine overdose occurred in 36.4% of patients with overdoses, compared with 4.3% with other cyclic antidepressant poisoning, indicating that amoxapine is particularly prone to CNS-related symptoms among cyclic antidepressants (Litovitz and Troutman, 1983). Tricyclic antidepressants are basic drugs that bind primarily to the α₁-acidic glycoprotein (AGP) in humans (Ferry et al., 1986). Since AGP is found only about 1/40–1/100 the level of albumin in plasma (Sun et al., 2010; Otagiri, 2009), saturation of the drug-AGP binding is likely to occur during intoxication. Furthermore, unbound drug concentrations in plasma could be increased rapidly with non-linear kinetics during saturation of protein binding.

Imipramine is hydroxylated by CYP2D6 and becomes inactive, but a part of it is demethylated by CYP1A2 and CYP2C19 and becomes the active metabolite, desipramine, in humans (Gardiner and Begg, 2006). It has been reported that desipramine is produced by a similar metabolic pathway in mice (Shen and Yu, 2009). The serum Cmax of imipramine and desipramine after administration of a therapeutic dose of imipramine (20 mg/kg) to mice was comparable to that reported by Yoo et al. (1996) (Imipramine 30 mg/kg, i.p.), and the brain Cmax of imipramine was comparable to that reported by O’Brien et al. (2012) (5 mg/kg, i.v.). Therefore, within the therapeutic dose range, the effect of the dose was not significant, and the blood and brain Cmax were considered to be similar. The Cmax of imipramine and desipramine in serum at the toxic dose (100 mg/kg) was approximately equivalent to the rate of dose increase (Fig. 3, Table 2). The Cmax of imipramine and desipramine in the brain at the toxic dose was approximately 10 times that at the therapeutic dose, which was approximately twice the dose increase rate (Table 2). An increase in the AUC ratio of brain to serum of unmetabolized imipramine was observed in the toxic dose of imipramine, indicating that the rate of imipramine transfer into the brain increases at the toxic dose (Fig. 3, Table 2). Therefore, it is suggested that brain imipramine and desipramine are more likely to increase than in the blood during imipramine intoxication. A possible cause of the increase in Cmax and AUC in the brain after imipramine intoxication is the saturation of protein binding. Since imipramine has a high protein binding rate of 60–96% (DRUGBANK 5.1.8), the concentration of unbound drug in plasma is likely to increase, leading to an increase in brain transfer. We measured the total drug concentrations in serum in the present study but measuring the unbound drug concentrations will further clarify the brain-transfer properties. These issues are currently under investigation and will be presented in a separate publication.

Although the imipramine concentration in the brain increased substantially with the toxic dose, its elimination from the brain was not affected by dose escalation. Imipramine is known to be a substrate for efflux transporters such as P-glycoprotein (P-gp) and BCRP (O’Brien et al., 2012, 2013) and therefore, could be actively eliminated from the brain. Phenobarbital has also been reported to be a substrate of P-gp (Liu et al., 2018), but it showed prolonged elimination from the brain at toxic doses. The brain concentration of phenobarbital was significantly higher than that of imipramine (Fig. 1b, Fig. 3b); therefore, the former could have saturated the efflux transporter, and the latter may not have saturated it. In this regard, it is necessary to compare the transport capacity of phenobarbital and imipramine by efflux transporters with their concentrations in the brain in mice.

The increase rates of Cmax and AUC in serum with toxic doses of amoxapine were lower than the dose escalation rate, and serum Tmax and half-life were similar to those at the therapeutic dose (Fig. 4, Table 2). However, the ratio of increase in Cmax and AUC of the brain with an increasing dose of amoxapine was substantially higher than the ratio dose increase (Table 2). The serum concentration after 24 hr of the toxic dose of amoxapine was not significantly different from that of the therapeutic dose (Fig. 4). The brain concentration after 24 hr of amoxapine at the toxic dose was considerably higher than at the therapeutic dose and Cmax was similar to the therapeutic dose (Fig. 4). In addition, the AUC of amoxapine in the brain was markedly increased at the intoxication dose and the rate of increase was much higher than that of imipramine. These results indicate that the toxicokinetics of amoxapine in the blood and brain are clearly different, with brain concentrations being highly susceptible to increase during intoxication. One possible reason for this is the saturation of protein binding. The protein binding rate of amoxapine is as high as 90% in humans (DRUGBANK 5.1.8), and most of it binds to AGP, which is easily saturated during overdose, resulting in a significant increase in the brain transfer rate and AUC in the brain. To date, there have been no reports of amoxapine as a substrate for efflux transporters. Therefore, it is suggested that amoxapine is easily transferred to and slowly eliminated from the brain. The common feature in chemical structure and other characteristics of drugs that exhibit specific brain toxicokinetics similar to those of amoxapine are unknown at this time. Future studies using a diverse range of drugs are expected to reveal chemical structures that share similar toxicological profiles.

In the present study, BTU was used to evaluate the brain transferability, and the order of this during overdose was as follows: amoxapine, imipramine, phenobarbital, flunitrazepam. The proposed BTU value seems to be particularly effective in extracting drugs that are expected to be transferred to the brain, and its usefulness must be verified in future studies. In addition, the combination of BTU values and pharmacodynamic parameters of drugs could be applied in the PBPK model to predict CNS toxicity. From this perspective, future investigations including species differences are desirable in this area.

In conclusion, the pharmacokinetic parameters of all drugs used in the present study differed significantly from those of the therapeutic dose during intoxication. In addition, the present study showed that there are certain drugs whose brain concentrations specifically increase during intoxication, which is consistent with the symptoms observed in clinical settings. Therefore, we hope that the experimental methodology used and the BTU value obtained in this study will be verified with more drugs and applied to new psychotropic agents, helping in the potential treatment of acute intoxication in the future.

Conflict of interest

The authors declare that there is no conflict of interest.

REFERENCES

•  Breckenridge,  A. (1974): Anticonvulsant pharmacokinetics. Proc. R. Soc. Med., 67, 63-64.
• Budavari, B., Maryadele, J.O., Ann, S., Patricia, E.H. and Joanne, F.K. (1996): Amoxapine, imipramine, phenobarbital. In: The merck index (Susan, B., ed.), pp.97, 4958, 7381-7390, Merck Research Laboratories, California.
•  Chuang,  V.T. and  Otagiri,  M. (2006): Stereoselective binding of human serum albumin. Chirality, 18, 159-166.
•  Depoortere,  H.,  Zivkovic,  B.,  Lloyd,  K.G.,  Sanger,  D.J.,  Perrault,  G.,  Langer,  S.Z. and  Bartholini,  G. (1986): Zolpidem, a novel nonbenzodiazepine hypnotic. I. Neuropharmacological and behavioral effects. J. Pharmacol. Exp. Ther., 237, 649-658.
• DRUGBANK 5.1.8: https://go.drugbank.com/
• Evaluation and licensing division, pharmaceutical and food safety bureau, ministry of health, labour and welfare (2014): Iyakuhinnkaihatsu ni okeru seitaisiryoutyuuyakubutunoudobunnsekihou (riganndoketugouhou) no baride-syonn ni kannsuru gaidorainn. (in Japanese), https://www.pmda.go.jp/files/000206206.pdf
•  Ferry,  D.G.,  Caplan,  N.B. and  Cubeddu,  L.X. (1986): Interaction between antidepressants and alpha 1-adrenergic receptor antagonists on the binding to alpha 1-acid glycoprotein. J. Pharm. Sci., 75, 146-149.
•  Gardiner,  S.J. and  Begg,  E.J. (2006): Pharmacogenetics, drug-metabolizing enzymes, and clinical practice. Pharmacol. Rev., 58, 521-590.
•  Gheshlaghi,  F.,  Eizadi-Mood,  N.,  Emamikhah-Abarghooeii,  S. and  Arzani-Shamsabadi,  M. (2012): Evaluation of serum sodium changes in tricyclic antidepressants toxicity and its correlation with electrocardiography, serum pH, and toxicity severity. Adv. Biomed. Res., 1, 68.
•  Ghosh,  L.,  Arunachalam,  G.,  Murugesan,  T.,  Pal,  M. and  Saha,  B.P. (2002): Studies on the psychopharmacological activities of Rumex nepalensis Spreng. root extract in rats and mice. Phytomedicine, 9, 202-206.
•  Giwa,  A. and  Oey,  E. (2018): The return of an old nemesis: survival after severe tricyclic antidepressant toxicity, a case report. Toxicol. Rep., 5, 357-362.
•  Hartmanshenn,  C.,  Scherholz,  M. and  Androulakis,  I.P. (2016): Physiologically-based pharmacokinetic models: approaches for enabling personalized medicine. J. Pharmacokinet. Pharmacodyn., 43, 481-504.
•  Hataoka,  K.,  Kaizaki-Mitsumoto,  A. and  Numazawa,  S. (2017): Alpha-PVP induces the rewarding effect via activating dopaminergic neuron. J. Toxicol. Sci., 42, 539-543.
•  Hikiji,  W.,  Okumura,  Y.,  Matsumoto,  T.,  Tanifuji,  T.,  Suzuki,  H.,  Takeshima,  T. and  Fukunaga,  T. (2016): Identification of psychotropic drugs attributed to fatal overdose -a case- control study by data from the Tokyo medical examiner’s office and prescriptions. Seishin Shinkeigaku Zasshi, 118, 3-13.
•  Ho,  A.K. and  Ho,  C.C. (1979): Toxic interactions of ethanol with other central depressants: antagonism by naloxone to narcosis and lethality. Pharmacol. Biochem. Behav., 11, 111-114.
•  Jana,  S.,  Chakravarty,  C.,  Taraphder,  A. and  Ramasubban,  S. (2014): Successful use of sustained low efficiency dialysis in a case of severe phenobarbital poisoning. Indian J. Crit. Care Med., 18, 530-532.
•  Kaizaki,  A.,  Tanaka,  S. and  Numazawa,  S. (2014): New recreational drug 1-phenyl-2-(1-pyrrolidinyl)-1-pentanone (alpha-PVP) activates central nervous system via dopaminergic neuron. J. Toxicol. Sci., 39, 1-6.
•  King,  J.D.,  Kern,  M.H. and  Jaar,  B.G. (2019): Extracorporeal removal of poisons and toxins. Clin. J. Am. Soc. Nephrol., 14, 1408-1415.
•  Kulig,  K.,  Rumack,  B.H.,  Sullivan,  J.B. Jr.,  Brandt,  H.,  Spyker,  D.A.,  Duffy,  J.P. and  Shipe,  J.R. (1982): Amoxapine overdose. Coma and seizures without cardiotoxic effects. JAMA, 248, 1092-1094.
•  Litovitz,  T.L. and  Troutman,  W.G. (1983): Amoxapine overdose. Seizures and fatalities. JAMA, 250, 1069-1071.
•  Liu,  L.,  Miao,  M.,  Chen,  Y.,  Wang,  Z.,  Sun,  B. and  Liu,  X. (2018): Altered function and expression of ABC transporters at the blood-brain barrier and increased brain distribution of phenobarbital in acute liver failure mice. Front. Pharmacol., 9, 190.
•  López-Muñoz,  F.,  Ucha-Udabe,  R. and  Alamo,  C. (2005): The history of barbiturates a century after their clinical introduction. Neuropsychiatr. Dis. Treat., 1, 329-343.
•  Malatynska,  E.,  Rapp,  R.,  Harrawood,  D. and  Tunnicliff,  G. (2005): Submissive behavior in mice as a test for antidepressant drug activity. Pharmacol. Biochem. Behav., 82, 306-313.
•  Mathiaux,  F.,  Dulaurent,  S.,  Julia,  F. and  Gaulier,  J.M. (2014): Case report of ivabradine intoxication. J. Anal. Toxicol., 38, 231-232.
• Matsumoto, T., Usami, T., Funada, D., Murakami, M., Okita, K., Tanibuchi, Y., Yamamoto, T. and Yamaguchi, S. (2020): Zennkoku no seisinnkairyousisetu ni okeru yakubutukannrennseisinnsikkann no jittaityousa. (in Japanese.), National Institute of Mental Health, National Center of Neurology and Psychiatry (Matsumoto, T. ed.): https://www.ncnp.go.jp/nimh/yakubutsu/report/pdf/J_NMHS_2020.pdf
•  McKenzie,  M.S. and  McFarland,  B.H. (2007): Trends in antidepressant overdoses. Pharmacoepidemiol. Drug Saf., 16, 513-523.
•  Merigian,  K.S.,  Browning,  R.G. and  Leeper,  K.V. (1995): Successful treatment of amoxapine-induced refractory status epilepticus with propofol (diprivan). Acad. Emerg. Med., 2, 128-133.
•  Miyagawa,  H.,  Watanabe,  T.,  Ogura,  K.,  Sumigama,  S.,  Nakanowatari,  J. and  Ohgou,  T. (1985): Flunitrazepam no kyuusei oyobi akyuusei dokuseisikenn. Clin. Rep., 19, 1277-1295. (in Japanese.)
•  Müller,  D. and  Desel,  H. (2013): Common causes of poisoning: etiology, diagnosis and treatment. Dtsch. Arztebl. Int., 110, 690-699, quiz, 700.
•  Nguyen,  H.,  Kidron,  A.,  Ghildyal,  C.,  Veluri,  S.,  Nguyen,  N.,  Nguyen,  Q. and  Nguyen,  H. (2021): Novel presentation of cardiotoxicity and other complications in tricyclic antidepressant poisoning. Cureus, 13, e17181.
•  O’Brien,  F.E.,  Clarke,  G.,  Dinan,  T.G.,  Cryan,  J.F. and  Griffin,  B.T. (2013): Human P-glycoprotein differentially affects antidepressant drug transport: relevance to blood-brain barrier permeability. Int. J. Neuropsychopharmacol., 16, 2259-2272.
•  O’Brien,  F.E.,  Clarke,  G.,  Fitzgerald,  P.,  Dinan,  T.G.,  Griffin,  B.T. and  Cryan,  J.F. (2012): Inhibition of P-glycoprotein enhances transport of imipramine across the blood-brain barrier: microdialysis studies in conscious freely moving rats. Br. J. Pharmacol., 166, 1333-1343.
•  Otagiri,  M. (2009): [Study on binding of drug to serum protein]. Yakugaku Zasshi, 129, 413-425.
•  Shen,  H.-W. and  Yu,  A.-M. (2009): Difference in desipramine metabolic profile between wild-type and CYP2D6-humanized mice. Drug Metab. Lett., 3, 234-241.
•  Shen,  Q.,  Wang,  L.,  Zhou,  H.,  Jiang,  H.D.,  Yu,  L.-S. and  Zeng,  S. (2013): Stereoselective binding of chiral drugs to plasma proteins. Acta Pharmacol. Sin., 34, 998-1006.
•  Sun,  D.-L.,  Huang,  S.-D.,  Wu,  P.-S.,  Li,  J.,  Ye,  Y.-J. and  Jiang,  H.-D. (2010): Stereoselective protein binding of tetrahydropalmatine enantiomers in human plasma, HSA, and AGP, but not in rat plasma. Chirality, 22, 618-623.
• Suzuki, O., Ono, Y., Suzaki, S. and Kikura-Hanajiri, R. ed. (2014): Yakudokubutugaku jyouhou indekkusu. (in Japanese.), pp.278-279, Japan medical journal, Tokyo.
•  Wróbel,  A.,  Serefko,  A.,  Wlaź,  P. and  Poleszak,  E. (2015): The effect of imipramine, ketamine, and zinc in the mouse model of depression. Metab. Brain Dis., 30, 1379-1386.
•  Yoo,  S.D.,  Holladay,  J.W.,  Fincher,  T.K.,  Baumann,  H. and  Dewey,  M.J. (1996): Altered disposition and antidepressant activity of imipramine in transgenic mice with elevated alpha-1-acid glycoprotein. J. Pharmacol. Exp. Ther., 276, 918-922.
•  Zhao,  J.,  Jung,  Y.H.,  Jang,  C.G.,  Chun,  K.H.,  Kwon,  S.W. and  Lee,  J. (2015): Metabolomic identification of biochemical changes induced by fluoxetine and imipramine in a chronic mild stress mouse model of depression. Sci. Rep., 5, 8890.