Changes in Biochemical Processes in Cerebellar Granule Cells of Mice Exposed To Methylmercury

Abstract

At postnatal day 34, male and female C57BL/6J mice were exposed orally once a day to a total of five doses totaling 1.0 or 5.0 mg/kg of methylmercuric chloride or sterile deionized water in moistened rodent chow. Eleven days after the last dose cerebellar granule cells were acutely isolated to measure reactive oxygen species (ROS) levels and mitochondrial membrane potential using CM-H₂DCFDA and TMRM dyes, respectively. For visualizing intracellular calcium ion distribution using transmission electron microscopy, mice were perfused 11 days after the last dose of methylmercury (MeHg) using the oxalate-pyroantimonate method. Cytosolic and mitochondrial protein fractions from acutely isolated granule cells were analyzed for cytochrome c content using Western blot analysis. Histochemistry (Fluoro-Jade dye) and immunohistochemistry (activated caspase 3) was performed on frozen serial cerebellar sections to label granule cell death and activation of caspase 3, respectively. Granule cells isolated from MeHg-treated mice showed elevated ROS levels and decreased mitochondrial membrane potential when compared to granule cells from control mice. Electron photomicrographs of MeHg-treated granule cells showed altered intracellular calcium ion homeostasis ([Ca²⁺]_i) when compared to control granule cells. However, in spite of these subcellular changes and moderate relocalization of cytochrome c into the cytosol, the concentrations of MeHg used in this study did not produce significant neuronal cell death/apoptosis at the time point examined, as evidenced by Fluoro-Jade and activated caspase 3 immunostaining, respectively. These results demonstrate that short-term in vivo exposure to total doses of 1.0 and 5.0 mg/kg MeHg through the most common exposure route (oral) can result in significant subcellular changes that are not accompanied by overt neuronal cell death.

Mercury is an ubiquitous chemical found in our environment as a result of natural events (volcanic eruptions) and anthropogenic sources (burning of fossil fuels, etc.). An organic form of mercury (methylmercury, MeHg) is highly neurotoxic by virtue of its lipophilicity. Based on two epidemics of MeHg poisoning (Minamata Bay, Japan, and Iraq), ingestion of contaminated food can be considered as the primary route of exposure (Bakir et al. 1973; Takeuchi et al. 1962). In both of these episodes, the primary signs of neurological dysfunction were cerebellar ataxia, generalized weakness of extremities, and sensory disturbances.

Generation of reactive oxygen species (ROS), through the electron transport chain attached to the inner mitochondrial membrane, is a part of normal cellular function. Excessive ROS formation can result from inefficient functioning of the electron transport chain. Yoshino, Mozai, and Nakao (1966) reported that mitochondria are the earliest sites of mercury accumulation in the brain following administration of MeHg, and O’Kusky (1983) reported changes in mitochondrial morphology in the developing rat brain. Several in vitro studies using cultured granule cells and brain slices demonstrated excessive ROS formation due to MeHg treatment (Gasso et al. 2000; Yee and Choi 1996).

MeHg has been associated with altered intracellular calcium ion homeostasis ([Ca²⁺]_i) in a number of cells including cultured rat cerebellar granule cells (Denny, Hare, and Atchison 1993; Marty and Atchison 1997; Mundy and Freudenrich 2000; Shenker et al. 1992, 1993). Recent in vitro studies using cultured cerebellar granule cells from rats demonstrated that MeHg causes the opening of the mitochondrial permeability transition pore, leading to apoptotic cell death (Limke and Atchison 2002; Limke, Otero-Montanez, and Atchison 2003). Opening of the mitochondrial permeability transition pore is associated with release of mitochondrial factors like cytochrome c, which in turn activates caspase 3 (InSug et al. 1997; Limke and Atchison 2002; Wigdal et al. 2002). Nagashima et al. (1996) demonstrated the process of apoptosis in cerebellar granule cells of rats exposed to a total oral dose of 40 mg/kg methylmercuric chloride divided into 10 equal doses given on alternate days over a period of 20 days.

Most of the previous studies demonstrated activation of the mitochondrial pathway of apoptosis either in cultured granule cells or exposing rodents to relatively high doses of MeHg, (i.e., total exposures greater than 10 mg/kg), whereas the U.S. Environmental Protection Agency (EPA) reference dose is 0.1 μg/kg body weight per day (U.S. EPA 1997; UNEP Chemicals 2002). The objective of this investigation is to determine whether short-term, low to moderate (1.0 to 5.0 mg/kg), in vivo exposure of rodents through ingestion of food containing MeHg (the most common route of exposure) will lead to subcellular events like elevation of ROS, alteration in [Ca²⁺]_i, and loss of mitochondrial membrane potential (MMP). Whether these changes lead to apoptosis of granule cells at low in vivo exposure levels will provide valuable insights with respect to in vivo toxicity studies.

In the present study male and female C57BL/6J mice were exposed at postnatal (P) days 34 to 38 to MeHg via food using a total dose of 1.0 and 5.0 mg/kg body weight and compared with age-matched control (given vehicle only) mice for subcellular changes (ROS production, alteration in [Ca²⁺]_i homeostasis, and loss of MMP). Western blot analysis was performed on protein isolated from cytosolic and mitochondrial fractions to determine levels of cytochrome c. In addition, Fluoro-Jade staining was used to label dying cerebellar neurons. Additional sections were immunolabeled with anti-activated caspase 3 antibody. It is well known from the literature that the central nervous system (CNS) effects of MeHg are typically delayed in onset (Rice 1996). In a previous study (Bellum et al. 2007a), we observed significant behavioral effects of MeHg exposure using a 5- to 11-day waiting period after cessation of MeHg exposure. In this study we chose to examine the cellular effects of MeHg exposure after an 11-day period during which MeHg-treated mice were no longer exposed to MeHg.

MATERIALS AND METHODS

Animals

Adult C57BL/6J wild type (+/+) male and female mice, originally obtained from The Jackson Laboratory (Bar Harbor, MA, USA), were bred to produce wild type offspring. A total of 43 male and 35 female mice were used in this study. Mice were further divided into three treatment groups as follows: control (n = 23), 1.0 mg/kg (n = 26), and 5.0 mg/kg (n = 29) methylmercuric chloride. All mice were housed at the Laboratory Animal Research and Resource building, Texas A&M University, in a constant temperature (21°C to 22°C) and humidity (45% to 50%) room with a 12-h light-dark cycle. The mice were weaned at 29 days of age and housed individually for the duration of dosing and experimentation. All procedures were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publication No. 85-23, revised 1996).

Chemicals and Dosing

Methylmercuric chloride (95% purity), obtained from Alfa Aesar (Ward Hill, MA, USA), was dissolved in sterile deionized water for addition to food. All methylmercuric chloride solutions were stored at 4°C until used. Mice at postnatal day (P) 29 were divided into three treatment groups and dosed as described earlier by Bellum et al. (2007b).

Acute Isolation of Cerebellar Granule Cells

Cerebellar granule cells were acutely isolated as described in Current Protocols in Toxicology (Oberdoerster 2001). At P50, mice were anesthetized using isoflurane and killed by decapitation. The brain was removed and the cerebellum separated from the rest of the brain. The meninges were removed, the cerebellum was chopped in six to eight pieces and transferred to a chilled 50-ml falcon tube containing minimum essential medium (MEM) with Earle’s salts (Life Technologies, Rockville, MD, USA). Cerebellar granule cells were then isolated using dissociation medium containing MEM and 1.5 U/ml protease (Sigma, St. Louis, MO, USA).

Measurement of Cell Viability

Aliquots of dissociated cells in minimum essential medium were stained with 0.08% Trypan blue (Sigma) dye to check viability using a hemocytometer. Trypan blue dye selectively stains dead cells. Cells were counted in four 1-mm² squares using an inverted compound microscope at 100× magnification. The cell number was determined by multiplying by the dilution factor (× 10,000). Based on this dye exclusion method cell viability was observed to be 90% or greater (data not shown).

Reactive Oxygen Species (ROS)

Cerebellar granule cells were acutely dissociated as described above and plated onto chambered slides (VWR International), then incubated in 95% O₂ and 5% CO₂ at 37°C for 25 min. The cells were loaded with an indicator dye, chloromethyldihydrodichlorofluorescein diacetate (CM-H₂DCFDA) (Molecular Probes, Eugene, OR, USA) at a concentration of 500 nM and incubated in 95% O₂ and 5% CO₂ at 37°C for 8 min. CM-H₂DCFDA is a redox-sensitive dye used to determine ROS levels (De Bernardo et al. 2004). Sequential time course fluorescent image capturing was performed for 22.5 min using a 90-s interval with a 20× objective on an Olympus 1X-70 microscope and a Hamamatsu ORCA-ER cooled charge-coupled device camera at excitation and emission of 490 and 520 nm, respectively. Image capturing and ROS levels were analyzed using Simple PCI version 5.0.0.1503 (Compix and Imaging System, Cranberry Township, PA, USA).

Intracellular Calcium Distribution

Control and MeHg-treated mice at P50 (three to four mice per treatment group) were deeply anesthetized to a surgical plane of anesthesia using xylazine and ketamine and perfused with 20 ml of 90 mM potassium oxalate (Sigma), followed by 20 ml of 3% glutaraldehyde (EMS, Fort Washington, PA, USA) containing 90 mM potassium oxalate (Silklos et al. 2000). The cerebella were sliced into 1-mm-thick sagittal sections and further fixed in the same oxalate containing fixative for 24 h at 4°C. The sections were rinsed in 7.5% sucrose containing 90 mM potassium oxalate and further fixed in a solution containing 2% potassium pyroantimonate (Sigma), 1% osmic acid (EMS), and 0.01 N acetic acid for 2 h at 4°C. Cerebellar sections were then rinsed, dehydrated, and embedded in araldite following the protocol of Rhyu et al. (1999). Cerebellar slices were cut at 100 α° and viewed using a Zeiss 10C electron microscope.

Mitochondrial Membrane Potential

Acutely dissociated cerebellar granule cells were plated onto cover slips and incubated at 37°C for 25 min in minimum essential medium. Cells were loaded with an indicator dye, tetramethyl rhodamine methyl ester (TMRM), at 150 nM and further incubated in 95% O₂ and 5% CO₂ at 37°C for 15 min (Kume et al. 2006). Fluorescent images were acquired using a 40× oil objective on an Olympus 1X-70 microscope and a Hamamatsu ORCA-ER cooled charge-coupled device camera at excitation and emission of 555 and 600 nm, respectively. Image capturing and membrane potential analysis were carried out using Simple PCI version 5.0.0.1503 (Compix and Imaging System).

Western Blot Analysis

Mice at P50 were anesthetized and decapitated for granule cell dissociation as described previously. Dissociated granule cells from two animals were pooled and processed using a Mitochondria Isolation Kit (Pierce, Rockford, IL, USA). Protein fractions were determined using BCA Protein Assay Reagent (Pierce). Whole cerebella from an additional three mice per treatment group also were collected and analyzed. Mice were anesthetized with isoflurane, killed by decapitation, brains rapidly removed and frozen with powdered dry ice, then stored at −70°C until used. Total protein from each cerebellum was extracted using M-PER (Pierce) and Complete, Mini, EDTA-free protease inhibitor (Roche Diagnostics, Indianapolis, IN, USA). Thirty micrograms of whole-cerebellum protein sample, 20 μg of the cytosolic fraction, and 10 μg of the mitochondrial fraction from each sample were denatured and run on a 4% to 20% gradient polyacrylamide gel. Proteins were transferred onto polyvinylidine fluoride (PVDF) blotting membrane (BioRad, Hercules, CA, USA) using a Mini Trans-Blot Electrophoretic Transfer Cell system (BioRad). Membranes were blocked then incubated with monoclonal mouse cytochrome c antibody (1.0 μg/ml; BD Pharmingen, San Diego, CA, USA) overnight at 4°C, followed by peroxidase-labeled secondary antibody for 2 h at room temperature. Immunoreactive bands were visualized using Super Signal West Femto Chemiluminescent substrate (Pierce) and an Alpha Innotech FluorChem 8800 Imaging System. The Western blots were run in duplicate. All membranes were also stained with Coomassie blue stain to verify equal loading and transfer of proteins during Western blotting.

Immunohistochemistry

Control and MeHg-treated mice at P50 were deeply anesthetized with ketamine/xylazine (given intraperitoneally) and perfused transcardially with 50 ml of Tyrode’s saline followed by 300 ml of 4% phosphate-buffered paraformaldehyde (pH 7.4). Brains were removed and cryoprotected with 20% sucrose in 0.1 M phosphate-buffered saline (PBS). Sagittal sections (20 μm) were cut using a cryostat and mounted on gelatin-coated slides. Sections were permeabilized using 0.3% Triton X-100 for an hour, followed by a 5-min incubation in 3.0% hydrogen peroxide in 0.1 M PBS to quench endogenous peroxidases. The sections were then blocked in 5% normal goat serum and incubated overnight at 4°C in anti-activated caspase 3 antibody (R&D Systems, Minneapolis, MN, USA) at a 1:50,000 dilution. Sections were incubated for 2 h in biotinylated goat anti-rabbit secondary antibody (1:400 dilution; Vector Laboratories, Burlingame, CA, USA), followed by peroxidase-labeled streptavidin (1:5000 dilution; Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA). Signal was detected with 0.024% 3,3-diaminobenzidine and 0.006% hydrogen peroxide in 0.05 M Tris-HCl buffer (pH 7.6). The reaction was stopped after achieving optimal signal by transferring the sections to 0.05 M Tris-HCl buffer. The slides were transferred to 0.1 M PBS and dehydrated in a graded series of ethanol followed by two changes of xylene and coverslipped with DPX mounting medium.

Fluoro-Jade Staining

At P50, mice were anesthetized with isoflurane and brains were removed, frozen with powdered dry ice, and stored at −70°C. Serial sagittal sections, 20 μm thick, were cut on a cryostat and stored at −70°C until they were stained with Fluoro-Jade (Histochem, Jefferson, AR, USA). The fixed frozen sections were thawed and dried thoroughly for 20 min in a 45°C oven. The slides were stained with Fluoro-Jade using the protocol described by Schmued, Albertson, and Slikker (1997) with a few modifications (Frank et al. 2003). Stained sections were examined using epifluorescence and images were acquired using a Zeiss Axioplot 2 research microscope (Carl Zeiss, Thornwood, NY, USA) and a fluorescein isothiocyanate (FITC) filter and equipped with a three-chip Hamamatsu video camera.

Statistical Analysis

Mitochondrial membrane potential (MMP), Fluoro-Jade, and activated caspase 3 data were analyzed using general linear model (GLM)–univariate analysis of variance at α = 0.05 (SPSS version 11.0 for Windows). ROS data were analyzed using GLM–repeated-measures analysis. Significant differences among treatment and control groups were interpreted using the Tukey’s honest significant difference (HSD) post hoc test.

RESULTS

Reactive Oxygen Species

ROS levels were examined in the three treatment groups included in this study: control, 1.0 mg/kg MeHg, and 5.0 mg/kg MeHg based on analysis of the intensity of fluorescence of CM-H₂DCFDA dye (Figure 1). GLM–multivariate analysis of variance (three-way analysis of variance [ANOVA]) indicated an overall significant difference between control and MeHg-treated mice (p < .001) and a significant treatment effect (p=.012) also was observed. No significant difference was observed with respect to gender (p=.239) or interaction between treatment and gender (p=.624). Starting with the fourth measurement, mouse granule cells exposed to 5.0 mg/kg MeHg in vivo exhibited significantly higher ROS levels compared to control granule cells. Starting with the sixteenth measurement, mouse granule cells exposed to 1.0 mg/kg MeHg in vivo also exhibited significantly higher ROS levels compared to control granule cells (Figure 1B ).

Intracellular Calcium Distribution

Qualitative analysis of the electron micrographs revealed alteration in intracellular calcium ion homeostasis in granule cells from both MeHg-exposed groups in which the mitochondrial calcium ion concentration was high compared to control granule cells (Figure 2). In addition, increased calcium ion concentrations in the nuclei of granule cells from both MeHg treatment groups were observed.

Mitochondrial Membrane Potential

Cerebellar granule cells were observed to contain clusters of mitochondria when loaded with TMRM dye, and the amount of dye found within individual cells is in proportion to the MMP. Figure 3 summarizes the relative mean mitochondrial membrane potential of cerebellar granule cells from control, 1.0 mg/kg, and 5.0 mg/kg MeHg-treated mice, based on analysis of fluorescence intensity of TMRM dye. GLM–univariate analysis of variance and Tukey’s HSD post hoc test indicated a significant difference between control and MeHg treated mice (p < .001). No significant difference was observed between males and females (p= .211). There also was no significant interaction between treatment and gender (p= .309).

Western Blot Analysis for Cytochrome c

Mitochondrial damage can alter the mitochondrial transition pore (MTP), an event characterized by decreased MMP (Scorrano, Petronilli, and Bernardi 1997). This can result in release of several factors from within the mitochondria into the cytosol, which induce apoptosis. These factors include cytochrome c and apoptosis-inducing factor. In normal cells, these factors are found in higher quantities in the mitochondrial fraction compared to the cytosolic fraction.

When cytochrome c levels were determined using total protein extracted from whole cerebella of control and MeHg-treated mice, no significant differences were observed (data not shown). To determine the distribution of cytochrome c between the mitochondrial and cytosolic compartments, granule cells were isolated then mitochondrial and cytosolic proteins were separated. Western blot analysis was performed using monoclonal cytochrome c antibody that recognized a 15-kDa cytochrome c in the mitochondrial protein fraction of all three treatment groups. The cytosolic fraction revealed a 15-kDa faint band only in granule cells from MeHg-treated mice (Figure 4A ). These results suggest that compared to control granule cells, MeHg-treated granule cells contained slightly elevated levels of cytosolic cytochrome c, indicating possible cytochrome c release from mitochondria into the cytosol.

Activated Caspase 3 Immunohistochemistry

To determine protein expression for activated caspase 3, four to eight serial sagittal sections from each cerebellum were immunostained with anti-activated caspase 3 antibody and immunopositive granule cells were counted by a researcher blinded to the treatment group of the sections during the counting process. Immunohistochemical staining of cerebella from both control and MeHg-treated mice showed very few activated caspase 3–immunoreactive granule cells in spite of the presence of cytochrome c in the cytosol of MeHg-treated mice granule cells (data not shown).

Fluoro-Jade Labeling of Dying Cells

Fluoro-Jade staining was performed to determine the extent of cell death and the possibility of activation of other pathways of apoptosis such as extrinsic via death receptors or intrinsic via release of mitochondrial associated proapoptotic proteins, including cytochrome c, apotosis-inducing factor (AIF), or Smac. Fluoro-Jade is an anionic fluorescein derivative that selectively stains degenerating neurons irrespective of the mechanism of cell death (apoptosis or necrosis) (Schmued, Albertson, and Slikker 1997; Frank et al. 2003). Fluoro-Jade–positive granule cells from four to eight serial sagittal sections from each mouse cerebellum were counted for each treatment group. GLM–univariate analysis of variance indicated no significant difference in the number of Fluoro-Jade–positive granule cells observed in the cerebella from control and MeHg-treated mice (p= .397) (data not shown).

DISCUSSION

The observations reported here demonstrate elevation of ROS levels, alteration in [Ca²⁺]_i homeostasis, and reduction in MMP in 1.0 and 5.0 mg/kg MeHg-treated mouse granule cells when compared to control granule cells. However, as shown by Western blot analysis for cytochrome c and immunohistochemical staining for activated caspase 3, the most common pathway of apoptosis discussed in the MeHg literature (Kroemer 2002), was not activated at the concentrations of MeHg used in this study and at the time point investigated. These particular doses (1.0 and 5.0 mg/kg total exposure doses) were selected because they are within a range that people could be exposed from consumption of MeHg-contaminated fish, especially the larger predatory fish such as king mackeral, pike, shark, swordfish, and large tuna (U.S. EPA 1997; UNEP Chemicals 2002).

Total cellular ROS levels were determined in control and MeHg-treated granule cells using CM-H₂DCFDA dye. There are no published data available to compare directly with the results obtained by the technique used in this study. However, results from several other laboratories demonstrated increased generation of ROS in MeHg-induced neurotoxicity using cultured cerebellar granule cells (Ishibashi et al. 2004; Limke and Atchison 2002; Limke, Otero-Montanez, and Atchison 2003; Marty and Atchison 1997, 1998; Mundy and Freudenrich 2000; Yee and Choi 1996) or immortalized cell lines (Belletti et al. 2002; Gatti et al. 2004). These in vitro studies support the observations reported here.

Several studies reported alteration of [Ca²⁺]_i homeostasis as a result of elevated ROS levels (Ishibashi et al. 2004; Marty and Atchison 1997, 1998). Qualitative analysis of transmission electron photomicrographs of oxalate-pyroantimonate–perfused cerebella from control and MeHg-treated mice in this study revealed excess calcium accumulation in mitochondria of granule cells from 1.0 and 5.0 mg/kg MeHg-treated mice. These observations suggest that mitochondria are buffering excess [Ca²⁺]_i, resulting in altered [Ca²⁺]_i homeostasis. Kamphuis et al. (1989) and Silklos et al. (2000) demonstrated altered [Ca²⁺]_i distribution in rat hippocampus and spinal and oculomotor neurons of superoxide dismutase (SOD1) knockout mice, respectively, using the oxalate-pyroantimonate technique. It is interesting to note that in addition to excess calcium ion concentrations in the mitochondrial lumen, nuclei of granule cells from both groups of MeHg-exposed mice also seem to accumulate more calcium in the nucleoplasm when compared to nuclei of control granule cells. The nuclear envelope has a relatively small volume but its outer membrane is well connected to the endoplasmic reticulum and maintains steady resting levels of 100 μM of calcium concentration within their lumen. Attached to the inner nuclear membrane are specific calcium release channels that can be activated by inositol trisphosphate (IP₃) or cyclic adenosine diphosphate (cADP) ribose that allows release of calcium into the nucleoplasm. In a normally functioning neuron, selective release of calcium into the nucleus helps to regulate gene expression (Hardingham and Bading 1998; Petersen et al. 1998). It is possible that excess cytosolic calcium in MeHg-treated granule cells is taken up by the nuclear envelope endoplasmic reticulum (NE-ER) continuum and eventually conveyed into the nucleoplasm. The excess nuclear calcium could alter gene expression patterns in MeHg-treated mice. Therefore additional studies to determine the gene expression levels using microarray technology are warranted. However, there are no additional published reports among MeHg toxicity studies to demonstrate calcium ion distribution in the neurons using the oxalate-pyroantimonate technique. Consistent with findings reported in the literature, we observed elevated ROS levels, which might be contributing to altered [Ca²⁺]_i homeostasis.

Cerebellar granule cells from 1.0 and 5.0 mg/kg MeHg-exposed mice showed a marked decrease in MMP when compared to control granule cells. Intracellular factors such as [Ca²⁺]_i and ROS levels or even damage to mitochondria by MeHg exposure can lead to decreased MMP (Limke and Atchison 2002). The elevated ROS levels and altered [Ca²⁺]_i homeostasis demonstrated in this study could be the cause of the observed decreased MMP in MeHg-exposed granule cells. Bernardi, Veronese, and Petronilli (1993) and Scorrano, Petronilli, and Bernardi (1997) demonstrated loss of MMP due to opening of a megapore called the mitochondrial permeability transition pore, which they determined could be the result of direct binding of MeHg or in response to either elevated ROS levels or altered [Ca²⁺]_i homeostasis.

Apoptosis is a process of cell death that is controlled internally by the dying cell (Hengartner 2000; Reed 2000). A cell dying due to apoptosis shows characteristic signs including cell shrinkage, nuclear pyknosis, and DNA fragmentation (Kroemer 1997; Mattson 2000). Loss of MMP is thought to occur in most mammalian cell types before apoptotic cell death (Kroemer 2002) and is usually associated with the release of mediators of apoptosis such as cytochrome c. Cytochrome c associates with an adaptor molecule, apaf-1, as well as caspase 9, resulting in activation of caspase 3. To assess whether the MeHg-exposed cerebellar granule cells were dying via a classic apoptotic cell death pathway, we performed Western blot analysis for cytochrome c levels in mitochondrial and cytosolic fractions, and performed immunohistochemistry to look for caspase 3 activation. Our results revealed that although there is a slight increase of cytochrome c in the cytosolic fraction of MeHg-treated granule cells, significant activation of caspase 3 is not occurring at the concentrations of MeHg used in this study or at the time point at which the immunohistochemistry was performed. To rule out the possibility of activation of other pathways of neuronal cell death, frozen serial sections of cerebella from control and MeHg-treated mice were stained for neuronal degeneration using Fluoro-Jade. However, very few Fluoro-Jade–positive cells were observed per section in all treatment groups. Consistent with our results, Fonfria et al. (2002) demonstrated absence of caspase 3 activation in cultured cerebellar granule cells with MeHg exposure, but showed a significant increase in the translocation of apoptosis-inducing factor (AIF) into the nucleus. AIF is a mediator of apoptosis that also is released from mitochondria following loss of MMP. AIF acts independently of the caspase cascade pathway by translocating into the nucleus, leading to DNA degradation and chromatin condensation (Castoldi et al. 2000; Fonfria et al. 2002). However, we do not expect to observe activation of this pathway, as no difference was observed with respect to the amount of granule cell death present using Fluoro-Jade staining in control and MMC-treated mice.

Using morphological evidence obtained with electron microscopy, Nagashima et al. (1996) reported the occurrence of apoptosis in cerebellar granule cells from rats exposed to a total MeHg dose of 40 mg/kg body weight. Using electron microscopy, cerebellar granule cells were not observed to show morphological evidence of apoptosis in any of the treatment groups. Thus, at the time of examination of the cerebellum, the concentrations of MeHg used in our study were not sufficient to kill granule neurons but the MeHg exposure was high enough to disturb cellular homeostasis.

In conclusion, these results show that short-term, in vivo exposures to total doses of 1.0 and 5.0 mg/kg MeHg through the most common exposure route (oral) can result in subcellular changes in terms of oxidative stress, altered [Ca²⁺]_i homeostasis, and loss of MMP in cerebellar granule cells. The doses used in this study were lower than the majority of studies that resulted in apoptosis in the rodent brain (Nagashima et al. 1996). We previously reported that MeHg at similar doses significantly impacts behavior related to cerebellar dysfunction (Bellum et al. 2007a). It was interesting to note that the behavior changes were subtle, suggesting a possible role of subcellular changes in the cerebellar granule cells reported here. But these changes may not be severe enough to induce apoptotic granule cell death or overt changes in behavior related to cerebellar dysfunction.