Brain, Behavior, and Immunity 40 (2014) 27–37

Contents lists available at ScienceDirect

Brain, Behavior, and Immunity journal homepage: www.elsevier.com/locate/ybrbi

Named Series: Diet, Inflammation and the Brain

Additive effects of maternal iron deficiency and prenatal immune activation on adult behaviors in rat offspring Louise Harvey, Patricia Boksa ⇑ Department of Psychiatry, McGill University, Douglas Mental Health University Institute, 6875 La Salle Blvd, Verdun H4H 1R3, Quebec, Canada

a r t i c l e

i n f o

Article history: Received 1 April 2014 Received in revised form 16 May 2014 Accepted 5 June 2014 Available online 12 June 2014 Keywords: Prenatal infection Maternal immune activation Maternal iron deficiency Behavior Schizophrenia Autism Maternal diet

a b s t r a c t Both iron deficiency (ID) and infection are common during pregnancy and studies have described altered brain development in offspring as a result of these individual maternal exposures. Given their high global incidence, these two insults may occur simultaneously during pregnancy. We recently described a rat model which pairs dietary ID during pregnancy and prenatal immune activation. Pregnant rats were placed on iron sufficient (IS) or ID diets from embryonic day 2 (E2) until postnatal day 7, and administered the bacterial endotoxin, lipopolysaccharide (LPS) or saline on E15/16. In this model, LPS administration on E15 caused greater induction of the pro-inflammatory cytokines, interleukin-6 and tumor necrosis factor-a, in ID dams compared to IS dams. This suggested that the combination of prenatal immune activation on a background of maternal ID might have more adverse neurodevelopmental consequences for the offspring than exposure to either insult alone. In this study we used this model to determine whether combined exposure to maternal ID and prenatal immune activation interact to affect juvenile and adult behaviors in the offspring. We assessed behaviors relevant to deficits in humans or animals that have been associated with exposure to either maternal ID or prenatal immune activation alone. Adult offspring from ID dams displayed significant deficits in pre-pulse inhibition of acoustic startle and in passive avoidance learning, together with increases in cytochrome oxidase immunohistochemistry, a marker of metabolic activity, in the ventral hippocampus immediately after passive avoidance testing. Offspring from LPS treated dams showed a significant increase in social behavior with unfamiliar rats, and subtle locomotor changes during exploration in an open field and in response to amphetamine. Surprisingly, there was no interaction between effects of the two insults on the behaviors assessed, and few observed alterations in juvenile behavior. Our findings show that long-term effects of maternal ID and prenatal LPS were additive, such that offspring exposed to both insults displayed more adult behavioral abnormalities than offspring exposed to one alone. Ó 2014 Elsevier Inc. All rights reserved.

1. Introduction Iron deficiency (ID) is a common and continuing public health issue during pregnancy and early childhood. An estimated 30% of women worldwide are iron deficient, and this percentage increases to a global incidence of 42% when considering ID during pregnancy (McLean et al., 2009). ID during pregnancy is associated with an increased risk of preterm birth and low birth weight, which increases the risk of disease during development and adulthood (Gambling et al., 2003; Haider et al., 2013; Radlowski and Johnson, 2013). Maternal ID also places the developing fetus at risk for neuronal changes, as iron is integral for myelination, mitochondrial efficiency and neurotransmitter metabolism (Connor and ⇑ Corresponding author. Tel.: +1 (514) 761 6131x5928; fax: +1 (514) 762 3034. E-mail address: [email protected] (P. Boksa). http://dx.doi.org/10.1016/j.bbi.2014.06.005 0889-1591/Ó 2014 Elsevier Inc. All rights reserved.

Menzies, 1996; Hare et al., 2013). Indeed maternal and early life ID are associated with reductions in motor skills, learning and cognition, increases in anxiety, depression, social problems and attentional alterations in childhood and adolescence, and in the risk for development of schizophrenia (Lozoff et al., 2006; Insel et al., 2008; Sorensen et al., 2010). Animal models of pre- and peri-natal ID have shown reductions and delays in myelination (Yu et al., 1986; Wu et al., 2008), altered brain monoamine content (Coe et al., 2009), and alterations in hippocampal size, neurochemistry and activity (de Deungria et al., 2000; Rao et al., 2011) as well as persistent deficits in sensorimotor and hippocampal-dependent behaviors (Kwik-Uribe et al., 2000; Felt et al., 2006) and increased anxiety and exploratory behavior (Eseh and Zimmerberg, 2005) in offspring. Similar to maternal ID, infections during pregnancy are varied and prevalent in developing and developed countries (Velu et al.,

28

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

2011) and are associated with an increased risk of adverse birth outcomes (Rasmussen et al., 2012), cerebral palsy (Miller et al., 2013), and juvenile and adult onset psychiatric disorders such as schizophrenia and autism (Brown, 2006, 2011; Khandaker et al., 2013). Animal models of prenatal immune activation have been used widely to model the effects of bacterial endotoxins, viral mimics, and peripheral infections on brain development and behavior. These studies have described structural and neurochemical changes in the brains of offspring, as well as significant behavioral alterations during development, adolescence and adulthood, including deficits in pre-pulse (PPI) inhibition of acoustic startle, increased sensitivity to amphetamine, and reductions in associative and spatial learning (for review see Boksa, 2010). Prenatal immune activation and maternal ID can individually affect neurodevelopment and health in later life, however, the interaction between these insults may be more significant. Exposure to a bacterial immune challenge during a state of active ID results in a potentiation of the pro-inflammatory cytokine response (Pagani et al., 2011). In turn, this enhanced cytokine induction results in a further decrease in serum iron (Kemna et al., 2005). The mechanisms of this circular relationship are still being investigated in non-pregnant animals, however, our objective was to understand the consequences of any interactions between these factors on a pregnant animal and her offspring. Therefore, given the prevalence of these two environmental factors during pregnancy, their associated effects on brain development, as well as the circular relationship between ID and inflammation, we recently described a rat model which pairs maternal ID during pregnancy and prenatal immune activation (Harvey and Boksa, 2014). Pregnant rat dams were placed on an ID diet from embryonic day 2 (E2) until postnatal day 7 (P7), and were also administered the bacterial endotoxin, lipopolysaccharide (LPS), on E15 and 16 to induce a state of prenatal immune activation. After just 13 days on an ID diet (E15), we confirmed that dams showed a significant 40% reduction in serum iron and 20% reduction in placental iron compared to dams on an iron sufficient (IS) diet (Harvey and Boksa, 2014). LPS administration on E15 induced an increase in serum levels of pro-inflammatory cytokines, interleukin-6 (IL-6) and tumor necrosis factor-a (TNF-a), in these dams. Most interestingly, we demonstrated that there was an interaction between maternal ID and prenatal immune activation, as there was significantly greater induction of IL-6 and TNF-a in response to LPS in ID dams compared to IS dams (Harvey and Boksa, 2014). Since IL-6 is thought to be one of the key mediators driving the alterations in brain development and behavior in animals as a result of prenatal immune activation (Smith et al., 2007), this suggested that the combination of prenatal immune activation on a background of maternal ID might have more adverse consequences than exposure to either insult alone. In this combined maternal ID/prenatal LPS model, the consequences of maternal ID on the offspring were significant. Offspring from ID dams showed 75–85% reduction in spleen and liver iron at P7 and 50% reduction in serum iron at P21, however, the increase in IL-6 in ID LPS dams did not result in more severe ID in the ID LPS offspring (Harvey and Boksa, 2014). Importantly, brain iron content of offspring from ID dams was 65% of control at P7 (Harvey and Boksa, 2014). Neurodevelopmental screening of perinatal offspring (P6–P18) in this combined model showed that pups from ID dams displayed abnormalities in forelimb grasp and acoustic startle, while pups from LPS treated dams displayed differences in grip ability, geotaxis reflex, cliff avoidance and acoustic startle (Harvey and Boksa, 2014). However, we did not observe any interaction between the two insults with respect to the markers of early neurodevelopment examined. Instead, our findings showed that the two insults produced an additive phenotype; offspring exposed to both maternal ID and prenatal LPS displayed the sum of

neurodevelopmental abnormalities produced by either factor alone. However, while we did not observe any interaction on markers of early neurodevelopment, we did not test more complex juvenile or adult behaviors of learning, cognition and attention which have previously been shown, in separate models of maternal ID or prenatal immune activation, to be altered by these early environmental insults. Thus, the aim of this study was to use our previously described rat model to determine whether there is an interaction between maternal ID and prenatal immune activation that alters and potentiates abnormalities in juvenile and adult behaviors in the offspring. We assessed a variety of behaviors chosen because of their relevance to described deficits in humans as a result of maternal and early life ID or prenatal infection. We assessed passive avoidance behavior as a measure of learning and memory and also measured a metabolic marker of neuronal activity in the brain following passive avoidance testing. We assessed pre-pulse inhibition of acoustic startle, amphetamine induced locomotion, social interaction and sensitivity to novelty, which are behaviors previously shown to be altered in some rodent models of prenatal immune activation and are relevant to neurodevelopmental psychiatric disorders such as schizophrenia and autism. Finally, in light of locomotor deficits we described previously in this model (Harvey and Boksa, 2014), we assessed motor performance, as well as exploratory activity in an open field. We hypothesized that maternal ID and prenatal LPS administration would interact to significantly alter behavior in juvenile and adult offspring, resulting in a more severe behavioral phenotype than the phenotype resulting from either prenatal insult alone.

2. Methods 2.1. Animals Timed pregnant Sprague Dawley rats were obtained from Charles River at E2 and housed individually at 21 ± 2 °C with a 12 h light/dark cycle (lights on at 8:00 a.m) and with food and water ad libitum. Animals were treated in accordance with guidelines from the Canadian Council on Animal Care (www.ccac.ca) and protocols approved by the McGill University Animal Care Committee. 2.1.1. Dietary manipulation of iron intake On E2, pregnant dams were placed on either an ID (background iron 3–8 ppm) or IS (200 mg/kg) modified AIN-93G diet (Harlan Teklad Diets, Madison WI, USA). Dams continued on their respective diets throughout pregnancy, until P7, at which time all dams were placed on Global 18% Protein Diet (iron content, 200 mg/kg, Harlan Teklad Diets). 2.1.2. LPS administration to pregnant rats On E15 and E16, dams on IS and ID diets received an injection of LPS (50 lg/kg, i.p., Escherichia coli serotype 0111:B4, Lot 42k4120, Sigma–Aldrich, Canada) or 0.9% saline (4 ml/kg, i.p.), between 10:00 a.m. and noon. Pregnancies were allowed to continue as normal. Previous experiments in our laboratory have shown that dosages of 50–100 lg/kg LPS administered between E15–E18 reliably induce fever, production of pro- and anti-inflammatory cytokines and increases in serum corticosterone in the dam (Ashdown et al., 2007; Cui et al., 2009, 2011; Harvey and Boksa, 2014). We have shown that LPS administered i.p. at E18 reaches the placenta, but not the fetus, and induces a pro-inflammatory cytokine response in the placenta but not in fetal brain or liver (Ashdown et al., 2006).

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

2.1.3. Offspring At birth, whole litters were maintained. There was no significant difference in litter size between groups (IS saline: 11.6 ± 0.3; IS LPS: 10.5 ± 0.5; ID saline: 10.2 ± 0.5; ID LPS: 10.4 ± 0.5; n = 20– 23 litters per group; means ± SEM) and no obvious miscarriage/ fetal loss. At P21, offspring were weaned, group housed (2–3 per cage) and fed standard rat chow (Global 18% Protein Diet, Harlan Teklad Diets). Only male offspring were used in this study. 2.2. Study design While the focus of this paper was to assess the adult behavioral phenotype of male offspring from maternal ID and prenatal LPS treated dams, we also tested juvenile offspring for passive avoidance behavior and response to amphetamine. Both of these behaviors are well described in juvenile rats and differences may be already emerging at a young age. We did not test social interaction in juvenile offspring as the style of testing we conducted may have been confounded by the juveniles’ propensity for play behavior. All male juvenile offspring were tested for passive avoidance, followed one week later by testing for amphetamine induced hyperlocomotion. A subset of naive juvenile offspring was assessed for Rotarod performance. (Given that initial experiments showed no change in Rotarod performance in juveniles, we only tested juvenile but not adult offspring on the Rotarod.) Adult male offspring were tested according to one of four Test Schedules. Schedule 1 animals were tested for sensitivity to novelty. Schedule 2 animals were tested for social interaction followed one week later by passive avoidance testing. Schedule 3 animals were tested for PPI followed one week later by amphetamine induced hyperlocomotion. Schedule 4 animals were tested for passive avoidance and a subset was tested for shock sensitivity. For each Test Schedule, offspring from several waves of dams generated at different times (but with each wave containing all 4 experimental groups) were tested, and data from all offspring were combined, to ensure that effects were not due to a single aberrant wave of dams. 2.3. Rotarod To test motor balance and coordination, juvenile rats (P35–P38) were trained to walk on the rotating textured drum (diameter = 9.5 cm) of a Rotarod apparatus (Rota-rod, Series 8, IITC Life Science) on day 1 over three training trials. On day 2, rats were assessed over three test trials. Each test session (maximum duration 5 min) began at 4 RPM and reached a maximum speed of 25 RPM over the first 150 s. Latency to fall was recorded and the best score used for analysis. 2.4. Passive avoidance and baseline shock sensitivity Baseline shock sensitivity was assessed using an opaque Perspex box (42 cm  21 cm  20 cm) resting on a metal grid. A series of increasing electric shocks was delivered through the metal grid, from 0 to 1.0 mA in increments of 0.05, over 4 min. The mA at which the rat exhibited audible vocalizations and a jump/flinch was recorded. We used a 9 day protocol to assess passive avoidance learning. A Perspex box, with two equal chambers (21 cm  21 cm  20 cm) with a gate between them (6 cm tall), was positioned on a metal grid connected to a shock box. One chamber with an opaque lid was designated ‘‘dark’’, and the other chamber, with clear Perspex lid, was ‘‘light’’. On day 1, rats were habituated to the apparatus (5 min, under red light) and were able to move freely between chambers. On day 2, the entire test room was dark except for a small lamp light shining directly into the light chamber. The rat

29

was placed in the light chamber and, after 10 s, the gate was opened. The latency for the rat to cross and place all four paws in the dark chamber was recorded (maximum 5 min), and then the gate was closed and an electric shock (juveniles: 0.8 mA for 2 s, adults: 0.8 mA for 3 s) was delivered. The rat was then returned to its home cage. In order to test learning, on day 3 (i.e., 24 h after the electric shock), and day 9 (i.e., 7 days after the electric shock), the rat was placed in the light chamber with the lamp on and the gate shut. After 10 s, the gate was opened and the latency for the rat to move to the dark chamber (maximum 10 min) was recorded. No electric shock was administered on day 3 or day 9. Some rats did not learn the association between the dark chamber and the electric shock, and were quicker to move to the dark chamber after 24 h than they were prior to experiencing the shock. A higher percentage of juveniles (21%) did not learn the task compared to adults who did not learn the task (14%). There was no effect of prenatal diet or LPS treatment on the numbers of juvenile rats who did not learn the task (IS-SAL: 10/37, IS-LPS: 6/31, ID-SAL: 5/28, ID-LPS: 6/24) but there did appear to be a pattern in the adult rats, in which more offspring from LPS treated dams did not learn the association (IS-LPS: 6/28, ID-LPS: 5/28) compared to offspring from saline treated dams. (IS-SAL: 3/29, ID-SAL: 2/28). Rats that did not learn the association were excluded from the data analysis, in order to determine effects of prenatal diet or LPS treatment on retention of the association. 2.4.1. Brain metabolic activity following passive avoidance – cytochrome oxidase histochemistry Cytochrome oxidase histochemistry was used as a measure of metabolic activity (Wong-Riley, 1989). Subsets of juvenile and adult rats were decapitated immediately following their performance on the 24 h passive avoidance test. Brains were frozen in isopentane at 40 °C and stored at 80 °C. Twenty lm cryostat sections were cut at 20 °C at the following bregma levels: 3.24 mm (cingulate [Cg], prelimbic [PL] and infralimbic cortex [IL]); 2.16 mm (nucleus accumbens [Acb]); 3.12 mm (central amygdala (CeA) and basolateral amygdala (BlA); 3.84 mm (dorsal hippocampal CA1, CA3, dentate gyrus [DG] and retrosplenial cortex [RSC]); and 5.0 mm (ventral hippocampus, VH). Sections were stained for cytochrome oxidase activity by incubating them for 2 h at 37 °C in the dark in cytochrome c solution (100 mg diaminobenzidine, 40 mg cytochrome c, 8 g sucrose, 36 mg catalase, 225 mg nickel chloride in 180 ml 0.1 M phosphate buffer, pH 7.0, all reagents Sigma–Aldrich). Slides were rinsed in 0.1 M phosphate buffer with 10% sucrose (w/v) (5 min, 21 °C), fixed in 10% buffered formalin (30 min, 21 °C), washed 5 times in 0.1 M phosphate buffer, dehydrated in ethanol (50%, 70%, 96%, 100%), incubated in xylene substitute (2  5 min) and coverslipped. Cytochrome oxidase activity was assessed by quantitative densitometric analysis using an MCID Elite image analysis system (Imaging Research, St. Catharines, Canada) and is expressed as relative optical density (OD). Values for background staining from sections incubated in the absence of cytochrome c were subtracted for each brain region. Regions in both right and left hemispheres were analyzed and the mean of values for the two hemispheres was used. For each brain region, 4 sections were analyzed for each animal and the mean of values from the 4 sections was used as the value for that animal. 2.5. Pre-pulse inhibition of acoustic startle Startle reactivity and PPI of acoustic startle were measured in adult male offspring (P64–P67) using SR-LAB startle chambers (San Diego Instruments, San Diego, USA), as described previously (Fortier et al., 2007). Briefly, test sessions began with a 7 min acclimatization period in the presence of 65 dB white noise.

30

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

Five orienting pulse alone trials (120 dB for 30 ms) were presented, followed by pseudo random presentation of the following: 11 pulse alone trials (120 dB for 30 ms), five null trials with no stimulus, and five pre-pulse + pulse trials at three different pre-pulse intensities. The pre-pulse + pulse trials consisted of a 30 ms pre-pulse (at 69, 73 or 77 dB), followed by a 70 ms delay, then a startle pulse (120 dB, 30 ms). Average inter-trial interval was 17 s (range: 9–29 s) and the test session lasted approximately 17 min. For each rat, a background startle value (average of 5 null trials) was subtracted from the startle amplitude of the pulse alone and pre-pulse + pulse trials. The percentage of pre-pulse inhibition is expressed as (1 [startle amplitude on pre-pulse + pulse trial/mean startle amplitude on pulse alone trials])  100. 2.6. Amphetamine induced hyperlocomotion Locomotor activity was assessed in juvenile (P35–P38) and adult (P70) male offspring in Perspex activity chambers (40  40  40 cm) with infrared sensors, using VersaMax software (Accuscan Instruments, Columbus, OH, USA). At least one rat from each experimental group was tested at the same time and testing was conducted under red light. Rats were habituated to the apparatus for 1 h per day, on two consecutive days prior to testing. On day 3, each test session began with 30 min of habituation, followed by saline injection, and another 30 min of recording. D-amphetamine sulphate (2 mg/kg, i.p.) was then administered and locomotor activity was recorded for an additional 100 min. Data are presented as total distance travelled in 10 min bins. 2.7. Social interaction 2.7.1. Habituation in large arena Rats were habituated to a large, wooden open field chamber (100 cm  100 cm  45 cm) for 1 h on the day prior to social interaction testing. A video camera (Everio G GZ-MG653, JVC, USA) recorded the rats’ movements and we used TopScan (CleverSys Inc, USA) to assess distance travelled, time spent travelling greater than 20 mm/s, and percent time in the center and corners of the open field. 2.7.2. Social interaction behavior To assess social interaction behavior, unrelated male Sprague Dawley rats (group housed, 2/cage) were used as naïve rats, for a maximum of two non-consecutive trials. Adult male offspring (P54–P58), ‘‘test’’ rats, were paired with naïve rats, 5–50 g lighter than the test rat. On the day of testing, a single test rat was placed in the chamber and a naïve rat introduced into the opposite corner. Sessions lasted 10 min. Interactions were recorded by video camera and behavior was quantitated using TopScan; measures included the total distance the test rat travelled; latency for the test rat to contact the naïve rat; and number of bouts of social contact, sniffing, approaching, following and leaving behavior exhibited by the test rat towards the naïve rat. 2.8. Sensitivity to novelty 2.8.1. Habituation in small arena Rats were habituated to a small, Perspex open field chamber (60 cm  60 cm  45 cm) for 1 h on the day prior to testing for sensitivity to novelty. Behavior was recorded and quantified as in 2.7.1. 2.8.2. Sensitivity to novelty Sensitivity to novel objects was assessed in adult male offspring (P54–P58). Objects (10–12 cm tall, 8–10 cm wide) were assessed for their suitability in pilot experiments. Offspring were tested in

two stages. During the first stage, ‘‘familiarization’’, rats were placed in the chamber with two novel objects of the same type. The total amount of time each rat spent actively exploring the objects was recorded and, once the rat had spent a cumulative total of 30 s exploring both objects (or maximum 10 min in the chamber was reached), the rat was removed from the chamber. One hour later, preference for the familiar object was tested by placing the rat in the chamber with one familiar object (the same type of object from the familiarization stage) and one novel object for 3 min. The amount of time each rat spent in the corners containing the novel and familiar objects was assessed using TopScan.

2.9. Statistical analyses To control for litter effects, in experiments where multiple offspring per dam were tested, data within litters were averaged to produce a single value per litter. The only exception to this was the analysis of the amphetamine induced locomotion data. Given individual differences may result due to drug administration, we analysed this data per rat. Unless otherwise specified, data were analyzed using two way analysis of variance (ANOVA) with maternal diet (IS or ID) and prenatal treatment (saline [Sal] or LPS) as between subject factors. For PPI data, three-way ANOVA was conducted, with maternal diet and prenatal treatment as betweensubject factor and pre-pulse intensity as within-subject measure. For the amphetamine-induced locomotion test, three way ANOVA was conducted, with maternal diet and prenatal treatment as between-subject factors and time (for habituation, saline, amphetamine treatment, respectively) as within-subject measure. Post hoc analyses were performed using Bonferroni’s test. Results were analysed using GraphPad Prism (GraphPad Software, San Diego, USA) and SPSS (SPSS Inc., Chicago, USA). Probabilities of p < 0.05 were considered statistically significant.

3. Results 3.1. Maternal ID decreases juvenile but not adult body weight There was a significant difference in juvenile (P30) body weight in male offspring due to maternal ID (F(1,60) = 18.22, p < 0.001) but no effect of prenatal LPS and no diet  prenatal LPS treatment interaction (n = 15–18 dams, 1–3 offspring per dam). Offspring from ID saline (67.5 ± 1.7 g) and ID LPS (70.5 ± 3.6 g) treated dams were significantly lighter than offspring from IS saline (81.0 ± 3.0 g) and IS LPS (81.0 ± 2.5 g) treated dams. There was no effect of maternal ID or prenatal LPS administration on adult (P60) male offspring weight (data not shown, n = 9–19 dams, 1–3 offspring per dam). We did not weigh offspring at birth, however, we have previously described a significant decrease in P0 body weight in offspring from ID dams in this model (Harvey and Boksa, 2014).

3.2. Maternal ID and prenatal LPS administration do not alter Rotarod performance There was no effect of maternal diet or prenatal LPS treatment on the juvenile offsprings’ ability to walk on an accelerating Rotarod apparatus, as measured by the maximum RPM and time (s) spent on the drum (IS saline: 16.3 ± 1.8 RPM, 91.2 ± 14.5 s; IS LPS: 14.1 ± 2.4 RPM, 71.6 ± 17.2 s; ID saline: 14.4 ± 1.1 RPM, 74.0 ± 7.8 s; ID LPS: 14.6 ± 1.8 RPM, 6.9 ± 13.8 s; means ± SEM).

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

31

Fig. 1. Passive avoidance (PA) learning in juveniles and adults. PA behavior was assessed in male offspring born from dams on either an iron sufficient (IS) or iron deficient (ID) diet from E2 to P7, and injected with either saline (SAL) or lipopolysaccharide (LPS, 50 lg/kg) on E15 and 16. Latency to cross to the dark chamber during training and testing (24 h and 7 days post shock) are shown for (A) juvenile (P25–P32) offspring and (C) adult (P63) offspring. A subset of juveniles (B) and adults (D) were collected immediately following the 24 h PA test and metabolic activity in the ventral hippocampus assessed using cytochrome oxidase histochemistry. Data are means ± SEM. For PA: juveniles, n = 11–17 litters/group and adults, n = 9–19 litters/group; data from 1 to 3 male offspring per dam were averaged to produce a single value per litter. For CytOx: n = 7–9 male offspring per group. #, ###Maternal ID offspring differ from maternal IS offspring at p < 0.05 and p < 0.001 respectively, due to a main effect of maternal diet.

3.3. Maternal ID decreases passive avoidance learning and increases metabolic activity in the ventral hippocampus after passive avoidance recall in adult offspring There was no difference in baseline shock sensitivity due to prenatal diet or prenatal LPS treatment in adult offspring as assessed by the threshold shock intensity (mA) at which animals first vocalised or flinched in response to shock (data not shown). Passive avoidance was assessed in juvenile (Fig. 1A) and adult offspring (Fig. 1C). Latency to cross from the light to the dark chamber was recorded during training and testing (24 h and 7 days post shock). There was no effect of maternal diet or prenatal LPS treatment on latency to cross during training or testing in juvenile offspring (Fig 1A). In adult offspring (Fig 1C), there was no effect of maternal diet or prenatal LPS treatment on latency to cross during training, indicating that the groups showed similar initial preference for the dark chamber before being shocked. Most interestingly, in adults there was a main effect of prenatal diet on the latency to cross during the 24 h test (F(1,66) = 16.76, p < 0.001), but no effect of prenatal LPS and no diet  prenatal LPS treatment interaction. At 24 h, offspring from ID dams displayed a decreased latency to cross compared to offspring from IS dams, indicating that offspring in the ID groups had learned the association between shock and the dark chamber less well. By 7 days post-shock, there were no longer any group differences. Metabolic activity was assessed in brain regions of rats sacrificed immediately after 24 h passive avoidance testing. There was

a significant main effect of maternal ID on metabolic activity in the VH of adult rats (F(1,27) = 4.55, p < 0.05, Fig 1D); offspring from ID dams displayed higher metabolic activity. This effect was not evident in the juveniles (Fig 1B). There was no effect of maternal ID or prenatal LPS treatment on metabolic activity in any other adult brain region assessed (Cg, PL, IL, Acb, CeA, BlA, CA1, CA3, DG or RSC, data not shown). 3.4. Maternal ID decreases PPI in adult offspring There were significant main effects of maternal diet (F(1,29) = 4.89, p < 0.05) and pre-pulse intensity (F(2,58) = 44.93, p < 0.001) on PPI in the adults (Fig. 2); offspring from ID dams showed reductions in PPI compared to offspring from IS dams across all pre-pulse intensities. There was no significant effect of prenatal LPS and no diet  prenatal LPS interaction on PPI. There was no effect of maternal diet or prenatal LPS treatment on baseline 120 dB startle responses or on startle responses to background noise (data not shown). 3.5. Prenatal LPS administration has small effects on amphetamine induced locomotion There was a trend for amphetamine induced locomotion to be greater in juvenile offspring from LPS treated dams compared to saline treated dams (p = 0.057) across all time points; however this missed statistical significance (Fig. 3A). There was no significant

32

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

Fig. 2. Pre-pulse inhibition in adults. Pre-pulse inhibition (PPI) of the acoustic startle reflex was assessed in adult male offspring (P64–P67) born from dams on either an iron sufficient (IS) or iron deficient (ID) diet from E2 to P7, and injected with either saline (SAL) or lipopolysaccharide (LPS) on E15 and 16. Data are expressed as percent of PPI at three different pre-pulse intensities [4, 8, 12 decibels (dB) above baseline] and are means ± SEM. n = 7 –9 litters/group; data from 1 to 3 male offspring per dam were averaged to produce a single value per litter. # Maternal ID differ from maternal IS offspring at p < 0.05 due to a main effect of maternal diet.

main effect of maternal ID and no diet  prenatal LPS interaction. After saline injection, juvenile offspring from LPS treated dams traveled further in the first ten minutes after the injection than controls (p < 0.01, Fig. 3A).

In adult offspring, there was no significant main effect of maternal ID or prenatal LPS and no diet  prenatal LPS interaction on amphetamine-induced locomotion. However, there was a significant interaction between time and prenatal LPS treatment (F(9,1071) = 2.03, p < 0.05, Fig. 3B). Post hoc analysis determined that amphetamine induced locomotion was significantly lower in adult offspring from LPS treated dams compared to saline treated dams during the first 30 min after amphetamine injection (p < 0.001) but there were no significant differences at later time points, suggesting that offspring from LPS treated dams were slower to respond to the amphetamine treatment. During the first 10 min of habituation, IS LPS offspring (p < 0.01) and ID saline offspring (p < 0.05) demonstrated small but significant decreases in locomotion compared to IS saline offspring.

3.6. Size of open field arena mediates subtle effects of maternal ID and prenatal LPS administration on exploratory behavior in adult offspring We assessed locomotor activity and behavior in a small and large open field arena. There was no significant main effect of maternal diet or prenatal LPS treatment and no diet  treatment interaction on total distance travelled in the small or large arenas. There were, however, effects on more subtle behavioral measures. In the small arena, there was a main effect of maternal ID on the time spent travelling at speeds of greater than 20 mm/s (F(1,41) = 4.66, p < 0.05,

Fig. 3. Amphetamine induced locomotion in juveniles and adults. Amphetamine induced locomotion was assessed in male offspring born from dams on either an iron sufficient (IS) or iron deficient (ID) diet from E2 to P7, and injected with either saline (SAL) or lipopolysaccharide (LPS) on E15 and 16. (A) Juvenile (P35–P38) and (B) adult offspring (P70) were habituated to the test chamber (1 h/day) for two days prior to testing. On the day of testing, activity was recorded during 30 min baseline, for 30 min following saline injection, and 100 min following injection with amphetamine (AMPH, 2 mg/kg). Data are expressed as distance travelled in 10 min bins, and are means ± SEM. Juveniles: n = 20–27 offspring/group; from 15 to 18 dams. Adults: n = 29–34 offspring/group; from 14 to 16 dams. ⁄⁄LPS treated differ from saline treated at p < 0.05 due to a main effect of LPS treatment. aOffspring from IS LPS treated dams (p < 0.01) and ID saline treated dams (p < 0.05) differ from offspring from IS saline treated dams. tLPS treated differ from saline treated (p < 0.001) at t = 70, 80 and 90 min.

33

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

Table 1 Behavior in small and large open field arenas in adults. Adult male offspring (P54–P58) born from dams on either an iron sufficient (IS) or iron deficient diet (ID) from embryonic day 2 (E2) to postnatal day 7 (P7) and injected with either saline or lipopolysaccharide (LPS) on E15 and 16 were tested for their exploratory behavior in a small (60 cm  60 cm) and large (100 cm  100 cm) open field arena for 10 min. Values are mean ± SEM. n = 10–13 litters/group for small arena and n = 13–16 litters/group for the large arena. Data from 1 to 3 male offspring per dam were averaged to produce a single value per litter. Iron sufficient

Iron deficient

Saline

LPS

Saline

LPS

SMALL ARENA Distance travelled (m) Travelling > 20 mm/s (% of time, s) In centre (% of time, s) In corners (% of time, s)

40.8 ± 1.5 65.5 ± 1.4 3.2 ± 0.8 51.9 ± 1.8

39.4 ± 1.7 63.8 ± 2.4 2.8 ± 0.6 50.4 ± 1.9

38.5 ± 1.9 61.8 ± 2.2# 4.4 ± 0.9 52.7 ± 1.1

35.1 ± 1.8 58.1 ± 2.3# 2.8 ± 0.6 52.6 ± 1.9

LARGE ARENA Distance travelled (m) Travelling > 20 mm/s (% of time, s) In centre (% of time, s) In corners (% of time, s)

57.6 ± 4.1 68.7 ± 2.9 3.0 ± 0.5 62.3 ± 2.2

58.2 ± 2.4 71.2 ± 0.9 3.1 ± 0.5 58.3 ± 1.6*

57.8 ± 3.5 70.2 ± 1.6 2.7 ± 0.5 65.9 ± 2.0

58.9 ± 2.5 71.9 ± 1.4 4.2 ± 0.6 59.6 ± 2.2*

# *

Maternal iron deficient differ from maternal iron sufficient offspring at p < 0.05 due to a main effect of maternal diet. LPS treated differ from saline treated at p < 0.05 due to a main effect of LPS treatment.

Table 1), with rats from ID dams spending less time moving ‘‘fast’’. This was reflected numerically in the total distance travelled, although this effect only approached statistical significance (p = 0.077). In the large arena, there was a main effect of prenatal LPS on the percentage of time spent in corners (F(1,55) = 6.17, p < 0.05, Table 1), with offspring from LPS treated dams spending less time in corners.

3.7. Prenatal LPS administration increases social behavior in adult offspring Prenatal LPS treatment increased social behavior in adult offspring. There were significant main effects of prenatal LPS on bouts of social contact (F(1,55) = 5.74, p < 0.05, Fig. 4A), social following (F(1,55) = 4.14, p < 0.05 Fig. 4B) and distance travelled (F(1,55) = 6.73,

p < 0.05, Fig. 4C). Offspring from LPS treated dams engaged in more bouts of social contact and of social following, and travelled further, than offspring from saline treated dams. There was also an effect of prenatal LPS on sniffing behavior (F(1,55) = 4.82, p < 0.05, Fig. 4D), with offspring from LPS dams spending more periods sniffing the stranger rat. There was also a main effect of maternal diet on the number of bouts of sniffing behavior (F(1,55) = 8.93, p < 0.01, Fig. 4D) and on the percent of total time spent sniffing (F(1,55) = 6.30, p < 0.05, data not shown), with offspring from ID dams spending less time sniffing the stranger rat. Finally, there was an interaction between maternal diet and prenatal LPS on the latency of the test rat to make social contact with the stranger rat (F(1,55) = 4.61, p < 0.05, Fig. 4E). Bonferroni post hoc analysis indicated that offspring from ID saline treated dams took longer to initiate social contact compared to offspring from IS saline treated dams (p < 0.05) and offspring from ID LPS treated dams (p < 0.01).

Fig. 4. Social interaction in adults. Adult male offspring (P54–P58) born from dams on either an iron sufficient (IS) or iron deficient diet (ID) from E2 to P7 and injected with either saline (SAL) or lipopolysaccharide (LPS) on E15 and 16 were tested for social interaction behavior with naïve, non-related rats for 10 min in a large 100  100 cm arena. Measurements shown are: (A) number of bouts of social contact between the test and naïve rat; (B) the number of times the test rat either approached, followed or left the naïve rat; (C) distance travelled by the test rat (mm); (D) the number of bouts the test rat spent sniffing the naïve rat; (E) latency for the test rat to make contact with the naïve rat (s). Values are means ± SEM. n = 13–16 litters/group; data from 1 to 3 male offspring per dam were averaged to produce a single value per litter. ⁄LPS treated differ from saline treated at p < 0.05 due to a main effect of LPS treatment. ###Maternal ID differ from maternal IS offspring at p < 0.001, respectively, due to a main effect of maternal diet. a Differs from IS-SAL and from ID-LPS at p < 0.05.

34

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

Fig. 5. Sensitivity to novelty in adults. Sensitivity to novelty was assessed in adult male offspring (P54–P58) born from dams on either an iron sufficient (IS) or iron deficient (ID) diet from E2 to P7, and injected with either saline (SAL) or lipopolysaccharide (LPS) on E15 and 16. Animals were habituated to a small arena for 10 min. The following day, two identical novel objects were placed in two corners of the arena and the time for the rat to explore both objects, for a combined 30 s, was recorded (A). One hour later, rats were placed in the arena which now contained one familiar object and one novel object. The time the rat spent in the novel and familiar corners was recorded over a three minute test period (B). Values are mean ± SEM. n = 13–16 litters/group; data from 1 to 3 male offspring per dam were averaged to produce a single value per litter. aaDiffers from ID-SAL at p < 0.01. bDiffers from IS-SAL at p < 0.05.

3.8. Sensitivity to novelty To investigate if our social behavior findings might be accounted for by the rats’ more general reaction to novelty, we assessed sensitivity to novelty using inanimate objects. When placed in an arena containing two identical novel objects, there was a maternal diet  prenatal LPS interaction on the time it took rats to explore both objects for a combined total of 30 s (F(1,52) = 5.23, p < 0.05, Fig. 5A). Rats from IS LPS treated dams took significantly less time to reach the 30 s criterion compared to offspring from IS saline dams (p < 0.05, Fig. 5A). One hour after familiarization to the initial objects, when rats were presented with one familiar and one novel object, there was a maternal diet  prenatal treatment interaction on the amount of time rats spent in the corner containing the familiar object (F(1,54) = 4.53, p < 0.05, Fig. 5B), with rats from ID LPS treated dams spending significantly less time in the familiar corner compared to rats from ID saline treated dams. There was no difference due to maternal diet or prenatal treatment on the time each rat spent actively exploring either object (nose to object, data not shown). 4. Discussion This study describes alterations in adult behaviors in rat offspring as a consequence of maternal ID and prenatal immune activation. As with our first study describing this model (Harvey and Boksa, 2014), we find a distinct lack of interaction between maternal ID and prenatal LPS exposure on our chosen behavioral markers. Instead, we describe an additive phenotype. In adult offspring, we describe deficits in passive avoidance and PPI as a result of maternal ID, and increases in social behavior as a result of prenatal LPS administration. We do not find any alterations in passive avoidance in juvenile offspring from ID or LPS treated dams, indicating that the deficit in passive avoidance observed in adult offspring following maternal ID is a late-developing abnormality. A significant strength of this study is the large number of litters that were produced, over time, in ‘‘waves’’ which allowed us to test multiple offspring per litter, and multiple litters over time. This study design promotes representative sampling and provides findings that are replicated within each data set. 4.1. Maternal ID produces deficits in body weight, memory and PPI In this study, we describe a significant weight deficit in juvenile but not adult offspring of ID dams. In our initial description of the model we found that adult male offspring from ID saline treated

dams were significantly lighter than controls, yet offspring from ID LPS treated dams appeared to have been protected from this effect (Harvey and Boksa, 2014). Both the current and previous study weighed multiple offspring per dam, from large numbers of dams, across a period of several months. Thus it appears as if subtle as yet undefined factors may influence whether early deficits in offspring weight due to maternal ID are maintained into adulthood. One difference between this study and the previous is that in the first study, offspring were weighed every week whereas in this study, offspring remained unhandled, aside from routine cage maintenance. We used a passive avoidance task to assess fear-motivated associative learning and subsequent short and long term memory retention in juvenile and adult offspring. We describe a significant deficit in memory after 24 h in adult offspring from ID dams, suggesting that maternal ID can lead to memory deficits in adulthood. Interestingly, we find a corresponding increase in cytochrome oxidase histochemistry, a metabolic marker of neuronal activity in the VH of offspring from ID dams sacrificed immediately after the 24 h passive avoidance test session. Previous lesion studies have indicated strong involvement of the VH in conditioned passive avoidance responses to painful stimuli (Ambrogi Lorenzini et al., 1997; Degroot and Treit, 2002; Trivedi and Coover, 2004). While at first glance one might expect decreased VH metabolic activity in animals with a deficit in passive avoidance, it is possible that increased metabolic activity in the VH represents either increased inhibitory activity or over-activity of an inefficient VH to attempt to meet the demands of the passive avoidance task in the offspring from ID dams. Regardless of the interpretation, the results implicate the VH as a possible region involved in memory abnormalities following maternal ID. To our knowledge, this is the first study to reports deficits in passive avoidance behavior in offspring from ID rat dams. Our finding of a long term deficit in passive avoidance memory due to maternal ID is consistent with reports of long term deficits in learning and memory in human children following maternal ID (Lozoff et al., 2006). In this study we described a deficit in PPI response, across all decibel levels, in adult offspring from ID dams. To our knowledge, this is the first study to report effects of maternal ID on PPI. Related measures, including alterations in hearing, auditory processing and baseline acoustic startle have previously been described in offspring in rodent models of maternal ID (Jougleux et al., 2011). 4.2. Prenatal LPS results in increases in social interaction and subtle changes in amphetamine induced locomotion We describe increases in multiple measures of social interaction in offspring from LPS treated dams, including total time spent in

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

social contact with a naïve rat, and increased time spent sniffing and following the naïve rat. This is the first study to describe an increase in social interaction in adult rat offspring as a result of prenatal LPS treatment. There have been some reports of reduced social interaction in rat (Kirsten et al., 2010) and mouse (Hava et al., 2006) offspring as a result of prenatal LPS, however, other studies reported no differences (Moser, 2013; Foley et al., 2014). Thus, it appears that alterations in social interaction following prenatal immune activation may be highly dependent on the species or on the immunogen, its dosage, and the gestational age at which it is administered. Given the experimental paradigm we used to assess social interaction is also commonly used to measure anxiety (File and Seth, 2003), our findings could suggest that offspring from LPS treated dams not only have a preference for social interaction, but also exhibit a less anxious phenotype. Past studies have found conflicting effects of prenatal immune activation on anxiety-like behaviors in adult offspring, however, recent studies suggest that the adult behavioral phenotype is dependent on the rodent strain and specific anxiety-related task tested (Babri et al., 2013; Lin et al., 2012). Our findings on sensitivity to novel objects are particularly interesting in their parallels to our social behavior findings. When presented with two novel inanimate objects in an arena, offspring from IS LPS treated dams interacted with these objects more quickly, taking less time to reach a criterion of 30 s. This is similar to their response in the social behavior experiment, where offspring from LPS treated dams displayed increased social behavior, that is, increased interaction with a novel animate object in their environment. In this study, we describe small, subtle differences in the amphetamine response due to prenatal LPS treatment, with a non-significant trend (p = 0.057) towards enhanced locomotor activity after amphetamine in juvenile offspring from LPS treated dams, and a time-dependent reduction in locomotor activity in adult offspring from LPS treated dams. These results are in contrast to previous findings from our laboratory which described an increase in amphetamine induced locomotion in adult offspring from dams treated with LPS at E18/19 (Fortier et al., 2004). This suggests that the precise timing of LPS administration during pregnancy may be an important determinant of effects on amphetamine-induced locomotion, since the current study administered LPS earlier in gestation on E15/16. We also saw no statistically significant change in PPI response in offspring from LPS treated dams. However, visual inspection suggests that LPS treatment could possibly contribute to the PPI deficit seen in offspring from ID LPS treated dams. In contrast to the current study, deficits in PPI have been reported in previously published studies of animal models of prenatal immune activation, both in our and others’ laboratories (Shi et al., 2003; Fortier et al., 2004, 2007; Smith et al., 2007; Wolff and Bilkey, 2008; Meyer et al., 2008; Romero et al., 2010). We used a lower mg/kg dose of LPS in the current combined ID /prenatal LPS model compared to our previous papers where we showed altered PPI in adult offspring using a model of prenatal LPS alone (Fortier et al., 2004, 2007). Given that deficits in PPI following prenatal polyinosinic:polycytidylic acid [poly (I:C)] administration are dependent on precise dosing (Meyer et al., 2005), it would not be surprising if this was also the case for LPS. There are also other differences between the current combined ID/prenatal LPS model and our previous prenatal LPS model. In the current study, litters consisting of both sexes were raised by their birth dams versus our previous studies, where male pups were cross fostered with surrogate dams, resulting in male only litters comprised of pups from different dams. Also, in this study we fed both control and ID dams a casein-based maternal diet designed to support gestation and lactation while a grain-based standard rat chow was used in the previous prenatal LPS model. Fat and sodium content in the dam’s diet have been reported to alter the offspring’s locomotor response

35

to amphetamine in later life (McBride et al., 2006; Naef et al., 2008). We also know that the PPI response is particularly sensitive to alteration by numerous environmental exposures, including prenatal stress (Koenig et al., 2005). Indeed, in models using sub-threshold doses of prenatal poly (I:C), PPI deficits may be absent unless offspring are subjected to additional environmental (Giovanoli et al., 2013) or genetic (Lipina et al., 2013) manipulations. We describe no alterations in Rotarod performance in the current model, suggesting the offspring’s gross motor function is unimpaired. However, using different arenas, we describe subtle effects, with a reduction in time spent in corners in the large arena, in offspring from LPS treated dams, an effect not evident in the smaller arena. Our findings suggest that the size of the arena is another variable that can directly influence the rats’ habituation, exploratory behavior, and locomotor activity; these considerations should be taken into account when interpreting data. 4.3. Assessing the maternal ID/prenatal LPS model We have used an animal model of moderate maternal iron deficiency that mimics ‘‘real world’’ ID in pregnant women and induces deficits in brain iron similar to those described in humans with perinatal ID (Rao et al., 2003). We have induced a transient, midgestation inflammatory response that correlates with common infections, such as bacterial vaginosis, that are prevalent during pregnancy (Deb et al., 2004). We have assessed juvenile and adult behaviors that bear similarity to human behavioral measures, including response to amphetamine, PPI, simple associative learning, and social interaction (Powell and Miyakawa, 2006). Thus, we have created a viable tool to model the effects of maternal ID and prenatal immune activation. Collectively, our results suggest that there are significant consequences to exposure to maternal ID and prenatal immune activation. We have previously shown that these two factors interact in the pregnant rat to potentiate the pro-inflammatory cytokine cascade induced by LPS administration (Harvey and Boksa, 2014). Given that either maternal iron deficiency or prenatal immune activation alone have been described to produce common alterations in dopamine and serotonin transmission, myelination, and hippocampal development and function (Boksa, 2010; Georgieff, 2011) we hypothesized that these factors would interact during neurodevelopment to worsen hippocampal- and dopamine-dependent juvenile and adult behaviors. However, though we assessed a variety of behaviors, representing a wide sample of neurotransmitter and regional function, we instead describe a striking lack of interaction between maternal ID and prenatal immune activation. Yet, this is not to say that the co-incidence of maternal ID and prenatal immune activation in the same animal, be it human or rodent, should be of no concern. Indeed, the breadth of behaviors that are altered in this model highlight important long-term consequences of these two common exposures. Rats that are exposed to both maternal ID and prenatal immune activation develop deficits and alterations in six different behaviors, each of which are associated with altered function in humans. This ‘‘additive’’ phenotype may also create vulnerabilities to other biological insults during adult life and predispose the offspring to additional behavioral or neurochemical abnormalities. Importantly, given the consistent lack of interaction between these factors in the offspring, we suggest that these exposures may affect brain function and behavior via independent mechanisms of action. Given the high prevalence of both iron deficiency and infections in pregnant women and women of childbearing age around the world, there is a relatively high likelihood that the two exposures will co-occur in human pregnancies. Our results suggest that prevention or treatment of these conditions may have a positive impact on neurodevelopmental outcomes and behavior in later life in offspring.

36

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37

5. Conflict of interest statement All authors declare that there are no conflicts of interest.

Acknowledgments We would like to thank Ying Zhang for technical assistance with the cytochrome oxidase histochemistry and Dominique Nouel for technical assistance with brain collection. This work was supported financially by the Canadian Institutes of Health Research.

References Ambrogi Lorenzini, C.G., Baldi, E., Bucherelli, C., Sacchetti, B., Tassoni, G., 1997. Role of ventral hippocampus in acquisition, consolidation and retrieval of rat’s passive avoidance response memory trace. Brain Res. 768 (1–2), 242–248. Ashdown, H., Dumont, Y., Ng, M., Poole, S., Boksa, P., Luheshi, G.N., 2006. The role of cytokines in mediating effects of prenatal infection on the fetus: implications for schizophrenia. Mol. Psychiatry 11, 47–55. Ashdown, H., Poole, S., Boksa, P., Luheshi, G.N., 2007. Interleukin-1 receptor antagonist as a modulator of gender differences in the febrile response to lipopolysaccharide in rats. Am. J. Physiol. Regul. Integr. Comp. Physiol. 292 (4), R1667–R1674. Babri, S., Doosti, M.H., Salari, A.A., 2013. Strain-dependent effects of prenatal maternal immune activation on anxiety- and depression-like behaviors in offspring. Brain Behav. Immun. 37, 164–176. Boksa, P., 2010. Effects of prenatal infection on brain development and behavior: a review of findings from animal models. Brain Behav. Immun. 24, 881–897. Brown, A.S., 2006. Prenatal infection as a risk factor for schizophrenia. Schizophr. Bull. 32, 200–202. Brown, A.S., 2011. The environment and susceptibility to schizophrenia. Prog. Neurobiol. 93, 23–58. Coe, C.L., Lubach, G.R., Bianco, L., Beard, J.L., 2009. A history of iron deficiency anemia during infancy alters brain monoamine activity later in juvenile monkeys. Dev. Psychobiol. 51, 301–309. Connor, J.R., Menzies, S.L., 1996. Relationship of iron to oligodendrocytes and myelination. Glia 17, 83–93. Cui, K., Ashdown, H., Luheshi, G.N., Boksa, P., 2009. Effects of prenatal immune activation on hippocampal neurogenesis in the rat. Schizophr. Res. 113 (2–3), 288–297. Cui, K., Luheshi, G.N., Boksa, P., 2011. Effects of endogenous glucocorticoid secretion on the interleukin-6 response to bacterial endotoxin in pregnant and nonpregnant rats. J. Endocrinol. 209 (1), 95–103. Deb, K., Chaturvedi, M.M., Jaiswal, Y.K., 2004. Comprehending the role of LPS in Gram-negative bacterial vaginosis: ogling into the causes of unfulfilled childwish. Arch. Gynecol. Obstet. 270 (3), 133–146. de Deungria, M., Rao, R., Wobken, J.D., Luciana, M., Nelson, C.A., Georgieff, M.K., 2000. Perinatal iron deficiency decreases cytochrome c oxidase (CytOx) activity in selected regions of neonatal rat brain. Pediatr. Res. 48, 169–176. Degroot, A., Treit, D., 2002. Dorsal and ventral hippocampal cholinergic systems modulate anxiety in the plus-maze and shock-probe tests. Brain Res. 949 (1–2), 60–70. Eseh, R., Zimmerberg, B., 2005. Age-dependent effects of gestational and lactational iron deficiency on anxiety behavior in rats. Behav. Brain Res. 164, 214–221. Felt, B.T., Beard, J.L., Schallert, T., Shao, J., Aldridge, J.W., Connor, J.R., Georgieff, M.K., Lozoff, B., 2006. Persistent neurochemical and behavioral abnormalities in adulthood despite early iron supplementation for perinatal iron deficiency anemia in rats. Behav. Brain Res. 171, 261–270. File, S.E., Seth, P., 2003. A review of 25 years of the social interaction test. Eur. J. Pharmacol. 463, 35–53. Foley, K.A., Macfabe, D.F., Vaz, A., Ossenkopp, K.P., Kavaliers, M., 2014. Sexually dimorphic effects of prenatal exposure to propionic acid and lipopolysaccharide on social behavior in neonatal, adolescent, and adult rats: Implications for autism spectrum disorders. Int. J. Dev. Neurosci. http://dx.doi.org/10.1016/ j.ijdevneu.2014.04.001. Fortier, M.E., Joober, R., Luheshi, G.N., Boksa, P., 2004. Maternal exposure to bacterial endotoxin during pregnancy enhances amphetamine-induced locomotion and startle responses in adult rat offspring. J. Psychiatr. Res. 38, 335–345. Fortier, M.E., Luheshi, G.N., Boksa, P., 2007. Effects of prenatal infection on prepulse inhibition in the rat depend on the nature of the infectious agent and the stage of pregnancy. Behav. Brain Res. 181, 270–277. Gambling, L., Danzeisen, R., Fosset, C., Andersen, H.S., Dunford, S., Srai, S.K.S., McArdle, H.J., 2003. Iron and copper interactions in development and the effect on pregnancy outcome. J. Nutr. 133, 1554S–1556S. Georgieff, M.K., 2011. Long-term brain and behavioral consequences of early iron deficiency. Nutr. Rev. 69 (Suppl 1), S43–S48. Giovanoli, S., Engler, H., Engler, A., Richetto, J., Voget, M., Willi, R., Winter, C., Riva, M.A., Mortensen, P.B., Feldon, J., Schedlowski, M., Meyer, U., 2013. Stress in puberty unmasks latent neuropathological consequences of prenatal immune activation in mice. Science 339 (6123), 1095–1099.

Haider, B.A., Olofin, I., Wang, M., Spiegelman, D., Ezzati, M., Fawzi, W.W.Nutrition Impact Model Study, G., 2013. Anaemia, prenatal iron use, and risk of adverse pregnancy outcomes: systematic review and meta-analysis. BMJ 346, f3443. Hare, D., Ayton, S., Bush, A., Lei, P., 2013. A delicate balance. Iron metabolism and diseases of the brain. Front. Aging Neurosci. 5, 34. Harvey, L., Boksa, P., 2014. Do prenatal immune activation and maternal iron deficiency interact to affect neurodevelopment and early behavior in rat offspring? Brain Behav. Immun. 35, 144–154. Hava, G., Vered, L., Yael, M., Mordechai, H., Mahoud, H., 2006. Alterations in behavior in adult offspring mice following maternal inflammation during pregnancy. Dev. Psychobiol. 48 (2), 162–168. Insel, B.J., Schaefer, C.A., McKeague, I.W., Susser, E.S., Brown, A.S., 2008. Maternal iron deficiency and the risk of schizophrenia in offspring. Arch. Gen. Psychiatry 65, 1136–1144. Jougleux, J.L., Rioux, F.M., Church, M.W., Fiset, S., Surette, M.E., 2011. Mild maternal iron deficiency anemia during pregnancy and lactation in guinea pigs causes abnormal auditory function in the offspring. J. Nutr. 141 (7), 1390–1395. Kemna, E., Pickkers, P., Nemeth, E., van der Hoeven, H., Swinkels, D., 2005. Timecourse analysis of hepcidin, serum iron, and plasma cytokine levels in humans injected with LPS. Blood 106, 1864–1866. Khandaker, G.M., Zimbron, J., Lewis, G., Jones, P.B., 2013. Prenatal maternal infection, neurodevelopment and adult schizophrenia: a systematic review of population-based studies. Psychol. Med. 43, 239–257. Kirsten, T.B., Taricano, M., Maiorka, P.C., Palermo-Neto, J., Bernardi, M.M., 2010. Prenatal lipopolysaccharide reduces social behavior in male offspring. NeuroImmunoModulation 17 (4), 240–251. Koenig, J.I., Elmer, G.I., Shepard, P.D., Lee, P.R., Mayo, C., Joy, B., Hercher, E., Brady, D.L., 2005. Prenatal exposure to a repeated variable stress paradigm elicits behavioral and neuroendocrinological changes in the adult offspring: potential relevance to schizophrenia. Behav. Brain Res. 156, 251–261. Kwik-Uribe, C.L., Gietzen, D., German, J.B., Golub, M.S., Keen, C.L., 2000. Chronic marginal iron intakes during early development in mice result in persistent changes in dopamine metabolism and myelin composition. J. Nutr. 130, 2821– 2830. Lin, Y.L., Lin, S.Y., Wang, S., 2012. Prenatal lipopolysaccharide exposure increases anxiety-like behaviors and enhances stress-induced corticosterone responses in adult rats. Brain Behav. Immun. 26 (3), 459–468. Lipina, T.V., Zai, C., Hlousek, D., Roder, J.C., Wong, A.H., 2013. Maternal immune activation during gestation interacts with Disc1 point mutation to exacerbate schizophrenia-related behaviors in mice. J. Neurosci. 33 (18), 7654–7666. Lozoff, B., Beard, J., Connor, J., Barbara, F., Georgieff, M., Schallert, T., 2006. Longlasting neural and behavioral effects of iron deficiency in infancy. Nutr. Rev. 64, S34–43, discussion S72–S91. McBride, S.M., Culver, B., Flynn, F.W., 2006. Prenatal and early postnatal dietary sodium restriction sensitizes the adult rat to amphetamines. Am. J. Physiol. Regul. Integr. Comp. Physiol. 291, R1192–R1199. McLean, E., Cogswell, M., Egli, I., Wojdyla, D., de Benoist, B., 2009. Worldwide prevalence of anaemia, WHO Vitamin and Mineral Nutrition Information System, 1993–2005. Public Health Nutrition 12, 444–454. Meyer, U., Feldon, J., Schedlowski, M., Yee, B.K., 2005. Towards an immunoprecipitated neurodevelopmental animal model of schizophrenia. Neurosci. Biobehav. Rev. 29 (6), 913–947. Meyer, U., Nyffeler, M., Yee, B.K., Knuesel, I., Feldon, J., 2008. Adult brain and behavioral pathological markers of prenatal immune challenge during early/ middle and late fetal development in mice. Brain Behav. Immun. 22, 469–486. Miller, J.E., Pedersen, L.H., Streja, E., Bech, B.H., Yeargin-Allsopp, M., Van Naarden Braun, K., Schendel, D.E., Christensen, D., Uldall, P., Olsen, J., 2013. Maternal infections during pregnancy and cerebral palsy: a population-based cohort study. Paediatr. Perinat. Epidemiol. 27, 542–552. Moser, P., 2013. Evaluating negative-symptom-like behavioural changes in developmental models of schizophrenia. Eur. Neuropsychopharmacol. 24 (5), 774–787. Naef, L., Srivastava, L., Gratton, A., Hendrickson, H., Owens, S.M., Walker, C.D., 2008. Maternal high fat diet during the perinatal period alters mesocorticolimbic dopamine in the adult rat offspring: reduction in the behavioral responses to repeated amphetamine administration. Psychopharmacology 197, 83–94. Pagani, A., Nai, A., Corna, G., Bosurgi, L., Rovere-Querini, P., Camaschella, C., Silvestri, L., 2011. Low hepcidin accounts for the proinflammatory status associated with iron deficiency. Blood 118, 736–746. Powell, C.M., Miyakawa, T., 2006. Schizophrenia-relevant behavioral testing in rodent models: a uniquely human disorder? Biol. Psychiatry 59 (12), 1198– 1207. Radlowski, E.C., Johnson, R.W., 2013. Perinatal iron deficiency and neurocognitive development. Front. Hum Neurosci. 7, 585. Rao, R., Tkac, I., Townsend, E.L., Gruetter, R., Georgieff, M.K., 2003. Perinatal iron deficiency alters the neurochemical profile of the developing rat hippocampus. J. Nutr. 133 (10), 3215–3221. Rao, R., Tkac, I., Schmidt, A.T., Georgieff, M.K., 2011. Fetal and neonatal iron deficiency causes volume loss and alters the neurochemical profile of the adult rat hippocampus. Nutr. Neurosci. 14, 59–65. Rasmussen, S.A., Jamieson, D.J., Uyeki, T.M., 2012. Effects of influenza on pregnant women and infants. Am. J. Obstet. Gynecol. 207, S3–S8. Romero, E., Guaza, C., Castellano, B., Borrell, J., 2010. Ontogeny of sensorimotor gating and immune impairment induced by prenatal immune challenge in rats: implications for the etiopathology of schizophrenia. Mol. Psychiatry 15, 372– 383.

L. Harvey, P. Boksa / Brain, Behavior, and Immunity 40 (2014) 27–37 Shi, L., Fatemi, S.H., Sidwell, R.W., Patterson, P.H., 2003. Maternal influenza infection causes marked behavioral and pharmacological changes in the offspring. J. Neurosci. 23, 297–302. Smith, S.E., Li, J., Garbett, K., Mirnics, K., Patterson, P.H., 2007. Maternal immune activation alters fetal brain development through interleukin-6. J. Neurosci. 27, 10695–10702. Sorensen, H.J., Nielsen, P.R., Pedersen, C.B., Mortensen, P.B., 2010. Association between prepartum maternal iron deficiency and offspring risk of schizophrenia: population-based cohort study with linkage of Danish national registers. Schizophr. Bull. 37, 982–987. Trivedi, M.A., Coover, G.D., 2004. Lesions of the ventral hippocampus, but not the dorsal hippocampus, impair conditioned fear expression and inhibitory avoidance on the elevated T-maze. Neurobiol. Learn. Mem. 81 (3), 172–184.

37

Velu, P.P., Gravett, C.A., Roberts, T.K., Wagner, T.A., Zhang, J.S., Rubens, C.E., Gravett, M.G., Campbell, H., Rudan, I., 2011. Epidemiology and aetiology of maternal bacterial and viral infections in low- and middle-income countries. J. Global Health 1, 171–188. Wolff, A.R., Bilkey, D.K., 2008. Immune activation during mid-gestation disrupts sensorimotor gating in rat offspring. Behav. Brain Res. 190, 156–159. Wong-Riley, M.T.T., 1989. Cytochrome oxidase: an endogenous metabolic marker for neuronal activity. Trends Neurosci. 12, 94–101. Wu, L.L., Zhang, L., Shao, J., Qin, Y.F., Yang, R.W., Zhao, Z.Y., 2008. Effect of perinatal iron deficiency on myelination and associated behaviors in rat pups. Behav. Brain Res. 188, 263–270. Yu, G.S., Steinkirchner, T.M., Rao, G.A., Larkin, E.C., 1986. Effect of prenatal iron deficiency on myelination in rat pups. Am. J. Pathol. 125, 620–624.