Persisting Neurobehavioral Consequences of Gestational exposure to Cadmium and Benzo[a]pyrene in Rats

Abstract

Neurotoxic risks in the environment come from many toxicants, which are often found together in complex mixtures. However, nearly all experimental studies evaluate one chemical at a time. Neurobehavioral effects of developmental exposure to heavy metals such as cadmium (Cd) and polycyclic aromatic hydrocarbons (PAHs) such as benzo[a]pyrene (BaP) have been well-studied, however their potential for non-additive or interactive effects are not well known. We recently reported that, in zebrafish, embryonic exposure to the PAH benzo[a]pyrene (BaP) and/or the heavy metal cadmium chloride (CdCl₂) led to selective sub-additive effects on behavior. The current study was performed in rats, to show whether such interactions translate to mammals and to better account for characteristics like biological sex. In this study, we exposed female rats to BAP (0.03 mg/kg/day), the metal salt CdCl₂ (0.3 mg/kg/day) or both via osmotic minipumps throughout gestation. Male and female offspring were assessed for bodily and reflex development, and locomotor, emotional and cognitive function. Cd treatment was associated with impaired sex differences in neonatal anogenital distance, enhanced negative geotaxis performance on PND7, reduced body weight at weaning, increased open-arm exploration in the elevated plus maze (females only), and reductions in sex differences in novel object recognition. Co-treatment with BaP attenuated those CdCl-effects on negative geotaxis and elevated plus maze. BaP was also associated with reduced metrics of food consumption in the novel environment suppressed feeding task, and with increases in errors during the initial phase of radial arm maze training (males only). As in zebrafish, persisting neurobehavioral effects are seen in rats after chronic developmental exposure to BaP and CdCl. However, these effects can differ between single-exposures and mixtures, which indicates a need for greater clarity on interactions within such mixtures.

1. Introduction

Developmental neurobehavioral toxicity of heavy metals and polycyclic aromatic hydrocarbons (PAHs) have been widely studied with individual exposures. But neurobehavioral consequences of combined exposures have not received much attention even though combined exposures are more common. Complex environmental mixtures often contain large numbers of compounds from varying classes of toxicants. This is particularly true of mixtures generated by combustion, as with cigarettes (Pappas, 2006; Richter, 2016), vehicle emissions (Alves, 2015; Chen, 2022; Coufalík, 2019), fly ash and other emissions from coal use (Donahue, 2006; Sarver, 2019; Verma, 2015), and wildfires (Boaggio, 2022). Although much of what is in these mixtures could be unique to the materials being combusted (e.g. plant material, coal, gasoline, etc.), these mixtures frequently contain an array of polycyclic aromatic hydrocarbons (PAH), multi-ring products of incomplete combustion, and metals which originated in that material. For example, wildfire smoke can contain a wide variety of metals, with the metal make-up of wildfire smoke being strongly associated with the number of man-made structures destroyed by those fires (Boaggio, 2022). Tobacco smoke, on the other hand, typically contains measurable amounts of cadmium (Cd), based on the tobacco plant’s high uptake of Cd from the soil (Lugon-Moulin, 2006). However, the concentrations of Cd and BaP found in tobacco are lower than doses used in the current study (Bukowska et al., 2022; Caruso et al., 2014). Little data is available directly comparing the toxicity of these different mixtures, although it has been proposed that differences in their relative PAH/metal content may influence their relative risk (Cui, 2019).

These two classes of toxicants have significant overlaps in their toxicological profiles, with developmental neurotoxicity being one of the best studied and most concerning areas of overlap. At the cellular and neurochemical levels, developmental exposures can modify key neurotransmitter systems, including monoamines, acetylcholine, GABA and glutamate (Abd El Naby, 2023; Bensoussan, 2009; Karri, 2016; Leret, 2002; Marchetti, 2005; McCallister, 2008; Slotkin, 2019a), and key aspects of cell/circuit development or plasticity, including neurogenesis, synaptic plasticity and long-term potentiation (Chow, 2008; Dou, 2011; Sadiq, 2012; Salehi, 2015; Zhang, 2022; Zhao, 2018). Unsurprisingly, based on these effects, developmental exposures to PAHs or metals have also been found to produce behavioral deficits in laboratory animals, particularly rodents. Such deficits include alterations in locomotor activity, anxiety-like behavior, and learning (Chen, 2012; Hawkey, 2019; Kamel, 2011; Lamtai, 2021; Sprowles, 2018; Zhang, 2016a), effects which provide support for clinical links between these toxicants and psychiatric symptoms in humans, including ADHD and other cognitive or learning impairments (Bauer, 2020; Chandravanshi, 2021; Jedrychowski, 2015; Rezaei Kalantary, 2020; Zhen, 2023).

As PAHs and metals are both generally associated with developmental neurotoxicology and neurobehavioral dysfunction, it would be reasonable to hypothesize that individuals exposed to both toxicants would predictably show more severe or more numerous symptoms, as compared to those with single-class exposures (e.g. PAH without metal, metal without PAH). The typical statement of this hypothesis is the “additivity hypothesis”, wherein toxic effects of mixtures are predicted by adding together the toxicity values provided by single-compound studies (Rider, 2018). However, prior mixture studies have reported that mixture effects can be unpredictable, with binary mixtures producing sub-additive, additive or supra-additive effects, depending on the endpoint and make-up of the mixture. Prominent examples include analyses of within-class mixtures containing PAHs (Gaskill, 2016; Pushparajah, 2018; Staal, 2007), polychlorinated biphenyls (PCBs) (Ramamoorthy, 1997), per- and polyfluoroalkyl substances (PFAS) (Fey, 2022; McCarthy, 2021), and metals (Barata, 2006; Utgikar, 2004). Currently, mixture interactions across-class remain poorly understood, although available data suggests that non-additive effects may also be observed in cross-class mixtures (Hawkey, 2019; Morais, 2018; Newsted, 1995; Slotkin, 2013; Yang, 2022).

We have recently reported non-additive effects of PAHs (benzo-a-pyrene, BaP; or fluoranthene, Fa) and the neurotoxic metal salt cadmium chloride (CdCl₂) in a zebrafish model (Stickler, 2024). When present, these were sub-additive, with one or more mixture groups failing to show behavioral deficits associated with the PAH or CdCl₂. Of note, these relationships were highly selective, with only certain behavioral tests and certain ages showing non-additive effects, while the majority of effects were either additive or linked to only one of the two exposures. Combinations of BaP and CdCl₂ tended to produce sub-additive effects on locomotor and sensorimotor tests, but only in relatively young animals. By contrast, combinations of Fa and CdCl₂ produced sub-additive effects on locomotor and fear-like tests, but only in relatively mature animals. These results suggest that interactions between PAHs and Cd are likely in the developing nervous system, but that these interactions may be relatively specific and complex. Even so, zebrafish are less complex than mammals in brain and behavior and have limitations in the testing of cognition and identification of sex differences (Vorhees, 2021). Thus, validation of these interactions in a mammalian model would be beneficial to expand upon those findings and aid in their translation to human environmental health.

The present study was performed to evaluate the long-term neurobehavioral effects of pre/perinatal exposure to BaP and/or CdCl₂ in a rat model. BaP was selected as the representative PAH, due to its common use in neurobehavioral toxicology studies and its “classic” PAH actions, which center on activation of the aryl-hydrocarbon receptor (Tartaglione, 2023). Cd was selected as the representative metal, due to its presence in PAH-containing tobacco smoke and established neurotoxicity (Malin, 2018). Our aim was to determine which aspects of behavior across maturation were adversely impacted by pre/perinatal exposure to these compounds, and the degree to which these effects may differ in animals which were co-exposed to both compounds. Based on prior studies, we hypothesized that both BaP and CdCl would impact behavior, but that specific aspects of behavior would display non-additive interactions between the two compounds. To do this, female rats were implanted with two osmotic mini pumps containing vehicle (DMSO), BaP (0.03 mg/kg/d) or CdCl₂ (0.3 mg/kg/d), then mated with untreated males to produce control litters and litters exposed to BaP, CdCl₂, or both. Offspring were then tested on a battery of growth and behavioral assessments beginning in the neonatal period and extending into adulthood. These included tests of physical growth, motor function, locomotion, affective-like behaviors, and cognitive performance.

2. Methods

2.1. Subjects and Housing

The animals used in this work were young adult female and male Sprague Dawley rats approximately three months of age obtained from Charles River Labs (Raleigh, NC, USA) and their offspring, which were bred in house in the Levin lab at Duke University. Standard housing consisted of social housing (2–3 rats per cage, except for dams post-mating) with corn cob bedding provided, as well as continuous ad libitum access to food/water, except when noted otherwise (food restricted during novel environment suppressed feeding and radial-arm maze testing). Housing rooms were maintained on a 12/12 reversed day:light cycle and all animal maintenance and testing were conducted during the dark phase of the cycle. All offspring were housed with same-age and -sex rats from their same treatment condition (2–3 rats per cage). All surgical, husbandry and testing procedures were conducted in compliance with federal guidelines for the use of rodent species in research and were approved by the Institutional Animal Care and Use Committee of Duke University.

2.2. Minipump Implantation, Mating and Exposure

Exposures and rat offspring were generated as described previously (Hawkey, 2019). Thirty-six adult female Sprague Dawley rats (3 cohorts of 12 breeders) and their male mates were obtained and acclimated to the new housing/light cycle conditions for one week prior to pump implantation surgery. Benzo[a]pyrene (>96% pure) and cadmium chloride (CdCl₂>98% pure) were obtained from Sigma-Aldrich, Inc. (St. Louis, MO, USA). Prior to surgery, these rats were weighed and the solutions for the pumps were made, such that the Alzet osmotic infusion pumps (2ML4, Durect Inc, Cupertino, CA, USA) would deliver weight-adjusted amounts of CdCl₂ (0 or 0.3 mg/kg/d CdCl₂) or BaP (0 or 0.03 mg/kg/d BaP) throughout the life of the pumps (Cd salt weight used in dosing calculations). CdCl₂ and BaP were dissolved in 100% DMSO. Each rat was then surgically implanted with two pumps (left and right side of the animal), subcutaneously under anesthesia (60mg/kg ketamine, Covetrus, Dublin, OH, USA, and 15mg/kg dexdormitor, Dechra, Overland Park, KS, USA). The manufacturer rating for these pumps is 4 weeks of consistent release (28 days), with the contents of the pump being fully exhausted in ~5 weeks. Incisions were located on the midline of the back, between the shoulder blades, and pumps were positioned along the ribcage on either side of the rat’s body. Groups were generated based on the solutions loaded into the pumps: control (DMSO vehicle, both pumps), BaP only (BaP in 1 pump, vehicle in the other), CdCl₂ only (CdCl₂ in 1 pump, vehicle in the other), and both (BaP in 1 pump, CdCl₂ in the other). These rats were allowed 72 h to recover from surgery and then housed with untreated males for a 5-day mating period. After parturition, developmental monitoring and behavioral testing proceeded, as described below.

On PN1, litters were weighed and culled to 10 pups, providing equal numbers of male and female pups as available. At weaning (PN21), one male and one female were randomly selected from each litter for behavioral testing. These were separated from their litter and housed with same-sex and same treatment individuals from other litters for the remainder of the experiment. All remaining siblings were retained for tissue collection for separate neurochemical studies.

2.3. Developmental Monitoring of Growth

Across the pre-weaning stage, offspring were measured for body weight and anogenital distance (AGD) as described previously (Hawkey, 2024), to provide metrics of physical growth. Briefly, pups were weighed using a laboratory balance (precision, 0.1 g) and AGD was taken with a digital caliper. These measurements were taken on the following dates: PN1, 2, 4, 7, 10, 17, and 21 to generate growth curves.

2.4. Behavioral Testing Design and Conditions

Behavioral testing consisted of a battery which began during the neonatal stage and extended into adulthood. Assessments were spaced out such that only one test was performed within a given week of development. As noted prior, testing occurred during the dark phase of the reversed day/night cycle (800 h–1700 h), and only low, ambient lighting was used during testing to minimize any effects of light on circadian rhythm.

2.4.1. Pre-weaning Reflex Development

Neonatal righting reflex testing was conducted, as described previously (Hawkey, 2024). This is an innate reflex, whereby rats will roll over into a prone position (on their bellies, when placed onto a flat surface in a supine position (on their backs). This motor reflex onsets 1–2 days after birth in rats (Rousset, 2013) and demonstrates gross motor function, as it requires the coordination of many muscle groups and all four limbs in order to complete. Testing occurred on PN 2, 4 and 6, to generate a developmental curve for the righting reflex. To conduct this test, the experimenter would place one pup down in a supine position onto the benchtop and record the latency of successful righting (achieving a prone position) using a stopwatch. Pups were allowed a maximum of 30 sec to complete this response. A failure to complete this response within 30 seconds (an omission, generally only seen on PN2) resulted in a 30 sec timeout, and up to two additional attempts. Three consecutive omissions resulted in a termination of testing for that animal, and a default score of 30 sec (as the maximum allowed time). The default score was used so that poor-performing individuals would not be excluded from the analysis, which would bias the results.

Neonatal negative geotaxis testing was conducted, as described previously (Hawkey, 2024). This is an innate reflex, whereby rats will reorient themselves to face up a slope, when placed onto a sloped surface facing down the slope. This motor reflex onsets late in the first week after birth in rats and reaches asymptote by ~13 days of age (Rousset, 2013). This motor reflex demonstrates gross and fine motor function, as it requires paw grip of the surface, and fine coordination/timing of all four limbs and paws in order to complete. This testing occurred on PN 7, 9, 11 and 13, to generate a developmental curve for the negative geotaxis reflex. To conduct this test, the experimenter would place one pup down in a prone position onto a wire mesh surface, set at a 45° angle, facing down the slope. Pups were allowed a maximum of 60 sec to reorient 180° to face up the slope. Omissions (generally only seen on PN7) were noted and treated the same as for righting reflex (30 sec timeout), with 60 sec as the maximum allowed time and default score.

2.4.2. Week 4: Elevated Plus Maze

The first test in the post-weaning phase of testing (4 weeks of age) was the elevated plus maze (Med Associates, St Albans, VT, USA). The maze consisted of a central hub (11.5 × 11.5cm) and 4 arms (11.5 × 50cm) extending out from that hub. Two arms were enclosed by 15-cm high black acrylic walls (closed arms), and the other two were left open, bordered by 2-cm railings. Rats were individually placed into the central hub, facing an open arm, and allowed to explore the maze for 5 min, while the experimenter manually scored the rats’ exploration. Scoring included the time spent in the open arms of the maze and the number of center crossings. Rats generally prefer the enclosed arms of the maze, a pattern which is often interpreted as anxiety-like behavior or risk aversion, as avoidance of open spaces would reduce the risk of predation in wild rats.

2.4.3. Week 5 and Adulthood: Figure-8 Locomotor Activity Test

The second test, the Figure-8 maze, was repeated twice during the post-weaning battery: once at 5 weeks of age, and again at 7 months of age (full adult). These two dates were selected to provide locomotor data within and following maturation, as in our previous study on BAP-containing mixtures (Hawkey, 2019). The figure-8 maze is a test of locomotor activity and locomotor habituation, wherein rats freely explore a completely enclosed set of tunnels (no open spaces) for 1 h. The figure-8 maze chamber (70cm × 42cm) was made of sheet metal walls arranged to create a continuous 10cm × 10cm alley in the shape of the number 8 with two side alleys. This chamber was closed with a clear Plexiglas lid, and housed on a partially enclosed (3 sides, floor and ceiling were sheet metal) shelf to minimize light exposure. Activity was recorded by photobeams placed at 8 equally spaced locations throughout the maze. Beam breaks were recorded and logged across 12, 5-min session blocks.

2.4.4. Week 6: Novelty Suppressed Feeding

As offspring entered adolescence (6 weeks of age), they were tested in the novel environment suppressed feeding test. This is a test of fear/stress responsivity in a novel, open environment, wherein responses generally suppress food consumption, even when the animals have been fasted beforehand. Rats were food restricted for 24h and then individually brought to an animal testing room which they had never entered before. The rat was then placed into a cage identical to its home cage, located on the floor in the center of the room and without a lid or bedding provided. Pre-weighed rat chow pellets were available on the floor of this cage, in 4 rows of 3 pellets. Rats remained in this cage for 10 min, and an experimenter seated across the room scored the animal’s eating behavior in real time: latency to first eat, total time spent eating, and number of bouts of eating. After the session ended, the experimenter then reweighed the pellets to determine the weight of food consumed during the session. For scoring purposes, eating was defined as chewing the food (visibly, or with an audible crunching sound), and did not include non-eating responses such as sniffing, holding or carrying the food pellets.

2.4.5. Week 7: Novel Object Recognition

The following week (week 7), adolescent offspring were tested in the novel object recognition test. This is a test of recognition memory, a basic memory function which does not require enhanced motivation (food restriction, etc.) or repetitive testing. The testing apparatus was a 10 cm × 41 cm black plastic box, which was either empty (acclimation phase) or contained two objects which could be explored. Prior to the testing sequence, rats were acclimated to the box twice (10 min per day for two days) with no objects provided. Following this, two sessions were performed: AA and AB. During the AA session, two identical objects were located in opposing corners at one end of the box, and rats were allowed to explore the box and the objects for 10 min. One hour later, rats were returned to this box for AB session. During this session, one of the provided objects was identical to the ones provided in AA (Object A), and the other was a novel object (Object B), which was matched to Object A in terms of size, coloration, and material (e.g. glass, ceramic, plastic). Rats were allowed to explore the box and the objects for 10 min while a camera mounted overhead recorded their exploration. The objects (4 pairs of objects) and object locations were pseudo-randomly assigned to each animal prior to testing, with this assignment counterbalanced across sexes and treatment groups. To remove odor cues, the objects and testing box were wiped with a 10% vinegar solution between trials. Videos of the sessions were post-scored for object investigation, specifically defined as sniffing the objects. All other contacts with the object (non-nose touching, climbing, etc.) were excluded. Scores were then used to generate the following endpoints: time spent investigating the familiar object, time spent investigating the novel object, and the difference in investigation between the two co-presented objects (preference). Scores were tabulated for the first and second 5-min blocks of the session, as novel object preference does tend to diminish over time and the change in each of these scores from the first to second 5 min block.

2.4.6. Weeks 8–11: Radial-Arm Maze

The final component of the main battery, completed from 8–11 weeks of age, was the 16-arm radial arm maze. This is a test of spatial learning and memory, which includes metrics of both working (within-session) and reference (across-session) memory. The maze consisted of a central hub (50 cm diameter) with 16, 10cm × 60 cm arms extending from it. During exploration of the maze, rats could retrieve pieces of sugar-coated cereal (Froot Loops^®; Kellogg’s Inc, Battle Creek, MI, USA) from cups located at the end of specific arms. Visual cues in the room were available to allow these rats to learn and navigate to the arms which contained these food reinforcers. Briefly, training and testing was laid out as follows. Rats were acclimated to the food and to the testing room, with food first provided in their home cages, then in the central hub (arms were blocked off) of the maze over two acclimation sessions (10 min per day for 2 days). Following this, twelve sessions (1 per day) were performed where the rat was allowed to explore the entire maze and retrieve the hidden food reinforcers. In each of these sessions, 4 randomly-assigned arms were left unbaited (no food), and these empty locations were kept uniform for a given rat across all of testing. As such, an entry to one of these arms was scored as a reference (long-term memory) error. The remaining 12 arms were baited (food provided), and within any given session, the food was not replaced after it was retrieved. An entry to one of these arms, after the food was already retrieved, was scored as a working (short-term memory) error. Sessions ended either when all food was retrieved, or when 10 min elapsed, whichever came first. All arm entries were logged by an experimenter and the two error types, as well as overall number of arm entries, which was used to calculate arm entry latency (arm entries/time, or the average interval in between each arm entry).

2.5. Data Analysis

All data were analyzed using analysis of variance (ANOVA) with litter (exposed dam) as the key unit of variance. CdCl₂ and BaP treatments were treated as between-litter variables and sex was treated as a within-litter variable (male and female siblings treated as repeated measures). Other repeated measures were included (e.g. time block, session, cue, etc.) as relevant to each test. For all main effects and post hoc tests, the threshold for statistical significance was set p<0.05 (two-tailed). For interactions, follow-up analyses of main effects were performed if the interaction term met a criterion of p<.10 (Snedecor, 1967), but the threshold for statistical significance was still set as p<0.05 for those post hoc tests. Post hoc comparisons were performed using Fisher’s Least Significant Difference test.

3. Results

Significant effects of CdCl₂ or BaP, and interactions between CdCl₂ and BaP are summarized in Table 1. The litter numbers per group were Control=10, CdCl₂=10, BaP=10 and CdCl₂+BaP=9. No significant effects were seen with CdCl₂ or BaP doses used on fertility rates. The doses used in this study were not observed to cause clinical signs of overt toxicity or dysmorphogenesis.

Table 1.

Summary of effects attributable to Cd or BaP, and interactions of Cd and BaP.

Age	Test	Cd	BaP	Interaction
PND 1–10	AGD	Cd reduced sex difference in AGD.	-	-
PND 2–6	RR	-	-	-
PND 7–13	NG	-	-	Cd-alone faster vs controls on day 7. Co-exposure not affected.
PND 21	Wt	Cd reduced pre-weaning body weight.	-	-
4 weeks	EPM	-	-	Cd (females) increased open arm time. Co-exposure not affected.
5 weeks	Fig-8	-	-	-
6 weeks	NES		BaP reduced food eaten and feeding bouts.
7 weeks	NOR	Cd reduced sex difference in NOR.	-	-
8–11 weeks	RAM	-	BaP increased total errors (males) during first half of testing (sessions 1–6).
7 months	Fig-8	-	-	-

Abbreviations: AGD, Anogenital distance; EPM, Elevated Plus Maze; Fig-8, Figure 8 maze; NES, Novel Environment Suppressed Feeding; NOR, Novel Object Recognition; NG, Negative Geotaxis; PND, Postnatal Day; RAM, Radial Arm Maze; RR, Righting Reflex; Wt, Body Weight.

3.1. Birth and Preweaning Morphometric Measurements

Litter size, sex ratio within litter or birthweight were not significantly affected by CdCl₂ and/or BaP treatment. Within neonatal growth measurements, a CdCl₂ x Sex interaction was observed for anogenital distance, F(1,34)=5.41, p<0.05, regardless of age. Developmental CdCl₂ exposure led to a significant (p<0.05) reduction in the typical sex difference in AGD (M>F), regardless of BaP exposure. Specifically, CdCl₂-treated offspring showed a 10.5% reduction in the AGD difference between males and females during the neonatal period (PND 1–10). Later in development, pre-weaning body weight (PND 21) was also found to be significantly reduced by CdCl₂ exposure (p<0.05) regardless of sex and BaP exposure.

3.2. Developmental Reflex Measurements

Neonatal righting reflex measurements were taken on PND 2, 4 and 6. The analysis of latency to right detected a BaP x Sex x Day interaction (F(2,62)=4.46, p<0.025). Post-hoc tests were then performed collapsing across CdCl₂-treatment. However, no comparisons reached significance within this analysis.

Neonatal negative geotaxis was measured on PND 7, 9, 11 and 13 (Fig. 1). The analysis of the latency for the pup to turn 180° to have their head turned upward detected an interaction of BaP x CdCl x day (F(3,102)=4.53, p<0.01) was followed-up by tests of CdCl₂ x BaP interactive effects at each day of testing. On PND 7, a significant CdCl₂ x BaP interaction (F(1,34)=5.58, p<0.01), was seen with CdCl₂ exposed offspring having significantly (p<0.025) shorter response latencies than controls. No additive effect of d and BaP was seen. In fact, the comparison between the CdCl₂ and CdCl₂+BaP groups showed a trend of attenuated effect that failed to reach significance (p<0.08). No significant effects were seen on the later days of testing.

Figure 1.

Negative geotaxis, response latency (mean±sem). There was a significant interaction of Cd x BaP x Day of testing (p < 0.025) on geotaxis latency. Cd-alone treatment led to improved scores (lower latency to geotaxis) on day 7 when compared to controls (p < 0.025). A similar trend relative to Cd+BaP failed to reach significance (p < 0.08).

3.3. Elevated Plus Maze

Percent of exploratory time spent in the open arm of the elevated plus maze is a measure of anxiety-like behavior, with lower percent open arm time indicating greater anxiety-like behavior. There was an interaction of BaP x CdCl₂ x Sex (F(1,35)=7.26, p<0.025) on this measure that prompted follow-up tests of the interactions of BaP and CdCl₂ in each sex. In females, CdCl-alone caused a significantly (p<0.05) higher percent open arm time than controls. No additive effects of CdCl₂+BaP exposure were seen. In fact, percent open-arm time in the combined CdCl₂+BaP exposure group showed a nearly significant difference (attenuation) (p<0.08) compared with the CdCl₂ exposed females (Fig. 2). Males did not exhibit any significant effects on this measure. With the measure of center crossings of the elevated plus maze, which represents the level of activity and exploration in the test, there was an interaction of BaP x sex (F(1,35)=3.07, p<0.09), that prompted follow-up tests of the simple main effects of BaP on center crossings in each sex. However, none of the group comparisons for each sex reached significance.

Figure 2.

Elevated plus maze, percent time on the open arms (mean±sem). There was a significant interaction of Cd x BaP x Sex (p < 0.025) on open arm time. No effects were observed in male offspring. Cd-alone treatment led to increased risk-taking-like behavior (increased open arm exploration) among females when compared to controls. A similar trend relative to Cd+BaP failed to reach significance.

3.4. Locomotor Activity in the Figure-8 Maze

In the overall analysis of locomotor activity of rats tested as adolescents and adults there was a significant interaction of CdCl₂ x sex x session time block (F(11,385)=2.38, p<0.01). Figure 3 shows the time-effect functions for male and female offspring in the control, CdCl₂, BaP and combined CdCl₂+BaP groups averaged across age of adolescent and adult age of testing. Follow-up tests of the simple main effects of CdCl₂ on sex differences at each time block did not provide evidence for significant differential effects of CdCl₂ exposure on sex differences in locomotor activity.

Figure 3.

Locomotor activity of offspring in the Figure-8 maze (mean±sem). There was a significant interaction Cd x sex x session time block (p < 0.01) for locomotor activity. Data is provided collapsed across age (adolescent and adult), however comparisons between Cd-treated and Cd-nontreated offspring were not significant when analyzing individual time blocks within the session.

3.5. Novelty Suppressed Feeding

In the analysis of the amount of food eaten, there was a main effect of BaP exposure (F(1,35)=4.17, p<0.025). BaP-exposed offspring consumed significantly less food (−18.7%) regardless of CdCl exposure (Fig. 4). Likewise, a main effect of BaP was observed for the number of feeding bouts (F(1,35)=6.08, p<0.025). BaP-exposed offspring showed significantly fewer feeding bouts (−27.1%) regardless of CdCl₂ exposure (BaP exposed=8.08±0.72, Non-BaP exposed=11.08±0.80). With the latency to begin feeding and total duration of feeding measure, no significant treatment effects were detected.

Figure 4.

Novelty suppressed feeding, grams eaten (mean±sem). There was a significant main effect of BaP treatment (p < 0.025) for the amount of food consumed during the test. BaP-treated offspring generally consumed less food than controls, regardless of Cd co-exposure.

3.6. Novel Object Recognition

Validating the veracity of the novel object recognition test, the analysis of novel vs. familiar object was significant, F(1,34)=31.17, p<0.0001 with greater investigation of the novel object. The familiar vs novel x 5-min time block interaction was also significant, F(1,34)=5.17, p<0.05. In the analysis of novel object recognition (Novel object - Familiar object investigation) there was an interaction of CdCl₂ x sex x 5-min time block, F(1,34)=2.99, p<0.10, that was followed up by tests of the simple main effects of CdCl₂ on sex differences during the first and second 5-min time block. As shown in figure 5, CdCl₂ significantly (p<0.05) reduced the sex differences in novel object recognition, regardless of BaP exposure, during the second time block.

Figure 5.

Novel object recognition, novel-familiar object investigation in seconds (mean±sem). There was a significant Cd x Sex x Time Block interaction (p < 0.10) for novel object preference. Cd-treatment significantly reduced the sex difference in novel object preference observed during the second half (min 6–10) of the session, regardless of BaP co-exposure.

3.7. Radial-Arm Maze

There was a significant main effect of sex (F(1,35)=8.64, p<0.01) with males (17.1±0.5) having fewer total errors/session than females (19.2±0.7). Evaluation of toxicant exposure effects in males and females showed that with males there was a significant interaction of BaP x session-block ((F1,35)=4.32, p<0.05), whereas with females no significant exposure effects seen. Follow-up tests of the simple main effects of BaP in males in the different session blocks, showed that BaP caused a significant increase in total errors/session in sessions 1–6 (p<−0.025) but not in sessions 7–12 (Fig. 6).

Figure 6.

Radial-arm maze total errors/session for sessions 1–6 (mean±sem). For males there was a significant interaction of BaP x session block (p<0.05). Male BaP-treated offspring made significantly (p<0.025) more errors during the initial training phase of testing (Sessions 1–6) compared to offspring not exposed to BaP, regardless of Cd co-exposure. No effects of exposure on radial-arm maze errors were observed in females.

With error type (repeated measure, working vs. reference memory), there was a CdCl₂ x Sex x error type x session block interaction, F(1,35)=6.95, p<0.025), that triggered follow-up tests of the simple main effects of CdCl₂ exposure for males and females for working and reference memory errors at each session block. These follow-up tests did not detect any significant effects of developmental CdCl₂ exposure at the individual session blocks for each type of error for male and female offspring.

Working memory errors measure showed an interaction of BaP x session block, F(1,35)=4.39, p<0.05. Follow-up tests of the simple main effects of BaP treatment at each session block did not detect any significant effects averaged across both sexes, although the difference for sessions 1–6 was nearly significant (p<0.06). The rats without developmental BaP had 17.4±0.7 total errors and those with developmental BaP exposure had 19.5±0.8 errors.

No significant treatment effects were seen with latency to respond in the radial-arm maze.

4. Discussion

The present study was performed to determine the neurobehavioral effects of early life exposure to CdCl₂ and/or BaP in rats across early development, testing the independence and interactive effects of these exposures. To produce this exposure, dams were implanted with osmotic minipumps prior to mating. These pumps provided constant exposure to these compounds throughout the prenatal period and were left in place throughout the postnatal nursing period. Pups underwent assessments for bodily growth, early motor reflexes, and a variety of locomotor, affective and cognitive functions extending from the early postweaning stage to adulthood. Based on prior literature, it was hypothesized that CdCl₂ and BaP would each modify neurobehavioral outcomes, but that interactions may be present wherein co-exposed groups show altered effects relative to single exposures. In the current study, the concentrations of these two toxicants were set to 0.3 mg/kg/day for CdCl₂ and 0.03 mg/kg/day for BaP, which is consistent with dose ranges adopted in prior animal studies in this field (reviewed below). These doses were selected to ensure that both toxicants would be bioactive in rodents, considering their higher metabolic rates and differential sensitivity to PAHs acting on Ah receptors and, realizing that non-additive effects, and sub-additive effects in particular, cannot be detected unless the compounds are both bioactive. As a result, these are higher than the levels associated with smoking (Bukowska et al., 2022; Caruso et al., 2014), and so should not be assumed to represent the specific effects and interactions that exist in the children of smokers. In addition, with tobacco smoke exposure there is always an over-riding neurotoxicological effect of nicotine exposure, which was not addressed in this work. Rather, these should be viewed as representing the degree to which overlaps in the bioactivities and toxicities of CdCl₂ and BaP could impact the predictability of their effects when in a mixture.

Expanding upon our hypotheses, our parallel work with the zebrafish model (Stickler, 2024) suggested that CdCl₂-BaP interactions may be highly selective, only impacting certain tests or at certain ages, and may represent sub-additive effects rather than supra-additive effects. Our data generally supported these hypotheses. The developmental test battery detected several specific effects of CdCl₂ or BaP on growth, motor, affective and cognitive function. CdCl₂ was uniquely associated with a reduction in typical sex differences in anogenital distance during the neonatal period, reduced body weight at weaning, and reduced sex-differences in novel object recognition. BaP was uniquely associated with reduced food consumption in the novel environment suppressed feeding test and increased errors among male offspring in the radial arm maze. Notably, multiple measures showed that the combination of CdCl₂ and BaP exposure produced less impairment than single exposure to either compound alone. These included a loss of CdCl₂-induced enhancements in early negative geotaxis performance and reduced anxiety-like behavior among female offspring in the elevated plus maze, due to BaP co-exposure. This generally indicated that these compounds produce an array of general compound-specific effects, alongside potentially sub-additive interactions between those same compounds.

A closer comparison between the preceding zebrafish study with CdCl₂ and BaP (Stickler, 2024) and the present data does reveal meaningful overlaps and distinctions between the two models. In both species, the only interactions detected were sub-additive, and these represented a small proportion of the total number of effects detected (i.e. they were highly specific). Additionally, the statistically identified interactions were exclusively detected in young animals (“neonatal”-equivalent to adolescence). Taken together, the two models appear to correspond well in terms of the direction and specificity of CdCl-BaP interactions. The most obvious difference concerns sex differences, which were prominent in this work but not the zebrafish study. Sex differences, though occasionally present, were not evaluated as a mediator of human-relevant toxicities, because zebrafish lack X and Y chromosomes and are governed by an alternative set of complex processes determining biological sex (Sandra, 2010). Beyond this, there were mixed similarities and differences in which tests displayed interactions. Negative geotaxis has no fish-equivalent test, and vice versa for the light-dark response component of the larval motility test, although some limited comparisons can be made, as both rely upon a reflex which onsets very early in life and show a comparable sub-additive interaction. The elevated plus maze roughly equates to the novel tank diving test, although zebrafish showed the sub-additive interaction (CdCl₂ reversed the BaP effect) on the locomotion metric instead of the anxiety-like metric. Taken together, the zebrafish project provided meaningful hypotheses and fair overall predictions but did so with limited ability to precisely predict specific interaction effects in rats.

Prominent among the results in this study were sex-specific effects and modifications of typical sex differences. These were as follows. During neonatal development, male vs female differences in anogenital distance (M>F), which is the primary metric used for identification of sex in young rodents, was reduced based on prior CdCl₂ exposure. This early metric was consistent with CdCl₂ effects later in development, specifically where prior CdCl₂ exposure (regardless of BaP exposure) attenuated sex differences in novel object preference and generated a sex difference not seen in controls in the open arm preference component of the elevated plus maze (females were selectively affected, but only when no BaP co-exposure occurred). Likewise for BaP, males were selectively impaired in the radial arm maze, generating performance which more closely resembled female performance within this task. Negative geotaxis and novelty suppressed feeding were largely spared from this pattern. Even so, a visual inspection of the data suggests that females contributed to the novelty suppressed feeding effects to a greater degree than males, so this finding may need to be interpreted cautiously. Taken together, this work suggests that CdCl₂ and BaP can exert developmental neurotoxicity, with biological sex and single- vs co-exposure conditions as major mediators of the resulting toxicity profile.

A number of prior studies have evaluated these or comparable endpoints in rodents with developmental exposure to Cd or BaP, with notable similarities and differences to the present results.

Prior rodent studies have inconsistently tested for sex differences, or powered their studies for analysis by sex, although they generally provide support for Cd or CD salts (Barański, 1983; Gumilar, 2019; Hudson, 2021; Lehotzky, 1990; Minetti, 2006) and BaP (Bouayed, 2009; Chen, 2012; Hawkey, 2019; Honaker, 2022; LüShengJie, 2019; Yang, 2019) influencing diverse behavioral endpoints during maturation or adulthood. With respect to Cd, prior work has observed sex as a driver of neurobehavioral toxicity. Interestingly, Baranski (1983) reported an additional sex difference in the response to CdCl₂ that was not observed in the present work. Female offspring with 0.4–4mg/kg Cd prenatally (gavage) showed reduced locomotor activity in an automated activity cage, whereas male offspring did not. This profile was altered in a parallel study using open field instead, and with older offspring (Barański, 1984), wherein males were hypoactive (inhalation of cadmium oxide - 0.16mg/m³) and females were hyperactive (inhalation of cadmium oxide - 0.02mg/m³). When compared with the current study, it is possible that the design of the activity chamber, in our case using the figure-8 maze, may be responsible for some notable differences in the effect of CdCl₂ on activity, as could dose/toxicokinetic differences mediated by varying Cd forms (CdCl₂ vs CdO) and routes of administration (osmotic pump vs gavage vs inhalation). Even so, sex appears to be a key mediator of those effects when they are present. Unrelated to locomotor activity, Gumilar and colleagues (2019) observed sex-selective effects, wherein prenatal CdCl₂ (0.6mg/kg/day) increased the number of reference memory errors in the RAM among male, but not female, offspring. Notably, this pattern was seen in a long-term retention phase of RAM testing, but not during the training/testing phase, which is consistent with what we report in the current study. We observed an elimination of sex differences in a different cognitive task, NOR. To our knowledge, this has not been previously assessed in models of pre/perinatal CdCl₂ exposure and represents a novel finding. NOR has been previously reported as a sensitive outcome for Cd-exposure in other portions of the lifespan (Batool, 2019; Enogieru, 2022; Pulido, 2019).

With respect to BaP, we and others have reported sex-relevant developmental, behavioral and cognitive effects of perinatal exposure. For example, we previously used the present dose and exposure paradigm for BaP (Hawkey, 2019), present alone or with nicotine, and found that perinatal BaP exposure eliminated sex differences in locomotor activity and an overall composite score of novelty-suppressed feeding. Generally, the potential for BaP exposure to task and sex-selectively impact performance is consistent across that and the current study, although the strength of baseline sex differences and more precise effects of BaP were less consistent. Although many more studies have examined BaP exposure in males only or in other periods of development, certain studies do support the general associations between perinatal exposure, sex, and behavior. For example, using a neonatal gavage method (0.02–2mg/kg, PND 5–11), Chen and colleagues (2012) found that this late exposure to BaP caused impairments in righting reflex and negative geotaxis (later in development, at PND 12–18) in a non-sex-selective manner, and sex-selective effects in anxiolytic-like effects in the elevated plus maze (females more sensitive) and reference memory in the Morris water maze (males more sensitive). These effects were generally only observed with elevated doses of BaP (0.2–2mg/kg/day). Likewise, Honaker and colleagues (2022) found that perinatal BaP exposure tended to produce motor effects in female offspring, but memory deficits in male offspring. This provides some support for the RAM data in the present study. Taken together, this work supports the association between these compounds and sex-specific risks or impairments in sexual dimorphisms.

As each of these compounds produce an array of effects, through an array of mechanisms, there are a variety of potential explanations for the specificity and interactive nature of Cd-BaP interactions within this study. Based on prior work, much of the potential for interactions may be based on overlapping neurotransmission targets. In our prior work (Slotkin, 2019a), we found that the current perinatal BaP exposure paradigm led to partially sex-specific effects in cholinergic markers (increased ChAT activity in select brain regions of males only, HC-3 binding generally reduced in females only) and sex-independent effects on serotonergic function (5HT-2 receptors elevated in select brain regions). A prior study by La’Nissa and colleagues (2007) indicated that comparable doses of BaP (0.025–0.15mg/kg) also have the potential to disrupt the expression of critical subunits within glutamate receptors, similar to that reported with higher levels of exposure (0.3mg/kg) (McCallister, 2008) These alterations provide points of potential interaction with Cd, which is also associated with alterations in serotonergic and glutamatergic neurotransmission. For example, Antonio and colleagues (2010) reported reduced hippocampal 5HT, reduced hippocampal/cortical 5HIAA, reduced GABA in the cortex and medio-basal hypothalamus, and altered glutamate levels (increased in the dorsal hippocampus, reduced in the medio-basal hypothalamus) in the offspring of dams that consumed ~1 mg/kg/day Cd (higher than the 0.3 mg/kg/day used here) during pregnancy. Likewise, Stolakis (2013) reported altered cholinesterase activity in offspring following prenatal Cd-consumption (increased in the cerebellum, decreased in the frontal cortex). The sex, regional, and neurotransmitter specificity of these varied effects would certainly allow these two toxicants to interact within certain individuals (male or female), brain structures, and circuits, but exert toxicity separately in others. Of course, this hypothesis needs to be evaluated directly in the context of this mixture. Brain tissue has been collected from the non-behaviorally tested siblings of the animals in this study, and analyses of those samples and comparable samples in future studies will be needed to reasonably evaluate the mechanisms involved in sub-additive BaP-Cd interactions.

Beyond effects on central neurotransmission, mechanistic work should also be conducted which target the kinetic factors and molecular initiating events associated with these compounds. Kinetically, this would include the potential for altered metabolism and excretion of these toxicants. For example, both Cd and BaP are believed to be inducers of CYP1a enzymes (Elbekai, 2007; Kou, 2024; Sidorova, 2016) which play a prominent role in PAH bioactivation/metabolism (Lu, 2020). Although this might enhance PAH metabolism and excretion in some contexts, they may actually have the opposite effect, Somewhat paradoxically, CYP1a induction has been reported to be reduced when multiple CYP1a inducers are provided in a mixture, as noted with Cd and BaP or TCDD (See Review, (Anwar-Mohamed, 2009)). This provides some basis for Cd and BaP being less neurotoxic in a mixture than alone, to the extent that CYP1a induction is necessary for these toxicants to produce behavioral toxicity. Given the relevance of sex differences to the present data, mechanisms like CYP1a enzymes, which typically metabolize estrogens, androgens and progestogens (Lu, 2020) may be of particular interest in future studies on this topic. Of course, sex hormones may also other roles in these effects and provide other opportunities for interactions as well, since Cd and BaP have been reported to be estrogenic prenatally or in young animals (Ali, 2010; Kummer, 2008) at levels of exposure comparable to those used in this study.

With respect to this work, certain limitations should be noted. The kinetic and mechanistic features of these exposures were not characterized, so although a generic dose can be reported and compared with other studies, it is not precisely known how the selected doses relate to broader understandings of Cd and BaP toxicity. The osmotic pumps used for compound delivery are rated to deliver consistent amounts of each compound across the pre- and perinatal period. However, quantification of maternal and pup exposures across this period would ultimately be required to more precisely identify the offspring doses, Cd:BaP ratios, and document potential kinetic interactions in offspring tissues. Correspondingly, although the doses selected here (below those which would impact reproductive success) were intended to minimize the potential for off-target (non-nervous system) effects, this potential has not yet been thoroughly investigated in this specific administration paradigm or under these mixture conditions. Generally, studies of Cd- or BaP-induced non-CNS pathologies have identified or relied upon considerably higher doses of these substances (Bekheet, 2011; Borman, 2000; De Jong, 1999; Knuckles, 2001; Sayed, 2014; Xu, 2019; Zou, 2020), although more subtle and sensitive endpoints could also be impacted within more comparable dose ranges (Long, 2018; Nair, 2013), particularly during development. Additionally, hypotheses can be generated as to the mechanistic basis of interactions, although this work did not measure the degree to which these doses and mixture impacted the primary molecular initiating events or corresponding changes in expression associated with these compounds (e.g. at AHr, metallothionein, CYP1a, ERα, etc). Therefore, these should be considered cautiously and examined more closely in the analysis of tissue from this experiment and in future studies.

The present work demonstrates the importance of testing cross-class toxicant mixtures in vivo, and an important scenario where the assumption of additivity is likely to be violated. Using one representative neurotoxic metal and one representative PAH, the current study demonstrated that these toxicants have the potential to produce fewer phenotypes when presented as a co-exposure than as individual exposures. This work also demonstrated the specificity of these sub-additive effects. Across a broad battery and a significant period of development, more of our affected outcomes could be attributed to a single toxicant than to the interactions between BaP and Cd. So, we conclude that it would be unreasonable to label the relationship between BaP and Cd as simply “additive” or “sub-additive”. Rather, these toxicants produce general profiles of behavioral toxicity, within which only select functions or symptoms exhibit the interaction. In this specific pairing, the co-presence of BaP blocked Cd’s acceleration of neonatal motor development and its suppression of anxiety-like responses in young females. Currently, it is not known whether other common metal-PAH pairings will exhibit similar or unique effects and interactions when compared to BaP and Cd, but given the commonality of such co-exposures, continuing efforts in this area will provide valuable insights into the nature and clinical impact of mixtures.

Highlights.

Persisting Neurobehavioral Consequences of Gestational exposure to Cadmium and Benzo[a]pyrene in Rats. Hawkey et al. 2024

Female rats were exposed to benzo[a]pyrene (BaP) and/or cadmium (Cd) during gestation.

Cd reduced sex differences in neonatal anogenital distance and novel object recognition.

Cd enhanced negative geotaxis and in females increased open-arm time in the elevated plus maze.

BaP co-treatment attenuated Cd-effects on negative geotaxis and elevated plus maze.

BaP reduced metrics of food consumption in the novel environment suppressed feeding test.

BaP increased errors during the initial phase of radial-arm maze training in males.