MODERATE PERINATAL ARSENIC EXPOSURE ALTERS NEUROENDOCRINE MARKERS ASSOCIATED WITH DEPRESSION AND INCREASES DEPRESSIVE-LIKE BEHAVIORS IN ADULT MOUSE OFFSPRING

Abstract

Arsenic is one of the most common heavy metal contaminants found in the environment, particularly in water. We examined the impact of perinatal exposure to relatively low levels of arsenic (50 parts per billion) on neuroendocrine markers associated with depression and depressive-like behaviors in affected adult C57BL/6J mouse offspring. Whereas most biomedical research on arsenic has focused on its carcinogenic potential, a few studies suggest that arsenic can adversely affect brain development and neural function.

Compared to controls, offspring exposed to 50 parts per billion arsenic during the perinatal period had significantly elevated serum corticosterone levels, reduced whole hippocampal CRFR₁ protein level and elevated dorsal hippocampal serotonin 5HT_1A receptor binding and receptor-effector coupling. 5HT_1A receptor binding and receptor-effector coupling were not different in the ventral hippocampal formation, entorhinal or parietal cortices, or inferior colliculus. Perinatal arsenic exposure also significantly increased learned helplessness and measures of immobility in a forced swim task.

Taken together, these results suggest that perinatal arsenic exposure may disrupt the regulatory interactions between the hypothalamic-pituitary-adrenal axis and the serotonergic system in the dorsal hippocampal formation in a manner that predisposes affected offspring to depressive-like behavior. These results are the first to demonstrate that relatively low levels of arsenic exposure during development can have long-lasting adverse effects on behavior and neurobiological markers associated with these behavioral changes.

INTRODUCTION

Failure to appropriately respond to external stressors has been implicated in a number of mental health disorders, particularly Major Depressive Disorder (Herman et al., 2003; Barden, 2004). Several lines of evidence suggest that HPA axis hyperactivity, due to excessive corticotrophin-releasing factor (CRF) release, combined with deficits in the negative feedback regulation of CRF release (Barden, 2004), may be a causative factor in the pathogenesis of depression (Pariente, 2004). In addition, elevated corticosterone levels have been associated with depression in a rodent model (Zhao et al, 2008). Elevated CRF release in response to stress is associated with onset of depression (Nemeroff 1992; Reul and Holsboer, 2002) and elevated cerebral spinal fluid levels of CRF have been reported in depressed patient populations (Nemeroff, 1988).

Several studies indicate functional interactions between CRF and the serotonergic system and have linked these interactions to depression (Leonard, 2005; Waselus and Van Bockstaele, 2007). In addition to the CRF receptors present in the raphe nucleus (Sakanaka et al, 1987; Swanson et al, 1983; Ruggiero et al, 1999; Kirby et al, 2000; Lowry et al, 2000; Valentino et al, 2001), CRF exerts a stimulatory effect on hippocampal serotonin (5HT) levels (Linthorst et al., 2002; Peñalva et al., 2002; Oshima et al., 2003). CRF receptor antagonists reduce 5HT levels in rat hippocampal formation (Isogawa et al, 2000; Linthorst et al, 2002). Homozygous CRFR₁ knockout mice showed enhanced hippocampal 5-hydroxyindoleacetic acid (5-HIAAA) under basal conditions and a greater rise in hippocampal 5HT during a forced swim task compared to wild type mice (Peñalva et al., 2002). The most abundant 5HT receptor in the hippocampus is the 5HT_1A receptor and its expression is regulated by HPA axis activity (Lopez et al., 1998); with expression being elevated during low corticosteroid conditions and decreased with high corticosteroid levels. Similarly, rats bred for high anxiety-related behavior (HAB) have lower 5HT_1A expression compared to the low anxiety-related behavior rats (Keck et al., 2004).

Arsenic is one of the most common environmental contaminants found in water, food and air (DeSesso et al., 1998). Human exposure has been associated with skin, lung, and bladder cancers, vascular diseases, hypertension, and diabetes (National Research Council, 2001; Kitchin, 2001). In addition, arsenic crosses the placental barrier (Jin et al, 2006) and can affect offspring during critical periods of brain development. Exposure has also been implicated in damaging the hypothalamic-pituitary-adrenal (HPA) axis (Delgado et al., 2000) leading to deficits in stress responding.

Given the reported effects of arsenic exposure on HPA axis responsiveness and the putative interactions between the HPA axis and the serotonergic system in depression, we examined the impact of moderate perinatal arsenic exposure on neuroendocrine markers associated with depression and depressive-like behaviors in mice. Because the perinatal exposure period encompasses the developmental period of the offspring from conception through embryonic stages, gestational stages, and postnatal stages until weaning, it is likely that a number of systems in the brain that are developing during these critical periods are affected. For example, 5HT_1A immunoreactivity has been shown as early as embryonic day 16 (Patel and Zhou, 2005) suggesting that this system is organized early on and could be affected in our model. We hypothesized that moderate arsenic exposure during the perinatal period would alter markers of the HPA axis and serotonergic neurotransmission and that these changes would be consistent with the appearance of depressive-like behavioral responses in affected offspring.

MATERIALS and METHODS

Perinatal Arsenic Exposure Paradigm

The arsenic exposure paradigm and behavioral tasks employed in these studies were approved by the UNM Health Sciences Center Institutional Animal Care and Use Committee. All mice were bred and maintained on a 12-hour light/dark cycle (lights on from 0700 to 1900) with food and water ad libitum in a temperature controlled (22 °C) room in the Animal Resource Facility. Female C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME) were assigned to either a control, arsenic-free water or 50 parts per billion (ppb) arsenic (sodium arsenate, Sigma) water treatment group. Arsenic-free water was prepared by reverse osmosis followed by a MilliQ water purification step. After a two-week acclimation period on the treated waters, male breeder mice were introduced into each female’s cage. Three days later, the males were removed from the cages and nesting material was placed in the female’s cage. Mouse dam water consumption was monitored throughout pregnancy and the dams continued to drink the treated waters until their offspring were weaned. Litter sizes and neonatal body weights were measured on postnatal day 7. Offspring were weaned at 23 days of age and maintained in same-sex, litter-mate housed cages with ad libitum access to untreated tap water and standard mouse chow until they were used in experimental procedures at 75 to 90 days of age. Tap water at UNM contains approximately 2 to 5 parts per billion arsenic. Mice (one per litter; 5–10 litters total per treatment group) were assigned to either one of the biochemical assay procedures or one of the two behavioral testing paradigms. Offspring used in behavioral studies were not used in biochemical studies.

Plasma Corticosterone

Blood samples were collected between 0800 and 1000 hours. Mice were decapitated and trunk blood collected into plastic SAFE-T-FILL capillary tubes prepared with liquid EDTA Di Potassium Salt (Ram Scientific Inc, Yonkers, NY). Tubes were immediately placed on ice and the plasma obtained by centrifugation (4000 rpm × 15 minutes) at 4°C. Supernatents were decanted into storage tubes and the samples frozen at −20 °C until assay. Plasma samples were then thawed and assayed for corticosterone using a rodent double antibody ¹²⁵I-corticosterone radioimmunoassay kit (MP Biomedicals, Orangeburg, NY), according to the manufacturer’s instructions. Samples were assayed in duplicate.

CRFR₁ Receptors

Tissue preparation

Adult offspring, 75 to 90 days of age, were sacrificed by decapitation and whole hippocampal formation rapidly dissected, rinsed, transferred to Dounce homogenizing tubes containing 0.5 mL homogenizing buffer (0.32 M sucrose, 1 mM EDTA, 20 mM Tris-HCL and 0.1 mM AEBSP; HB buffer) and homogenized by five loose and five tight up and down strokes. Homogenates were centrifuged at 1,000×g for 6 minutes, the supernatant fraction (S1) collected and centrifuged at 200,000×g for 30 min. The resulting pellet (P2) was resuspended with 0.5 mL HB buffer with 75 mM NaCl, 75 mM KCl and 1% Triton X-100 added, re-homogenized and centrifuged at 200,000×g for 30 min. to produce detergent-soluble proteins (S3). Aliquots of S3 were snap-frozen in liquid nitrogen and stored at −80 °C until assay.

Western blotting for CRFR₁ receptor levels

The S3 membrane extracts were thawed on ice, diluted in 4X SDS-PAGE sample buffer and heated at 70 °C for 10 minutes. Samples (1ug protein per well) were then separated using 7% NuPAGE Tris-acetate gels (Invitrogen, Carlsbad, CA) and transferred to 0.45-µm-think nitrocellulose membranes. The membranes were blocked with 5% nonfat dry milk in TBS-T (25 mM Tris-HCL pH 7.2, 150 mM NaCl and 0.05% Tween-20) overnight at 4 °C. Blots were then incubated with a polyclonal primary antibody to CRFR₁ (1:200; Santa Cruz H-215) in 5% nonfat dry milk and TBS-T for 2 hours at room temperature. The reaction was stopped with six consecutive 10-minute washes in TBS-T. A goat anti-rabbit IgG (H+L):HRP (1:5000; Pierce) was used for the secondary antibody incubation in 5% nonfat dry milk and TBS-T for one hour at room temperature. The reaction was stopped with six consecutive ten-minute washes in TBS-T. Membranes were then incubated in Supersignal West Pico Working Solution (Pierce) for five minutes and exposed to CL-Exposure film (Pierce). Film was developed in Kodak D-19 developer then washed and fixed in Kodak fixer. The developed film was scanned (Hewlett Packard Scan Jet 5P) and immunoreactivities quantified by measurements of optical densities using BioRad Quantity-One analysis software. Each protein sample was run in duplicate and the average optical density determined. B-actin was used as a loading control. The optical density measurements, obtained from internal standards, were linear over the range of 10 to 50 µg of whole brain homogenate. For experimental samples with optical densities in the linear range of the standard curve, the amount of immunoreactivity was defined by dividing the optical density of the sample by the optical density of one unit of immunoreactivity as determined by the standard curve. The antigen protein level in the sample was then determined by dividing the units of antigen present by the total µg of protein loaded on the SDS-PAGE gel.

Serotonin 5HT_1A Receptor Number and Receptor-Effector Coupling

Histological Sectioning

Adult offspring, 75 to 90 days of age, were sacrificed by decapitation. Whole brains were rapidly dissected, frozen in isopentane, chilled in a dry ice/methanol bath and then stored in airtight containers at −80 °C until sectioning. Ten-µm-thick coronal sections were collected at atlas coordinates, interaural 2.10 mm, bregma −1.70 mm, according to mouse stereotaxic atlas of Franklin and Paxinos (1997; atlas Figure 45). The brain was then rotated ninety degrees to collect horizontal sections containing the ventral hippocampal formation, entorhinal cortex and inferior colliculus. The level of sectioning in each plane was verified by examination of Nissl-stained sections prior to the collection of sections for the binding assays. The sections were thaw-mounted onto precleaned Superfrost-Plus® microscope slides (VWR Scientific, West Chester, PA) and stored at −80 °C in airtight containers until assay.

[³H]-DPAT Binding Assay

The binding of [³H]-(+)-8-Hydroxy-2-dipropylaminotetralin hydrobromide ([³H]-DPAT) to 5HT_1A receptors was conducted according the methods reported by Hensler et al., (1991). Sections were preincubated for 30 minutes in incubation buffer (150 mM Tris-HCl, pH 7.6 at 25 °C). Sections were then incubated with 2 nM [³H]-DPAT (200 Ci/mmole, Amersham, Piscataway, NJ) for 60 minutes at 25 °C in the absence or presence of 20 µM unlabelled DPAT. After incubation, sections were rinsed twice for five minutes each in ice-cold incubation buffer, quickly dipped in ice-cold distilled water, dried under a stream of cool air, vacuum desiccated overnight and then loaded into x-ray cassettes along with tritium standards. A 20.3 × 25.4 cm sheet of Biomax MR Film (Kodak, Rochester, NY) was apposed to the standards and sections and exposed for eight weeks. After exposure, the film was developed in Kodak D-19 developer (1:1) at 18 °C for 5 minutes, fixed for 5 minutes in Kodak Fixer, rinsed and dried. For ease of handling and protection of the emulsion, the film was cut into 2 × 5 cm strips and mounted onto clean microscope slides with Permount.

DPAT-Stimulated [³⁵S]-GTPγS Binding Assay

Agonist-stimulated [³⁵S]-guanosine 5’-(γ-thio) triphosphate ([³⁵S]-GTPγS) binding was conducted using the methods originally reported by Sim et al. (1995). Tissue sections were preincubated for 10 minutes in incubation buffer (50 mM Tris-HCl, 100 mM NaCl, 5 mM MgCl₂, pH 7.4 at 25 °C) containing 1 mM DL-dithiothreitol, 0.2 mM EGTA and 2 mM guanosine 5'-diphosphate. Sections were then incubated with 100 pM [³⁵S]-GTPγS (specific activity = 1250 Ci/mmole; Perkin Elmer Life Sciences, Boston, MA) for 90 minutes in the absence or presence of 10 µM unlabeled GTPγS). DPAT-stimulated [³⁵S]-GTPγS binding was assessed by incubating sections in the presence of 2 µM or 20 µM unlabelled DPAT. After incubation, sections were rinsed twice for 15 seconds each in fresh incubation buffer at 4 °C, dipped for one second in 4 °C distilled water, dried under a stream of cool air and then vacuum desiccated overnight. Sections were then loaded into x-ray cassettes along with ¹⁴carbon standards and a sheet of Biomax MR Film was apposed to the standards and sections and exposed for four days. Film development, fixation and subsequent handling were the same as described for the [³H]-DPAT assay above.

Quantitative microdensitometry

Microdensitometry was performed using Media Cybernetics Image Pro Plus® (Silver Spring, MD) on an Olympus BH-2 microscope at a total image magnification of 3.125X. In each assay, an optical density standard curve, expressed in picoCuries/10⁵ µm², was established based on the autoradiograms of the standards. In the coronal sections, optical density measurements of ligand binding were made in two brain regions, the dorsal hippocampal CA1 and in the parietal cortex (Layers I–III) immediately above the dorsal hippocampal region. In horizontal sections, binding measurements were made in the ventral hippocampal CA1, the medial entorhinal cortex (Layers I–III) and in the inferior colliculus.

In the [³H]-DPAT binding assay, total [³H]-DPAT binding in each brain region of interest was measured in quadruplicate sections incubated in the absence of excess unlabelled DPAT. Non-specific [³H]-DPAT binding was measured in duplicate sections incubated in the presence of 20 µM unlabelled DPAT. Specific [³H]-DPAT binding in each brain region of interest was defined as the difference between total and non-specific [³H]-DPAT binding.

In the [³⁵S]-GTPγS binding assay, total [³⁵S]-GTPγS binding in each brain region of interest was measured in quadruplicate sections incubated in the absence of excess unlabelled GTPγS. Non-specific [³⁵S]-GTPγS binding was measured in duplicate sections incubated in the presence of 10 µM unlabelled GTPγS. Basal [³⁵S]-GTPγS binding in each brain region of interest was defined as the difference between total and non-specific [³⁵S]-GTPγS binding. DPAT-stimulated [³⁵S]-GTPγS binding was measured in triplicate sections incubated in the presence of one of two concentrations of unlabelled DPAT. DPAT-stimulated [³⁵S]-GTPγS binding was defined as the difference between [³⁵S]-GTPγS binding in the presence of DPAT and basal [³⁵S]-GTPγS binding.

In a preliminary [³⁵S]-GTPγS binding study, brain sections were incubated with twelve different concentrations of DPAT ranging from 50 nM to 30 µM. DPAT increased [³⁵S]-GTPγS binding in a concentration-dependent manner between 200 nM and 20 µM DPAT (data not shown). At 20 µM DPAT, [³⁵S]-GTPγS binding in the dorsal hippocampal formation was seven-fold higher than basal [³⁵S]-GTPγS binding. The apparent EC₅₀ for DPAT-stimulated [³⁵S]-GTPγS binding was 2 µM. Based on this study, 2 µM and 20 µM DPAT were selected as EC₅₀ and EC₁₀₀ concentrations for the study of 5HT_1A agonist-stimulated [³⁵S]-GTPγS binding in perinatal arsenic-exposed offspring.

Learned Helplessness Task

The learned helplessness task was performed in a Coulbourn™ Habitest© shuttle box apparatus using training and testing procedures modified from a method described by Caldarone et al. (2000). Each experimental group consisted of six animals (one mouse per litter per group) in each of eight experimental groups: Two perinatal treatments (control or arsenic) × two genders (male or female) X two training groups (shock or no shock). Mice in the learned helplessness (LH-trained) groups received 120 uncontrollable and unpredictable foot shocks (0.5 mA, 5 second duration) over the course of one hour. The probability for delivery of a foot shock was 50% for every 15 second interval. The mouse was removed from the shuttle box 30 seconds after the delivery of the 120^th footshock. The non-shocked (NS-trained) groups were placed into the shuttle box apparatus for one hour during which no foot shocks were delivered. Two mice were trained at a time, one LH-trained and one NS-trained. Mice were returned to their home cages after the one-hour training session. Twenty-four hours later, escape latencies were measured using an active avoidance procedure. Animals were placed into the shuttle box and given 30 trials with an inter-trial interval of 30 seconds. Latency for the mouse to escape through the door was measured as the time from the door rising to the time it closed. If no escape was made 24 seconds after the start of the trial, the shock terminated and the door closed.

Forced Swim Task

Eight animals from each group (arsenic-exposed and control) were tested. Mice were placed in a 30 cm diameter, 46 cm tall cylinder of water (22–25°C, depth 26 cm) and forced to swim for 5 minutes as described previously (Sunal et al., 1994; Porsolt et al, 1977). The combination of floating, twitching and paddling behaviors represented immobility indicative of depressive-like behavior. Time spent in immobile behaviors was scored during the three minute period. The sum of thrashing, climbing and swimming behaviors were defined as escape-directed behaviors and were not actively scored. Data from the first 3 minutes of the test were used in the analysis because most mice chose to float in the water during the last few minutes.

Statistical Analysis

All data were analyzed either by t-test comparing the perinatal arsenic to the control on each of the dependent measures (Table 1 and Figure 1, Figure 2, and Figure 7) or one (comparing perinatal arsenic exposed to control for each of the brain regions, Figure 5) or two-way ANOVA (2 levels of perinatal treatment and the two genders (male and female, Figure 6). Analysis of the escape latency in the no shock condition was analyzed similarly. Significance was set at p≤ 0.05. As needed, t-tests were used for the post-hoc test. All analyses were conducted using SPSS Version 14 (Chicago, IL).

Table 1.

Effects of perinatal exposure to 50 parts per billion arsenic in drinking water on litter size and offspring body weight on postnatal day 7.

Experimental Group	# of Litters	Litter Sizea	Body Weightb
Control	8	6.0 ± 0.78	4.66 ± 0.39
Sodium Arsenate	9	4.6 ± 0.86	4.29 ± 0.46

^aMean + S.E.M. number of pups per litter

^bMean + S.E.M. pup body weight in grams

Figure 1.

The effect of perinatal arsenic exposure on plasma corticosterone levels. Data bars represent the mean ± SEM nanograms corticosterone per mL plasma from five or six mice per perinatal treatment group. Asterisk denotes data significantly elevated compared to the control water group (two-tailed t-test, t(9) = 2.491, p = 0.03).

Figure 2.

The effect of perinatal arsenic exposure on hippocampal CRF receptor protein levels in adult offspring. A: Representative immunoblots of triplicate control and arsenic-exposed samples and a standard curve. B: Data expressed as immunoreactivity units calculated from a protein standard curve using control hippocampal tissue. Data bars represent the mean ± SEM of 12 mice in each perinatal treatment group. Beta-actin was used as a loading control. Asterisk denotes data significantly different compared to the control water group (t (22) = 3.64, p<0.001).

Figure 7.

The effect of perinatal arsenic exposure on forced swim task in adult offspring. Data bars represent the mean ± the SEM of eight mice in each group. Asterisks denote a significant elevation in time spent in immobile behaviors compared to the control water group (two-tailed t-test, t(14) = 8.11, p < 0.001).

Figure 5.

The effects of perinatal arsenic exposure on 5HT_1A receptor density and 5HT_1A receptor-effector coupling in five regions of adult mouse brain. A: Specific [³H]-DPAT binding. B: Basal [³⁵S]-GTPγS binding. C: 2 µM DPAT-stimulated [³⁵S]-GTPγS binding. D: 20 µM DPAT-stimulated [³⁵S]-GTPγS binding. Data bars represent the mean ± SEM of 5 to 11 mice in each group. Asterisks denote data significantly increased compared to the untreated control group (multivariate, one-way ANOVA F(1,12)=9.8 p<0.001 in Figure 5A; F(1,10)=6.6 p<0.03 in Figure 5C).

Figure 6.

The effect of perinatal arsenic exposure on learned helplessness in adult male and female offspring. Data bars represent the mean ± the SEM of five mice in each group. Asterisks denote significant differences between control and arsenic-exposed offspring. A two-way ANOVA was performed with a t-test as a post-hoc. There was a significant effect of perinatal arsenic F(1,20) = 81.85 p<0.0001. There was no effect of gender F(1,20)= 3.043 p=0.096 (not significant) and no interaction between perinatal treatment and gender in learned helplessness training F(1,20)= 0.012 p=0.913, not significant.

RESULTS

Perinatal Arsenic Exposure Paradigm

The perinatal arsenic exposure paradigm produced no significant differences in litter size or offspring body weight at seven days of age (Table 1). Furthermore offspring body weight, within the same gender, were not different between the treatment groups (t(15) = 1.92; p=0.8, not significant).

Plasma Corticosterone Levels

Plasma corticosterone levels were significantly elevated compared to control offspring. Figure 1 represents the results from the RIA analysis, showing that corticosterone levels in arsenic offspring were higher than controls by approximately a 2-fold increase (t(9) = 2.491; p = 0.03).

CRFR₁ Receptor Protein

Figure 2A illustrates a representative Western blot analysis of hippocampal CRFR₁ receptor protein in whole hippocampal homogenates from arsenic-exposed and control offspring. As shown in Figure 2A, the optical densities of the bands from the arsenic-exposed mice were lower than the controls. A summary of the Western blot data is provided in Figure 2B. Perinatal arsenic exposure significantly reduced the quantity of CRFR₁ receptor protein by approximately 32% in hippocampal formation compared to the control water group (33 +/− 1.8 for control vs. 22 +/− 1.8 for arsenic samples). Asterisk denotes data significantly different compared to the control water group (t (22) = 3.64, p<0.001).

Specific [³H]-DPAT Binding and DPAT-stimulated [³⁵S]-GTPγS Binding

Figure 3 and Figure 4 illustrate representative autoradiograms depicting the distribution of total and non-specific [³H]-DPAT binding to 5HT_1A receptors (panels A & B) and 2 µM DPAT-stimulated and basal [³⁵S]-GTPγS binding (panels C & D) in coronal and horizontal sections from the brain of an untreated control mouse. Total [³H]-DPAT binding site densities were highest in the hippocampal formation and cortical brain regions (Figure 3A & Figure 4A). The distribution of DPAT-stimulated [³⁵S]-GTPγS binding (Figure 3C & Figure 4C) was similar to the distribution of [³H]-DPAT binding (Figure 3A & Figure 4A).

Figure 3.

Autoradiograms of [³H]-DPAT binding and DPAT-stimulated [³⁵S]-GTPγS binding in coronal sections of brain from an untreated control mouse. A: Total [³H]-DPAT binding. B: Non-specific [³H]-DPAT binding (binding in the presence of 20 µM unlabelled DPAT). C: [³⁵S]-GTPγS binding in the presence of 2 µM DPAT. D: Basal [³⁵S]-GTPγS binding (binding in the absence of added DPAT). Binding measurements were made in the dorsal hippocampal CA1 stratum radiatum and the parietal cortex (Layers 1–3) immediately dorsal to the hippocampal formation. The horizontal bar at the bottom of panel D denotes a distance of 1 mm.

Figure 4.

Autoradiograms of [³H]-DPAT binding and DPAT-stimulated [³⁵S]-GTPγS binding in horizontal sections of brain from an untreated control mouse. A: Total [³H]-DPAT binding. B: Nonspecific [³H]-DPAT binding (binding in the presence of 20 µM unlabelled DPAT). C: [³⁵S]-GTPγS binding in the presence of 2 µM DPAT. D: Basal [³⁵S]-GTPγS binding (in the absence of added DPAT). Binding measurements were made in the ventral hippocampal CA1 stratum radiatum, medial entorhinal cortex, (Layers 1–3) and the inferior colliculus. The horizontal bar at the bottom of panel D denotes a distance of 1 mm.

The effects of perinatal arsenic exposure on specific [³H]-DPAT binding, basal and DPAT-stimulated [³⁵S]-GTPγS binding in adult offspring are summarized in Figure 5. Specific [³H]-DPAT binding was significantly increased by approximately 77% in the dorsal hippocampal formation of arsenic-exposed mice (0.57 +/−0.06) compared to the control group (0.31+/− 0.05) (Figure 5A). In contrast, specific [³H]-DPAT binding was not different in the ventral hippocampal formation, the parietal cortex, entorhinal cortex or the inferior colliculus. Basal [³⁵S]-GTPγS binding was not different in any of the brain regions measured (Figure 5B). In the presence of a half-maximally effective (EC₅₀) concentration of DPAT (2 µM) dorsal hippocampal DPAT-stimulated [³⁵S]-GTPγS binding was 64% greater in arsenic-exposed offspring (0.35 +/−0.07) than control (0.23 +/− 0.03) (Figure 5C). Asterisks denote data significantly increased compared to the untreated control group (multivariate, one-way ANOVA F(1,12)=9.8 p<0.001 in Figure 5A; F(1,10)=6.6 p<0.03 in Figure 5C). As was the case for specific [³H]-DPAT binding (Figure 5A), 2 µM DPAT-stimulated [³⁵S]-GTPγS binding was not different between the two groups in the other four brain regions examined (Figure 5C). In the presence of a maximally effective concentration of DPAT (20 µM), DPAT-stimulated [³⁵S]-GTPγS binding was not different in any of the five brain regions examined (Figure 5D).

Learned Helplessness Task

Perinatal arsenic exposure produced a significant increase in escape latencies in both male and female adult mice (Figure 6). A two-way ANOVA was performed with a t-test as a post-hoc. There was a significant effect of perinatal arsenic F(1,20) = 81.85 p<0.0001. There was no effect of gender F(1,20)= 3.043 p=0.096 (not significant) and no interaction between perinatal treatment and gender in learned helplessness training F(1,20)= 0.012 p=0.913, not significant. No observed escape differences among the no shock groups (data not shown) were observed. The ANOVA results for escape latency in the no shock training condition were: F(1,2)= 2.0, p= 0.169, for gender, F(1,20)=2.67, p=0.118 for perinatal treatment comparision and F(1,20) = 0.931, p=0.346 for the gender by perinatal interaction. These data indicate that there were no extreme differences between the perinatal groups in general activity level or their ability to navigate the learned helplessness escape task. No differences (t(10)= 1.03, p= 0.4,) were observed between the groups in a pain threshold test measuring vocal response to footshock intensity on a separate group of littermates (data not shown).

Forced Swim Task

Perinatal arsenic exposure produced a significant increase in immobility in exposed mice compared to controls (Figure 7). There was a significant effect of arsenic on immobility t(14)=8.14, p<0.001.

DISCUSSION

The present findings demonstrate that perinatal exposure to relatively low levels (50 ppb) of arsenic in drinking water can have long-lasting biochemical and behavioral effects on adult offspring. Perhaps the most striking effect observed in these studies was the impact of perinatal arsenic exposure on plasma corticosterone (Figure 1) which was assessed under baseline, non-stressed circumstances (Figure 1). Qualitatively similar elevations in corticosterone levels have been noted also in perinatal lead-exposed rats (Cory-Slechta et al, 2008). Increased serum corticosterone levels suggest that the HPA axis is overactive in these arsenic-exposed offspring.

While it is not possible to discern from these experiments whether elevated plasma corticosterone is a primary teratogenic effect of perinatal arsenic exposure or a secondary, perhaps compensatory consequence of effects on systems that regulate the HPA axis, it is noteworthy that the other biochemical alterations (Figure 1, Figure 2, Figure 5) and behavioral effects (Figure 6, Figure 7) observed in these studies are consistent with sustained elevations in HPA axis activity. For example, perinatal arsenic exposure decreased hippocampal CRFR₁ levels compared to controls (Figure 2). While hippocampal CRF levels were not measured in this study, it is reasonable to speculate that if CRF levels are elevated to a similar degree as plasma corticosterone (Figure 1), the decrease in hippocampal CRFR₁ in the perinatal arsenic mice may represent a compensatory down-regulation in response to heightened CRF activation. Several studies have reported increased CRF in response to intense or prolonged stress. Four hours of restraint stress in mice resulted in a significant increase of CRF mRNA expression within the paraventricular nucleus (Nomura et al., 2003) and similar increases in CRF were seen using in an earlier study using crowded rearing conditions as a stressor (Albeck et al., 1997) in rats.

Perinatal arsenic exposure elevated specific [³H]-DPAT binding to 5HT_1A receptors (Figure 5A) in dorsal hippocampal formation. This study utilized a [³H]-DPAT concentration near the half-maximally saturating (Kd) concentration (Hensler et al., 1991). This effect of perinatal arsenic exposure on specific [³H]-DPAT binding was accompanied by increased DPAT-stimulated [³⁵S]-GTPγS binding at a half-maximally effective concentration of agonist (Figure 5C), but not at a saturating concentration of the agonist (Figure 5D). Taken together, these data suggest that perinatal arsenic exposure increases the sensitivity of dorsal hippocampal 5HT_1A receptors to serotonin without altering the total number of receptors present. Further, these results suggest that basal [³⁵S]-GTPγS binding (Figure 5B) and 5HT_1A receptor-effector coupling were not affected in perinatal arsenic exposed offspring. More detailed saturation of specific [³H]-DPAT binding and DPAT-stimulated [³⁵S]-GTPγS binding would be required to substantiate this interpretation. However, the inability to collect an adequate number of histological sections in specific regions of mouse brain limits more extensive saturation of binding studies.

It is noteworthy that elevations in 5HT_1A receptor sensitivity were not observed in other brain regions (Figure 5A and 5C) and were observed only in the dorsal hippocampal formation. This suggests that the hippocampal formation is at least one of the brain regions most susceptible to the consequences of perinatal arsenic exposure. Further, while differences between the dorsal and ventral hippocampal formation have been identified (see Moser and Moser, 1998), the basis for differential dorsal-ventral hippocampal sensitivity in arsenic-exposed mice is not known. However, qualitatively similar outcomes have been observed in the study of hippocampal glutamate receptors in prenatal ethanol-exposed offspring (Farr et al., 1988; Savage et al., 1991). The mechanistic basis for elevated 5HT_1A receptor sensitivity in perinatal arsenic-exposed mice is not known. Chronic elevated HPA axis activity that has been associated with diminished firing of raphe serotonergic neurons and a reduction in serotonin turnover and release (Davis et al., 1995; Myint et al., 2007), including a decrease in 5HT_1A mRNA following 20 days of chronic unpredicted footshock stress (Xu et al., 2007) and decrease in 5HT_1A activity in the dorsal raphe, following chronic social defeat stress (Cornelisse et al., 2007). Thus, in our study a diminished serotonergic activity following the chronic elevation of corticosterone produced by perinatal arsenic may have resulted in a compensatory up-regulation of 5HT_1A receptor sensitivity in the hippocampus.

The effects of perinatal arsenic exposure on biochemical measures associated with depression were accompanied by alterations in two measures of depressive-like behaviors, learned helplessness and immobility during forced swim. Both male and female perinatal arsenic-exposed mice displayed greater learned helplessness behavior following inescapable shock exposure (Figure 6). No gender effects were noted between groups. Forced swim task data (Figure 7) revealed that arsenic-exposed animals show more immobility, which is indicative of depressive-like behavior. These data suggest an increased susceptibility to depression and are consistent with human studies suggesting that elevated cortisol in response to chronic stress is associated with increased depressive symptoms (see Southwick et al., 2005 for review). While perinatal arsenic may not be considered a typical stressor, our data suggest that the prolonged elevation of corticosterone produced by perinatal arsenic may have produced physiological state mimicking chronic stress.

While there are no studies directly demonstrating arsenic induced 5HT_1A receptor changes, arsenic does interfere with glucocorticoid signaling pathways (Stancato et al., 1993; Kaltreider et al., 2001; Bodwell et al., 2004) which, in turn, affect serotonin neurotransmission, acute stress increasing activity and chronic stress attenuating activity (Meijer and DeKloet, 1998). Arsenic directly interferes with the glucocorticoid receptor complex and inhibits steroid binding to glucocorticoid receptors (Simons et al., 1990). Our model suggests that this would prevent the normal operation of the negative feedback mechanism, increasing plasma corticosterone levels, a phenomenon we observed. We also observed lower levels of CRFR₁ in the perinatal arsenic-exposed animals, perhaps as a compensatory response to elevated CRF. Lower activity by CRFR₁ would attenuate serotonin release in the hippocampus (Oshima et al., 2003) resulting in a compensatory increase in 5HT_1A receptor affinity.

Previous research has outlined a role for 5-HT_1A receptor in antidepressant response. Patients who respond to antidepressant treatment require a desensitization of 5-HT_1A receptor (Rausch et al., 2006). Studies also suggest that a polymorphism in the 5-HT_1A gene resulting in non-repression of the receptor has been associated with poor response to antidepressant therapy in unipolar and bipolar subjects (Lemonde et al., 2004; Serretti et al., 2004). Further evidence suggests that there is an increased expression of the 5-HT_1A receptor and mRNA in brains of suicide victims (Escriba et al., 2004). Recently, it has been suggested that an alteration of the hippocampal function may be involved in the etiology of depression (Warner-Schmidt and Duman, 2006) and that changes in the hippocampal serotonergic system are, in part, responsible for its pathogenesis (Iritani et al., 2006). All of the previous research on serotonin ties into our depressive-like behavioral analysis of perinatal arsenic-exposed mice.

Finally, a negative impact of perinatal arsenic on learning has previously been reported (Rodriguez et al. 2002). In a study using rats, they showed perinatal exposure to arsenate (36.70 mg/L in drinking water) from gestation day 15 (GD 15) resulted in increased spontaneous locomotor activity and an increased number of errors in a delayed alternation task in comparison to the control group (Rodriguez et al., 2002). Although we are studying a mouse model, it is important to note that the whole body retention of inorganic arsenic in rats has been shown to be 20 times higher than that in similarly exposed mice due to retention of arsenic in erythrocytes, (Vahter, 1981) therefore the arsenic effects observed in our mouse model are at a dose that is much lower than those used in the Rodriguez study. Also, the Rodriguez study dosing schedule occurs at GD15 suggesting that effects we have seen at lower doses may be taking place during the embryonic stages when the brain is forming. One alternative interpretation of our data is that failure to escape in the learned helplessness task may be due to arsenic-induced deficits in learning and memory, and not an indication of increased depressive-like behavior. While this is possible, the arsenic animals failed to demonstrate escape-directed behaviors in the forced swim task which is not dependent upon learning or memory to perform. Further, there were no differences in non-shock escape latencies in the non-shock control groups between the two perinatal exposure conditions in either male or female mice (data not shown). These results suggest that the perinatal arsenic-exposure did not affect the ability to escape in the absence of learned helplessness training. Thus, while arsenic has been demonstrated to produce deficits in learning performance, the deficits in learned helplessness seen in our study were not the result of an inability to perform the task or to understand association between escape and the termination of the foot shock since the perinatal arsenic-exposed mice were as successful as control mice in escaping the shock in the non-shocked control condition.

These results demonstrate that there are neurochemical and behavioral consequences of heavy metal exposure during brain development. The perinatal period encompasses the key developmental periods in most species. Synaptic organization, which sets up functional development of neurotransmitter and neuromodulatory systems, is occurring during this time. This period is critical because it influences how the animals will respond to their external environment throughout their lives. Administration of toxins during this critical period can have lifelong consequences due to their deleterious effects during major brain maturation periods. Future research should aim to address the consequences of arsenic in the developing CNS.