Characterization of Developmental Neurotoxicity of As, Cd, and Pb Mixture: Synergistic Action of Metal Mixture in Glial and Neuronal Functions

Abstract

Neurotoxicity of individual metals is well investigated but that of metal mixture (MM), an environmental reality, in the developing brain is relatively obscure. We investigated the combinatorial effect of arsenic (As), cadmium (Cd), and lead (Pb) on rat brain development, spanning in utero to postnatal development. MM was administered by gavage to pregnant and lactating rats, and to postweaning pups till 2 months. The pups exhibited behavioral disturbances characterized by hyperlocomotion, increased grip strength, and learning-memory deficit. Disruption of the blood-brain barrier (BBB) was associated with dose-dependent increase in deposition of the metals in developing brain. Astrocytes were affected by MM treatment as evident from their reduced density, area, perimeter, compactness, and number of processes, and increased apoptosis in cerebral cortex and cerebellum. The metals induced synergistic reduction in glial fibrillary acidic protein (GFAP) expression during brain development; however, postweaning withdrawal of MM partially restored the levels of GFAP in adults. To characterize the toxic mechanism, we treated rat primary astrocytes with MM at concentrations ranging from lethal concentration (LC)₁₀ to LC₇₅ of the metals. We observed synergistic downregulation in viability and increase in apoptosis of the astrocytes, which were induced by proximal activation of extra cellular signal-regulated kinase (ERK) signaling and downstream activation of Jun N-terminal kinase (JNK) pathway. Furthermore, rise in intracellular calcium ion ([Ca²⁺]_i) and reactive oxygen species generation promoted apoptosis in the astrocytes. Taken together, these observations are the first to show that mixture of As, Cd, and Pb has the capacity to induce synergistic toxicity in astrocytes that may compromise the BBB and may cause behavioral dysfunction in developing rats.

The developing brain is vulnerable to injury from toxic metals that interfere with the critical developmental processes, i.e., cellular proliferation, migration, differentiation, synaptogenesis, myelination, and apoptosis in the central nervous system (CNS) (Rice and Barone, 2000). The limited capacity of the developing CNS to compensate for the cell loss and the disruptions in neural networking results in compromised neuronal functions (Bayer, 1989) and increased risk of neurodegeneration (Grandjean and Landrigan, 2006).

Heavy metals including arsenic (As), cadmium (Cd), and lead (Pb) have received attention as both environmental contaminants and potential neurotoxicological hazards (Brender et al., 2006; Fowler et al., 2004; Jadhav et al., 2007). Exposure to the metals in utero and in infancy is associated with risk of impaired cognitive development (Hu, 2000; Landrigan et al., 1975), subclinical brain dysfunction (Lanphear et al., 2005), and behavioral anomalies (Tsai et al., 2003; Wright et al., 2006). Studies with single metal exposure have demonstrated that As, Cd, or Pb infiltrate the immature blood-brain barrier (BBB) and accumulate in developing brain (Lidsky and Schneider, 2003; Valkonen et al., 1983; Wang et al., 2007a; Xi et al., 2010). Pb uptake through the BBB disrupts Ca²⁺ transport mechanism (Marchetti, 2003) and promotes activation of mitogen-activated protein (MAP) kinases in apoptotic glial cells (Posser et al., 2007). The sequestration of Pb at the level of the choroid plexus undermines brain growth and affects learning and cognitive functions of CNS (Marchetti, 2003). In vivo and in vitro studies have revealed that acute or chronic Cd exposure enhances oxidative stress in astrocytes and accumulates reactive oxygen species (ROS) that induces astrocytic death (Yang et al., 2008). Perinatal exposure to Cd induces anxiety (Minetti and Reale, 2006) and reduces learning ability of offspring (Ishitobi et al., 2007). Chronic exposure to As, even at a submicromolar concentration, promotes oxidative stress (Garcia-Chavez et al., 2006) and induces neuroglial damage in human brain (Jin et al., 2004). Intoxication with As presents deficits in spontaneous locomotor activity (SLA) and alterations in learning-memory task during postnatal development (Rodriguez et al., 2002).

Current knowledge of metal-induced neurocellular damage, however, is primarily confined to single metal exposure, and there has been increasing demand for cumulative hazard assessment of metals in mixture in the brain (Rodriguez et al., 1998; Wright and Baccarelli, 2007). The effect may be either dose additive, interactive (synergistic or antagonistic), or independent of each other. Of the very few reports on metal mixture (MM), studies with early-life low doses of Pb + Cd–associated exposure have revealed greater oxidative stress than with single Pb or Cd (Zhang et al., 2009). The mixture altered cerebellar and striatal functions that related to changes in motor activity (Antonio et al., 2002) and anxiety (Leret et al., 2003) in the long term. The synergistic toxic effect of Pb and Cd absorption in the brain cells of cerebellum, cortex, and hippocampus is accountable for the enhanced CNS damage in the mixture (Gu et al., 2009). Therefore, we hypothesized that concurrent exposure to As along with Cd and Pb in drinking water may have greater-than-additive/synergistic toxic responses to brain development.

Here, we demonstrated the toxic effect of As, Cd, and Pb on rat brain development. We investigated the behavioral impairments induced by the MM in developing rats. Because abnormal blood-brain communication is a key mechanism underlying neuronal dysfunction (Shalev et al., 2009), we examined the effect of the MM on BBB integrity and glial fibrillary acidic protein (GFAP) levels. We further focused on the toxic mechanism of action (additive, synergistic, or antagonistic) of the MM on the GFAP-expressing rat astrocytes. Collectively, our data strongly suggested that single metal risk assessment underestimates the toxic impact of the metals present in mixture and sheds new light on the harmful role of environmental metal contaminants in pediatric and long-term CNS complications.

MATERIALS AND METHODS

Reagents and Antibodies

Na-arsenite, Pb-acetate, Cd-chloride, Na-orthovanadate, NaF, Ponceau S stain, Evans blue (EB), PMSF, protease inhibitor cocktail, MTT [3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide], 1,2-Bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis(acetoxymethyl ester) (BAPTA-AM), α-tocopherol, Hoechst 33258 stain, poly-L-lysine, dichlorofluorescein diacetate (DCF-DA), Fluo3 AM, and mammalian tissue protein extraction reagent were procured from Sigma Chemical Co. (St Louis, MO). Fluo3 AM was obtained from Molecular Probes (Carlsbad, CA). The Alexa Fluor secondary antibodies, PD98059, SB203580, LY294002, and SP600125, cell culture reagents, sample loading buffer for Western blotting, and protein markers were purchased from Invitrogen (Carlsbad, CA). The supersignal west femto maximum sensitivity substrate for Western blotting was purchased from PIERCE Biotechnology (Rockford, IL). Rabbit polyclonal antibodies to extracellular signal-regulated kinase (ERK1/2) and Jun N-terminal kinase (JNK1/2), and phospho-ERK1/2 and phospho-JNK1/2 forms were purchased from Cell Signaling Technology (Danvers, MA). Mouse monoclonal antibody to GFAP was obtained from Millipore (Temecula, CA). Mouse monoclonal antibodies to β-actin, peroxidise-conjugated secondary antibodies were from Sigma Chemical Co.. The terminal deoxynucleotidyl-transferase (TdT)-mediated dUTP nick end labeling (TUNEL) kit was purchased from Roche (Indianapolis, IN). Diaminobenzidine tetrahydrochloride (DAB) substrate kit and Vectashield medium Elite ABC kit were purchased from Vector Laboratories (Burlingame, CA). Omniscript RT Kit and SYBR-green qPCR Kit were from Qiagen (Valencia, CA).

Animals and Treatments

All animal-handling procedures were carried out following the regulations of Institutional Animal Ethics Committee and with their prior approval for using the animals. The pregnant female Wistar rats were procured from Indian Institute of Toxicology Research and were housed in a 12-h day and light cycle environment with ad libitum availability of diet and water.

The pregnant female rats were divided into nine groups and treated with the metals (Table 1) through gavage from gestation day 5. The treatment was continued in the lactating rats and postweaning pups till 2 months. The MM was treated at two different concentrations (1× and 10×) to observe the dose-dependent effect of the MM on rat brain development, 1× being the most frequently occurring concentration of the metals in groundwater sources of India (Jadhav et al., 2007). The single metals were treated at two different concentrations. One was the same as that in 10× and the other was three times of it.

The number of pregnant rats per treatment group was 30. After standardization of litters (culling), equal numbers of male and female pups were taken for each experiment, and pups from different litters were independent subjects (Holson et al., 2008).

To determine the toxic effect of metals on glial damage and behavioral aberrations during rat brain development, studies were carried out at postnatal day (P) 16 and adult P60 rats.

Behavioral Study

Spontaneous locomotor activity.

Locomotive behavior in rats was studied using the computerized Opto-Varimex (Columbus Instruments, Columbus, OH) system as previously described (Ali et al., 1990). The locomotor markers monitored were the distance traveled, number of stereotypic movements and rearings, and time moving (in minutes).

Grip strength.

Vehicle and MM-treated rats were subjected to forelimb grip strength test using a digital grip strength meter (Columbus Instruments) following the standard procedure as described previously (Terry et al., 2003). Each rat was tested five times, with a 10-s period between two successive trials. The five recordings were averaged to obtain a final reading for each individual.

Y-maze.

Learning-memory test for vehicle- and MM-treated rats was carried out as described previously using Y-maze (Wetzel and Matthies, 1982) with minor modifications. The training apparatus was a Y-maze with electrifiable grid-floored three alleys and 15-W light bulb at the end of the alleys. When the rats were being tested, only one arm with its light on (bright arm) was a safe area without footshock, whereas the other two arms with the lights off (dark arms) were unsafe areas with electric footshock (1–5 mA). The safe and unsafe areas were arbitrarily shifted during testing. The training session consisted of 30 trials per animal. Running into the dark alley of the Y-maze was counted as an error (E). Retention of the brightness discrimination was tested after 24 h and after 7 days of initial learning using a 30-trial relearning session; performance in relearning was considered as test for memory and expressed as percent (%) saving (Serota, 1971).

Analysis of EB Extravasation in Brain

Rats were anesthetized and 3% EB in saline was injected slowly through the femoral vein. The rats were then transcardially perfused and the brain isolated and incubated in formamide as previously described (Lin et al., 2010). The supernatant was collected, and the optical density (OD) at 620 nm was measured using a SPECTRA max PLUS³⁸⁴ spectrophotometer (Molecular devices, Sunnyvale, CA) to determine the relative amount of EB in each sample. The OD ratio was derived from the OD of the MM-treated animals over that of the vehicle-treated animals.

Fluorescence study of EB extravasation was carried out as described previously (Duran-Vilaregut et al., 2009). Briefly, EB-injected rats were perfused; cerebral cortex and cerebellum dissected, postfixed in 4% paraformaldehyde (PFA), and cryoprotected with 30% sucrose; and 20-μm-thick cryostatic sections obtained using cryomicrotome (Microm HM 520; Labcon, Germany). The cryostat sections were visualized under fluorescence microscope (Nikon Instech Co. Ltd, Kawasaki, Kanagawa, Japan) after being coverslipped on Vectashield medium (Vector Laboratories).

Evaluation of Metals in Brain

For determining the levels of As, Cd, and Pb in brain, whole-brain samples were snap frozen in liquid nitrogen and kept in a −80°C freezer till analysis. Samples for Pb, Cd, and As estimation were prepared by acid digestion as previously described (Ong et al., 2006; Singh and Rana, 2007), and analysis was carried out using atomic absorption spectrophotometer equipped with a vapor generation assembly (Varian AAS 250+ coupled with VGA 77; Varian Australia Pvt Ltd [manufacturing site], Mulgave, Australia). Detection limit of the instrument was 1 ppb (Behari and Prakash, 2006).

Immunohistochemistry and Quantitative Estimation of GFAP-Immunoreactive Astrocytes

Immunoperoxidase staining with 5-μm cryostat sections of cerebellum (transverse) and cerebral cortex (coronal) was carried out for GFAP antibody. Briefly, four pups from four different litters were taken at the developmental stages, anesthetized, and perfused and the brain was fixed and cryoprotected as described previously (Sinha et al., 2009; Zhu et al., 2001). Five-micron transverse sections were made from the cerebellum and coronal sections of the cerebral cortex using cryomicrotome (Microm HM 520; Labcon). The sections were then mounted on 3-aminopropyltriethoxysilane-coated slides. Immunoperoxidase staining for GFAP (monoclonal, 1:400) was carried out with DAB chromogen and ABC kit as described previously (Otani et al., 1999) and visualized under optical microscope (Nikon Instech Co. Ltd). Ten fields in each section were captured using ×40 objective for evaluation. Image-Pro Plus 5.1 software (Media Cybernetics Inc., Silver Spring, MD) was used for image capturing.

The images were then imported into Image-J 1.42q (http://rsb.info.nih.gov/ij/; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). The number of GFAP-immunoreactive (ir) cells and the area, perimeter, compactness, and number of processes in the GFAP-ir cells were quantified with the Shape descriptors plugin of the software.

Detection of Apoptosis (in vivo) in GFAP-ir Astrocytes

In situ detection of apoptosis was carried out by TUNEL assay in P16 and P60 rats. Briefly, four pups from four different litters were taken at the developmental stages, anesthetized, and perfused and the brain was fixed and cryoprotected as described previously (Sinha et al., 2009; Zhu et al., 2001). Five-micron transverse sections from the cerebellum and coronal sections from the cerebral cortex were made using cryomicrotome (Microm HM 520; Labcon). For the TUNEL assay, a labeling reaction was carried out with fluorescein-labeled dUTP in the presence of TdT at 37°C for 1 h. To investigate whether the apoptotic cells were astrocytes, the sections were immunostained with anti-mouse GFAP antibody (1:400 dilution in TBST [10mM Tris, pH 8.0, 150mM NaCl, 0.01% Tween 20]) according to manufacturer’s protocol. The sections were then incubated with Alexa Fluor 546–conjugated (fluorescent color: red; Abs/Em: 555/565) goat anti-mouse antibody (1:200 dilution) as previously described (Bandyopadhyay et al., 2007); counterstained with Hoechst 33258 (0.2mM) for 10 min; and visualized under a fluorescence microscope (Nikon Instech Co. Ltd) after being coverslipped on Vectashield medium (Vector Laboratories).

Protein Extraction and Western Blotting

Cerebral cortex and cerebellar tissues from five to seven postnatal rats were harvested, snap frozen in liquid nitrogen, and stored at −80° C until further investigation. SDS polyacrylamide gel electrophoresis and Western blotting were done with the tissues following an optimized protocol (Sinha et al., 2009) with GFAP and β-actin antibodies. To determine the phosphorylation of JNK1/2 and ERK1/2, previously described protocol was followed (Bandyopadhyay et al., 2006). The working dilutions for the primary antibodies were as follows: GFAP (monoclonal, 1:1000), β-actin (monoclonal, 1:1000), JNK1/2 (polyclonal, 1:1000), ERK1/2 (polyclonal, 1:1000), phospho-JNK1/2 (polyclonal, 1:1000), and phospho-ERK1/2 (polyclonal, 1:1000). The working dilutions for secondary anti-rabbit IgG (for the polyclonal primary antibodies) and anti-mouse IgG (for the monoclonal primary antibodies) conjugated to horseradish peroxidase were 1:1000 in PBS plus 0.2% Triton X-100. The samples were detected by chemiluminescence with the supersignal west femto maximum sensitivity substrate. Relative expression of each protein was determined by densitometric quantification of blots using the VersaDoc Gel Imaging System (BioRad, Hercules, CA).

Cell Culture

Astrocytes.

Rat pups (P1) were decapitated, the brain was removed, and astrocytes were cultured from the rat brain as described previously (Tanaka et al., 1998). The purity of astrocytes was 94–97% as determined by the immunofluorescence staining of GFAP.

Treatment of Cells with Metals and Cell Signaling Inhibitors

The astrocytes were grown to 80% confluence, pre-incubated in reduced serum (0.5% fetal bovine serum [FBS]) medium for 2 h, and then treated with As, Cd, Pb, or MM at concentrations ranging from 0.01 to 200μM for 18 h in a humidified tissue culture incubator at 37°C with 5% CO₂-95% air.

The involvement of signaling pathways, such as PI3 kinase, MEK1/2, P38-MAPK, and JNK1/2, in MM-induced astrocyte toxicity was determined by incubating the astrocytes with the MM and cotreating with LY294002 (10μM), PD98059 (10μM), SB203580 (10μM), or SP600125 (10μM), which are inhibitors to PI3 kinase, MEK1/2, P38-MAPK, and JNK1/2 pathways, respectively.

The time course of phosphorylation of MEK1/2 (assessed by ERK1/2 phosphorylation) and JNK1/2 in MM-treated astrocytes was determined by incubating the cells with MM for 0, 5, 10, 15, 30, and 60 min.

Cytotoxicity Assay

MTT cell viability assay, a colorimetric assay, was used to determine astrocyte viability. The astrocytes were grown to 80% confluence, pre-incubated in reduced serum (0.5% FBS) medium for 2 h, and then treated with As, Cd, Pb, or MM at concentrations ranging from 0.01 to 200μM for 18 h in reduced serum medium in a humidified tissue culture incubator at 37°C with 5% CO₂-95% air. The cells were then incubated with 10 μl MTT (10.4 mg/ml) and an optimized protocol was followed to determine cell viability (Sanders et al., 2000). Absorbance was measured at 595 nm, with background subtraction at 655 nm. The LC of the metals on astrocytes was analyzed by the GraphPad Prism 3.0 software.

TUNEL Assay

The in situ detection of apoptosis in astrocytes was carried out with the In Situ Cell Death Detection fluorescein kit (Roche Applied Science) for TUNEL assay, as per the manufacturer’s instruction and visualized under a fluorescence microscope (Nikon Instech Co. Ltd). Briefly, the astrocytes were grown to 80% confluence, pre-incubated in reduced serum medium for 2 h, and then treated with As, Cd, Pb, or MM in reduced serum medium at a concentration range of LC₁₀ to LC₁₀₀ of the metals for 18 h in a humidified tissue culture incubator at 37°C with 5% CO₂-95% air. The cells were counterstained with Hoechst 33258. The TUNEL-positive cells were counted in five randomly selected fields. Around 10,000 cells in each cover slip were scored. The apoptotic index was expressed as the number of TUNEL-positive cells per 100 nuclei (Hoechst stained). Image-Pro Plus 5.1 software (Media Cybernetics Inc.) was used for cell counting.

Assay for [Ca²⁺]_i

The astrocytes were grown to 80% confluence at 4000 cells/well in 96-well poly D-lysine-coated black view plates (VWR no. 62406–036; Falcon), pre-incubated in reduced serum medium for 2 h, and then treated with MM in reduced serum medium for 2 min, 5 min, 10 min, 30 min, 1 h, 2 h, and 24 h in a humidified tissue culture incubator at 37°C with 5% CO₂-95% air. An increase in [Ca²⁺]_i was measured using Fluo3 AM as an indicator dye after the addition of metals (single or in mixture) to the culture wells following an optimized protocol (Arey et al., 2005). The fluorescent signals were read by fluorescence imaging plate reader Synergy HT (BioTek, Winooski, VT).

Assay for ROS Generation

The production of ROS, mainly H₂O₂, was assessed using DCF-DA, followed by semiquantitative fluorometric measurements as described previously (Izawa et al., 2009). The astrocytes were grown to 80% confluence at 4000 cells/well in 96-well poly D-lysine-coated black view plates (VWR no. 62406–036; Falcon); pre-incubated in reduced serum medium for 2 h; treated with MM for 0 min, 30 min, 60 min, 2 h, or overnight; and then incubated with DCF-DA to a working concentration of 20 μg/ml. The fluorescent level indicating intracellular ROS production was measured using a fluorescent microplate reader with excitation at 485 nm and emission at 530 nm.

Evaluation of Metal(s) Interaction in Combination

To characterize the interaction between the heavy metals for their effects on astrocyte viability, apoptotic index and GFAP level, a combination index (CI) was calculated using the software Calcusyn (Biosoft, Manchester, United Kingdom). CI values less than 1.0 indicated synergism (Zhao et al., 2004), which suggests greater than expected based on effect addition.

The U.S. Environmental Protection Agency has selected dose addition as the no-interaction definition for mixture risk assessment, so that synergism would represent toxic effects that exceed those predicted from dose addition (Hertzberg and MacDonell, 2002). Synergistic interactions have implications of gain (Berenbaum, 1985) and the combined effects are greater than the additive effect of the components (Wessinger, 1986).

Statistical Analysis

Data are presented as mean ± SE of the indicated number of experiments. Statistical analysis was carried out in SPSS 9.0 software (SPSS Inc., Chicago, IL). Data were analyzed by one-way ANOVA, followed by Student-Newman, Keuls post hoc test or Student’s t-test when appropriate.

RESULTS

Effect of MM on SLA, Grip Strength, and Learning-Memory Performance in Developing Rats

It has been previously reported that exposure to inorganic As (Rodriguez et al., 2002), Cd (Ishitobi et al., 2007), or Pb (Marchetti, 2003) caused behavioral alterations in developing rats. We, therefore, investigated whether MM treatment altered SLA, grip strength, and learning-memory performance in developing rats. The rats exhibited dose-dependent increase in the distance traveled, number of stereotypic movements, number of rearings, movement time, and grip strength (Table 2). MM treatment also demonstrated dose-dependent reduction in learning-memory performance (Table 2).

Effect of MM on BBB Permeability in Developing Rat Brain

Increase in BBB permeability is a potentially important cause of brain dysfunction including behavioral disorders (Shalev et al., 2009). Moreover, it has been reported that As and Cd disrupted the integrity of BBB (Zheng, 2001) and exposure to Pb disrupted the BBB (Lidsky and Schneider, 2003). We, therefore, determined the effect of MM on BBB permeability by measuring EB extravasation. We observed dose-dependent increase in EB extravasation in MM-treated postnatal rat brain (Figs. 1A and 1B). Transverse section of cerebellum (Figs. 1C and 1D) and coronal section of cerebral cortex (Fig. 1E) exhibited marked BBB damage, as evident from the increased red fluorescence (indicative of EB leakage) in adult rats.

Levels of As, Cd, or Pb in Metal-Treated Developing Rat Brain

Previous studies have demonstrated that exposure to inorganic As (Garcia-Chavez et al., 2006), Cd (Lafuente et al., 2001), or Pb (Benitez et al., 2009; Rader et al., 1981) led to deposition of the metals in rat brain. Because the metals could reach the fetus through placenta (Benitez et al., 2009; Xi et al., 2010), and lactating off-springs through mother’s milk (Counter et al., 2007; Kippler et al., 2009; Samanta et al., 2007), we investigated whether metal(s) treatment (groups 1–9) caused their deposition in developing brain. A dose-dependent increase in the levels of As, Cd, and Pb was observed in the postnatal rat brain (Table 3).

Effect of MM on the Number, Area, Perimeter, Compactness, Processes, and Apoptosis of Astrocytes in Developing Rat Brain

Astrocytes are involved in the maintenance of BBB (Choi and Kim, 2008; Pekny et al., 1998), and loss of BBB integrity is associated with astrocyte damage under neuropathological conditions (Prior et al., 2004; Willis et al., 2004). Because we observed that MM promoted permeability of the developing rat BBB, we studied the response of the astrocytes to MM. We observed reduction in the GFAP-ir astrocyte count, area, perimeter, compactness, and number of processes in the cerebral cortex and cerebellum (Figs. 2A–D, Table 4). We further observed that the MM promoted apoptosis in the developing cerebral and cerebellar astrocytes (Figs. 2E–H).

Effect of MM on GFAP Level in Developing Rat Brain

We studied the effect of MM on the GFAP levels in the developing cerebral cortex and cerebellum. We observed dose-dependent decrease in the level of GFAP that persisted till adulthood (Figs. 3A and 3B). Upon withdrawal of MM treatment from weaning, the GFAP level was significantly restored in adults (Figs. 3C and 3D). This suggests that the decrease in level of GFAP was primarily caused by MM treatment.

We further investigated whether MM had synergistic or additive toxic effect on the level of GFAP in the developing rat brain. We determined the effect of the single metals (groups 4–9 in the “Materials and Methods” section) or of MM (10× concentration) on the GFAP level. We observed that the sum of the fold-reduction in GFAP level by single metals was less than the fold reduction in GFAP by MM (Figs. 3E–H), suggesting synergistic GFAP suppression by MM.

Effect of MM on Viability and Apoptosis of Rat Primary Astrocyte Culture

We investigated the toxic mechanism of action of MM in the rat primary astrocytes.

We first determined the effect of As, Cd, or Pb on the astrocyte viability. The individual metals induced cell death, and the lethal concentrations (LC₁₀, LC₂₅, LC₅₀, and LC₇₅) of the individual metals were determined (Table 5). We then treated the cells with MM and observed that MM induced greater cell death than the sum of cell death induced by individual metals (Table 6), suggesting that As, Cd, and Pb induced synergistic reduction in astrocyte viability.

We determined the effect of As, Cd, Pb, or of MM on the astrocyte apoptotic index. We observed that the MM synergistically promoted apoptosis in the astrocytes (Table 6).

Effect of MM on Activation of the MAPK Signaling Pathways in the Rat Primary Astrocytes

Possible involvement of the signaling pathways, such as PI3 kinase, MEK1/2, P38-MAPK, and JNK1/2, was determined in MM-induced toxicity in the astrocytes. We incubated the astrocytes with the MM and cotreated the cells with LY294002 (10μM), PD98059 (10μM), SB203580 (10μM), or SP600125 (10μM), which are inhibitors to PI3 kinase, MEK1/2, P38-MAPK, and JNK1/2 pathways, respectively. Inhibitors themselves were nontoxic to astrocytes. Treatment with PD98059 or SP600125 promoted viability of MM-treated astrocytes (Fig. 4A) but LY294002 or SB203580 did not, suggesting the involvement of MEK1/2 and JNK1/2 pathways in MM-induced toxicity.

We then studied the time course of phosphorylation of MEK1/2 (assessed by ERK1/2 phosphorylation) and JNK1/2. We observed that phosphorylation of ERK1/2 started before 5 min, reached its peak at 10 min, and declined to less than the basal level in 30 min (Fig. 4B). The phosphorylation of JNK1/2 started at 10 min, reached its peak at 15 min, and declined thereafter to the basal level in 30 min (Fig. 4C). Because MM-induced phosphorylation of ERK1/2 preceded phosphorylations of JNK1/2, we determined whether activation of JNK1/2 was ERK1/2 dependent. We simultaneously incubated the primary astrocytes with PD98059 and MM, and observed that PD98059 blocked MM-stimulated phosphorylation of JNK1/2 (Fig. 4D). But upon incubation of the cells with SP600125 (10μM) and MM, the former failed to block MM-stimulated phosphorylation of ERK1/2 (Fig. 4E). These data suggested that MM toxicity in rat primary astrocytes involves an upstream MEK1/2 activation followed by JNK1/2.

Effect of MM on [Ca²⁺]_i and ROS Generation in the Rat Primary Astrocytes

It has been previously reported that induction of apoptosis by metals correlates with [Ca²⁺]_i release and ROS generation (Yang et al., 2008) in astrocytes, and therefore, we investigated the effect of MM on them.

We treated the astrocytes with MM and observed that the MM triggered [Ca²⁺]_i release. The [Ca²⁺]_i release reached its peak after 30 min of MM treatment (Fig. 5A). Similarly, MM triggered ROS generation, and the ROS generation reached its peak after 1 h of MM treatment (Fig. 5B).

To investigate whether the [Ca²⁺]_i release was ROS, ERK1/2, or JNK1/2 –dependent, we incubated the MM-treated astrocytes with an antioxidant (α-tocopherol, 200 μg/ml), PD98059 (10μM), or SP600125 (10μM). α-Tocopherol itself was nontoxic. We observed that PD98059 (10μM) or SP600125 (10μM) suppressed [Ca²⁺]_i release, but α-tocopherol (200 μg/ml) did not (Fig. 5C). This suggested that [Ca²⁺]_i release in MM-treated astrocytes was ERK1/2 and JNK1/2 dependent.

To investigate whether the ROS generation was [Ca²⁺]_i, ERK1/2, or JNK1/2 dependent, we incubated the MM-treated astrocytes with [Ca²⁺]_i chelator (BAPTA-AM, 5μM), PD98059 (10μM), or SP600125 (10μM). BAPTA-AM itself was nontoxic to astrocytes. We observed that BAPTA-AM, PD98059, or SP600125 suppressed ROS generation in MM-treated rat primary astrocytes (Fig. 5D), suggesting that ROS generation was [Ca²⁺]_i, ERK1/2, and JNK1/2 dependent.

We next investigated whether ERK1/2, JNK1/2, [Ca²⁺]_i and ROS signaling resulted in apoptosis. We treated the MM-treated astrocytes with α-tocopherol (200 μg/ml), PD98059 (10μM), BAPTA-AM (5μM), or SP600125 (10μM) and observed that they all suppressed apoptosis (Fig. 5E). This suggested that activation of ERK1/2 and JNK1/2, followed by increased [Ca²⁺]_i and ROS generation, resulted in apoptosis in the MM-treated astrocytes.

DISCUSSION

Understanding the cellular mechanisms leading to pathological alterations in developing rat brain by a mixture of heavy metals containing As, Cd, and Pb prompted us to undertake this study. Major findings of our study are that MM affects neurobehavioral parameters in developing rats that may be triggered by a compromised blood-brain communication. The metals act in synergy to reduce the expression of GFAP, an important protein component of BBB (McCall et al., 1996), accompanied by increase in apoptosis and morphological alterations in GFAP-expressing astrocytes. The synergistic apoptotic effect on astrocytes involves initial activation of ERK, which in turn activates the JNK pathway, followed by an increase in Ca²⁺_i release and ROS generation.

We investigated neurobehavioral impact of MM during rat brain development. Our data show that MM treatment induced hyperlocomotion, enhanced grip strength, and reduced learning-memory functions in rats. The abnormalities in neurobehavior can be explained by altered glia-neuron interactions (Sykova, 2001) resulting from reduced GFAP level that plays a vital role in modulating synaptic efficacy in the CNS (Garcia et al., 2003; McCall et al., 1996). Moreover, the synapse being tightly ensheathed by astrocyte processes (Sykova, 2001, 2005), change in the number of processes may modulate synaptic activity (McCall et al., 1996) leading to behavioral impairments.

The synthesis of γ-amino butyric acid (GABA; the chief inhibitory neurotransmitter in the mammalian CNS) in neurons is closely associated with astrocyte metabolism (Sonnewald et al., 1993). Thus, it can be suggested that the damaged astrocytes induced by MM may play an important role in the pathophysiology of GABA hypofunction (Kondziella et al., 2006)–induced hyperactivity (hyperlocomotion and enhanced grip strength). Moreover, marked leakage of BBB in the white matter, corpus callosum, and posterior lobe of cerebral cortex and cerebellum may cause profound modulation in coordinated stimulation and integrated information from different neural circuits (Habas, 2001; Poliak and Peles, 2003) leading to the observed behavioral alterations.

Astrocytes play an important role in long-term potentiation (LTP) (McCall et al., 1996) that is widely considered as one of the major mechanisms that underlies learning and memory (Bliss and Collingridge, 1993; Cooke and Bliss, 2006). The astrocytes are a source of neurotrophic factors (Muller et al., 1995; Rudge et al., 1992), and reduced GFAP level interferes with their production (McCall et al., 1996). Moreover, GFAP directly participates in the change in arborization of astrocytic processes during LTP (Wenzel et al., 1991). Thus, reduced GFAP expression by MM may interfere in the necessary signaling between astrocytes and neurons or synapses, and impair learning-memory performance in developing rats.

An immature or disrupted BBB appeared to be vulnerable to metal permeation (Lidsky and Schneider, 2003; Wang et al., 2007b) that led to significant dose-dependent rise in levels of As, Cd, and Pb. The rise in metal content in developing rat brain confirmed metal passage to the postnatal rats from lactating mothers. Prolonged metal treatment sustained BBB damage in the P60 rats indicating reduced resistance to the effect of the metals even in adults.

GFAP is an established marker of astrocyte maturation and reactivity (Gomes et al., 1999). We observed that MM induced reduction in cortical and cerebellar GFAP levels, and increased apoptosis in GFAP-ir astrocytes. MM exposure reduced GFAP-expressing astrocyte count, size, and number of projections, thus suggesting that it adversely impacts astrocyte function. Our data extend previous observation of reduced glial differentiation by either Pb or Cd in developing rat brain (Chow et al., 2008; Zawia and Harry, 1996). In addition, our data show that the reduction in GFAP levels by MM exposure is more than the sum of effects of individual metal, thus suggesting higher risk of the metals in mixture than expected or theoretically predicted. Therefore, addition of As enhances previously reported synergistic toxic effect of Pb and Cd in the brain cells of cerebellum and cortex (Gu et al., 2009).

Inhibition of GFAP immunoreactivity by MM in developing brain appears to be caused by astrocyte apoptosis. In primary cultures of astrocytes, our data show that MM synergistically induced apoptosis. This was manifested by the activation of MEK/ERK, followed by the activation of JNK pathways, which then enhanced intracellular Ca²⁺ levels and subsequently ROS generation. There are reports on the activation of JNK1/2 downstream to ERK1/2 in pluripotent HT29 intestinal cell line (Salh et al., 2000) and human melanoma (Lopez-Bergami et al., 2007). Cd is reported to induce apoptosis via the activation of JNK-mediated signaling where Ca²⁺ and oxidative stress act as the pivotal mediators of apoptotic signaling in skin epidermal cell line (Son et al., 2010). But we are not aware of any report describing our observed sequences of events leading to astrocyte apoptosis by metals.

Astrocytes are networked together by a series of gap junctions permitting to propagate Ca²⁺ waves through the linked network (Lobsiger and Cleveland, 2007), and Ca²⁺-mediated intercellular communication is a mechanism by which astrocytes communicate with each other and modulate the activity of adjacent cells (Verderio and Matteoli, 2001). MM-induced alteration in astrocyte morphology may influence [Ca²⁺]_i (Barres et al., 1989); in contrast, an increase in [Ca²⁺]_i may also play a key role in altering astrocyte cytoskeleton, affecting the glia-neuron interaction (Shelton and McCarthy, 2000).

Previous reports show that As (Jin et al., 2004) and Pb (Zhang et al., 2008) induced apoptosis in astrocyte culture, and Ca²⁺ chelation and increase in antioxidant defense system protected astrocytes from Cd insults (Yang et al., 2008). Our data, for the first time, shed new light on the toxic mechanism of the three metals in combination involving synergistic loss in viability and enhanced apoptosis in rat primary astrocytes.

Overall, this study emphasizes that MM exposure during brain development has greater than additive response on astrocyte toxicity with adverse impacts on BBB function that culminate into neurological deficits in developing rats. This study underscores the importance of synergistic action of metals to that compared individually and hence would warrant revision of the threshold of metal toxicity and the current cutoff values of such environmental contaminants in assessing relative neurodevelopmental risks.

FUNDING

SERC Fast Track Scheme For Young Scientists of the Department of Science and Technology, Government of India; Supra Institutional Project of the Council of Scientific and Industrial Research, India.

We thank Dr Naibedya Chattopadhyay (scientist, Central Drug Research Institute), Dr Rohit A. Sinha (researcher, Duke-Nus Graduate Medical School), Dr A. K. Agarwal (emeritus scientist, Indian Institute of Toxicology Research), and Dr Preeti Srivastava (scientist, Indian Institute of Toxicology Research) for helpful suggestions and insightful comments.

References