Maternal-Fetal Disposition of Domoic Acid Following Repeated Oral Dosing during Pregnancy in Nonhuman Primate

Abstract

Domoic acid (DA) is a marine algal toxin that causes acute and chronic neurotoxicity in animals and humans. Prenatal exposure to DA has been associated with neuronal damage and cognitive and behavioral deficits in juvenile California sea lions, cynomolgus monkeys and rodents. Yet, the toxicokinetics (TK) of DA during pregnancy and the maternal-fetal disposition of DA have not been fully elucidated. In this study, we investigated the TK before, during and after pregnancy and the maternal-fetal disposition of DA in 22 cynomolgus monkeys following daily oral doses of 0.075 or 0.15 mg/kg/day of DA. The AUC_0-τ of DA was not changed while the renal clearance of DA was increased by 30 – 90% during and after pregnancy when compared to the pre-pregnancy values. DA was detected in the infant plasma and the amniotic fluid at delivery. The infant plasma concentrations correlated positively with both the maternal plasma and the amniotic fluid concentrations. The paired infant-to-maternal plasma DA concentration ratios ranged from 0.3 to 0.6 and increased as a function of time which suggests placental efflux and longer apparent fetal half-life than the maternal half-life. The paired amniotic fluid-to-infant plasma DA concentration ratios ranged from 4.5 to 7.5 which indicates significant accumulation of DA in the amniotic fluid. A maternal-fetal TK model was developed to explore the processes that give the observed maternal-fetal disposition of DA. The final model suggests that placental transport and recirculation of DA between the fetus and amniotic fluid are major determining factors of the maternal-fetal TK of DA.

INTRODUCTION

Domoic acid (DA) is a naturally occurring neurotoxin produced by species of the marine phytoplankton, Pseudo-nitzschia.¹ It is a potent non-NMDA glutamate receptor agonist² which potentiates the NMDA receptor^3,4 to cause severe excitotoxicity in humans and animals.^5–8 Over the past decade, an increasing number of DA-producing species of Pseudo-nitzschia have been identified around the globe; making this a prominent global problem.⁹ To protect humans from DA toxicity, regulatory agencies have established regulatory limit of 20 μg of DA per gram of shellfish¹⁰ and proposed a tolerable daily intake (acute TDI) level of 0.075 – 0.1 mg/kg/day.^11,12 While this limit has successfully protected humans from acute DA toxicity, emerging evidence shows that chronic asymptomatic exposure to DA may cause detrimental health effects in humans and animals.^8,13–22 Chronic low-level DA exposure from the consumption of contaminated shellfish in humans has recently been associated with adverse changes in memory that impacts daily living skills.^13,15 DA caused memory impairments in adults have also been documented in preclinical rodent models and wild populations of California sea lions.^17,21 In animals, the effect of DA on the mammalian brain includes neuronal death, increased oxidative stress, alter glutamatergic transmission and neuronal connectivity in the hippocampus, and altered gene transcription in the central nervous system.^{14,17,19,21,22}

There are compelling data from animal studies that the fetus may be particularly sensitive to the effects of DA.^1,23 Laboratory studies in rodents have shown that DA causes neuronal damage in fetuses and impairments in the behavioral development of young animals following single or repeated dosing.^24–28 In California sea lions, DA exposure during pregnancy has been linked to reproductive failure, fetal death, and significant neurological injuries in exposed offspring.^8,29,30 These observations have raised a concern that low-level maternal DA exposure in humans could have significant consequences on fetal brain development²³ similar to other environmental chemicals such as methylmercury, polychlorinated biphenyls (PCBs) or lead.^31,32 To address the concern of reproductive and fetal toxicity following chronic oral DA exposure, our group initiated a long-term neurotoxicity study in cynomolgus monkeys (Macaca fascicularis),³³ in which adult female monkeys received daily DA doses of the proposed TDI (0.075 mg/kg/day) or two-times the proposed TDI (0.15 mg/kg/day) or vehicle control before and throughout pregnancy. While DA did not cause reproductive toxicity, our studies demonstrated that chronic subacute DA exposure in adult female monkeys resulted in a significant increase in observations of intentional-tremor and altered brain morphometry.^33,34 Infants born to the female monkeys in the 0.15 mg/kg/day DA group scored poorly on some aspects of a standardized memory test, suggesting that chronic fetal exposure to DA may impact developing cognitive processes.³⁵

Our group has previously reported the toxicokinetics (TK) of DA following single intravenous and oral doses in non-pregnant female cynomolgus monkeys³⁶ but the TK of DA following repeated oral doses and TK changes during pregnancy have not been reported. We have shown that in monkeys, DA follows flip-flop kinetics after a single oral dose and consequently has a terminal half-life of about 10 hours³⁶. Based on this half-life, little accumulation (<1.2-fold) of DA is expected following repeated daily oral doses, and plasma concentrations of DA after chronic dosing were predicted to be similar to those after a single dose. DA is mainly eliminated by the kidney with about 40 – 70% of intravenously dosed DA excreted unchanged in the urine in cynomolgus monkeys^36,37. It is well recognized that the renal clearance of exogenous and endogenous compounds increases during pregnancy in humans due to increased renal blood flow and glomerular filtration³⁸. Hence, renal clearance of DA may increase during pregnancy which may lead to an increase in the systemic clearance and a decrease in the area under the plasma concentration-time curve (AUC). Despite the polarity, ionization at physiological pH, and poor permeability of DA, it has been shown to cross the placenta and distribute to the fetus in California sea lions and rodents.^29,39–41 The cognitive effects we observed in this study also support these findings. One study in rats showed that the apparent half-life of DA in fetal rats was twice of the dams’ and the fetal-to-maternal plasma AUC ratio was 0.3 following a single intravenous dose to pregnant rats.⁴¹ The residence time of DA in amniotic fluid was longer than in fetal plasma⁴¹ and re-exposure through fetus ingestion of amniotic fluid has been attributed to the long apparent fetal half-life of DA.^39,41

The goal of this study was to describe the maternal-fetal TK of DA following repeated oral doses of DA around the TDI during pregnancy in a nonhuman primate model. We herein report the maternal TK of DA before, during, and after pregnancy and the fetal disposition of DA at delivery, and provide a kinetic model to explore the mechanisms of maternal-fetal distribution kinetics of DA. Based on the known TK of DA and physiological changes during pregnancy in the nonhuman primate, we hypothesized that the renal clearance of DA would increase and the AUC of DA would decrease during pregnancy and that DA would accumulate in the fetus following repeated DA doses. This is the first report to describe maternal-fetal disposition of DA following chronic oral exposure, a relevant route of exposure in humans. The findings provide new insights into the health risks associated with exposure to this increasingly prevalent neurotoxin, particularly in the sensitive pregnant population.

MATERIALS AND METHODS

Chemicals and Reagents

DA powder purified from biological sources was purchased from BioVectra (Charlottetown, PE, Canada) to prepare dosing solutions. Purity of this DA powder was measured by HPLC-MS/MS as described below and the powder was found to be 94% pure. A certified calibration standard solution of DA (332 μM) in acetonitrile/water (v/v 1:19) was purchased from National Research Council Canada (Halifax, NS, Canada). HPLC solvents including Optima LC/MS grade water, methanol, acetonitrile, and formic acid were purchased from Fisher Scientific (Pittsburgh, PA). The analytical internal standard, tetrahydrodomoic acid (THDA), was synthesized as previously described ⁴². Frozen treatment-naïve monkey plasma and urine were obtained from the Washington National Primate Research Center (WaNPRC) at the University of Washington (Seattle, WA).

Animal Study Protocol

All the monkeys in this toxicokinetic (TK) study were part of a double-blinded randomized vehicle controlled long-term toxicology study of the effects of DA on reproduction and fetal neurodevelopment^33–35. The animal procedures followed the Animal Welfare Act and the Guide for Care and Use of Laboratory Animals of the National Research Council. All protocols were previously described and approved by the University of Washington Institutional Animal Care and Use Committee to meet the highest standards of ethical conduct and compassionate use of animals in research. Briefly, thirty-two DA-naïve adult female cynomolgus macaques (Macaca fascicularis) were randomized to either the vehicle-dosed control group (n=10), or two groups receiving daily oral doses of 0.075 mg/kg/day (TDI) (n=11) or 0.15 mg/kg/day (2-times TDI) (n=11) (Figure 1). The average age and weight of the monkeys at randomization were 7 years (range: 5.5 – 11 years) and 3.5 kg (range: 2.8 – 4.2 kg), respectively. The monkeys were monitored daily to assess their health and behavior throughout the study. The health and behavior assessments were initiated immediately after randomization and dosing was initiated two months after randomization to allow for time to capture baseline values. Individual dosing solutions were prepared weekly according to the weight of each animal. Powder of DA was dissolved in tap water with 5% (w/v) sucrose, sonicated for 10 minutes, and the individual dosing solution was filter-sterilized into a glass vial and stored at 4°C until dosing. All dosing solutions were analyzed by LC-MS/MS to confirm DA concentration. After two months of dosing, the female monkeys were paired with DA-naïve male monkeys for breeding. All monkeys, except for two in the 0.075 mg/kg/day group, conceived between one to seven months after initiation of breeding. After confirmation of pregnancy, the dosing solutions were prepared according to the last recorded pre-pregnancy weight³³. The pregnant monkeys delivered either naturally or through caesarean sections after 21 to 25 weeks of gestation. Health and behavior monitoring of the infants was initiated after birth. DA dosing continued for the adult female monkeys according to their pre-pregnancy weight until necropsy.

Figure 1. Toxicokinetic study design.

Five TK studies were performed over the course of the study (Figure 1): on the first day of dosing (Day1), on the 56^th day of dosing (Day56), on a day during gestational week eight to ten (GW8–10), on a day during gestational week 16 to 18 (GW16–18), and on a day during two to four weeks post-partum (PP2–4) for each animal. Only Day1 and Day56 TK studies were performed for monkeys who did not get pregnant. The PP2–4 TK study was not performed in three monkeys, two in the 0.075 mg/kg/day group and one in the 0.15 mg/kg/day group. On each TK study day, blood samples were collected from the saphenous vein into sodium heparin tubes at pre-dose and 1, 2, 4, 6, 8, 10, 12, and 24 hours post-dose. Extensive behavioral training with positive reinforcement allowed the blood samples to be collected without anesthesia. Collected blood samples were centrifuged at 3,000 g for 15 minutes to isolate plasma within an hour of collection. Plasma samples were aliquoted and stored at −20°C until analysis. At each TK day, all the urine was collected at designated time points over a dosing interval (24 hours) from a metabolic pan placed at the bottom of the cage. Volume of the urine collected was recorded and urine samples were aliquoted and stored at −20°C until analysis. At delivery, maternal blood and infant blood were sampled and processed as described above and amniotic fluid was sampled and stored at −20°C until analysis. Maternal blood was not sampled in one monkey in the 0.15 mg/kg/day group and infant blood was not sampled in one infant in the 0.075 mg/kg/day group and two infants in the 0.15 mg/kg/day group. Amniotic fluid was sampled following c-section delivery in two monkeys in the 0.075 mg/kg/day and five monkeys in the 0.15 mg/k/day group.

HPLC-MS/MS Analysis

DA concentrations in plasma, urine, and amniotic fluid were measured using a published validated HPLC-MS/MS method⁴². Briefly, plasma, urine, and amniotic fluid samples were thawed at room temperature on the day of analysis. For each plasma and amniotic fluid sample, 120 μL of methanol containing 5 nM THDA (internal standard) was added to 60 μL of sample. The samples were vortexed briefly, centrifuged at 16,100 g for an hour at room temperature, and supernatant was collected. For each urine sample, 490 μL of water containing 10 nM THDA (internal standard) was added to 10 μL of sample. The samples were vortexed briefly, centrifuged at 16,100 g for 15 minutes at room temperature, and supernatant was collected. Calibration standards were prepared in DA-naïve plasma and urine with concentrations ranging between 0.16 and 15.6 ng/mL for plasma and 7.8 and 1000 ng/mL for urine. Amniotic fluid concentrations were measured using the plasma calibration curve. For each urine sample with preliminary measured concentration >1000 ng/mL, 10 μL of sample was first diluted with 90 μL of water before processing as above. Supernatants were analyzed using a Shimadzu UFLC XR DGU-20A5 (Kyoto, Japan) equipped with a Phenomenex Synergi Hydro-RP 100 Å (2.5 μm, 50 × 2 mm²) LC column and a guard cartridge (2 × 2.1 mm², sub 2 μm) (Torrange, CA) linked to a Sciex 6500 QTRAP system (Foster City, CA). The limit of detection (LOD) and the lower limit of quantitation (LLOQ) in plasma were 0.16 ng/mL and 0.31 ng/mL, respectively⁴². Plasma concentrations below LLOQ but above LOD were assigned a value of 0.23 ng/mL (mid-point between LOD and LLOQ) and plasma concentrations below LOD were assigned as 0 ng/mL. The LLOQ in urine was 7.8 ng/mL⁴². All urine concentrations were above the LLOQ.

Creatinine Analysis

One plasma and one urine sample from each animal on each TK study were analyzed for creatinine concentrations. The measurement was done by clinical laboratory tests at the Department of Laboratory Medicine at the University of Washington Medical Center. Creatinine concentrations were measured in all but two monkeys at GW8–10 and PP2–4 in the 0.075 mg/kg/day group, one monkey at GW16–18 in the 0.075 mg/kg/day group, and one monkey at PP2–4 in the 0.15 mg/kg/day group.

Toxicokinetic Analysis

TK parameters including steady-state area under the plasma concentration time curve over a dosing interval (AUC_0-τ) and oral clearance (CL/F) at each TK study for each monkey were calculated by noncompartmental analysis (NCA) using Phoenix WinNonlin (St Louis, MO). AUC_0-τ was calculated by the linear up log down method and CL/F was calculated using the following equation:

$$CL/F=\frac{D_{\tau }}{AU{C}_{0-\tau }}$$

where τ is a dosing interval (24 hours) and D_τ is the amount of daily dose received. The rest of the TK parameters including renal clearance (CL_R), creatinine clearance (CL_creatinine), ratio between CL_R and CL_creatinine, and fraction of the dose excreted in urine unchanged (f_e) at each TK study for each animal were calculated using Microsoft Excel 2013 (Redmond, WA). Renal clearance (CL_R) was calculated using the following equation:

$$C{L}_R=\frac{{\mathrm{A}}_{e,0-\tau }}{AU{C}_{0-\tau }}$$

where A_e,0-τ is the total amount of DA excreted in urine unchanged over a dosing interval and was calculated by summation of urine DA concentrations multiplied by urine volumes for cumulative urine samples collected. Fraction of the oral dose excreted unchanged in urine (f_e) was calculated using the following equation:

$$f_e(\%)=\frac{A_{e,0-\tau }}{D_{\tau }}\times 100\%$$

Creatinine clearance (CL_creatinine) was calculated using the following equation:

$$C{L}_{creatinine}=\frac{\frac{A_{e,creatinine}}{\mathrm{\Delta }t}}{C_{creatinine}}$$

where A_e,creatinine is the amount of creatinine excreted in urine over a collection period calculated by multiplying the urine creatinine concentration by the urine volume, ∆t is the duration of the urine collection period, and C_creatinine is the plasma creatinine concentration. Renal clearance to creatinine clearance (CL_R/CL_creatinine) ratio was calculated by dividing CL_R by CL_creatinine. Infant-to-maternal plasma concentration ratio was calculated by dividing the infant plasma concentration by the maternal plasma concentration at the time of delivery when both concentrations were above the LLOQ. The geometric mean and the 95% confidence interval for each TK parameter were calculated using GraphPad Prism version 5 (San Diego, CA).

Power Analysis

Intra-individual variability of the plasma AUC before and during pregnancy was assessed using the biweekly 5-hour-post-dose plasma concentrations sampled prior to and during pregnancy, respectively, from each monkey³³. A power analysis using the averaged intra-individual variability before and during pregnancy showed that 11 animals per group would allow detection of 70% increase or 40% decrease in AUC_0-τ of DA during pregnancy with 80% power (β = 0.2) and significance level of α = 0.017.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism version 5 (San Diego, CA). Differences in the natural log-transformed AUC_0-τ, CL/F, CL_R, and CL_creatinine during pregnancy (GW8–10 and GW16–18), and after pregnancy (PP2–4) in comparison to before pregnancy (Day56) were tested for each dose group by paired t-tests with p-value <0.017 (Bonferroni correction for multiple comparisons, α = 0.05/3) considered significant. The correlations between (1) maternal plasma and infant plasma concentration at delivery, (2) infant plasma and amniotic fluid concentration at delivery, (3) maternal plasma concentration at delivery and infant-to-maternal plasma concentration ratio, and (4) infant plasma concentration at delivery and infant-to-maternal plasma concentration ratio, were tested by simple linear regression. The deviation of the slope of the linear regression line from zero was tested using t-test with a p-value ≤ 0.05 considered significant.

Maternal-Fetal Toxicokinetic Model of DA

Model structure

A maternal-fetal TK model of DA was developed using SimBiology R2018b (Natick, MA) to simulate late pregnancy maternal-fetal DA disposition. This model is structured to include compartments of the maternal whole-body, fetal whole-body, and amniotic fluid. Each compartment has a fixed volume of distribution assigned for a 4 kg cynomolgus monkey. The maternal volume of distribution was assigned to be 0.6 L according to the observed volume of distribution of DA in cynomolgus monkeys following an IV dose (0.15 L/kg bodyweight).³⁶ The fetal volume of distribution was assigned to be 0.06 L according to the mean birthweight of the infant monkeys observed in this study (0.4 kg) and the observed volume of distribution of DA in adult cynomolgus monkeys (0.15 L/kg bodyweight) assuming fetal DA distribution is similar to adult monkeys.

Absorption

The steady-state maternal and fetal DA disposition following repeated oral DA doses during pregnancy was simulated by applying fifty daily (every 24 hour) DA doses of 0.075 mg/kg/day to the maternal compartment. DA absorption was assumed to be first-order defined by an oral absorption rate constant (k_a). The initial k_a was assigned as 0.07 h⁻¹ as previously reported in adult non-pregnant monkeys.³⁶ The bioavailability of DA was set at 7.5% as previously reported in the same study.³⁶

Distribution

DA distribution from the maternal compartment to the fetal compartment (k_mf) and vice versa (k_fm) was modeled according to first-order kinetics. The initial distribution rate constant k_mf was calculated by dividing the transplacental clearance from mother to fetus (CL_mf) with the maternal volume of distribution (0.6 L). The initial CL_mf was assigned as 0.1 L/h, a value estimated by multiplying the reported Caco2 permeability⁴³ with the extrapolated placental microvillus surface area based on human physiology^44,45. The initial distribution rate k_fm was calculated by dividing the transplacental clearance from fetus to mother (CL_fm) with the fetal volume of distribution (0.06 L). The initial CL_fm was assigned to be the same as CL_mf. (0.1 L/h). DA distribution from the fetal compartment to the amniotic fluid compartment (through fetal urination) was assumed to be according to first order kinetics with a first order rate constant (k_e, amniotic fluid). The k_{e, amniotic fluid} was calculated by dividing fetal renal clearance (CL_{e, fetal}) by the fetal volume of distribution (0.06 L). The initial CL_{e, fetal} was assigned as 0.015 L/h, a value calculated assuming fetal renal clearance is 2.5%^46–48 of the maternal renal clearance observed in this study. DA distribution from the amniotic fluid compartment to the fetal compartment (through fetal swallowing of amniotic fluid) was modeled according to first order kinetics with a first order rate constant (k_{a, amniotic fluid}). The initial k_{a, amniotic fluid} was assigned to be the same as k_a (0.07 h⁻¹).

Elimination

DA elimination from the maternal compartment was assumed to be first order with a first order elimination rate constant (k_e). The initial k_e was calculated by dividing the maternal systemic clearance (CL_{e, maternal}) by the maternal volume of distribution (0.6 L). The initial CL_{e, maternal} was assigned as 1.66 L/h, which was calculated by multiplying the observed oral clearance at GW16–18 in the 0.075 mg/kg/day group with the 7.5% bioavailability reported previously.³⁶

Simulations

Steady-state (after the 50^th dose) maternal-fetal disposition was first simulated using the initial parameters described above. Local sensitivity analyses were then conducted to explore how specific TK parameters (k_a, k_a, _{amniotic fluid}, CL_e, _maternal, CL_e, _fetal, CL_mf, and CL_fm) impact simulation results of maternal fetal disposition of DA. Sensitivity analyses were conducted by altering the value of each of the above parameters individually within the range of 10% to 10-fold of the initial values. Based on the sensitivity analyses, specific parameters for maternal-fetal disposition (CL_mf, CL_fm, and CL_{e, fetal}) were optimized simultaneously to capture the time course of the observed infant-to-maternal plasma concentration ratio. All other parameters were kept as described above for the model. The simulated maternal and fetal steady-state AUC_0-τ (after the 50^th dose, from 1200 – 1224 hour) and terminal half-lives (t_1/2, after distribution equilibrium is reached) were analyzed using Phoenix WinNonlin (St Louis, MO). The fetal-to-maternal AUC_0-τ and plasma concentration ratios were calculated using Microsoft Excel 2013 (Redmond, WA). The maternal DA concentrations observed at GW16–18 were compared to the simulated maternal DA concentrations using GraphPad Prism version 8 (San Diego, CA). Similarly, observed maternal and infant concentrations at delivery were compared to the simulated maternal and fetal concentrations at term using GraphPad Prism version 8 (San Diego, CA).

RESULTS

Toxicokinetics (TK) of DA following Daily Oral Doses in Adult Female Monkeys Before, During, and After Pregnancy

TK of DA following daily oral doses in adult female monkeys before, during, and after pregnancy is reported in Table 1 and Figure 2. The AUC_0-τ on Day56 following a dose of 0.075 mg/kg/day was 20 h*ng/mL (95% CI: 16 – 25 h*ng/mL), when the dose was increased by 2-fold to 0.15 mg/kg/day, the AUC_0-τ increased by 3.3-fold to 65 h*ng/mL (95% CI: 48 – 87 h*ng/mL). This more than dose-proportional increase was observed throughout the study, with the AUC_0-τ in the 0.15 mg/kg/day group being 2.5 to 3.3-fold greater than that of the 0.075 mg/kg/day group. The weight-normalized CL/F of DA was 22 to 41% lower in the 0.15 mg/kg/day group in comparison to the 0.075 mg/kg/day group throughout the study. The CL_R of DA, CL_creatinine, and the CL_R/CL_creatinine ratio were not different between the two dose groups on any study day. However, the CL_creatinine of the control group appeared to be consistently lower than the CL_creatinine in the 0.15 mg/kg/day group throughout the study (Table 1). Both the CL_R of DA and the CL_creatinine showed a trend to increase during and after pregnancy in both dose groups when compared to the pre-pregnancy values (Figure 2e–h), but these increases did not translate to a change in AUC_0-τ and CL/F of DA during and after pregnancy (Figure 2a–d). The CL_R/CL_creatinine was <1 throughout the study and was not altered by pregnancy.

Table 1. Toxicokinetic (TK) parameters of DA measured following daily oral doses of 0.075 mg/kg/day and 0.15 mg/kg/day before, during, and after pregnancy in adult female monkeys.

The creatinine clearance values measured in vehicle control animals are also reported for the study days. TK parameters are reported as geometric means (95% confidence interval). Pregnancy (GW8–10 and GW16–18) and post-partum (PP2–4) parameters were compared to pre-pregnancy (Day56) parameters and a p-value of 0.017 was considered significant (Significance level adjusted for multiple comparisons). Values that are significantly different from values on Day56 are bolded.

		n=	aAUC0-τ (ng/mL*h)	aCL/F (mL/min/kg)	aCLR (mL/min)	bCLcreatinine (mL/min)	c$\frac{C{L}_R}{C{L}_{creatinine}}$	fe (%)
0.075 mg/kg/day	Day1	11	19 (14 – 27)	--	5.2 (2.8 – 9.9)	5.9 (4.0 – 8.7)	0.8 (0.4 – 1.7)	--
	Day56	11	20 (16 – 25)	61 (49 – 76)	6.2 (5.3 – 7.4)	8.1 (4.9 – 13)	0.8 (0.4 – 1.5)	2.7 (2.1 – 3.4)
	GW8–10	9	19 (13 – 28) p = 0.54	67 (44 – 100) p = 0.54	7.7 (4.6 – 13) p = 0.34	15 (11 – 19) p = 0.06	0.7 (0.4 – 1.1) p = 0.93	3.3 (1.9 – 5.9) p = 0.40
	GW16–18	9	14 (8.5 – 22) p = 0.04	92 (58 – 146) p = 0.04	11 (7.3 – 17) p = 0.014	16 (12 – 22) p = 0.07	0.6 (0.5 – 0.8) p = 0.94	3.2 (2.5 – 4.1) p = 0.19
	PP2–4	7	17 (8.8 – 34) p = 0.41	72 (37 – 142) p = 0.41	10 (5.1 – 20) p = 0.11	12 (10 – 14) p = 0.15	0.9 (0.4 – 1.9) p = 0.60	3.5 (1.8 – 6.9) p = 0.20
0.15 mg/kg/day	Day1	11	44 (34 – 57)	--	6.4 (4.9 – 8.2)	12 (7.5 – 20)	0.5 (0.3 – 0.9)	--
	Day56	11	65 (48 – 87)	39 (29 – 52)	6.2 (4.7 – 8.2)	11 (7.9 – 14)	0.6 (0.5 – 0.8)	4.2 (3.0 – 5.9)
	GW8–10	11	48 (38 – 62) p = 0.06	52 (40 – 67) p = 0.06	10 (8.2 – 13) p <0.001	17 (14 – 20) p = 0.003	0.6 (0.5 – 0.8) p = 0.73	5.1 (3.9 – 6.8) p = 0.30
	GW16–18	11	46 (33 – 65) p = 0.04	54 (38 – 76) p = 0.04	9.4 (6.6 – 13) p = 0.018	17 (15 – 20) p = 0.006	0.5 (0.4 – 0.8) p = 0.54	4.5 (3.2 – 6.3) p = 0.75
	PP2–4	10	51 (39 – 68) p = 0.13	49 (37 – 65) p = 0.13	9.1 (6.9 – 12) p = 0.002	17 (14 – 21) p = 0.002	0.5 (0.4 – 0.6) p = 0.56	4.9 (3.8 – 6.2) p = 0.13
Control	Day1	10	--	--	--	5.7 (4.0 – 8.0)	--	--
	Day56	10	--	--	--	6.5 (5.3 – 8.0)	--	--
	GW8–10	8	--	--	--	12 (8.7 – 16) p = 0.005	--	--
	GW16–18	8	--	--	--	11 (8.2 – 16) p = 0.002	--	--
	PP2–4	7	--	--	--	11 (7.8 – 14) p = 0.006	--	--

^a.More than 50% of plasma samples collected from one monkey in the 0.075 mg/kg/day group were below BLQ (0.3 ng/mL), hence, AUC_0-τ, CL/F, and CL_R were not calculated for that monkey for the whole study (i.e. Day1, n=10; Day56, n=10; GW8–10, n=8; GW16–18, n=8; PP2–4, n=6)

^b.For the 0.075 mg/kg/day group, CL_creatinine was measured in n=7 monkeys at GW8–10 and n=8 monkeys at GW16–18

^c.For the 0.075 mg/kg/day group CL_R/CL_creatinine ratio was calculated only when both values were availbale (i.e. Day1, n=10; Day56, n=10; GW8–10, n=6; GW16–18, n=7; PP2–4, n=6)

Figure 2. Toxicokinetic parameters of DA measured in individual animals before (Day 56), during (GW 8–10 and GW 16–18) and after (PP 2–4) pregnancy.

Panel (a, b) AUC_0-τ, panel (c, d) oral clearance (CL/F), panel (e, f) renal clearance (CL_R), and panel (g, h) creatinine clearance (CL_creatinine). Each open circle represents the observed value in an individual monkey and the values observed in each monkey over the course of the study are linked with lines. (* p <0.017, ** p <0.003, *** p <0.001). Black symbols are for the 0.075 mg/kg group and red symbols are for the 0.15 mg/kg group.

Maternal Plasma, Infant Plasma and Amniotic Fluid DA Concentrations at Delivery

The DA concentrations in maternal plasma, infant plasma, and amniotic fluid were measured at delivery to assess maternal-fetal disposition of DA (Figure 3a–b). The infant plasma concentration correlated positively with both the maternal plasma (Figure 3c) and the amniotic fluid concentrations (Figure 3d), and the infant plasma DA concentrations were consistently lower than the maternal plasma and amniotic fluid DA concentrations. The infant-to-maternal plasma concentration ratios were 0.3 (range: 0.2 – 0.5) in the 0.075 mg/kg/day group and 0.6 (range: 0.3 – 1.0) in the 0.15 mg/kg/day group while the amniotic fluid-to-infant plasma concentration ratios were 4.5 (range: 3.9 – 5.1) in the 0.075 mg/kg/day group and 7.5 (range: 2.7 – 13) in the 0.15 mg/kg/day group (Figure 3e–f). Correlations between the infant-to-maternal plasma concentration ratios and the maternal and infant plasma concentrations were tested to explore whether any potential nonlinearity in the maternal-fetal distribution could be observed but no correlation was detected (Figure 3g–h). Instead, the infant-to-maternal plasma concentration ratio appeared to vary with time after dosing of DA, with lower ratios observed in monkeys who delivered shortly (<12 hours) after dosing and higher ratios observed in monkeys who delivered later (>12 hours) after dosing.

Figure 3. Maternal-fetal disposition of DA at delivery.

Paired maternal plasma, infant plasma, and amniotic fluid DA concentrations of (a) 0.075 mg/kg/day group and (b) 0.15 mg/kg/day group. Open circles represent concentrations measured in each individual animal, and data for each maternal-infant pair are linked with lines. Data presented as (*) represents samples where DA was detected but concentrations were less than LOQ. Mean and range of concentrations in the maternal, infant and amniotic fluid samples are listed on the graph above corresponding datapoints. Panel (c) shows the correlation between maternal and infant plasma DA concentrations and panel (d) shows the correlation between amniotic fluid and infant plasma DA concentrations. Open black circles and red circles represent data from the 0.075 mg/kg/day and 0.15 mg/kg/day groups, respectively. Panel (e) shows the infant-to-maternal plasma DA concentration ratios and panel (f) shows the amniotic fluid-to-infant plasma DA concentration ratios with the mean values and range of ratios listed next to each dataset. The correlation between infant-to-maternal plasma concentration ratio and maternal plasma DA concentration is shown in panel (g) while panel (h) shows the correlation between the infant-to-maternal plasma concentration ratio and infant plasma DA concentration at delivery. The mean values and ranges of maternal and fetal DA concentrations at delivery indicated with # were reported previously³³.

DA Maternal-Fetal Toxicokinetic Modeling and Simulations

To explore the potential mechanisms controlling maternal-fetal disposition of DA, a TK model was developed to simulate DA maternal-fetal disposition at the time of delivery (Figure 4). The initial parameter values were assigned either as values observed in this and previous studies or estimated based on the known physiology of pregnant monkeys at late gestation as described in materials and methods. The steady-state maternal-fetal disposition of DA (after the 50^th dose) was first simulated using the initial parameter values. These simulations showed an increase in fetal-to-maternal plasma concentration ratio with time after dosing, but the simulated fetal-to-maternal plasma concentration ratio over-predicted the observed ratios at delivery (Figure 5a). To address this over-prediction, local sensitivity analyses were conducted to define the impact of each individual model parameter on the simulated maternal-fetal disposition of DA (Table 2). Local sensitivity analyses were conducted by changing one parameter at a time while keeping the rest of the parameters at their initial values. Decreasing maternal absorption rate (k_a) increased both the fetal and maternal half-lives of DA (Table 2) due to the flip-flop kinetics (k_e > k_a, elimination is absorption rate limited) of DA, but decreasing k_a had a minimal effect on the fetal-to-maternal plasma concentration ratio. Increasing k_a had no impact on the fetal and maternal half-lives of DA once distribution equilibrium was achieved (Table 2) as under this scenario the reabsorption from the amniotic fluid becomes rate-limiting. Increasing k_a shortened the initial half-life and also increased the fetal-to-maternal plasma concentration ratio.

Figure 4. Structure and parameters of the final DA maternal-fetal TK model.

The two CL_{e, maternal} values used in the model were assigned based on the observed oral clearance values in the two different DA dose groups at GW16–18.

Figure 5. Simulated steady-state (after the 50^th dose) maternal-fetal disposition of DA overlaid with observed maternal and infant plasma concentrations at delivery.

(a) Simulated fetal-to-maternal plasma concentration ratios over time under the initial parameter values (k_a = 0.07 h⁻¹, k_e = 2.76 h⁻¹ (CL_e,maternal = 1.66 L/h), k_mf = 0.17 h⁻¹ (CL_mf = 0.1 L/h), k_fm = 1.7 h⁻¹ (CL_fm = 0.1 L/h), k_{e, amniotic fluid} = 0.25 h⁻¹ (CL_{e, fetal} = 0.015 L/h) and k_{a, amniotic fluid} = 0.07 h⁻¹), and with decreased k_mf, increased k_fm, or increased k_{e,amniotic fluid} as described for the sensitivity analyses. The simulations are overlaid with observed infant-to-maternal ratios from individual monkeys at delivery. (b) Simulated (using the final model) maternal plasma concentration versus time profiles overlaid with observed maternal concentration-time profiles from individual monkeys at GW16–18. (c) Simulated maternal and fetal plasma concentration profiles overlaid with observed maternal and infant concentrations at delivery. (d) Simulated fetal-to-maternal plasma concentration ratio over time using the final kinetic model (k_mf = 0.01 h⁻¹ (CL_mf = 0.006 L/h), k_fm = 0.3 h⁻¹ (CL_fm = 0.018 L/h), k_{e, amniotic fluid} = 1 h⁻¹ (CL_{e, fetal} = 0.06 L/h), and initial values for k_a, k_e, and k_{e, amniotic fluid})

Table 2. Sensitivity analyses to illustrate the effect of altering individual TK parameters on steady state (after the 50^th dose) maternal-fetal DA disposition.

Simulations were performed by changing only one parameter while keeping the other parameters at their initial values. Initial values were as described in materials and methods: k_a = 0.07 h⁻¹, k_e = 2.76 h⁻¹ (CL_e,maternal = 1.66 L/h), k_mf = 0.17 h⁻¹ (CL_mf = 0.1 L/h), k_fm = 1.7 h⁻¹ (CL_fm = 0.1 L/h), k_{a, amniotic fluid} = 0.25 h⁻¹ (CL_{e, fetal} = 0.015 L/h) and k_{e, amniotic fluid} = 0.07 h⁻¹. Greater than 30% changes in the simulated AUC_0-τ, terminal half-life (t_1/2), and fetal-to-maternal (F/M) AUC_0-τ ratio with altered individual TK parameters compared to initial condition are bolded.

Altered TK parameter	New k (h−1)	New CL (L/h)	Maternal		Fetal		F/M AUC0-τ ratio
			t1/2(h)	AUC0-τ(h*ng/mL)	t1/2(h)	AUC0-τ(h*ng/mL)
Nona (Initial parameters	–	–	11	14	12	14	1.0
ka	0.007	–	99	14	99	14	1.0
	0.02	–	35	14	35	14	1.0
	0.2	–	11	14	11	14	1.0
	0.7	–	11	14	11	14	1.0
ke	0.3	0.18	13	130	13	130	1.0
	1.0	0.6	12	37	12	37	1.0
	8.0	4.8	11	4.7	11	4.7	1.0
	25	15	10	1.5	11	1.5	1.0
kmf	0.017	0.01	10	14	11	1.4	0.1
	0.051	0.03	11	14	11	4.1	0.3
	0.51	0.3	12	14	12	41	3.0
	1.7	1.0	12	14	12	140	10
kfm	0.17	0.01	28	14	28	140	10
	0.51	0.03	16	14	16	45	3.3
	5.1	0.3	10	14	10	4.5	0.3
	17	1.0	10	14	10	1.4	0.1
ke, amniotic fluid	0.025	0.0015	10	14	10	14	1.0
	0.08	0.0048	10	14	10	14	1.0
	0.8	0.048	15	14	15	14	1.0
	2.5	0.15	26	14	26	14	1.0
ka, amniotic fluid	0.0	–	10	13	10	12	0.9
	0.007	–	115	14	115	14	1.0
	0.02	–	40	14	40	14	1.0
	0.2	–	10	14	10	14	1.0
	0.7	–	10	14	10	14	1.0

As expected, fetal and maternal AUCs of DA correlated positively with maternal clearance (CL_{e, maternal} and k_e) (Table 2) but changing CL_{e, maternal} had no impact on the fetal-to-maternal plasma concentration ratio. Conversely, decreasing maternal-to-fetal transfer clearance (CL_mf) or increasing fetal-to-maternal transfer clearance (CL_fm), as would happen in the presence of active transport in the placenta, decreased fetal AUC but had no effect on maternal AUC (Table 2) and resulted in a considerable decrease in both the fetal-to-maternal AUC and plasma concentration ratio. Decreasing fetal reabsorption rate of DA from amniotic fluid (k_{a, amniotic fluid}) increased the fetal and maternal half-lives (Table 2) and increased the fetal-to-maternal plasma concentration ratio (Figure 6). It is interesting to note that fetal and maternal half-lives were the shortest in the absence of fetal reabsorption (k_{a, amniotic fluid} = 0) (Figure 6a) due to the switch of amniotic fluid from a distribution compartment to an elimination compartment under this scenario. Increasing fetal renal clearance of DA into the amniotic fluid (CL_{e, fetal} and k_{e, amniotic fluid}) increased the fetal and maternal half-lives (Table 2) and increased the fetal-to-maternal plasma concentration ratio (Figure 5f–j, 6a). These simulations demonstrate that the CL_mf, CL_fm, k_{a, amniotic fluid}, and CL_{e, fetal} have significant impact on the maternal-fetal distribution kinetics of DA. Moreover, altering each parameter alone did not fully replicate the observed infant-to-maternal plasma concentration ratio over time (Figure 5a). Therefore, the kinetic model parameters were optimized to define the combination of model parameters that can capture the observed data.

Figure 6. Sensitivity analyses to illustrate the effect of altering k_{a, amniotic fluid} and k_{e, amniotic} fluid on steady-state maternal and fetal plasma concentration time profiles up to 200 hours post last (the 50^th) dose.

Simulated maternal and fetal plasma concentration time profiles with (a) in the absence of k_{a, amniotic fluid}, (b-e) altering k_{a, amniotic fluid} from 0.007 to 0.7 h⁻¹, (f-j) altering k_{e, amniotic fluid} from 0.007 to 0.7 h⁻¹. Simulations were performed by changing either k_{a, amniotic fluid} or k_{e, amniotic fluid} while keeping the other parameters at their initial values (k_a = 0.07 h⁻¹, k_e = 2.76 h⁻¹ (CL_e,maternal = 1.66 L/h), k_mf = 0.17 h⁻¹ (CL_mf = 0.1 L/h), k_fm = 1.7 h⁻¹ (CL_fm = 0.1 L/h), k_{e, amniotic fluid} = 0.25 h⁻¹ (CL_{e, fetal} = 0.015 L/h) and k_{a, amniotic fluid} = 0.07 h⁻¹).

Since the maternal-fetal model was developed using observed maternal TK during pregnancy and the sensitivity analysis showed that k_a and k_e have minimal effect on maternal-fetal distribution kinetics of DA, we focused on optimizing CL_mf, CL_fm, and CL_{e, fetal} to capture the maternal-fetal disposition of DA. The optimized, physiologically plausible model incorporated a 4-fold higher CL_{e, fetal} together with an 82% lower CL_fm and a 94% lower CL_mf compared to the initial values (Figure 4). This optimized model captured the observed maternal plasma concentration-time profiles during pregnancy (GW16–18) well (Figure 5b). However, it slightly underpredicted the maternal and fetal concentrations at delivery. Hence, a bioavailability of 12% was applied (initial value was 7.5%) in the final model to predict the maternal and fetal plasma concentration-time profiles at delivery. The final model captured the maternal and fetal plasma concentration time profiles at delivery (Figure 5c) and the time course of the maternal-to-fetal concentration ratios observed in this study (Figure 5d) with a simulated fetal-to-maternal AUC ratio of 0.35.

DISCUSSION

Despite the abundance of evidence that DA acts as a developmental neurotoxin,^8,24–29 only a limited number of studies have reported TK of DA during pregnancy and fetal disposition following in utero exposure.^29,39–41 A previous study in rats showed that DA crosses the placenta and distributes to the fetus with a fetal-to-maternal plasma AUC ratio of 0.3 following a single intravenous dose.⁴¹ However, since DA follows flip-flop kinetics after oral doses in monkeys and possibly in humans, administration route could have a profound impact on maternal-fetal TK due to different maximum concentrations (C_max) and duration of exposure. As previously demonstrated through physiologically-based pharmacokinetic (PBPK) modeling, the C_max and concentration-time profile of DA in the brain varied drastically following intravenous or oral dosing.³⁶ To date, fetal TK analyses following oral DA exposure have been limited to opportunistic studies in California sea lions.^29,39 Based on these observations, a maternal-fetal TK model of DA was proposed which suggests that the fetus is continuously re-exposed to DA through swallowing of amniotic fluid which increased fetal exposure.³⁹ However, the extent and duration of DA exposure in the fetus cannot be ascertained in these opportunistic studies such that verification of the proposed model is not possible. This study is the first to describe the TK of DA during pregnancy and maternal-fetal disposition following oral doses in any species. In this study, we show that the renal clearance of DA is increased by 30 – 90% during pregnancy, however, this renal clearance increase did not alter the AUC of DA. The AUC of DA was not significantly different before, during, and after pregnancy. We also show that the fetal plasma DA concentrations were lower than, but positively correlated with, maternal plasma and amniotic fluid concentrations. A maternal-fetal TK model was developed to simulate maternal-fetal distribution kinetics of DA. The kinetic simulations suggested that the placental transport and that the recirculation of DA between the fetus and amniotic fluid are the main determining factors of the maternal-fetal distribution kinetics of DA.

The pre-pregnancy weight-normalized mean oral clearance following 0.075 or 0.15 mg/kg/day of DA observed in this study was 40 – 90% higher than that observed in our previous crossover study where three non-pregnant female monkeys were given a single oral dose of 0.075 or 0.15 mg/kg of DA.³⁶ This difference can be due to the inter-individual variability of DA disposition. A greater than dose proportional increase of AUC (~ 3-fold) was observed between the groups that received 0.075 or 0.15 mg/kg/day DA throughout the study, consistent with the greater than dose proportional increase observed in our previous crossover study.³⁶ Since the CL_R of DA and the CL_R/CL_creatinine were not different between the two dose groups in this study, it suggests that the renal clearance and the renal clearance processes (filtration and active reabsorption) of DA are not affected by the dose. Together, the data suggests saturation of either extra-renal elimination or absorption processes of DA.

As we hypothesized, the renal clearance of DA was increased during and after pregnancy when compared to the pre-pregnancy values. The increase in DA renal clearance was likely due to the increase in GFR as the CL_R/CL_creatinine values were not different throughout pregnancy when compared to pre-pregnancy. Although renal clearance is a major elimination pathway of DA (f_e = 0.4 – 0.7), the increased renal clearance during and after pregnancy did not alter the AUC of DA during pregnancy likely due to the presence of other elimination pathways and the large intraindividual variability³⁶. Hence, the dose-exposure relationship determined in non-pregnant monkeys is likely to predict the relationship in pregnant monkeys. While we cannot rule out the possibility that pregnant women are pharmacologically more susceptible to DA toxicity, our results suggest that pregnant women are exposed to similar concentrations of DA as non-pregnant individuals following the same dose.

Consistent with previous studies in other species,^29,39–41 our study shows that DA crosses the placenta and distributes to the fetus. The infant-to-maternal plasma concentration ratio at delivery was between 0.3 and 0.6, which agrees with the fetal-to-maternal AUC ratio of 0.3 observed previously in rats following a single intravenous dose.⁴¹ As demonstrated through the sensitivity analyses, either a decrease in CL_mf or an increase in CL_fm (both can occur in the presence of placental efflux) leads to the <1 fetal-to-maternal AUC ratio. This <1 fetal-to-maternal AUC ratio suggests that the placenta is partially protecting the fetus from exposing to DA, likely via active transport. DA has been shown to be transported by MRP5,⁴⁹ an ABC efflux transporter that is expressed in the syncytiotrophoblast,^50,51 and other anion transporters⁴³ which may be present in the placenta and decrease DA distribution to the fetus. Although irreversible fetal elimination may also decrease fetal-to-maternal AUC ratio, such clearance pathway is not likely because there is no evidence of DA metabolism. Hence, our results suggest that the placenta acts as a partial protective barrier that decreases the risk of fetal toxicity. Despite the placenta acting as a partial barrier to protect the fetus from DA, fetal AUC is driven by maternal AUC as shown by the positive correlation between fetal and maternal plasma concentrations at delivery, and by the sensitivity analyses that both maternal and fetal DA AUCs significantly changed with altered maternal absorption and elimination. Therefore, an increase in maternal AUC will increase the risk of DA toxicity in both the mother and the fetus, consistent with the observed dose-dependent increase in tremor rates in the adult female monkeys and dose-dependent impact on the developing cognitive processes in the infants.^33,35

As observed previously in rats,⁴¹ the apparent fetal half-life was longer than the maternal half-life in this study as evidenced by the infant-to-maternal plasma concentration ratio increase over time after dosing. The longer apparent fetal half-life has been attributed to fetal re-exposure through swallowing of amniotic fluid.⁴¹ We observed a 4.5 – 7.5-fold higher DA concentration in amniotic fluid compared to fetal plasma which, in agreement with the previous studies,^39,41 suggests that DA accumulates and is eliminated slowly from the amniotic fluid. We demonstrated using the maternal-fetal TK modeling and simulations, that placental transport and the recirculation between fetus and amniotic fluid are the main determining factors of the maternal-fetal distribution kinetics of DA. We proposed a final model to simulate a kinetic scenario of maternal-fetal disposition of DA that can explain the observed maternal and infant plasma concentrations at delivery. The final model incorporated the presence of placental efflux which resulted in a CL_mf to CL_fm ratio of 0.3, rapid fetal elimination (fetal renal clearance equals 10% of maternal renal clearance), and fetal reabsorption of DA through swallowing of amniotic fluid. It is important to note that as shown through simulation, that the maternal and fetal terminal half-lives are the same, however, the apparent fetal half-life (0–24 hour) was longer than the apparent maternal half-life due to delay in reaching distribution equilibrium (Figure 5c). As we hypothesized, the prolonged apparent half-life may increase the duration of fetal exposure to DA and hence, the risk of fetal toxicity. Nevertheless, the observed <1 infant-to-maternal plasma concentration ratio suggests that the fetus is exposed to lower plasma concentrations of DA than the mother.

In conclusion, this study shows that the plasma AUC of DA is not significantly changed by pregnancy, the fetal AUC of DA is less than the maternal AUC likely due to placental efflux, and that the maternal fetal distribution kinetics increases the duration of DA exposure in the fetus. Overall, our study suggests that pregnant women and developing fetuses are not subjected to additional exposure than the non-pregnant population based on the maternal-fetal TK. To fully assess the risks of DA toxicity, future studies to elucidate the toxicological response and exposure-toxicity relationship in pregnant women and developing fetuses are warranted.

HIGHLIGHTS.

1. Toxicokinetics of DA after oral exposure is not altered by pregnancy

2. In utero exposure to DA correlates with maternal DA exposure

3. The placenta limits fetal exposure to DA

4. The amniotic fluid is a significant distribution compartment for DA