Maternal organophosphate flame retardant exposure alters the developing mesencephalic dopamine system in fetal rat

Abstract

Organophosphate flame retardants (OPFRs) have become the predominant substitution for legacy brominated flame retardants but there is concern about their potential developmental neurotoxicity (DNT). OPFRs readily dissociate from the fireproofed substrate to the environment, and they (or their metabolites) have been detected in diverse matrices including air, water, soil, and biota, including human urine and breastmilk. Given this ubiquitous contamination, it becomes increasingly important to understand the potential effects of OPFRs on the developing nervous system. We have previously shown that maternal exposure to OPFRs results in neuroendocrine disruption, alterations to developmental metabolism of serotonin (5-HT) and axonal extension in male fetal rats, and potentiates adult anxiety-like behaviors. The development of the serotonin and dopamine systems occur in parallel and interact, therefore, we first sought to enhance our prior 5-HT work by first examining the ascending 5-HT system on embryonic day 14 using whole mount clearing of fetal heads and 3-dimensional (3D) brain imaging. We also investigated the effects of maternal OPFR exposure on the development of the mesocortical dopamine system in the same animals through 2-dimensional and 3D analysis following immunohistochemistry for tyrosine hydroxylase (TH). Maternal OPFR exposure induced morphological changes to the putative ventral tegmental area and substantia nigra in both sexes and reduced the overall volume of this structure in males, whereas 5-HT nuclei were unchanged. Additionally, dopaminergic axogenesis was disrupted in OPFR exposed animals, as the dorsoventral spread of ventral telencephalic TH afferents were greater at embryonic day 14, while sparing 5-HT fibers. These results indicate maternal exposure to OPFRs alters the development trajectory of the embryonic dopaminergic system and adds to growing evidence of OPFR DNT.

Use of organophosphate ester (OPE) compounds as flame retardants (OPFRs) and plasticizers in consumer goods and building materials has continuously increased over the past 2 decades. Global production of OPE compounds grew from 168 000 tons in 2001 to 620 000 tons by 2013, approximately 70% of which were used as flame retardants (FRs) (Fu et al., 2021; Wei et al., 2015). This dramatic increase in OPFR use coincided with the discontinuation of many brominated and polybrominated diphenyl ether (PBDE) FRs due to their toxic effects, and OPFRs have emerged as the predominant alternative to legacy FR use. Despite the structural similarity of OPEs to known neurotoxic compounds such as organophosphate (OP) insecticides, the potential neurodevelopmental and neuroendocrine disrupting effects of OPFRs remain largely uncharacterized (Doherty et al., 2019a). Consequently, the US EPA categorized some as high priority compounds in critical need of more toxicological, particularly developmental neurotoxicological, studies (Bajard et al., 2019). We previously found and reported that, in rats, gestational OPFR exposure impairs the development of serotonergic projections in the fetal rat forebrain, particularly in males (Rock et al., 2020). Using sections from the same fetal animals, plus materials from 2 additional experiments and a mix of 2-dimensional (2D) and 3-dimensional (3D) approaches, here we tested the hypothesis that dopaminergic development could also be vulnerable and sought to expand upon our previous results demonstrating disruption of 5-HTergic projections to the fetal forebrain by analyzing 5-HT fetal architecture in 3D.

Experimental evidence from a variety of systems suggests that OPEs are developmental neurotoxicants (DNTs). Using a combination of database mining and other computational approaches (Bajard et al., 2019), 97 OPFRs have been identified having potential toxicity (Blum et al., 2019). Although the primary mode of action for OP pesticides is acetylcholine esterase (AchE) inhibition, the OPFRs of interest were specifically designed not to act via this mechanism, yet they appear to have other DNT effects. We examined the OPFR components of the commercial mixture Firemaster 550, which include triphenyl phosphate (TPHP), 2-isopropylphenyl diphenylphosphate (2IPPDPP), and a mixture of isopropylated triarylphosphate isomers.

These compounds or the mixture have previously been shown to have deleterious effects. For example, TPHP exposure increased cell death in human neuroectoderm cells in vitro (Belcher et al., 2014), increased apoptosis in cortex and hippocampus, and increased microglial infiltration in adult mice in vivo (Liu et al., 2020). In zebrafish, developmental exposure to TPHP disrupted transcription and protein levels of critical mediators of neuroendocrine development in brain including estrogen receptor α, estrogen receptor β, and androgen receptor (Liu et al., 2013). Evidence of hyperactivity, has also been reported in zebrafish exposed to tris(1,3-dichloroisopropyl)phosphate (TDCIPP) and at least one study has found evidence of OPE toxicity in zebrafish at levels within the range of human exposures (Alzualde et al., 2018). We have also reported evidence of hyperactivity (Witchey et al., 2020) in rats as well as higher anxiety, with transcriptomic and lipidomic data from the newborn cortex indicating impaired cellular respiration along with dysregulated axonal transport and cholinergic synapse function in males (Witchey et al., 2022). An earlier National Toxicology Program led study reported similar effects in a battery of cell-based, nematode, and zebrafish assays with many OPEs having comparable activity to at least 2 brominated flame retardants (BFRs) known to be neurotoxic (Behl et al., 2015). In rats, we have previously shown that gestational exposure to the same OPFR exposure used herein (Table 1) significantly increased the length of forebrain serotonergic (5-HTergic) axons at embryonic day 14 (E14) in male fetuses compared with unexposed controls with evidence of less profound, but dose-responsive, effects in females (Rock et al., 2020). Emerging evidence indicates that the developing 5-HT system interacts with the dopaminergic system because it develops in parallel, including during forebrain innervation and axogenesis (Cunningham et al., 2005; Niederkofler et al., 2015); a relationship supporting the hypothesis that exposure to the same mixture of OPFRs may also alter the development of midbrain dopamine nuclei and their projections.

Probing the potential for OPFRs to be DNTs is significant because human exposure is ubiquitous. OPEs are used as FRs and plasticizers in myriad household products including furniture foam, textiles, baby products, and electronics, as well as commercial building materials such as PVC, glues, lacquers, and paints (van der Veen and de Boer, 2012; Wei et al., 2015). Because OPFRs are not chemically bound to the substrate, over time they transfer to the environment via abrasion, dissolution and/or volatilization (Möller et al., 2011; Wensing et al., 2005). Their widespread use and tendency to leach into the environment have resulted in OPFRs being ubiquitous contaminants commonly detected in household and office dust suggesting unavoidable human exposure, with some found at higher levels than legacy FRs in urban, rural, and background environments (Dodson et al., 2012; Rauert et al., 2018; Stapleton et al., 2009). OPE compounds also migrate away from areas of human activity as they have been detected at higher levels than PBDEs in some of the most remote regions of the globe such as the North Sea (Möller et al., 2011), and in different matrices in remote polar regions including atmosphere, soil, sediment, surface water, fish, and seabirds (Fu et al., 2021).

Developmental exposure to OPFRs is particularly concerning given the poorly understood toxicological profile of these compounds, and their structural similarity to OP pesticides, which have been shown to disrupt neurodevelopmental and neuroendocrine function, even at low-dose exposures (Patisaul et al., 2021). In humans, OPEs are typically metabolized to mono- or diesters and primarily excreted in urine, although conjugation and hydroxylation are possible (Doherty et al., 2019a). Biomonitoring studies routinely detect OPFRs and their metabolites in urine and other human biological samples, with the half-life in rats on the order of hours to days (Doherty et al., 2019a). For instance, OPFR metabolites have been detected in adult and toddler urine (Butt et al., 2014; Hoffman et al., 2014), and human breast milk samples (Kim et al., 2014), and there is growing evidence that placental accumulation and transplacental transfer is possible. Worryingly, children have consistently been shown to have higher levels of OPFR exposure than adults, likely due to frequent hand-mouth contact and the high levels of OPFRs in household dust (Butt et al., 2016; Gibson et al., 2019; Ospina et al., 2018; Percy et al., 2020). In humans, although data are limited, developmental OPFR exposure has been linked with cognitive deficits and behavioral problems in some (Doherty et al., 2019a; Vuong et al., 2020) but not all studies (Percy et al., 2021).

The present studies were conducted using exposures meant to model a human-relevant range and close to estimated point of departure (POD) values (Doherty et al., 2019a). The US EPA reference dose for OPE exposure is 15 000 ng/kg bw/day with biomonitoring data from the NHANES database, published by the Centers for Disease Control, revealing urine levels in the 1.5–5 ng/ml range for the most common parent compounds and metabolites; levels that can exceed excreted levels of PBDEs and other legacy FRs. Modeling and other data reveal that for some OPEs the in vivo rodent POD lies close to the range of human exposure (Blum et al., 2019). Estimates of human exposure posit inadvertent ingestion of dust as the major source of OPE exposure, followed by dermal absorption and inhalation of contaminated air (Gbadamosi et al., 2021; Kim et al., 2019) and foodstuffs (Li et al., 2019).

We chose E14 as our fetal time point for analysis because major events in development of both the midbrain dopamine nuclei and 5-HT nuclei and their respective connections occur during this time. Between E10.5 and E18, 5-HT and DA neurons differentiate from neuroblasts, undergo neuronal migration, and begin axogenesis and cortical innervation (Gates et al., 2006; Kalsbeek et al., 1988; Niederkofler et al., 2015). A prior study from our lab, Rock et al. (2020), using 2D brain sections showed no effect of maternal OPFR exposure on 5-HT cell body number or spread, but significant changes in 5-HT axogenesis at E14. Using adjacent sections from the same fetuses, here we immunolabeled the rate-limiting enzyme in DA synthesis, tyrosine hydroxylase (TH) to assess DA system development. Corticopetal 5-HT and DA fibers of the medial forebrain bundle should both be in transit to the forebrain at this time, allowing us to visualize any potential difference in both DA and 5-HT neural migration and pathway morphology. The mature midbrain DA system is comprised of 3 nuclei: the substantia nigra pars compacta (SNc), ventral tegmental area (VTA), and retrorubral field, all typified by their synthesis of the neurotransmitter DA and their expression of TH. The mature central 5-HT system comprises a subset of 5-HT expressing neurons situated in the ventral hindbrain collectively known as the raphe nucleus. On E14, both systems are still undergoing neuronal migration and early axogenesis.

These relatively small nuclei situated in the ventral mid- and hindbrain have an outsized effect on brain activity and behavior, as DA and 5-HT are potent neuromodulators of cognition, mood, and motor function. Fibers from these nuclei innervate broad segments of the forebrain, notably the medial prefrontal cortex (mPFC), ventral and dorsal striatum, hippocampus, and amygdala (for review see, Brignani and Pasterkamp, 2017). DAergic innervation of the mPFC specifically, originates in the VTA and creates symmetrical synapses largely in the deep layers of the prelimbic (PL), infralimbic (IL), and anterior cingulate cortex (ACC) (Islam et al., 2021). These connections mediate behaviors associated with executive function, including sensory gating of attention, working memory, and control of motor activity. These behaviors are critical for flexible adaptive responses to diverse stimuli and contexts (Ott and Nieder, 2019). 5-HT innervation of the forebrain arises largely from the dorsal and medial raphe and broadly innervates the cerebral cortex, as well as the hippocampus, amygdala, septum, and other structures. These connections are diffuse, with few synapses, and act as neuromodulators for a variety of cognitive functions (some overlapping with DA) including executive function and cognitive flexibility, emotional learning and memory, affective behaviors, and anxiety (Alvarez et al., 2021; Deneris and Gaspar, 2018; Ögren et al., 2008). The formation of both DA and 5-HT connections largely occur during the embryonic and postnatal periods. Thus, to begin to understand the potential impact of developmental OPFR exposure on these systems, we examined them at E14; a point before the DA neurons have migrated to their final destinations and the major midbrain DA nuclei are fully formed and defined, and when 5-HT and DA axonal fibers of the medial forebrain bundle are in transit to the cortex. We also assessed DA fiber density in the adult mPFC as a single, initial measure of long-term organizational effects.

As is common for these kinds of morphological studies, we first identified and quantified midbrain-fated DAergic neurons by TH immunolabeling in fetal sections cut in the sagittal and coronal planes to assess neuronal spread in 2D. The approach was akin to the 2D analysis performed for 5-HT by our lab in a previous experiment using the same animals (Rock et al., 2020). We then also cleared and immunolabeled a subset of whole E14 male brains to assess the complex of both monoaminergic structures in 3D using light sheet microscopy. Although we typically always examine both sexes, for the 3D imaging we only assessed males because they were found to be most impacted in the already published 2D analyses of the 5-HT system, and because the high dimensional nature of the approach limits the number of animals that can be assessed. We hypothesized that maternal OPFR exposure during the first 2 weeks of gestation alters axogenesis and cytoarchitecture of the developing mesencephalic DA system and ascending 5-HT system, with males at greater risk. Finally, to begin to investigate the long-term effects of developmental exposure to OPFRs on the DA signaling network, we measured DA fiber density in the adult rat mPFC of both sexes. We hypothesized any fetal sex differences would persist.

Materials and methods

Animals

As detailed below, the fetal tissues used for “2-dimensional” (2D) analyses in Experiment 1, and the adult “2D” tissues used in Experiment 3, were obtained from 2 prior, published studies (Rock et al., 2020; Witchey et al., 2020). Whole cleared fetal heads (3D) were generated specifically for Experiment 2 of this project (Figure 1). Regardless of source, housing, maintenance, and experimental treatment of all animals in the study were approved by the North Carolina State University (NCSU) Institutional Animal Care and Use Committee (IACUC) and met the standards of Animal Welfare Act and U.S. Department of Health and Human Services “Guide for the care and the use of Laboratory Animals.” For each experiment, Wistar rats obtained from Charles River (Raleigh, North Carolina) were bred in-house at the biological resource facility at NCSU. As in our prior studies (Rock et al., 2020; Witchey et al., 2020) animals were housed in conditions to minimize inadvertent exposure to endocrine disrupting chemicals (EDC), including use of glass water bottles with metal spouts, water filtered by reverse osmosis, wood chip bedding, soy-free diet (Teklad2020), and polysulfone caging, in accordance with recommendations for EDC research (Li et al., 2008).

Dosing preparation, animal husbandry and exposure

A stock solution of concentrated OPEs in ethanol (approximately 100 mg/ml) including a mixture of OPEs commonly found in commercial FR mixtures was prepared by the laboratory of Dr. Heather Stapleton at Duke University. This mixture was prepared with the same compounds and relative concentrations as prior studies with the Stapleton lab and is the same OPFR mixture used in 2 prior studies examining the neuroendocrine disrupting effects of OPFRs (Phillips et al., 2017; Rock et al., 2020).

The chemical components included TPHP (Cas No. 115-86-6), 2-isopropylphenyl diphenylphosphate (2IPPDPP; Cas No. 28108-99-8; 93925-53-2; 64532-94-1), 3-isopropylphenyl diphenyl phosphate (3IPPDPP), 4-isopropylphenyl diphenyl phosphate (4IPPDPP), 2,4-diisopropylphenyldiphenyl phosphate (24DIPPDPP), bis(2-isopropylphenyl) phenylphosphate (B2IPPPP; Cas No. 69500-29-4), bis(3-isopropylphenyl)phenyl phosphate (B3IPPPP), bis(4-isopropylphenyl) phenylphosphate (B4IPPPP), and tris(3-isopropylphenyl) phosphate (T3IPPP; Cas No. 72668-27-0) (Table 1). All OPE solutions were then transferred to NCSU, and stored in light blocking glass bottles at 4°C. To ensure experimental blinding, for all 3 experiments the person who dosed the animals was independent from the team performing the experimental work.

At the time of dosing, the Patisaul lab prepared dosing solutions by diluting the appropriate volume of concentrated stock solution with 100% ethanol (0, 25, 50, and 100 mg/ml). Twenty microliters of this diluted solution were then pipetted onto one-fourth of a soy-free food treat pellet (apple flavored AIN-76A Rodent Diet Test Tabs, Test Diet, Richmond, Indiana) resulting final exposure dose of (0 µg OPFRs—control, 500 µg OPFRs—low, 1000 µg OPFRs—mid, 2000 µg OPFRs—high). The food treat was administered only after it was completely dry, and dams were observed to ensure complete consumption of the treat. Oral exposure to the dam via food treat greatly reduces handling and stress compared with oral gavage (Bonnichsen et al., 2005; Brown et al., 2000; Cao et al., 2013; Walker et al., 2012), and therefore reduces the potential confounding variable of prenatal stress. Individual dams were not dosed by individual body weight but rather by an average mass of a typical dam at day of conception of 250 g, resulting in an approximate final dose per kilogram of bodyweight per day of 0, 2, 4, and 8 mg/kg. These doses were chosen because we have previously documented neuroendocrine toxicity in this range (Rock et al., 2020; Witchey et al., 2020). For 2D fetal TH-immunohistochemistry (IHC), all 4 doses were utilized. For fetal 3D whole brain clearing and adult 2D imaging only the middle dose of OPFRs was used (1000 µg) because of the technical constraints of whole brain clearing, immunolabeling and imaging, and because the adult tissue was obtained from a prior study (Witchey et al., 2020) and thus limited.

Experiment 1—Fetal 2D TH immunohistochemistry

2D fetal tissue preparation

Fetal brain sections cut in the sagittal and coronal planes were used for 2D analysis of the developing dopaminergic system. In the sagittal sections, the developmental trajectory of dopaminergic axons in the medial forebrain bundle was also evaluated. For all litters, fetal sex was determined by PCR for the sry gene as described previously (Rock et al., 2020).

Fetal brains sectioned in the sagittal plane were generated for a prior study (Rock et al., 2020), the details of which can be found therein. Briefly, a total of 80 female, and 20 male adult wistar rats were obtained from Charles River (PND 70–90) in 4 cohorts and randomly assigned to an exposure group, each cohort contained animals from each exposure group. For breeding, 2 females were paired with one male. Receptivity and pregnancy were determined by vaginal cytology. If a female was not found to be pregnant after 72 h paired with a male, this dam was then added to the next cohort as part of the same dose group. The total number of successfully impregnated and dosed dams across all 4 cohorts was 70. No differences in breeding success, fecundity, or pregnancy phenotype were observed in this study. Dams were exposed once per day for 14 days across E1–14 and body weight averaged 250 g producing exposures of approximately 0, 2 (low), 4 (medium), and 8 (high) mg/kg bw/day OPFRs. Dosing occurred at approximately 9 am. On E14, 4 h after final dosing dams were euthanized by CO₂ asphyxiation and rapid decapitation. Embryos were removed from the dam, the heads drop fixed in 4% paraformaldehyde, and stored at −80°C until cryosectioning. A total of 60 fetal heads were sectioned in the sagittal plane at 20 µm (male controls = 7, female controls = 10, male low OPFR = 8, female low OPFR = 7, male mid OPFR = 7, female mid OPFR = 7, male high OPFR = 8, female high OPFR = 6).

An additional 50 whole fixed fetal heads generated by the Rock (2020) study were specifically selected and sectioned at 20 µm in the coronal plane for this experiment. Some groups had more available heads than others, thus sample size varied (male controls = 9, female controls = 13, male low OPFR = 7, female low OPFR = 4, male mid OPFR = 4, female mid OPFR = 5, male high OPFR = 3, female high OPFR = 5). All sections were cut into 3 serial sets on a Leica CM1900 cryostat, slide mounted (Superfrost Plus, Fisher, Pittsburg, Pennsylvania) and stored at −80°C until immunolabeling (Rock et al., 2020).

2D fetal brain immunohistochemistry

For both the sagittal and coronal sections, the anti-TH polyclonal antibody raised in rabbits (AB152, Millipore, RRID: AB_390204) was used to label DA neurons, as we have done previously (Gillera et al., 2021). This antibody is listed in the Antibody Registry (http://antibodyregistry.org) and has been validated for IHC in brain (Geerling and Loewy, 2006). In line with our previously published methods (Rock et al., 2020) one slide per animal from a serial set (containing the mesencephalic dopamine nuclei and medial forebrain bundles) was used. Briefly, slides were allowed to come to room temperature and dry for 20 min. Once dry, slides were removed from KPBS and a hydrophobic barrier was drawn around the perimeter of each slide with a PAP pen. The slides were then placed in 100% methanol for 10 min and washed 3× for 10 min in 0.02 M KPBS. Blocking buffer (LKPBS; 3% Triton X-100, 2% Normal Goat Serum in 0.02 M KPBS) was carefully pipetted onto the slide covering all sections and incubated at 4°C overnight in a moisturized chamber to prevent evaporation. Blocking solution was gently poured off and 500 µl of LKPBS containing rabbit polyclonal anti-TH primary antibody (1:5000, AB152, Millipore) was carefully pipetted over each slide to cover all sections. The slides were then incubated a humidified chamber at 4°C overnight and washed in 0.02 M KPBS 6× for 10 min. Secondary antibody (1:200, goat anti-rabbit Alexa fluor 568, Invitrogen) was pipetted on each slide and the sections incubated for 1 h at 4°C, then washed 6× for 10 min in KPBS. Finally, the slides were counterstained with Hoechst 33258 (Invitrogen), rinsed 5× in KPBS and cover slipped (VWR micro cover glass, Radnor, Pennsylvania) with glycerol mountant.

Analysis of fetal TH-ir in sagittal sections

Anatomically matched sagittal sections were identified using well-defined anatomical landmarks and approximately equate to Figure 66 in the Atlas of the Developing Rat Nervous System, fourth edition (Paxinos and Ashwell, 2018). For measurement of the TH-ir clustered cells, sections were chosen where the fovea of the isthmus separating the cerebral aqueduct and fourth ventricle had only just closed. A Leica DM6B epifluorescent microscope was used to image TH-ir cells and fibers dorsal and rostral to the cephalic flexure. Height of the TH-ir cell cluster was measured by drawing a straight line from the most dorsal TH-ir cell orthogonal to the cephalic flexure. The span of the nucleus was taken from the most rostral TH-ir cell to the most caudal (Figure 2). Quantification of TH-ir cell cluster dimensions (putative VTA and SNc) and TH-ir fiber length and width achieved using the Leica Application Suite X software measurement tool.

For TH-ir axon fibers, similar measurements of length and spread were obtained. Because the leading TH-ir axons in all animals passed anterior to the intraventricular foramen of the lateral ventricle, this was chosen as a consistent and easily recognizable point to establish axon length. TH-ir axon length was measured from the intraventricular foramen (anterior to choroid plexus) to the most anterior axon tip across sections. To measure the spread of the developing medial forebrain bundle, the distance spanned by TH-ir axons was measured at its widest point dorsal to ventral within the ventral telencephalon (Figure 2A). Images of the medial forebrain bundle at this point were taken and used to calculate density of TH-ir axons using FIJI image software. TH-ir was thresholded and binarized, then skeletonized to control for fiber thickness and brightness. Total-ir was then calculated using the measure function.

Analysis of sagittal “2D” imaging was carried out by 2 independent observers both blind to treatment condition and sex, and their results then averaged together for statistical analysis to decrease experimental variability and ensure reproducibility.

Analysis of fetal TH-ir in coronal sections

Coronally sectioned tissue was matched to Figures 56 (coronal TH-ir cell cluster), and 47 (coronal TH-ir axons) of the Atlas of the Developing Rat Nervous System fourth edition (Paxinos and Ashwell, 2018) using anatomical landmarks. For measurement of the TH-ir midbrain cell cluster in coronal sections, sections were chosen where the mammillary recess of the third ventricle was closing but still present. TH-ir cells ventral to the mesencephalic aqueduct were imaged with an epiflourescent microscope (Leica DM6B). Height of the TH-ir cell cluster was measured from the most ventral cell body to the most dorsal in coronal sections, whereas the width of the nucleus was measured between the lateral most TH-ir cells in each hemisphere (Figs. 2D and 2E), using the Leica Application Suite X software tool. To gauge the rostral progress of TH-ir axons in coronal sections at a consistent anatomically landmark, the number of TH-ir axons were counted where the caudate putamen first began to protrude into the lateral ventricle, and where the anterior cerebral arteries were visible (approximately Figure 47 in the Atlas of the Developing Rat Nervous System (Paxinos and Ashwell).

Experiment 2—Fetal 3D TH immunofluorescence

Collection and preparation of fetal whole brains for 3D imaging

The mature dopamine and 5-HTergic systems are complex 3D structures composed of several ornately shaped midbrain nuclei (VTA, SN, raphe, and subregions of each) as well as the intricate forebrain bundle coursing a nonlinear path rostrally. To better capture the intricacies of these structures as it emerges in fetal life, and in 3 dimensions we employed whole mount tissue clearing and immunolabeling for TH and performed the analysis using light sheet microscopy and Imaris software. Additionally, to confirm previous findings regarding 5-HT axogenesis conducted in 2D (Rock et al., 2020) we also immunolabeled these fetal heads for 5-HT and performed a similar 3D analysis.

To generate the fetal brains used for whole mount clearing and light sheet microscopy, 1 dam was paired with 1 male (10 dams, 4 sires total) on the day of proestrus (determined by vaginal cytology). If sperm was detected 24 h later, the female was placed in her home cage and assigned to an experimental group, 5 dams were assigned to OPFR, 5 to control. Forty-eight hours after sperm detection was considered E1 and dosing (1000 µg OPFRs) began, with an average dam weight of 250 g this resulted in an approximate 4 mg/kg bw OPFR per day. Dams in which sperm was not detected after 24 h were paired with a different male when they next entered proestrus. Dosing was daily at 9 am from E1–14, with sacrifice approximately 4 h after the final dose to control for time post-exposure. Dams were anesthetized by CO₂ asphyxiation, weighed, and sacrificed by rapid decapitation. Fetal pups were dissected, and heads removed. Approximately half the fetal heads were drop fixed in 4% paraformaldehyde in 0.1 M PBS, and the other half flash frozen on dry ice (to be used for future studies). As we have done previously, fetal age was assessed by developmental morphology according to Witschi, Thelier, and Carnegie criteria to confirm embryonic day of each litter and only those determined to be E14 were included in the study. A single paw was taken from each fetus for sex determination by PCR as previously described (Rock et al., 2020). After 24 h in PBS buffered 4% paraformaldehyde, drop fixed fetal heads were transferred to 0.02M KPBS and stored at 4°C until processing for IHC and iDISCO+. Because 3D analysis is resource intensive, and because no pronounced sex differences were identified in the 2D analyses, only one sex was used for 3D clearing. Males were chosen because we had sufficient numbers in both experimental groups to use litter as the statistical unit.

Whole brain clearing for 3D imaging

A total of 11 male E14 heads (N = 6 OPFR, N = 5 control), underwent pretreatment with methanol, immunolabeling, and clearing as outlined by Renier et al. (2014). Our goal for the group composition of this study was one animal per sex per dam, however due to the technical challenges associated with whole fetal head clearing and imaging we predicted we would lose individuals, and so included extra heads in the clearing process. Due to this, one dam is represented twice among the OPFR individuals. Whole head was used to preserve the structural integrity of the brain. Heads were kept in solution on an orbital shaker for the entire protocol. All steps were performed in fully filled, closed tubes to prevent oxidation. Secured tubes were wrapped in parafilm before entering the hybridization oven.

Pretreatment began with increasing series of methanol incubation (20% ×1, 40% ×1, 60% ×1, 80% ×1, 100% ×2, each for 1 h) in ddH₂O, at RT to dehydrate the tissue. Samples were then chilled at 4°C in 100% methanol (1× 1 h) and incubated at RT in 66% dichloromethane, 33% methanol overnight. The following day, tissue was rinsed with 100% methanol at RT (2× 10 min) before chilling samples at 4°C (1× 1 h). Tissue was incubated overnight with fresh 5% hydrogen peroxide in methanol at 4°C to decolorize and quench endogenous fluorescence. Samples were then incubated in serial decreasing methanol concentration solutions (100/80/60/40/20% methanol in ddH2O) at RT for 1 h per wash. Samples were placed in PBS at RT for an additional hour. Tissue was dilapidated by incubation in 0.2% TritonX-100 in PBS (PTx.2) at RT (2× 1 h). Fetal heads were then permeabilized in 2.3% glycine, 20% dimethyl sulfoxide 0.2% in PTx.2 for 2 days at 37°C in a hybridization oven (Renier et al., 2014; Tainaka et al., 2016).

Subsequently, fetal heads were incubated in blocking solution (84% PTx.2, 6% Donkey Serum, 10% dimethyl sulfoxide) for 2 days at 37°C in a hybridization oven. Samples were then incubated with a cocktail of anti-TH primary antibody made in rabbit AB152, Millipore, RRID: AB_390204 and anti-5-HT primary antibody made in goat (each at 1:1000 μl) in 0.2% Tween-20, 0.1% of 10 mg/ml heparin, 5% DMSO, 3% donkey serum in PBS (PTwH) for 5 days at 37°C in a hybridization oven. Fetal heads were washed with PTwH 3× for 2 h then once overnight. Samples were then incubated with fluorescent secondary antibodies (Donkey anti-rabbit Alexafluor 555 Invitrogen, donkey anti-goat Alexafluor 647 Invitrogen) for 5 days at 37°C in a hybridization oven. Fetal heads were washed with PTwH 3× for 2 h then once overnight (Renier et al., 2014). Tissue was dehydrated in a series of methanol/ddH2O (20/40/60/80/100%, each for 1 h) at RT. Fetal heads were left in 100% methanol at RT overnight. A 66% dichloromethane, 33% methanol wash (1× 3 h) was administered before the tissue underwent washes in 100% dichloromethane (2× 15 min). Lastly, fetal heads were stored in dibenzyl ether at RT until imaging. The sample and solution were briefly mixed by inverting before the tubes were placed in a closed box to limit light exposure (Renier et al., 2014; Tainaka et al., 2016).

Light sheet microscopy

The heads were imaged on a custom-built light sheet fluorescence microscope designed based on previous protocols (Greenbaum et al., 2017; Li et al., 2021, 2022; Royer et al., 2018). Two control heads were damaged during imaging and were excluded, leaving a total of 9 male fetal heads (n = 6 OPFR, n = 3 control) for analysis. Samples were imaged using illumination of 561 nm, with a voxel size of 0.65 × 0.65 × 10 μm³. Datasets were stitched by TeraStitcher (v1.10.18). Imaris (v9.5.1) software’s surface function was used to render 3D models of dopaminergic (TH-ir) and 5-HTergic (5-HT-ir) axonal projections and cell clusters and calculate volume. Light sheet sectioned images in the sagittal plane were used to measure the length (rostro-caudal) and height (dorsal-ventral) of cell cluster (putative VTA and SN nuclei) and the spread of TH-ir axons in the dorsal-ventral axis, and the length and volume of 5-HT axonal projections and total volume of 5-HT cell cluster (putative raphe).

Experiment 3—Adult 2D TH immunohistochemistry

2D adult tissue preparation

To begin to determine the potential long-term effects of developmental OPFR exposure on cortical innervation by TH-ir axons, the total and average density of TH-ir fibers in the PL, IL, and ACC of the mPFC in adult rats were compared between rats exposed to OPFRs (1000 µg—mid dose) throughout gestation and lactation, and controls.

All adult rat brains were generated for a prior study, the details of which are contained therein (Witchey et al., 2020). Briefly, 56 female and 16 male adult rats were obtained from Charles River and females were randomly assigned to treatment groups. Two females were paired with a single male in 2 cohorts for one week (24 in one and 32 in the other cohort). Following pairing, pregnant dams were singly housed for the rest of the study. Dosing occurred at approximately 9 am daily beginning 72 h after pairing and continued until weaning on (postnatal day [P] 21). OPFR solution containing 1000 µg was pipetted onto a wafer cookie (Nilla; Nabisco), resulting in an approximate dose of 3.3 mg/kg bw OPFR per day (based on 300 g average dam weight). Of the control and OPFR dams utilized in the current study, 1 control female and 3 OPFR females did not conceive, resulting in 11 control dams and 9 OPFR exposed dams. Pups were weaned on P21 and housed in same sex pairs until sacrifice on or around PND 80. The sections used herein came from the siblings of animals that underwent behavioral testing in the published study, but did not undergo behavioral testing themselves, and all females were sacrificed on estrus (Witchey et al., 2020).

Adult animals were deeply anesthetized with sodium pentobarbital (i.p.) and serially perfused through the heart with 0.9% saline in PBS and 4% paraformaldehyde in PBS. Brains were then removed and postfixed for 24 h in 4% paraformaldehyde, then cryoprotected in 30% sucrose in PBS and stored at −80°C until cryosectioning on a freezing sliding microtome or cryostat for immunohistochemistry.

Adult 2D TH immunohistochemistry

Adult rat brains (N = 44, male controls = 12, female controls = 12, male OPFR = 9, female OPFR = 11) were sectioned coronally at 40 µm on a freezing sliding microtome (Leica 2010R) into 4 serial sets and stored in antifreeze solution at −20°C until immunolabeling. Although our goal was to use 1 animal per sex per dam, there was not sufficient tissue to do that, thus 1 control dam was represented twice and 2 OPFR dams (for female rats) were represented twice. Sections of prefrontal cortex (PFC) were selected from one serial set and free floating IHC was performed for TH according to our previously published methods (Gillera et al., 2021). Cryoprotectant solution was first rinsed off in 0.02M KPBS 3× for 10 min. Sections were then placed in blocking solution (3% Triton X-100, 2% Normal Donkey Serum in 0.02 M KPBS) overnight at 4°C. Following blocking, sections were incubated with primary antibody for TH at 1:5000 (AB152, Millipore, same as for fetal heads) for 72 h at 4°C. Sections were then thoroughly rinsed in 0.02 M KPBS 6× 10 min before being incubated for 90 min with anti-rabbit conjugated fluorescent antibody (donkey anti-rabbit Alexafluor 568, Invitrogen) at 1:200. Sections were then washed in 0.02M KPBS 6× 10 min and counterstained with Hoechst 33258 (Invitrogen) before a final wash series in KPBS (5× 5 min). Sections were then mounted on Superfrost Plus slides (Fisher) and cover slipped with glycerol mountant.

Analysis of adult mPFC TH-ir

Images of TH-ir in adult rat medial mPFC were obtained on a Leica SPE confocal microscope. Three subregions of the mPFC were analyzed: the PL, IL, and ACC. For TH-ir analysis in the PL, anatomically matched sections corresponded to Figure 11 in The Rat Brain in Stereotaxic Coordinates, seventh Edition (Paxinos and Watson, 2014), whereas sections corresponding to Figure 13 were selected for IL and ACC. For each region, 2 µm z stacks at 20× were obtained and the central 15 images of each used for the analysis. To ensure consistent sampling across animals, the edge of the field of view was aligned with the medial edge of the forceps minor of the corpus callosum to capture the basal cortical layers (IV–VI), where the majority of TH-ir was observed. Images were then exported to Fiji for analysis where TH-ir density of each of the 3 mPFC regions was analyzed using the thresholding tool. Each thresholded z-stack of 15 images was then binarized and skeletonized to control for fiber thickness and brightness. The total number of black voxels was then counted across all 15 images to obtain a total value for each z stack using the Voxel Counter function.

Statistical analysis

Statistical analysis was performed using SigmaPlot 14.5 with statistical significance set at p < .05. For the fetal and adult brain sections, differences in TH-ir were first assessed by 2-way ANOVA with sex and exposure as factors. Because no sex differences or significant interactions between sex and exposure were found for any measure, we collapsed the data across sex and compared exposure groups to controls via 1-way ANOVA. Differences between specific group means were assessed by pre-planned post hoc comparisons using the Holm-Sidak test. For volumetric data from 3D models generated from whole mount fetal heads, t-tests were run on control and OPFR-treated cell clusters and axonal projections. Effect sizes were calculated for all statistically significant tests using eta squared (η2) ANOVAs and hedge’s g for post hoc comparisons of group means with unequal sample sizes. We used Cohen’s recommendations for effect size descriptions (η2, 0.01 = small, 0.06 = medium, 0.14 = large, and for g, 0.2 = small, 0.5 = medium, 0.8 = large, [Cohen, 1992]). Spearman’s correlation (rs) was calculated as a measure of inter-observer reliability for sagittal sectioned 2D analysis.

Results

Experiment 1—Fetal 2D TH immunohistochemistry

Effect of OPFRs on morphology of developing TH-ir cell cluster

In the sagittal sections, there was a significant effect of exposure on the span (rostral-caudal) of the mesencephalic TH-ir cell cluster(F(₃,₅₉) = 4.908, p ≤ .004, η² = 0.211, large effect, r_s = 0.61; Figure 3), post hoc analysis revealed that rostral-caudal span in the low (p ≤ .007, g = 1.11, large effect), and mid (p ≤ .018, g = 1.003, large effect) dose groups were significantly larger than controls in both sexes.

Comparison of the height, or dorsal-ventral dimension of the mesencephalic TH-ir cell cluster measured in the sagittal sections revealed a significant effect of exposure (F_(3,58) = 4.840, p ≤ .005, η² = 0.202, large effect, r_s = 0.66; Figure 4). The mid (p ≤ .010, g = 1.525, large effect) and high (p ≤ .011, g = 1.003 large effect) dose groups had significantly greater height than controls.

In the coronal sections, no significant differences on TH-ir dimensionality were detected between OPFR dose groups, either within sex, or after the data from both sexes were combined to increase statistical power (data not shown).

Impact of OPFRs on TH-ir axogenesis in fetal rats

In the sagittal sections, length, spread, and density of TH-ir axons in the diencephalon were compared across groups. There was no significant main effect of dose or sex on the length or density of TH-ir axons in their progress rostrally. However, the spread between the most dorsal and ventral axon was significantly different between OPFR exposed and control animals (F_(3,52) = 4.247, p ≤ .009, η² = 0.185, large effect, r_s = 0.90; Figure 5), with the low (p ≤ .037, g = 1.016) and high (p ≤ .008, g = 1.313) doses having significantly greater spread than the controls. Although the mid dose was appreciably greater however did not reach statistical significance (p ≤ .176, g = 0.869 large effect).

In the coronal sections, the number of TH-ir axons present in the ventral diencephalon was quantified. No significant difference in the number of TH-ir axons which had progressed to the anterior ventral diencephalon was observed between exposed and control animals.

Experiment 2—Fetal 3D TH immunofluorescence

Effect of OPFRs on the 3D structure of the fetal DA and 5-HT systems

We first sought to reproduce the 2D assessment procedure of Experiment 1 by digitally sectioning the whole mounts sagittally and assessing the span (rostral-caudal) and height (dorsal-ventral) of the midbrain dopaminergic cell cluster as well as the length and spread of leading TH-ir axons. In the whole brains, there was no difference in span of the TH-ir cell cluster (data not shown), or length of TH-ir afferents in the ventral diencephalon. However, height (dorsal-ventral) of the TH-ir cell cluster in the sagittal plane was significantly greater in OPFR exposed males than control males (t(8) = 2.54, p ≤ .038, g = 1.23; Figure 6A) and spread of axons in the dorsal-ventral axial plane was significantly wider (t(8) = 3.37, p ≤ .01, g = 1.63; Figure 6B) in OPFR exposed males than controls, confirming the observations and measurements of Experiment 1.

Additionally, 3D surface volumes were created for both TH and 5-HT immunoreactive cell clusters and axon pathways in OPFR exposed and control whole fetal heads (Figure 7). Analysis of 3D surface volumes revealed a significant decrease in total volume of the TH-ir cell cluster in OPFR exposed males compared with controls (t(8) = −2.90, p ≤ .023, g = 1.41; Figure 8). However, there was no significant difference in the volume of TH-ir axon pathways (t(8) = 0.68, p ≤ .52, g = 0.33; Figure 9). In 3D surface renderings based on 5-HT-ir there was no difference in volume of cell clusters (t(8) = 0.485, p ≤ .641, g = 0.235; Figure 10A), and no difference in the volume (t(8) = 0.734, p ≤ .488, g = 0.356; Figure 10B) or length (t(8) = 0.564, p ≤ .598, g = 0.274; Figure 10C) of 5-HTergic fibers originating from the 5-HT-ir cell cluster.

Experiment 3—Adult 2D TH immunohistochemistry

No effect of developmental OPFR exposure on adult mPFC TH-ir fibers

To begin to determine the potential long-term effects of developmental OPFR exposure on cortical innervation by TH-ir axons, the total and average density of TH-ir fibers in the PL, IL, and ACC subregions of the mPFC of exposed and unexposed adult rats was compared. There were no significant main effects of exposure or sex and no significant interactions between the factors for any of the 3 regions examined.

Discussion

E14 offspring of both sexes from dams exposed to OPFRs during gestation exhibited significant alterations to the developmental trajectory of the mesencephalic dopamine system, with the 3D analysis complementing and augmenting the 2D findings in males. Both “traditional” 2D IHC and 3D whole brain analysis revealed the developing mesencephalic DAergic neuron cluster in the OPFR exposed fetuses was larger in the dorso-ventral dimension, whereas 2D analysis also showed a significant increase in the rostral-caudal span. 3D volumetric analysis revealed that the overall volume of the TH-ir cell cluster was significantly smaller in OPFR exposed males than controls. Additionally, the course of dopamine axogenesis was significantly altered by OPFR exposure in both sexes. The dorso-ventral spread of TH-ir axons at the leading edge of the developing medial forebrain bundle was found to be significantly broader in OPFR exposed male and female fetuses than in controls in both the 2D and 3D analyses. The length and density of the axons, however, and the volume of the medial forebrain bundle were unchanged. By contrast, gestational OPFR exposure was not found to alter male ascending 5-HTergic development in the 3D analysis, a result which is incongruent with our prior, 2D study which used both sexes, more doses, and was more robustly powered (Rock et al., 2020). 3D analysis of the developing 5-HTergic nuclei and corticopetal fibers showed no significant difference between OPFR exposed offspring and controls in cell cluster volume, axon bundle volume or length. Collectively, the experiments in this study reveal how inclusion of whole brain analysis can significantly augment conventional 2D analysis by revealing morphological changes to the full structure not readily discernable in either the coronal or sagittal planes. Finally, although changes in the developing mesencephalic DA system were observed in the embryonic brain, the density of TH-ir fibers in the adult mPFC, one of the main forebrain targets of fetal TH-ir axon innervation, was not significantly altered by perinatal OPFR exposure. This, however, is not an exhaustive assessment of adult DA morphology or function, therefore it remains unclear what, if any, aspects of disrupted dopaminergic innervation persist postnatally. Nonetheless, these results strongly indicate that the ontogeny of the DAergic system is vulnerable to the effects of maternal OPFR exposure during fetal brain development. Capacity for OPFR exposure to disrupt the ontogeny of monoaminergic systems merits further inquiry.

OPFRs impinged on the morphology of the midbrain dopamine cell cluster

Fetal OPFR exposure resulted in significant changes to the morphology of the developing midbrain DAergic nuclei, as revealed by “traditional” 2D immunohistochemistry, particularly in the medial aspect (putative, emerging VTA). At low and mid doses (500 and 1000 µg), the distance between the most rostral and most caudal TH-ir cell body in the medial midbrain cell cluster was significantly longer in OPFR animals of both sexes than in controls. The rostral-caudal span of the TH-ir cell cluster in digitally “collected” sagittal sections from the immunolabeled whole male brains (dosed at 1000 µg), did not reach statistical significance but was also appreciably longer in OPFR exposed animals. This effect would likely reach significance with higher statistical power, but sample size was necessarily small because of the technical limitations of performing whole brain analysis, which is a limitation of the present study. Additionally, the distance between the most dorsal TH-ir cells and the cephalic flexure was significantly greater in sagittal sections of OPFR exposed fetal brains of both sexes than in controls in the 2D analysis. This height increase was confirmed by digital sagittal “sectioning” and 3D volumetric analysis. Importantly, E14 is before the prenatal androgen surge in rats, which is the primary driver of brain sexual differentiation (Clarkson et al., 2014). This is likely why no appreciable sex differences were observed in the 2D analysis herein, but OPFR exposure has been shown by us and others to have sex-specific effects on development and neuroendocrine function. Future studies should address this limitation by examining impacts to the DA system at later fetal ages, particularly after the postnatal androgen surge, and into postnatal development.

The most parsimonious conclusion regarding the increased height and length detected in the 2D and whole brain slice analysis is that the TH-ir cell cluster is larger in OPFR exposed animals. Interestingly, however, the 3D volumetric analysis revealed that the volume of the TH-ir cell cluster was actually significantly smaller in OPFR exposed males than controls. These results reflect the complexity of assessing the totality of neuronal migration in fixed planes because neuronal subsets are differentiating, separating, and moving in different dimensions to establish their eventual subregions (in this case the SNc and VTA). The 3D data suggest that while the medial portion of the TH-ir cell cluster (which will ultimately form the VTA) is larger in the dorsal-ventral dimension in OPFR exposed animals (based on the medial region from where the measures were made), it is smaller in other dimensions, thus reducing the total volume of the TH-ir cell cluster in the 3D analysis. Our approach demonstrates that while assessing morphological impacts to such a complex and irregularly shaped structure is achievable using classic approaches in “2D” thin sections, the level of detail afforded by whole brain clearing and light sheet imaging can reveal subtle changes to overall structure that are unappreciable in serial sections and can greatly augment data interpretation.

The mechanisms by which maternal OPFR exposure altered the shape and composition of the TH-ir cluster at this point of fetal development remains to be determined. In typical development, rostro-caudal patterning of the VTA/SNc is coordinated in part by the secreted morphogens wingless-int1 (Wnt1) and fibroblast growth factor 8 (Fgf8). The combined action and concentration of Wnt1 and Fgf8 organize the midbrain/hindbrain boundary and determine the permissive region for DA cell migration (Arenas et al., 2015). Cells in the developing midbrain gain rostro-caudal information from the concentration gradients of these morphogens. Interestingly, exposure to the OPE tris (2-butoxyethyl) phosphate has been shown to impair Wnt signaling in embryonic zebra fish heart, demonstrating some OPFRs can interact with this signaling cascade (Xiong et al., 2021). Future studies should examine whether OPFRs change the extracellular levels or distribution of these critical proteins and, consequently, alter the permissive region for DA cell cluster patterning thereby allowing anterior-posterior expansion of the dopaminergic cell cluster as we observed herein. The dorso-ventral patterning of the neural axis depends on (Briscoe and Ericson, 1999) precise cellular migration. DA fated neuroblasts migrate radially from the ventricular zone at E11 (Yang et al., 2013) along radial glia scaffolding that provide a guide for the migrating DA neurons. Recent findings suggest that radial glial migration is significantly retarded by OPFR exposure (Klose et al., 2021). If the formation of the radial glial scaffold, or timing of this process is disrupted by OPFR exposure, changes in the dorso-ventral dimensions of the DA cell cluster might occur.

OPFRs altered DAergic axogenesis

TH-ir fibers originating from the DA cell cluster in animals exposed to the low or high OPFR dose showed a significantly wider spread at the leading edge of axon growth at E14 than in controls in 2D analysis. This finding was reproduced in analogous sagittal sections of whole mount fetal heads obtained from dams fed the middle (1000 µg) dose. Volumetric measurement of TH-ir axons in the medial forebrain bundle revealed no difference in total volume between exposed males and controls. These results indicate that while the dorsal-ventral spread was larger in OPFR exposed animals, the volume of DA axons remained the same, at least in males. This may indicate a concurrent decrease in some other dimension of axon spread not accounted for in our measurements to result in a similar volume. Assessment at a later age, as the discrete subnuclei become more distinct would help resolve this.

Multiple molecular signals orchestrate mesocortical dopamine axon guidance and previous work has implicated semaphorin 3F (Sema3F) and its cognate receptor neuropilin 2 (Nrp2) in regulating DA axon passage in the ventral di-and telencephalon. Indeed, developmental exposure to the OPE TPHP decreased transcription of axon guidance genes in mouse hippocampus, suggesting OPFRs may interact with axon guidance signals in the brain (Zhong et al., 2021). Additionally, we recently demonstrated that gestational OPFR exposure significantly downregulated expression of axogenesis and axon guidance genes in male rat P1 cortex (Witchey et al., 2022). OPFR-related changes to the levels of expression or timing of guidance cues, including altered distribution, concentration, expression, or activity, might result in the disruption of DAergic axon guidance. Future studies examining multiple ages and guidance markers will be needed to further elucidate how OPFRs may be disrupting the ontogeny of the mesencephalic DA system and, in a broader sense, the DA system may be vulnerable to chemical disruption during fetal development. In later stages of development immunolabeling of DA neuronal subpopulations and their projections would also help identify impacted systems and thus potential, consequential, functional deficits.

No evidence of altered 5-HTergic cell cluster or axons by 3D analysis

The DA and 5-HT systems develop in parallel and are known to interact at different stages of development. Axons of both 5-HT and DA course together in the medial forebrain bundle, and both monoamine system projections target the same regions of the mPFC, and their innervation of this region is interdependent (Niederkofler et al., 2015). Of critical importance in this ontogeny is the timing and progress of typical DA and 5-HT axon growth. 5-HT ascending axogenesis begins a half day earlier than DA axogenesis (approximately E11 vs E11.5). Because, 5-HT efferents begin migration from a more caudal position than DA axons, the leading edge of 5-HT axons typically remains behind DA axons (Niederkofler et al., 2015). Given prior findings demonstrating that 5-HTergic axons of the medial forebrain bundle were significantly longer in OPFR exposed males compared with controls at E14 (Rock et al., 2020), we had hypothesized that 5-HT axons might impinge on the domain of growing DAergic fibers and displace them, resulting in the increased axon spread we observed. However, 3D volumetric analysis revealed the leading edge of 5-HT axons clearly trail those of DA (Figure 7), and we observed no change in 5-HTergic axon bundle volume or length.

This apparent discrepancy with prior findings could be a result of differences in the method of measurement of 5-HT fiber length between studies, as the previous study measured 5-HT length in reference to other fibers (Netg1a expressing fibers) which had variable, though not significantly different, length, and in this study, we measured the length of the fibers from their origin at the 5-HT-ir cell cluster. It is also possible we failed to detect a difference in 5-HT axon length due to a combination of a small effect size on 5-HT length and the small sample size used for the whole brain imaging. Although powerful, whole brain imaging has technical and other limitations that make performance on anything other than small sample sizes extremely challenging. 5-HTergic hindbrain nuclei (putative raphe) revealed no significant difference in total volume between exposed and control male offspring. These results comport with the 2D findings of our lab’s previous work demonstrating no change in the spread of 5-HT cell bodies, and support our previous findings that developmental exposure to OPFRs does not impact the gross neuroanatomy or volume of the fetal 5-HT nuclei. Significantly, these results suggest the observed change in DA cell cluster morphology and axon width are not a generalized effect in one dimension or another, but specific morphological effect to TH-ir structures.

No observed effect of OPFRs on TH-ir in adult mPFC

Using tissues from an extant study where animals were exposed to OPFRs across the perinatal period, we found no evidence that the density of TH-ir fibers in adult mPFC subregions (ACC, PL, IL) was appreciably different compared with controls. This may be due to the marked changes that occur in PFC innervation and circuitry during development. Across the postnatal and adolescent period, the PFC undergoes significant remodeling, reducing total volume (Casey et al., 2000) and number of synapses (Drzewiecki et al., 2016), whereas other measures of DA innervation increase over the adolescent period (Willing et al., 2017). Given the extensive remodeling of mPFC/DA-connectivity during the adolescent period, it seems possible that differences observed in early development (including E14) might be obviated by this process, resulting in a similar TH-ir density in adulthood. This remains to be definitively determined given that the adult tissues examined herein came from animals perinatally exposed to only a single dose. Plus, the dams in that project were slightly heavier (300 vs 250 g on average) than those used for the fetal measures, meaning fetal exposure was slightly lower compared with the tissues used for the fetal measurements herein.

Critically, adolescence may represent an additional critical period in which DA innervation is vulnerable to OPFR exposure and embryonic effects could be exacerbated. The cytoarchitecture of the VTA and SN in young and adult brain should be evaluated given the changes observed in OPFR exposed fetuses prior to the full formation of those regions. Other targets of the mesencephalic dopamine system should also be examined including hippocampus, striatum, and amygdala, particularly in light of published evidence that developmental exposure to some OPFRs, or FR mixtures containing them, can impact relevant locomotor-dependent behaviors including startle responses and general motor activity (Baldwin et al., 2017; Oliveri et al., 2015; Patisaul et al., 2021; Wiersielis et al., 2020; Witchey et al., 2020). 5-TH also modulates these and other behaviors. DA cells of the medial midbrain (VTA) and their innervation of the mPFC are essential for the exercise of executive functions and related behaviors such as cognitive control, or the ability to direct appropriate behavior in reference to a goal (Ott and Nieder, 2019). Cognitive control is comprised of several cognitive abilities including sensory gating and attention, working memory, rule switching, and execution of motor function. When combined, these behaviors allow for adaptive, purposeful responses to diverse stimuli. Each of these behaviors relies on the appropriate functioning of neuromodulatory effects (enhancing or suppressing synaptic signals in the mPFC) of the mesocortical DA circuit, and dysfunction in these connections can result in impairment of cognitive control (Arnsten and Li, 2005; Clark and Noudoost, 2014; Vijayraghavan et al., 2017). Dysregulated prefrontal network activity is implicated in the pathology of multiple neurodevelopmental disorders including schizophrenia, autism, attention deficit disorder, and bipolar disorder, along with mental health conditions with a developmental component such as post-traumatic stress disorder and depression (Ranganath and Jacob, 2016; Kosillo and Bateup, 2021). Unregulated DA signaling in the mPFC is a hallmark of many of these disorders. As such, future studies should examine the potential behavioral effects of gestational OPFR exposure on mPFC-mediated behaviors. This is particularly important given the growing evidence that OPEs, including those examined herein, could adversely impact child cognitive, fine motor, and language function (Doherty et al., 2019a,b,c).

Limitations to generalizability

There are several factors which limit the generalizability of the current results. First, because we used existing tissue generated for prior studies, and some 3D samples were inadvertently damaged in the analysis, we had unequal sample sizes, potentially leading to a general reduction in power but, importantly, not unequal variances. Risk of higher type 1 error rates hinges more on the latter than the former (Keppel, 1993), and our chosen statistical approach was appropriately conservative and adjusted for the minor sample size differences. Additionally, only 4 sires were used in the 3D experiment which could theoretically contribute variance based on paternity, but we found no evidence of that. Obviously, the results are limited to the specific regions we examined, and the analysis was not exhaustive, particularly in the adults. Several other regions innervated by DA and 5-HT, including striatum and hypothalamus, should be a part of future work. Finally, as noted above, only fetal and adult animals were used in this study. A substantial portion of monoaminergic development in the cortex occurs postnatally and during adolescence. OPFR exposure could potentially have effects on the development of these systems during that time, thus to gain a full picture of the effects of OPFR exposure on these systems these time points must be investigated. Overall, these factors may limit some of the generalizability of our results, particularly across developmental stage and region.

Conclusions

OPFR use in consumer and commercial goods is increasing as legacy BFRs are discontinued, and as such, they and their metabolites have become a ubiquitous presence in the human and wild environment. This unavoidable exposure coupled with their apparent effects on the developing dopamine system demonstrates the pressing need to assess the potentially deleterious effects of OPFRs on human neurodevelopment and behavior. Here, we demonstrated that environmentally relevant OPFR exposure during gestation disrupts development of the rat mesencephalic DA system. Dam oral exposure from day of conception to E14 produced a significant increase in the dimensions both rostral-caudal and dorsal-ventral aspects of the medial midbrain DA cell cluster in embryos of both sexes, whereas the spread of TH-ir axons as they coursed to the ventral telencephalon was wider, as observed with traditional “2D” IHC. 3D whole fetal head imaging supported and further informed our findings, indicating that the volume of midbrain DA cell cluster is smaller in males exposed to gestational OPFRs and spread of TH-ir axons in the ventral telencephalon was wider in the dorsal-ventral axis, but of a similar volume. Conversely, we observed no significant effect of OPFR exposure on 5-HTergic hindbrain nuclei or ascending fiber volume or length, confirming that the OPFR effects on the DA system reported herein are system-specific. Taken together, these results suggest gestational OPFR exposure disrupts the developmental processes governing DA cell cluster cytoarchitecture and axogenesis and highlight the added explanatory nuance conferred by whole mount clearing and 3D visualization techniques. The present findings expand on our previous research on 5-HT morphology in the same animals by demonstrating gestational OPFR exposure likely perturbs developing monoamine neurotransmitter systems generally. The persistence and functional consequences of the observed disruptions to DA axogenesis remain to be shown.

Acknowledgments

The authors gratefully acknowledge Heather Stapleton and her lab for preparing the dosing solutions as well as the animal care staff at the Biological Research Facility at NC State for providing animal care and husbandry.

Funding

Comparative Medicine Institute of North Carolina State University; and National Institute of Environmental Health Sciences (ES028110 to HBP and P30ES025128)

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References