Neonatal Exposure to Ultrafine Iron but Not Combined Iron and Sulfur Aerosols Recapitulates Air Pollution-Induced Impulsivity in Mice

Abstract

Air pollution (AP) is becoming recognized as a major threat to neurological health across the lifespan with increased risk of both neurodevelopmental and neurodegenerative disorders. AP is a complex mixture of gases and particulate matter, with adsorbed contaminants including metals and trace elements, which may differentially contribute to its neurodevelopmental impacts. Iron (Fe) is one of the most abundant metals found in AP, and Fe concentrations may drive some behavioral deficits observed in children. Furthermore, brains of neonate mice exposed to concentrated ambient ultrafine particulate matter (UFP) show significant brain accumulation of Fe and sulfur (S) supporting the hypothesis that AP exposure may lead to brain metal dyshomeostasis. The current study determined the extent to which behavioral effects of UFP, namely memory deficits and impulsive-like behavior, could be recapitulated with exposure to Fe aerosols with or without concomitant SO₂. Male and female neonate mice were either exposed to filtered air or spark discharge-generated ultrafine Fe particles with or without SO₂ gas (n = 12/exposure/sex). Inhalation exposures occurred from postnatal day (PND) 4-7 and 10-13 for 4 hr/day, mirroring our previous UFP exposures. Mice were aged to adulthood prior to behavioral testing. While Fe or Fe + SO₂ exposure did not affect gross locomotor behavior, Fe + SO₂-exposed females displayed consistent thigmotaxis during locomotor testing. Neither exposure affected novel object memory. Fe or Fe + SO₂ exposure produced differential outcomes on a fixed-interval reinforcement schedule with males showing higher (Fe-only) or lower (Fe + SO₂) response rates and postreinforcement pauses (PRP) and females showing higher (Fe-only) PRP. Lastly, Fe-exposed, but not Fe + SO₂-exposed, males showed increased impulsive-like behavior in tasks requiring response inhibition with no such effects in female mice. These findings suggest that: 1) exposure to realistic concentrations of Fe aerosols can recapitulate behavioral effects of UFP exposure, 2) the presence of SO₂ can modulate behavioral effects of Fe inhalation, and 3) brain metal dyshomeostasis may be an important factor in AP neurotoxicity.

1. Introduction

Air pollution (AP) is recognized as one of the largest threats to human health [1]. Though much of the research on AP has been directed at cardio-pulmonary effects, there is substantial evidence that the brain is a major target of AP exposure [2]. Numerous epidemiological studies link prolonged AP exposure to neurodevelopmental disorders (NDDs) and neurodegenerative disorders (NDGDs) such as Autism-Spectrum Disorder (ASD) and Alzheimer’s Disease (AD), respectively [3-8]. In children, AP exposure is associated with poorer cognitive function [9], including increased impulsive action [10]. These cognitive impairments are often driven by exposure during the 3^rd trimester [11], which is also the critical window most associated with increased risk of ASD due to AP exposure [12]. These data suggest that extended exposure to AP has the potential to induce clinically relevant disruptions in brain development and cognition.

AP is a complex mixture of particulate matter (PM) and gases largely a result of industrial emissions and vehicle exhaust. Health effects of AP are related to PM size defined as coarse (<10μm, or PM₁₀), fine (<2.5μm or PM_2.5), or ultrafine (UFPs; <100nm or 0.1 um), with UFPs considered as the most reactive component of AP per unit mass [13, 14]. While levels of PM₁₀ and PM_2.5 have been declining, this may not be true of UFP, which does not correlate consistently with PM_2.5 levels [15]. Additionally, UFP does not currently have any federal regulation largely due to inadequate measurement protocols and data availability, although advances are being made [16]. Thus, it is critically important to outline adverse consequences associated with UFP exposure and its specific components, especially when considering its role in NDD and NDGD risk.

Our studies and others have repeatedly shown that UFP exposure in rodent models can recapitulate some of the neurobehavioral pathologies observed in NDDs and NDGDs [17-21]. For example, mice exposed to UFP from postnatal day (PND) 4-7 and 10-13, considered the human 3^rd trimester brain development equivalent; [22, 23], show increased neuroinflammatory markers [18], elevated glutamate/GABA ratios [24], and heightened impulsivity and cognitive inflexibility in adulthood using sophisticated behavioral tasks [20, 25]. These effects also tend to be primarily detected in males, mirroring higher male:female ratios of NDDs [26].

Given the parallel findings from experimental and epidemiological studies regarding developmental AP neurotoxicity, there has been recent interest in identifying UFP components that could underlie the associated neurotoxic effects. One component that could contribute to these neurobehavioral pathologies is the various inorganic contaminants found in AP, including metals and trace elements [27]. Indeed, a National Emissions Inventory study from 2009 showed that various metals including, iron (Fe), aluminum (Al), silicon (Si), sulfur (S), and copper (Cu) can be detected in the atmosphere across the United States with emission estimates of 103,000, 140,000, 382,000, 79,000, and 1710 tons/year, respectively, depending on geographical region [28]. These findings align with x-ray fluorescence (XRF) analyses of exposure chamber filters from real-world UFP exposures in our studies in Rochester and Toledo, New York showing metals and other trace elements detected in the UFP aerosol, notably Fe, Al, S, zinc (Zn), and Cu [29]. Furthermore, these metal profiles are also present in the brains of mice following early-life UFP exposure as shown by laser ablation inductively coupled plasma mass spectrometry (ICPMS) analysis confirming elevated levels of S, Cu, and Fe throughout the brain [29].

Metals found in AP that are essential for normal brain development are of particular concern, which are also the metals often found most in air (e.g., Fe) [30]. Exposure to and possible brain accumulation of redox-active metals may produce metal dyshomeostasis leading to oxidative stress and inflammation [31]. This is particularly true of Fe, which not only regulates many aspects of the oxidative stress response [31], but also is critical for proper hippocampal and cortical development [32]. While Fe does accumulate naturally throughout life, tight regulation of Fe homeostasis does not begin until 6-9 months in children [33] and PND 20 in rodents [34], allowing excess Fe to potentially enter the brain and influence brain development. In fact, epidemiological data across 4 European cohorts (Netherlands, Italy, Spain, and Germany) showed that atmospheric Fe concentrations at birth were associated with psychomotor impairment in early childhood [35]. However, the precise mechanisms of this association and the extent to which Fe inhalation could produce similar effects under controlled settings is unclear. Much of our understanding of Fe neurotoxicity via canonical respiratory portals of entry rely on high-dose, intranasal instillation [36] somewhat limiting extrapolation to real-world, whole-body inhalation preparations. Furthermore, Fe inhalation studies have primarily focused on its cardio-pulmonary effects again leaving neurotoxic effects of Fe inhalation relatively unexplored [37].

Based on its abundance in AP [28, 30] and its association with psychomotor dysfunction in children [35], the purpose of the present study was to assess the extent to which early-life ultrafine Fe inhalation could recapitulate neurobehavioral effects of ambient UFP exposure in mice. In previous studies, we have shown that UFP exposure from PND 4-7 and 10-13 in mice produces impairment across several psychomotor domains including open-field locomotor behavior, object reference memory, fixed-interval schedule acquisition, motor impulsivity, and delay tolerance [17, 20, 38, 39]. Given the association between AP exposure and NDDs, behavioral endpoints were focused upon due to the reliance of behavioral indicators in the diagnosis of NDDs [40]. In ambient air, Fe concentrations are correlated with sulfate content, based on sulfate’s ability to mobilize Fe from its oxide form [41]. Further, an early study [42] reported that sulfur dioxide (SO₂) increased the uptake of Fe into the central nervous system (CNS) and altered its distribution among different cell types. Consequently, in the current studies separate cohorts of mice were exposed to Fe oxide nanoparticle exposures alone or with SO₂ exposure (Fe + SO₂). Exposure parameters for Fe aerosols were selected to mirror, inasmuch as possible, realistic atmospheric conditions [43-46].

2. Methods

2.1. Animals and Fe exposures

Male and female mice C57BL6/J mice were purchased from Jackson Laboratories (Bar Harbor, ME) at 8 weeks of age and allowed to acclimate in a temperature- (71-74°F) and humidity-controlled (35-40%) colony room operating on a 12 h light/dark cycle (lights on at 0600) for one week prior to breeding. Males and females were group housed prior to breeding with dirty bedding from males to induce female estrous cycle synchronization [47]. Mice were then bred monogamously for 3 days, after which males were removed the cage and dams remained singly housed with their pups until weaning at postnatal day (PND) 21.

Pups were randomly assigned to be exposed to ultrafine Fe aerosols alone, or in a separate exposure to ultrafine aerosols concurrently with SO₂, or HEPA-filtered air. To preclude litter effects, only one randomly selected pup per sex, per litter were included in each exposure group for behavioral testing. The breeding and exposure assignments yielded group sizes of sufficient magnitude to detect small-medium effect sizes across behavioral paradigms (n = 12). One Fe+SO₂-exposed male was euthanized before the end of behavioral testing due to health problems and all data from this mouse are omitted. Exposures (described below) were conducted 4 h/day from 0900-1300 h between PND 4-7 and 10-13 as in our previous studies (Cory-Slechta et al., 2018). This postnatal window in rodents is roughly equivalent to 3^rd trimester brain development in humans [22]. Pups were returned to the corresponding dam after each exposure session.

Mice were exposed in exposure cages via whole body inhalation. For this study, the intended SO₂ concentration was 1.31 mg/m³, and the intended Fe concentration was 1.0 μg/m³. The Fe concentration was chosen to be within the range of values cited for outdoor Fe levels [44-46, 48]. The SO₂ concentration was based on the U. S. Environmental Protection Agency secondary standard for SO₂. Fe-oxide UFP particles were generated by electric spark discharge between two 99.9% pure iron rods (3N5 Purity, ESPI Metals, Ashland, OR, USA) using a GFG-1000 Palas generator (Palas GmbH, Karlshrue, Germany), and fed into a compartmentalized whole-body mouse exposure chamber, while HEPA-filtered air was delivered to the control chamber, as in prior studies in our inhalation facility [49]. Passing the airborne particles through a deionizer (Isotope Po-210, model P-2031, NRD, Grand Island, NY, USA) was used to bring particle charge to Boltzmann equilibrium. Particle number concentration was adjusted by altering electric spark discharge frequency. Aerosol number concentration and particle size were monitored in real-time using a Condensation Particle Counter (CPC, model 3022, TSI Inc, St Paul, MN, USA) and Scanning Mobility Analyzer (SMPS, model 3934 TSI Inc, St Paul, MN, USA) respectively. The Fe-oxide particles were generated by adding a low flow of oxygen (~50 mL/min) into the argon flow (~5 L/min) entering the spark discharge chamber. The oxygen concentration of 21% in the exposure chamber was verified by an O₂ sensor (MAXO2 −250E, Maxtec, Salt Lake City, UT, USA). This procedure produced particle sizes exclusively in the ultrafine size range with a count median diameter (CMD) of approximately 12-14 nm. Mass concentrations were determined by ICP-OES analysis of Fe on nitrocellulose membrane filters (0.8 micron, AAWP02500, Millipore Ltd., Tullagreen, Cork, IRL) collected daily (5 L/min for 60 min., 300L total volume) from the filtered air and ultrafine Fe-oxide particle exposure chambers. For the concurrent SO₂ exposures, SO₂, compressed in gas cylinders (EPA Protocol Standard, 50 ppm, Airgas East, Radnor, PA, USA), was diluted with filtered air and then bled into (200 ml/min) the Fe-oxide containing conduit to achieve final desired concentrations for Fe + SO₂ exposures. This Fe-oxide/SO₂ mixture was fed into the compartmentalized whole-body exposure chamber at 25-30 liters per minute. SO₂ concentrations were continuously monitored and recorded with an SO₂ gas monitor (model 43C, Thermo Environmental Instruments Inc., Franklin, MA, USA).

After weaning at PND 21, mice were pair-housed by sex and exposure group with standard rodent chow and water freely available until testing protocols required food restriction (see section 2.3.3.). Mice were allowed to mature in their home cage until behavioral testing in adulthood (PND 65). Behavioral testing took place at approximately the same time of day across all testing paradigms. All mice were treated humanely and with regard to the alleviation of suffering, and all study protocols and procedures were approved by the University of Rochester Institutional Animal Care and Use Committee.

2.3.1. Locomotor Activity (Fe-only = PND 70; Fe+SO₂ = PND 65)

Locomotor testing was conducted as described previously on PND 70 for Fe-only and PND 65 for Fe+SO₂ cohorts [20]. Briefly, locomotor activity was assessed using automated chambers containing an array of 48 infrared photo beams located along the x, y, and z axis (Med Associates Inc., St. Albans, VT). Photo beam breaks were recorded continuously during the 60-minute locomotor session to assess horizontal and vertical movements. Ambulatory counts were defined as beam breaks across successive beams while in ambulatory movement. Ambulatory distance was calculated based on the sum of all ambulatory episodes defined as breaking at least 3 successive photo beams. Vertical activity was defined as movement that broke photo beams in the z-axis. Stereotypical movement was defined as the number of beam breaks in a 2 x 2 beam box that were non-ambulatory.

2.3.2. Novel Object Recognition (NOR; Fe-only = PND 75; Fe+SO₂ = PND 86)

NOR was conducted beginning across two sessions in an open Plexiglas arena (30.5 cm × 30.5 cm × 30.5 cm) as previously described on PND 75 for Fe-only and PND 86 for Fe+SO₂ cohorts [21]. During the first session, mice were placed individually into the test chamber containing two small round white knobs secured to the chamber floor, and were allowed to explore the chamber and objects for 10 min. The purpose of this initial session was to allow mice to become familiar with the sample objects and for assessment of side preference and overall exploration patterns across exposure groups. The second session occurred 24 h after the first session to assess memory of the sample objects premised on the ability of the mice to detect novel stimuli. During the second session, mice were returned to the testing chamber, which now contained one small round, white knob (sample object) and one small square, black knob (novel object). Position (right or left) of the novel object was counterbalanced across treatments and subjects to preclude side bias. Both sessions were videotaped and scored using Noldus software (The Observer XT, Noldus) by a trained observed blinded to treatment condition. Object exploration was defined as a mouse being oriented toward the object with its head crossing a pre-marked 2 cm circle surrounding the object. Object recognition was primarily analyzed using three different indices, which control for differences in overall exploration across mice: duration index (total novel exploration time / [total novel time + total sample time]), bout index (total novel bouts / [total novel bouts + total sample bouts]), and time-per-bout index (average novel time per bout / [average novel time per bout + average sample time per bout]).

2.3.3. Fixed Interval (FI) 60-s Schedule (PND Fe-only = 110-142; Fe+SO₂ = PND 119-155)

Following NOR testing, food intake was gradually restricted to maintain mice at 85% of free-feeding weight prior to lever press training with water freely available in the home cage. Lever pressing was established using a 6-hr autoshaping procedure [50] in mouse operant-conditioning chambers (ENV-307W, Med Associates, St. Albans, VT) housed in sound-attenuating cabinets with fans for ventilation. Three response levers were arranged horizontally along the back panel of each chamber with a pellet dispenser and food receptacle on the opposite panel. A houselight centered above the 3 response levers illuminated the chambers during sessions. Once lever pressing was reliable, the FI 60-s schedule began. On this reinforcement schedule, a food pellet was delivered for the first response occurring after 60 s had elapsed. Responses prior to the 60-s interval were recorded but had no other programmed consequence. Reward delivery initiated the subsequent 60-s interval. Sessions were 30 min in duration and occurred approximately between 0900-1200 each day Monday-Friday. Fixed interval testing occurred between PND 110-142 for Fe-only and between PND 119-155 for Fe+SO₂ cohorts. A total of 25 sessions were conducted (Fe only) or 29 (Fe + SO₂). Dependent measures included response rate (lever presses / session time), post-reinforcement pause (PRP; latency between reward delivery and next response) and run rate (lever presses / [session time – cumulative PRP]).

2.3.4. Differential reinforcement of low rate (DRL) schedule (Fe-only = PND 143-180; Fe+SO₂ = PND 239-271)

The day following the final FI 60-s session, mice were switched to a DRL schedule. Because of a break in behavioral testing due to COVID-19 pandemic logistics, 10 additional FI 60-s sessions were conducted prior to DRL testing in the Fe + SO₂ cohort. This was done to mimic the sudden FI-to-DRL transition experienced by the Fe-only cohort. DRL schedules only reinforce inter-response times (IRTs) of a minimum duration. For example, a DRL 10-s schedule would only provide reinforcement for a response that occurred at least 10 s after the previous response. IRTs that do not meet this minimum interval requirement re-initiate the required time interval. The initial DRL requirement was 10 s (7 sessions), which was subsequently increased across sessions to 20 s (8 sessions), 28 s (4 sessions), and 36 s (8 sessions). Sessions were 30 min in length occurring daily from Monday-Friday. DRL testing occurred between PND 143-180 for Fe-only and between PND 239-271 for Fe+SO₂ cohorts. Dependent measures included response rate, burst responses (IRTs < 2 s), and responses per reinforcer.

2.3.5. Fixed-ratio (FR) Wait Behavior (Fe-only = PND 185-200)

Following DRL testing, Fe only condition mice were also tested in a FR-wait task as an additional measure of impulsivity or delay tolerance [51]; COVID-19 pandemic logistics prevented evaluation in Fe + SO₂ groups. FR-wait testing occurred between PND 185-200 for the Fe-only cohort. Training for the FR-wait task began with five sessions of a FR-25 schedule in which 25 lever presses were required to earn a food pellet. FR-25 sessions terminated following 30 pellet deliveries or 30 minutes elapsing, whichever occurred first. After the fifth FR-25 session, a wait component was added. The wait component allowed mice to obtain “free” pellets following completion of each FR component during a session. However, “free” pellets were delivered at increasing delay intervals (10 s, 20 s, 30 s, etc.) as long as no lever presses occurred. If a lever press occurred during the wait component (termed a “reset”), the FR component was re-introduced. To increase probability of waiting behavior, an adjusting intertrial interval (ITI) was used such that mice could not maximize reinforcement by completing successive FR components without entering the wait component. The ITI adjusted based on how long the mouse had waited during a given wait component. For example, if a mouse did not wait for a “free” pellet in a given component, there was a 60-s ITI. However, if a mouse earned 2 “free” pellets during a wait component in 10-s wait session (10 s + 20 s = 30 s of wait time), the ITI would be 30 s. Two different wait interval increments were used across 8 total FR-wait testing sessions. A 10-s wait increment was in place during the first four sessions with a 15-s wait increment in place for the last four sessions. Dependent measures included total resets, no-wait resets (resets in which no “free” pellets were delivered), maximum wait time (the longest wait interval reached in session), and response rate. Response rate during the initial FR-25 training was also analyzed to assess pre-FR-wait response differences between groups.

2.4. Statistical Analysis

All repeated-measures data (locomotor, FI 60, DRL, and FR-wait) were analyzed using repeated-measures ANOVA with exposure group as the between-subjects factor and either time bin (locomotor) or session (FI, DRL, FR-wait) as the within-subjects factor. Due to consistent sex differences detected in previous studies regarding UFP neurotoxicity [20, 21], analyses were also stratified by sex. Additionally, analyses for DRL and FR-wait outcomes were separated by experimental condition (e.g., DRL 10 s & 20 s were analyzed separately). Lastly, due to many repeated observations in the FI condition (i.e., 25), FI outcomes were analyzed by week in 5-session increments to detect week-by-week changes in FI schedule acquisition. All analyses were conducted using JMP (version 16) Differences were considered statistically significant if p ≤ 0.05; marginal changes (p≤0.10) are also shown where relevant.

3. Results

3.1. Exposure characterization

Data from Fe only and Fe + SO₂ aerosol exposures have previously been reported [52]. Overall, Fe particle mass concentrations during Fe-only exposures were slightly above the intended 1 μg/m³ target on average (8-day average = 1.36 μg/m³) with variation across the 8 days (range = 0.68-2.3 μg/m³). Particle diameter was consistently between 13-14 nm (mean = 13.6 nm) indicating a consistent ultrafine aerosol, as intended. On average, particle count concentrations were consistent around 200,000 particles/cm³ with some variability across days (range = 36,890 – 856,050). For Fe + SO₂ exposure, Fe mass concentrations across days ranged from 0.75 to 2.37 μg/m³ over the exposure period, producing an average concentration of 1.51 μg/m³, Count Median Diameter (CMD) particle size varied from 11.2 to 13.6 nm, with an average Geometrical Standard Deviation (GSD) of 1.4, which is within the ultrafine size range. The daily particle number concentration average +/− standard deviation was 2.16E+05 +/− 0.17E+05 part/cm³.

3.2. Locomotor activity

While neither Fe nor Fe + SO₂ exposures produced consistent significant changes in total locomotor activity throughout the entire session, analyses by center vs. edge of the apparatus revealed numerous changes in Fe + SO₂-exposed females relative to their corresponding filtered air controls, as shown in Figure 1. These included significant or marginally significant reductions in ambulatory distance, ambulatory time, ambulatory counts, stereotypic time, stereotypic counts, rest time, vertical counts and time in zone in the center of the apparatus (F(1,21)=5.19, p=0.033; F(1,21)=4.05, p=0.0573; F(1,21)=4.59, p=0.044; F(1,21)=3.72, p=0.067, F(1,21)=3.85, p=0.063; F(1,21)=4.49, p=0.046, F(1,21)=3.51, p=0.075; F(1,21)=4.61, p=0.0436, respectively), while significant increases were seen in stereotypic time, stereotypic counts, rest time and time in the edge zone of the apparatus (F(1,21)=6.5, p=0.0186, F(1,21)=6.63, p=0.0177; F(1,21)=4.79, p=0.040; F(1,21)=4.61, p=0.0436, respectively).

Figure 1.

Locomotor performance of Fe-only and Fe-SO₂-exposed mice with respect to the center and edge of the locomotor chamber. Data represent mean percent of respective air controls ± SEM condensed across the entire session. * indicates significant difference from air controls (p < 0.05); ~ indicates marginal difference from air controls (p < 0.1).

3.3. Learning and memory behavioral paradigms

3.3.1. Novel object recognition

Overall, there were no exposure-related differences observed in NOR performance in either male or female mice in either the Fe or the Fe + SO₂-exposed mice as shown for the total session duration (5 min) or for the first 2 min of the session (Figure 2 and 3, respectively). As expected, each object recognition index (duration, bout, and duration per bout) was above 0.5, indicating greater interaction with the novel object on average (See Table 1 for a summary of locomotor and NOR outcomes).

Figure 2.

Performance during the first two minutes of novel object recognition testing in Fe-only and Fe-SO₂-exposed mice. Data represent mean ± SEM.

Figure 3.

Performance during the entire 5-minute test novel object recognition assessment in Fe-only and Fe-SO₂-exposed mice. Data represent mean ± SEM.

Table 1.

Summary of locomotor testing outcomes across groups.

	Fe-only		Fe+SO2
Locomotor	Female	Male	Female	Male
Ambulatory Distance Center	No Δ	No Δ	↓	No Δ
Ambulatory Time Center	No Δ	~↓	↓	No Δ
Ambulatory Counts Center	No Δ	~↓	↓	No Δ
Ambulatory Distance Edge	No Δ	No Δ	No Δ	No Δ
Ambulatory Time Edge	No Δ	No Δ	No Δ	No Δ
Ambulatory Count Edge	No Δ	No Δ	No Δ	No Δ
Stereotypical Time Center	No Δ	No Δ	~↓	No Δ
Stereotypical Counts Center	No Δ	No Δ	~↓	No Δ
Stereotypical Time Edge	No Δ	No Δ	↑	No Δ
Stereotypical Counts Edge	No Δ	No Δ	↑	No Δ
Rest Time Center	No Δ	No Δ	↓	No Δ
Rest Time Edge	No Δ	No Δ	↑	No Δ
Vertical Count Center	No Δ	↓	↓	No Δ
Vertical Count Edge	No Δ	No Δ	No Δ	No Δ
Vertical Time Center	No Δ	No Δ	↓	No Δ
Vertical Time Edge	No Δ	No Δ	No Δ	No Δ
Center Zone Entries	No Δ	No Δ	No Δ	No Δ
Edge Zone Entries	No Δ	No Δ	No Δ	No Δ
Novel Object Recognition
Duration Index - 2-min	No Δ	No Δ	No Δ	No Δ
Bout Index - 2-min	No Δ	No Δ	No Δ	No Δ
Time/Approach Index - 2-min	No Δ	No Δ	No Δ	No Δ
Duration Index - 5-min	No Δ	No Δ	No Δ	No Δ
Bout Index - 5-min	No Δ	No Δ	No Δ	No Δ
Time/Approach Index - 5-min	No Δ	No Δ	No Δ	No Δ

Symbols: No Δ = no difference between groups; ↓ = significant decrease; ~↓ = marginal decrease; ↑ = significant increase; ~↑ = marginal increase.

3.3.2. FI 60-s schedule

Prior to initiating the FI 60-s schedule, there were no exposure-related differences in the number of sessions to complete lever press training in male or female mice in either the Fe only or the Fe + SO₂-exposed groups (data not shown).

As anticipated, FI response rates increased across sessions (Figure 4). FI response rates were marginally and not consistently influenced in males exposed to Fe + SO_2, where decrements in response rate were observed primarily after 10 sessions, with marginal reductions produced by Fe + SO₂ during sessions 11-15 (treatment x group, F(4,18)=2.5, p=0.078 and significant reductions during sessions 21-25 (treatment x group, F(4,18)=4.01, p=0.017) and sessions 26-29 (treatment x group, F(4,18)=3.14, p=0.0496), as well as marginal reductions as analyzed across all sessions (treatment x group, F(1,21)=2.85, p=0.053). These reductions in rate were mirrored in reduced run rates (Figure 5) in males exposed to Fe + SO₂, although these reductions were only statistically significant during sessions 21-25 (treatment x group, F(4,18)=3.16, p=0.039).

Figure 4.

Response rate outcomes across sessions during FI 60-s training in Fe-only and Fe-SO₂-exposed mice. Data represent mean ± SEM. * indicates significant difference from air controls (p < 0.05).

Figure 5.

Run rate outcomes across sessions during FI 60-s training in Fe-only and Fe-SO₂-exposed mice. Data represent mean ± SEM. * indicates significant difference from air controls (p < 0.05).

Notably, changes in postreinforcement pause time were seen in females exposed to Fe only, and to males exposed to Fe + SO_2, with changes that were opposite in direction (Figure 6). For Fe-exposed females, increases in postreinforcement pause time emerged after approximately 10 sessions, with marginal increases from sessions 6-10 (group, F(1,22)=3.24, p=0.085) and significant increases across sessions 11-15 (group, F(1,22)=10.88, p=0.003), 16-20 (group, F(1,22)=4.37, p=0.48) and 21-25 (group, F(1,22)=5.54, p=0.028). No differences in postreinforcement pauses were found in Fe + SO₂-exposed females. Fe-exposed males showed initial reductions in postreinforcement pause time during the first 5 sessions (group, F(1,22)=5.28, p=0.031) and increases in sessions 11-15 (group, F(1,22)=4.44, p=0.04), but persistent changes were not evident. Shorter postreinforcement pause times were seen in Fe + SO₂-exposed males that were evident across sessions 16-25 (16-20: group, F(1,21)=5.88, p=0.024 and 21-25: treatment x group, (F(4,18)=3.25, p=0.036) (See Table 2 for a summary of FI 60-s outcomes).

Figure 6.

Postreinforcement pause (s) across sessions during FI 60-s training in Fe-only and Fe-SO₂-exposed mice. Data represent mean ± SEM. * indicates significant difference from air controls (p < 0.05).

Table 2.

Summary of FI 60-s testing outcomes across groups.

	Fe-only		Fe+SO2
Fixed Interval 60-s	Female	Male	Female	Male
Response Rate
Session 1-5	No Δ	↑	No Δ	No Δ
Session 6-10	No Δ	No Δ	No Δ	No Δ
Session 11-15	No Δ	No Δ	No Δ	~↓
Session 16-20	No Δ	No Δ	No Δ	~↓
Session 21-25	No Δ	No Δ	No Δ	No Δ
Session 26-29	--	--	No Δ	No Δ
Run Rate
Session 1-5	No Δ	No Δ	No Δ	No Δ
Session 6-10	No Δ	No Δ	No Δ	No Δ
Session 11-15	No Δ	No Δ	No Δ	No Δ
Session 16-20	No Δ	No Δ	No Δ	No Δ
Session 21-25	No Δ	No Δ	No Δ	↓
Session 26-29	--	--	No Δ	No Δ
Postreinforcement Pause
Session 1-5	No Δ	↓	No Δ	No Δ
Session 6-10	~↑	No Δ	No Δ	No Δ
Session 11-15	↑	↑	No Δ	No Δ
Session 16-20	↑	No Δ	No Δ	↓
Session 21-25	↑	No Δ	No Δ	↓
Session 26-29	--	--	No Δ	No Δ

Symbols: No Δ = no difference between groups; ↓ = significant decrease; ~↓ = marginal decrease; ↑ = significant increase; ~↑ = marginal increase; -- = not tested.

No consistent changes in either median interresponse times or in total errors (responses on levers that were not associated with the FI schedule) or any outcomes during the 10 extra FI 60-s sessions in the Fe + SO₂ cohort were found (data not shown).

3.4. Impulsivity and delay tolerance behavioral paradigms

3.4.1. DRL testing

DRL testing was initiated the day after the final session of the FI 60-s schedule beginning with the DRL 10-s schedule. DRL response rates were consistently increased by Fe exposure in males, with some similar but marginal effects in males exposed to Fe + SO₂ (Figure 7), whereas no consistent changes were observed in females. Increases in DRL response rates in Fe-exposed males were significant during the DRL 10-s schedule (F(1,22)=5.12, p=0.034) and while slightly increased during both the DRL 20-s and DRL 28-s schedules, re-emerged as significantly greater during the DRL 36-s component (F(1,22)=5.84, p=0.0245). Marginally significant increases were seen during the DRL 28-s and DRL 36-s schedule in Fe + SO₂-exposed males (F(1,21)=3.36, p=0.081 and F(1,21)=4.02, p=0.058, respectively).

Figure 7.

Response rate measures across DRL schedules in Fe-only and Fe-SO₂-exposed mice. Data represent mean ± SEM. * indicates significant difference from air controls (p < 0.05); ~ indicates marginal difference from air controls (p < 0.1).

Consistent with these effects were a similar pattern of burst responses (Figure 8) which, like DRL response rates, were markedly increased in Fe-exposed males during the DRL 10-s schedule (F(1,22)=4.71, p=0.041) and again during the DRL 36-s schedule (F(1,22)=6.1, p=0.0218), with higher but non-significant mean values during both the DRL 20-s and DRL 28-s schedules. Here too, changes were less pronounced in Fe + SO₂-exposed males, but again emerged during the DRL 28-s (F(1,21)=6.59, p=0.018) and marginally during the DRL 36-s schedule (F(1,21)=3.94, p=0.06). No differences in burst responses were observed in females.

Figure 8.

Burst responses (IRT < 2 s) across DRL schedules in Fe-only and Fe-SO₂-exposed mice. Data represent mean ± SEM. * indicates significant difference from air controls (p < 0.05); ~ indicates marginal difference from air controls (p < 0.1).

When calculated as responses per reinforcer (Figure 9), effects were restricted to Fe-exposed males, where significant increases were found during the DRL 10-s schedule (F(1,22)=5.17, p=0.033) and particularly increased levels seen on the DRL 36-s schedule (F(1,22)=9.88, p=0.0047), where increases of greater than 100% in numbers of responses were observed. Again, no differences were found in females.

Figure 9.

Response efficiency measured as responses per reinforcer across DRL schedules in Fe-only and Fe-SO₂-exposed mice. Data represent mean ± SEM. * indicates significant difference from air controls (p < 0.05).

3.4.2. FR-wait testing.

Following the final session of the DRL 36-s schedule, Fe-exposed mice were trained on a FR 25 schedule for five days. During FR 25 training, there were no group differences in response rate for males (p = 0.33) or females (p = 0.12) (data not shown).

In males, Fe-exposed mice showed greater no-wait resets during the 10-s wait phase (Figure 10; F(1,22) = 4.50, p = 0.045). No group differences in response rate, total resets, or maximum wait were observed in males in the 10-s wait phase. During the 15-s wait phase, Fe-exposed male mice displayed elevated response rate (F(1,22) = 8.74, p = 0.007) and total resets (F(1,22) = 7.36, p = 0.013), along with decreased maximum wait time (F(1,22) = 5.98, p = 0.023). No differences were observed in no-wait resets for males in the 15-s wait phase. While data from female mice resembled the overall pattern to that of male mice, no exposure-related differences were observed in either the 10-s or 15-s wait phases (See Table 3 for a summary of DRL and FR-Wait outcomes).

Figure 10.

FR-Wait outcomes (response rate, total resets, no-wait resets, and maximum wait time (s)) across the 10-s and 15-s FR-Wait conditions in Fe-exposed mice. Data represent mean ± SEM. * indicates significant difference from air controls (p < 0.05).

Table 3.

Summary of DRL and FR-Wait testing outcomes across groups.

	Fe-only		Fe+SO2
DRL 10-20-28-36	Female	Male	Female	Male
Response Rate
DRL 10	No Δ	↑	No Δ	No Δ
DRL 20	No Δ	No Δ	No Δ	No Δ
DRL 28	No Δ	No Δ	No Δ	~↑
DRL 36	No Δ	↑	No Δ	~↑
Burst Responses
DRL 10	No Δ	↑	No Δ	No Δ
DRL 20	No Δ	No Δ	No Δ	No Δ
DRL 28	No Δ	No Δ	No Δ	↑
DRL 36	No Δ	↑	No Δ	~↑
Responses per Reinforcer
DRL 10	No Δ	↑	No Δ	No Δ
DRL 20	No Δ	No Δ	No Δ	No Δ
DRL 28	No Δ	No Δ	No Δ	No Δ
DRL 36	No Δ	↑	No Δ	No Δ
FR-Wait 10-15
Response Rate
10-s	No Δ	No Δ	--	--
15-s	No Δ	↑	--	--
Resets
10-s	No Δ	No Δ	--	--
15-s	No Δ	↑	--	--
No-Wait Reinforcers
10-s	No Δ	↑	--	--
15-s	No Δ	No Δ	--	--
Maximum Wait Time
10-s	No Δ	No Δ	--	--
15-s	No Δ	↓	--	--

Symbols: No Δ = no difference between groups; ↓ = significant decrease; ~↓ = marginal decrease; ↑ = significant increase; ~↑ = marginal increase; -- = not tested.

4. Discussion

The goal of the current study was to determine the extent to which developmental exposure to Fe aerosols, a significant metal found in AP, could produce behavioral impairment similar to that caused by developmental exposure to UFP in our previous studies [17, 20, 21, 39]. Fe was of primary interest for several reasons including: 1) its abundance in AP relative to other metals and trace elements [30], 2) the essential role of Fe in normal development with Fe regulatory mechanisms developing after AP exposure can occur [33, 53-55], and 3) the concurrence of significant Fe concentrations found both on UFP exposure filters and in the brains of neonate mice following UFP exposure in our prior studies, supporting the role of brain metal dyshomeostasis in UFP neurotoxicity [27, 29]. Overall, we found that neonatal exposure to ultrafine Fe aerosols, either Fe only or Fe + SO₂ during the first two postnatal weeks (equivalent to human third trimester brain development; [22, 23] produced effects that were dependent upon the administration of Fe alone vs. Fe in a mixture, with effects that were also predominantly male-specific, and included impairments in impulsive-like and timing-based behaviors in adulthood. While these results mirror some of the behavioral deficits we have observed following UFP exposures, particularly regarding male specificity, there are important differences to explore.

While we did not observe locomotor disruption or impairment following Fe or Fe + SO₂ exposure, this is consistent with a lack of consistent locomotor effects in our previous UFP studies [20, 21]. UFP exposure has induced locomotor suppression in mice when exposures occurred at both PND 4-13 and PND 56-60 at relatively high UFP concentrations (>100 ug/m³;[38]) or after PND 4-13 exposure to a lower UFP concentration (~50 ug/m³) when tested at 7 months of age [20]. Importantly, when locomotor behavior has been disrupted by neonatal UFP exposure alone, males displayed elevated ambulatory counts throughout the locomotor session suggesting reduced habituation to the testing environment rather than an inability to ambulate. Similar habituation-related locomotor effects have been detected following high-dose (7 mg/kg, oral) neonatal Fe exposure in rats in which locomotor counts were elevated when tested at 3 months of age [56].

It was found that Fe+SO₂-exposed females consistently spent more time and engaged in more behavior in the edge portions of the locomotor chamber relative to the center. In the open-field test, greater time spent in the edge of the chamber is typically interpreted as anxiety-like behavior [57]. One major difference in testing environments that limits an interpretation of anxiety-like behavior in Fe+SO₂-exposed females is that open-field testing occurs in brightly lit conditions, which produces avoidance of the open area of the chamber. Locomotor assessment in the current study occurred in a darkened chamber, which does not necessarily elicit avoidance of the center of the chamber making interpretation of increased anxiety-like behavior difficult. Thus, increased edge time in Fe+SO₂-exposed females may reflect general thigmotaxis rather than increased anxiety, per se. Indeed, intranasal administration of high-dose iron-oxide (10 mg/kg) had no effect on anxiety-like behavior (elevated plus maze) in a small sample of male rats despite elevated brain Fe levels [58]. There are data to suggest that short- and long-term exposure to air pollution may affect emotionality (e.g., depression and anxiety) outcomes in adult populations [59]. However, specific epidemiological associations with anxiety are minimal and preliminary. Additional studies will be necessary to identify consistent changes in emotionality-related outcomes in animal models and human populations induced by exposure to air pollution or air pollution constituents.

Recognition memory was also not affected by either neonatal inhalation of Fe or Fe + SO₂. This outcome was unexpected as we have observed consistent disruptions in NOR performance in prior UFP studies, even to the extent that higher UFP concentrations produce larger disruptions in object recognition memory [18, 20]. NOR is a largely hippocampal-dependent task [60], and we have observed several features of hippocampal disruption following UFP exposure that could contribute to NOR disruption, including glutamate imbalance and heightened microglia activity [24, 61], which are currently unknown following Fe inhalation. In rats, neonatal Fe exposure (10 mg/kg, oral) can produce NOR deficits coupled with elevated hippocampal ROS activity after relatively high oral exposures [62]. Thus, it is possible that the current concentrations of Fe aerosols were not sufficient to induce hippocampal damage through ROS or another inflammatory mechanism. It is also important to note that other sensitive windows of AP exposure in adolescence and adulthood were not assessed in this study. Apart from Fe-specific effects, there is also a possibility of floor effects in NOR performance in the current study, as air control recognition indices were slightly above indifference (0.55-0.60). However, similar air control recognition indices have been observed in our prior studies with UFP-exposed mice showing indices sufficiently below indifference (~0.45) such that they differed from controls.

Several exposure-related differences were found in FI 60-s performance; however, effects differed by sex and exposure. In Fe-exposed males, transient increases in response rate were detected early in FI 60-s testing as in prior studies which dissipated across sessions. An opposite pattern was observed in Fe-SO₂-exposed males in which response rates showed transient decreases across sessions, as observed previously [38] that then sharply increased as training continued relative to air-exposed males in that cohort, which show consistently gradual elevations in response throughout training. Coupled with decreased postreinforcement pauses in Fe- and Fe-SO₂-exposed males in similar training blocks to response rate changes, these changes may reflect disrupted waiting behavior and/or an uncertainty in when to initiate responding. Opposite effects were observed in females in which Fe-exposed female mice showed consistently elevated postreinforcement pauses relative to air-exposed females. This could be interpreted as enhanced waiting behavior in Fe-exposed females. However, these changes did not carry over to DRL or FR-Wait testing, which explicitly tested for waiting behavior or response inhibition. Exploratory analyses of temporal estimation curves (data not shown) used to assess how well mice time the 60-s interval showed no differences across exposure groups in male or female mice suggesting that hippocampal-dependent timing processes were not affected by Fe or Fe + SO2 exposures.

While Fe exposure did not affect performance in memory-based tasks, consistent male-specific impairments were observed in tasks targeting impulsive-like behavior. These deficits were observed both after the sudden transition from the FI 60-s schedule to the DRL 10-s schedule and in later phases of the DRL and FR-wait tasks. During the first session of DRL 10 s, Fe-exposed males responded at nearly twice the rate of air-exposed males. This increase in rate appears to be due to periods of rapid responding, as evidenced by extremely high burst responses, or lever presses with less than 2 s between each press. This resulted in highly inefficient DRL performance with Fe-exposed males emitting 10 responses per reinforcer on average during the DRL 10-s phase relative to 4-5 responses per reinforcer in air-exposed males. Interestingly, similar trends emerged during the final 36-s DRL testing phase in which response rates and burst responses of air-exposed males continued to decrease whereas Fe-exposed males plateaued at DRL 28-s levels. Similar male-specific disruption in DRL responses has been observed following neonatal UFP exposure [20]. However, UFP-induced deficits were only apparent at specific DRL transitions (e.g., from 6 s to 12 s requirement) in that previous study. The current study was designed such that a task producing high-rate responding (i.e., FI 60 s) was abruptly followed by a task producing low-rate responding (i.e., DRL 10 s). Thus, this specific effect could be conceptualized as an inability to adjust to changes in contingencies of reinforcement [63], sometimes referred to as behavioral inflexibility or perseveration [64], a process generally disrupted by UFP exposure [20, 25]. These male-specific, impulsive-like deficits carried over to the FR-wait paradigm where Fe-exposed males showed higher response rates, more FR resets, and reduced maximum wait times all of which are indications of possible disruptions in delay tolerance. Importantly, Fe exposure did not affect responding on the FR-25 pre-training procedure prior to the FR-wait paradigm. However, once a delay to “free” pellets was introduced, Fe-exposed males showed almost immediate evidence of delay intolerance via increases in no-wait FR resets during the 10-s phase. These data are also consistent with our previous study showing neonatal UFP exposure produced similar disruptions in FR-wait behavior via increased FR resets in UFP-exposed male mice [17]. Not only are these consistent with our previous reports of UFP neurotoxicity, but impulsivity, delay intolerance, and a general tendency to perseverate are hallmarks of NDDs, including ASD [65-67].

Given the limited data on Fe inhalation neurotoxicity as it relates to effects of UFP exposure, the current study lends preliminary support that metals found in AP can produce similar neurobehavioral effects to UFP exposure. However, precise mechanisms for these neurobehavioral impairments are currently unclear. As with AP, primary toxicological effects of Fe inhalation center on inflammatory mechanisms [37]. Unlike the gut, which has regulatory mechanisms for absorbing and sequestering unbound Fe and other minerals [68], the lungs lack proper mechanisms to process Fe leading to Fe dyshomeostasis following smoke PM exposure [69]. The presence of excess lung Fe can then produce a systemic inflammatory response through increased cytokine production and subsequent systemic distribution in the blood [70, 71] representing an indirect mechanism that increases neuroinflammation as seen in AP neurotoxicity [6]. However, direct particle translocation to the brain via olfactory nerve fibers is another mechanism often observed in AP neurotoxicity that could contribute to the behavioral effects of Fe in the current study [72, 73]. Indeed, intranasal instillation of 20 ug of radiolabeled Fe nanoparticles (NPs) produced detectable Fe content in the brain with highest concentrations in the olfactory bulb followed by hippocampus, striatum and cortex, supporting olfactory transport of Fe particles to the brain [36]. Olfactory transport of Fe particles has also been confirmed following whole-body inhalation of iron-soot aerosols along with increased microglial activation in areas of olfactory nerve accumulation [74]. This is particularly concerning as overall Fe turnover in the brain is slow leading to natural brain Fe accumulation [75], which could be exacerbated by both early-life and extended exposure to Fe through AP. Once in the brain, evidence suggests that subchronic Fe instillation (e.g., 7 days) can produce elevations in whole-brain dopamine and norepinephrine [58] and increases in hydrogen peroxide production in the striatum and hippocampus 7 days following exposure [36]. Related to Fe-related toxicity, recent data based on a different subset of mice from the same exposure cohorts as in this study showed sex-dependent alterations in frontal cortex trans-sulfuration markers and serum inflammatory markers [52]. While most exposure effects were detected in Fe+SO₂ mice, females showed elevated serum glutathione and IL1-a whereas males did not show a similar inflammatory response. Similarly, Fe+SO2-exposed males showed increased glutamine/GABA ratios with an opposite effect in Fe+SO2-exposed females [52]. This outcome is particularly interesting given that elevated glutamate, in addition to being neurotoxic alone, can also facilitate glutathione depletion and Fe-mediated oxidative damage and ultimately ferroptotic cell death [76]. Thus, sex-dependent changes in ferroptosis/glutamate interactions could partially account for sex differences noted here. Those recent data [52], however, primarily focused on PND 14 mice making any conclusions regarding possible sex-dependent neurotoxic mechanisms in adulthood tentative at best.

While the current study suggests that neonatal Fe inhalation at the current concentration can recapitulate male impulsive-like behavioral effects of UFP exposure, some limitations warrant discussion. First, the current Fe inhalation exposure preparation may not fully replicate respirable Fe in ambient AP. The Fe concentration selected for the current study (1 ug/m³) is similar to what has been reported from ambient AP in different geographical locations including the Middle East (0.42-2.85 ug/m³) [77] and Switzerland (0.07 – 1.37 ug/m³ [78]) . The current 1 ug/m³ target also underestimates exposures that can occur in restricted air systems like subways where Fe concentrations can reach as high as 2.67-47.88 ug/m³ [79]. Second, a related limitation regarding exposure metrics is the natural variability in AP constituents in ambient air, including Fe [80], that is lacking in the current protocol. While modeling variation in ambient Fe concentrations was not an aim of the current study, it has been a key advantage in our prior studies of ambient UFP that provided a natural method to build concentration-effect data on UFP neurotoxicity [29]. The current method of generating a stable Fe aerosol concentration was primarily chosen for purposes of replication and consistency across exposures. Third, differences between rodent and human respiration patterns along with relatively larger olfactory bulb size in rodents could impact particle deposition and retention in the airways and olfactory bulb as well as potential particle translocation to the brain. Fourth, third-trimester equivalent exposure regimens in rodents are qualitatively distinct from third trimester exposures in humans. In altricial species like rodents, the human third trimester equivalent window of brain development occurs outside the womb [22]. Therefore, “third-trimester” rodents directly inhale aerosols whereas third-trimester humans would only be exposed indirectly via the placental blood supply. Fifth, it is unclear the extent to which Fe inhalation primarily affects impulsive-like behavior as observed here or if the effects are related to the order of testing or age of testing. Prior UFP studies have been designed with sufficient sample sizes to address both test order and age of testing [20]. For example, it is possible that, at the current concentration, behavioral deficits are related to age such that effects are not observed until mid-adulthood perhaps due to accelerated Fe accumulation and accompanying neuropathology regardless of the behavioral process tested. Likewise, the increased burst responses on the DRL task may not have been observed in Fe-exposed males if the abrupt change from high-rate responding to low-rate responding was not present. However, this strategy has been used by others to uncover otherwise unobserved metal neurotoxicity representing the utility of such procedures, particularly at low toxicant exposure levels [81]. Sixth, based on our recent data showing that Fe+SO₂ exposure can increase inflammatory markers in blood serum and frontal cortex at PND 14, the lapse in behavioral testing that occurred with the Fe+SO₂ cohort in the current study may have masked possible behavioral differences.

In conclusion, exposure to ultrafine Fe aerosols during the third-trimester equivalent in mice produced male-predominant behavioral impairments in impulsive-like behavior. Similarly, exposure to Fe+SO₂ resulted in transiently reduced or enhanced pausing on an interval timing task in males and females, respectively. These data not only contribute to the hypothesis that metal dyshomeostasis may be driving neurodevelopment effects of AP exposure [27], but also contribute to our understanding of male vulnerability to AP as it relates to third trimester brain development and NDDs [12]. Furthermore, the current results point to Fe as a possible contributing factor to AP neurotoxicity helping to pinpoint regulatory efforts to those sources of AP that generate Fe aerosols into the atmosphere. Lastly, these data outline the importance of considering constituent mixtures of air pollution as the introduction of SO₂, which reportedly enhances tissue Fe uptake [42], resulted in less male-specific behavioral impairment as Fe alone.

Highlights.

• Air pollution contains trace contaminants and metals, including iron and sulfur

• Neurodevelopmental effects of air pollution may involve brain metal dyshomeostasis

• Neonatal iron inhalation produced male-specific impulsive behavior in adulthood

• Memory and motor function were not impaired by iron exposure

• Trace metals, including iron, may drive neurodevelopmental effects of air pollution