EFFECTS OF METHYLMERCURY ON CEREBELLAR GRANULE CELLS OF THE TOTTERING MOUSE By Brenda Marie Marrero-Rosado A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Genetics – Doctor of Philosophy 2013 ABSTRACT EFFECTS OF METHYLMERCURY ON CEREBELLAR GRANULE CELLS OF THE TOTTERING MOUSE By Brenda Marie Marrero-Rosado Methylmercury (MeHg) is an organic highly toxic form of mercury and a persistent environmental neurotoxicant. The most common manner in which humans are exposed to MeHg is by consumption of contaminated fish. Studies from cases of chronic and acute human poisonings have revealed that MeHg causes a massive loss of cerebellar granule cells (CGCs), causing the characteristic dysarthria and ataxic signs. Previous research has found that acute treatment of rat CGCs with MeHg in vitro causes a time- and concentration-dependent increase in intracellular Ca 2+ 2+ ([Ca ]i) that is sufficient to cause CGC death. Voltage-gated Ca 2+ channels (VGCCs) are believed to play a role in the mechanism of MeHg-induced cytotoxicity by possibly facilitating access for the metal to intracellular targets. Many subtypes of VGCCs are expressed in CGCs, each characterized by different pharmacological and kinetic properties. The endeavor of the present studies was to investigate the effects of MeHg on CGCs of a mouse model of human Cav2.1 (P/Q-type) channelopathy. The tottering (tg) mouse is the result of a non-lethal deleterious point mutation in the α1A pore-forming subunit of the P/Q-type Ca 2+ channel. This VGCC subtype plays a crucial role in the process of neurotransmitter release in + mature CGCs of humans and animal models. Data will show that, in low-K conditions (closely mimicking their mature state) CGCs isolated from mice homozygous (tg/tg) and heterozygous 2+ (+/tg) for the tg mutation present a delay in the onset of MeHg-induced [Ca ]i increase in vitro. On the other hand, when they are acutely exposed to MeHg under depolarizing environments (mimicking their state in early development) tg/tg CGCs show increased susceptibility to cytotoxicity. Cerebellar organotypic slices from postnatal day (PND) 23-25 (after onset of ataxia) mice were also used to study the response of CGCs to MeHg taking into account the entire local cerebellar circuitry. Interestingly, +/tg cerebellar organotypic slices showed a higher percentage of CGC death than WT and tg/tg. The effect of MeHg on +/tg CGCs was partially eliminated by pretreating the slices with ω-conotoxin GVIA, an N-type VGCC antagonist. This suggests an important role of N-type VGCCs in the greater susceptibility of mature +/tg CGCs to MeHg. Of great importance is that tg-like mutations have been linked to human disorders including episodic ataxia type 2, familial hemiplegic migraine and spinocerebellar ataxia type 6. This work presents evidence of a genotype-environment interaction that could potentially identify human populations with higher risk for the neurotoxic effects of MeHg. In dedication to my family; My life accomplishments are possible because of your infinite love and constant support. iv ACKNOWLEDGEMENTS I would like to thank my mentor, Dr. William Atchison, for his guidance, patience, and support. Many thanks to my committee members: Drs. John Fyfe, Colleen Hegg, John Goudreau, and Ke Dong. Their invaluable constructive criticism and guidance helped shape me as a scientist. Thank you, Dr. Ravindra Hajela. You believed in me from the moment I stepped foot in the Atchison lab, and gave me the tools to succeed. You are an exemplary scientist, but an even better friend. To my dear friends: Elizabeth Molina-Campos, Alexandra Colón-Rodriguez, Sara Fox, Aaron Bradford, Dawn Autio, all the Atchison lab, and Jessica Switzenberg. Thank you for your support and for lending me an ear whenever I needed advice. I could not ask for a better group of people to call my friends. I am very lucky to have you all in my life! Thanks to the students that helped in many aspects of the experiments presented in this dissertation, and that I was very fortunate to mentor: Katiria González-Rivera, Pamela S. Vega-Pérez, and Kaye L. Long-Roldán. I am immensely grateful to my husband, Duane Bish, for his love and patience during these challenging Ph.D. years. You kept me sane and encouraged me to always give my best and keep pushing forward, especially at times when I thought of giving up. You were my rock! I love you. v Last, but not least, I would like to give many thanks to my family who always cheered me on and offered their unconditional support. Thanks to my mother, Doraida Rosado, for sacrificing so much in order to provide my brother and me all the resources needed to succeed; and to my brother, Josué Marrero-Rosado, I thank you for always believing in your big sister and showing me your support every day. vi TABLE OF CONTENTS LIST OF TABLES………………………………………………………………………………...x LIST OF FIGURES……………………………………………………………………………....xi LIST OF ABBREVIATIONS…………………………………………………………………..xiii CHAPTER ONE: INTRODUCTION……………………………………………………………..1 A. Methylmercury………………………………………………………………………........2 2+ B. Voltage-gated Ca channels……………………………………………………………...8 2+ 1. Effects of methylmercury on voltage-gated Ca channel function………………9 C. The cerebellum…………………………………………………………………………...16 1. Development of cerebellar granule cells…………………………………………18 2. Effects of methylmercury on cerebellar granule cells…………………………...20 D. The tottering (tg) mouse……………………………………………………………........25 E. Specific aims……………………………………………………………………………..28 2+ CHAPTER TWO: EFFECT OF METHYLMERCURY ON INTRACELLULAR Ca HOMEOSTASIS OF TOTTERING CEREBELLAR GRANULE CELLS…………………………………………………………………………………………...41 A. Abstract………………………………………………………………………………….42 B. Introduction……………………………………………………………………………...44 C. Materials and methods…………………………………………………………………..46 1. Mice……………………………………………………………………………...46 2. Cerebellar granule cell isolation…………………………………………………46 2+ 3. Single-cell measurement of changes in [Ca ]i…………………………………47 4. Immunohistochemistry…………………………………………………………..48 5. Statistical analysis………………………………………………………………..50 D. Results……………………………………………………………………………………51 1. VGCC subtype composition at each DIV in WT, +/tg, and tg/tg CGCs cultured in + 25 mM K ………………………………………………………………………..51 2+ 2. Effect of acute MeHg exposure on [Ca ]i homeostasis in WT, +/tg, and tg/tg CGCs at DIV 4, 6, 8, and 10……………………………………………………..52 3. Effect of the tg mutation on the duration (min) of the first phase of MeHg-induced 2+ [Ca ]i increase…………………………………………………………………..53 E. Discussion………………………………………………………………………………..54 CHAPTER THREE: EFFECT OF METHYLMERCURY ON TOTTERING CEREBELLAR GRANULE CELL VIABILITY IN VITRO……………………………………………………...70 A. Abstract…………………………………………………………………………………..71 vii B. Introduction………………………………………………………………………………73 C. Materials and methods…………………………………………………………………...75 1. Cerebellar granule cell isolation…………………………………………………75 2. MeHg exposure…………………………………………………………………..75 3. Calcein-AM and ethidium homodimer-1 cytotoxicity assay……………………75 4. Automated cell quantification of CGCs with ImageJ……………………………76 5. Statistical analysis………………………………………………………………..77 D. Results……………………………………………………………………………………78 1. Effect of acute exposure to MeHg on the viability of WT, +/tg, and tg/tg CGCs + cultured in 25 mM K ……………………………………………………………78 E. Discussion………………………………………………………………………………..80 CHAPTER FOUR: EFFECT OF METHYLMERCURY ON CEREBELLAR GRANULE CELL VIABILITY IN CEREBELLAR ORGANOTYPIC SLICES…………………………………...87 A. Abstract…………………………………………………………………………………..88 B. Introduction………………………………………………………………………………89 C. Materials and methods…………………………………………………………………...91 1. Organotypic slices………………………………………………………………..91 2. MeHg exposure of organotypic slices……………………………………………92 3. ω-conotoxin GVIA treatment……………………………………………………92 4. Calcein-AM and ethidium homodimer-1 cytotoxicity assay…………………….92 5. Confocal microscopy and image analysis………………………………………..93 6. Statistical analysis………………………………………………………………..93 D. Results……………………………………………………………………………………95 1. CGC viability in WT, +/tg, and tg/tg organotypic slices after 24-hr exposure to MeHg…………………………………………………………………………….95 2. CGC viability in WT, +/tg, and tg/tg organotypic slices treated with ω-conotoxin GVIA prior to 24-hr exposure to MeHg…………………………………………95 E. Discussion………………………………………………………………………………..97 CHAPTER FIVE: SUMMARY AND DISCUSSION…………………………………………109 APPENDIX: EFFECT OF METHYLMERCURY ON LETHARGIC CEREBELLAR GRANULE CELL VIABILITY IN VITRO ……………………………………………………119 A. Abstract…………………………………………………………………………………120 B. Introduction……………………………………………………………………………..122 C. Materials and methods………………………………………………………………….123 1. Cerebellar granule cell isolation………………………………………………..123 2. MeHg exposure…………………………………………………………………123 3. Calcein-AM and ethidium homodimer-1 cytotoxicity assay…………………...123 4. Automated quantification of CGCs with ImageJ……………………………….123 5. Statistical analysis………………………………………………………………123 D. Results…………………………………………………………………………………..124 1. Effect of acute exposure to MeHg on the viability of WT, +/lh, and lh/lh CGCs + cultured in 25 mM K …………………………………………………………..124 viii E. Discussion………………………………………………………………………………125 BIBLIOGRAPHY………………………………………………………………………………131 ix LIST OF TABLES Table 1: Characteristics of voltage-gated Ca 2+ channels………………………………………..32 x LIST OF FIGURES Figure 1.1: Schematic representation of voltage-gated Ca 2+ channel quaternary structure…….30 Figure 1.2: The layers and neuronal circuitry of the adult mouse cerebellar cortex…………….34 Figure 1.3: Cerebellar granule cell migration in the mouse cerebellum………………………....36 Figure 1.4: Representative tracings of fura-2 signals during acute MeHg treatment: F340, F380, and ratio (F340/F380)……………………………………………………………………………..37 2+ Figure 1.5: Schematic drawing of MeHg-induced dysregulation of [Ca ]i homeostasis.……...39 Figure 1.6: Location of the tottering mutation in the α1A subunit of the P/Q-type voltage-gated Ca 2+ channel……………………………………………………………………………………..40 Figure 2.1: Voltage-gated Ca 2+ channel subtype composition in tottering cerebellar granule cells + cultured in 25 mM K ……………………………………………………………………............61 2+ Figure 2.2: Mean time-to-onset (min) ± SEM of first and second phase of MeHg-induced [Ca ]i + increase in tottering cerebellar granule cells cultured in 25 mM K , at DIV 4, 6, 8, and 10…………………………………………………………………………………………………65 2+ Figure 2.3: Mean duration of phase 1 (min.) ± SEM of MeHg-induced [Ca ]i increase in + tottering cerebellar granule cells cultured in 25 mM K ………………………………………...69 + Figure 3.1: Tottering cerebellar granule cell (cultured in 25 mM K ) viability after acute exposure to MeHg….....................................................................................................................83 + Figure 3.2: Tottering cerebellar granule cell (cultured in 5 mM K ) viability after acute exposure to MeHg………………………………………………………………………………………….86 Figure 4.1: Representative images of sagittal cerebellar organotypic slices stained with calceinAM and Ethd-1 after 24-hour treatment with MeHg...................................................................101 Figure 4.2: Cerebellar granule cell viability in cerebellar organotypic slices exposed in vitro to MeHg………...............................................................................................................................106 Figure 4.3: Cerebellar granule cell viability in cerebellar organotypic slices exposed in vitro to MeHg after pretreatment with 500 nM ω-conotoxin GVIA……………………………………108 xi Figure 5.1: MeHg-induced intracellular cell death pathways………………………………….117 Figure A.1: Location of the lethargic mutation in the β4 subunit of P/Q- and N-type VGCCs………………………………………………………………………………………….128 + Figure A.2: Lethargic cerebellar granule cell (cultured in 25 mM K ) viability after acute exposure to MeHg………………………………………………………………………………129 xii LIST OF ABBREVIATIONS +/tg heterozygotes for tottering mutation AIF apoptosis inducing factor ALS amyotrophic lateral sclerosis Apaf-1 apoptotic protease activating factor 1 Ara-C cytosine arabinoside BC basket cell 2+ [Ca ]e 2+ extracellular calcium concentration [Ca ]i intracellular calcium concentration CaM calmodulin CBD Ca -binding domain CGC cerebellar granule cell Cl - 2+ chloride DIV day in vitro DMEM Dulbecco’s modified eagle medium DMM dimethylmercury EGL external germinal layer EPSC excitatory post-synaptic current EPP end-plate potential EPSP excitatory post-synaptic potential FBS fetal bovine serum xiii FITC fluorescein isothiocyanate Fura-2AM fura-2 acetoxymethyl ester GABA γ-aminobutyric acid HEK 293 human embryonic kidney 293 cells HEPES N-[2-hydroxyethyl] piperazine-N’-[2-ethanesulfonic acid] Hg mercury IGL internal granular layer IO inferior olive nucleus IP3 inositol 1,4,5-trisphosphate IPSC inhibitory postsynaptic current HVA high voltage-activated lh lethargic LTD long-term depression LTP long-term potentiation LVA low voltage-activated MeHg methylmercury MEPP miniature end-plate potential min minutes ML molecular layer MMP mitochondrial membrane potential MTP mitochondrial transition pore NMDA N-methyl-D aspartic acid xiv NMJ neuromuscular junction Pb lead PC12 pheochromocytoma cell PCL Purkinje cell layer PLC phospholipase C PKC protein kinase C PMT photomultiplier tube PND postnatal day ROS reactive oxygen species RRP readily-releasable pool SC stellate cells S.E.M standard error of the mean SER smooth endoplasmic reticulum SOD1 superoxide dismutase 1 tg tottering VGCC voltage-gated Ca WT wild type 2+ channel xv CHAPTER ONE INTRODUCTION 1 A. Methylmercury 2+ The mercuric (Hg ) form of inorganic mercury is most frequently found in the environment bound to organic groups such as methyl, ethyl, or phenyl groups (Bridges and Zalups, 2010). Methylmercury (MeHg), an organic and highly toxic form of mercury, is a potent environmental neurotoxicant that has received much attention because it has caused many deaths + over decades. In its ionic form, CH3Hg , MeHg has high affinity to thiol groups present in a wide variety of proteins. MeHg taken in the diet is absorbed by the gastrointestinal tract and distributed to other tissues via the bloodstream. It forms a complex with cysteine present in the plasma, which allows it to cross the blood-brain and placental barriers by mimicking methionine (Clarkson, 1995). The MeHg that is not absorbed by tissues is secreted in bile and subsequently eliminated through feces. The entire process, from absorption to excretion, can take approximately 3 days in humans (Clarkson et al., 2007). Large-scale cases of chronic and acute exposure to MeHg have been reported and extensively studied. One of the first major cases of chronic poisoning with MeHg occurred during the 1950-60s in the area of Minamata Bay, Japan, where a chemical plant discharged toxic waste over a period of years into the ocean bay (Clarkson, 1995; Harada, 1995; Tsuda et al., 2009). Children and adults showed neurological signs similar to those of brain inflammation, which brought about an investigation for fear of an epidemic. Affected individuals experienced sensory dysfunction (i.e., constriction of peripheral vision, hearing loss, and impaired speech), followed by more severe signs including tremors and ataxia (Ceccatelli et al., 2010); this syndrome became known as Minamata Disease. In cases of severe poisoning, these signs were followed by incapacitation and death (Ceccatelli et al., 2010). A group of scientists appointed by 2 the government to find the cause of the outbreak found a high incidence of the disease in fishing families, specifically those who fished in the Minamata Bay area (Tsuda et al., 2009). The contaminant was not obvious, but many of the signs were similar to those caused by heavy metal poisoning. MeHg was confirmed as the etiologic agent when it was found in waste of the local chemical factory, and fish and shellfish from the immediate bay area contained high levels of mercury (Tsuda et al., 2009). In addition to the reported cases of affected children and adults, there was a high incidence of congenital cerebral palsy among newborns in the Minamata Bay area (Harada, 1995). Both mothers and congenital patients were found to have high levels of mercury in their hair samples; they ranged from 1.8-191 ppm in the mothers and 5.2-110 ppm in affected children. Interestingly, the mothers of the children with congenital disorder only presented mild sensory disturbances, whereas their affected children showed mental retardation, cerebellar-based ataxia, and dysarthria, among other signs (Harada, 1995). Since the first report of Minamata disease, more than 2,200 affected individuals have been officially identified in that area. In 1969 it was discovered that aquatic microorganisms have the ability to methylate mercury pollutants available in the environment (Jensen and Jernelov, 1969). This is currently a health concern in regions where mercury is released into the atmosphere due to anthropogenic sources including gold mining, combustion of fossil fuel, and iron and steel production, among others. Inorganic mercury deposited in water bodies is converted to MeHg and absorbed by zooplankton, which in turn feeds the fish population (Clarkson, 1995; Sanfeliu et al., 2003). Since the amount of MeHg an organism consumes is more than it can eliminate over a period of time, MeHg bioaccumulates from fish-eating fish to humans (Clarkson, 1995). This leads to a 3 biomagnification of MeHg so that humans, at the top of the food chain, consume fish with higher levels of MeHg. In the Brazilian Amazon, the use of mercury in gold mining is an anthropogenic source of Hg contamination. Tons of mercury are released into the environment as vapor, which subsequently reaches the waterways and is converted to MeHg. Studies also demonstrated that the Amazonian soils naturally contain high amounts of mercury which can be deposited into water bodies by floods, soil erosion and deforestation (Passos and Mergler, 2008). Consequently, exposure to mercury is highly common in mining and fish-eating populations along the Amazon river. The riparian villages (immediately adjacent to the rivers and streams of the Amazon river) have the highest mean hair mercury levels at approximately 38 µg/g, compared to urban fish consumers with a mean of approximately 2 µg/g (Passos and Mergler, 2008). Interestingly, not all populations that consume MeHg-contaminated fish present with the neurological dysfunctions associated with mercury poisoning. The Seychelles Islands, located off the eastern coast of Africa, have been the subject of investigation for many years. Habitants of these islands are mainly dependent on fish for their dietary needs, but there seems to be no correlation between in utero exposure to MeHg (range of 0.5-26.7 µg/g of mercury in hair samples of mothers) and abnormal development of children (Clarkson and Strain, 2003; Nuttall, 2006). In contrast, cohort studies from habitants in the Faroe Islands, located north of the United Kingdom, did find a significant correlation between prenatal exposure of mercury (levels in hair samples from mothers averaging 4.08 ppm) and mental retardation, motor dysfunctions, and speech impairment in children (Sanfeliu et al., 2003). Many factors have been hypothesized as reasons of the differences between these two studies, including ethnic background, and type of fish being 4 consumed, among others. Studies are currently underway to identify potential nutrients in certain fish (or other sources) that may counteract the effects of MeHg. There have also been cases of acute MeHg poisoning resulting in similar sensory and motor disturbances as those reported in chronic exposures. MeHg was discovered to be a very effective fungicide in the early 20 th century and it was, therefore, used to treat seeds before planting; the growing plant does not accumulate the compound and was believed to be safe for consumption. During the winter of 1971-72, rural villages of Iraq were given cereal grains that were intended solely for planting. However, some of the MeHg-treated grain was used for baking bread which was eaten daily for a period of months (Clarkson and Strain, 2003). Because MeHg is odorless, tasteless, and symptoms of poisoning are not immediately evident, villagers continued to consume the MeHg-poisoned bread. Affected individuals started experiencing the same signs as those reported for Minamata disease (i.e. paresthesia, ataxia, dysarthria, etc.). Since toxicologists and epidemiologists had been on high alert after the Minamata incident, they immediately recognized the cause of the symptoms. Even though the etiologic agent was the same, the period of exposure was different: months in Iraq versus years in Japan. Analysis of mercury concentration in hair samples from affected Iraqi adults provided the first dose-response data: asymptomatic individuals corresponded to < 300 µg/g, mild symptoms (i.e., slight tremor, mild ataxia, and blurred vision) to 120-600 µg/g, moderate symptoms (i.e., partial paralysis, loss of peripheral vision, and difficulty hearing) to 200-800 µg/g, and severe symptoms (coma, paralysis, blindness, hearing loss, and inability to speak) to 400-1,600 µg/g (Nuttall, 2006). Similar to the Minamata poisoning, Iraqi children exposed to MeHg in utero failed to reach developmental milestones, and showed signs of mental retardation, impaired motor skills, and delayed onset of speech (Sanfeliu et al., 2003). 5 An accidental case of acute dimethylmercury (DMM) poisoning was reported in the United States in 1997. DMM, another organomercuric compound, is composed of two methyl groups that increases its ability to permeate the skin, and confers a higher level of toxicity than MeHg. A chemistry professor accidentally spilled a drop of concentrated DMM on her latex glove. Just five months later, she was hospitalized with the trademark symptoms of mercury poisoning: impaired speech, ataxia, and loss of peripheral vision (Sanfeliu et al., 2003; Nuttall, 2006). Three weeks after her hospitalization she went into a coma and died just 298 days after her exposure. Examination of her cerebellum showed the characteristic reduction in cortical thickness due to a robust loss of CGCs, Purkinje cells, and inhibitory basket neurons (Nierenberg et al., 1998). Her maximum hair mercury level was reported to be 1,100 µg/g (Nuttall, 2006). Given the neurological nature of the signs presented in humans chronically or acutely exposed to MeHg, scientists reckoned that this compound caused degeneration of cells in the nervous system. Indeed, in 1954 pathologic details of a human subject that died as a result of occupational MeHg exposure were published (Hunter and Russell, 1954). No histological changes were discovered in the brainstem and spinal cord, so scientists suspected that the ataxia must be the consequence of detrimental changes in the cerebellar cortex. Upon microscopic examination of the cerebellum, a massive loss of cerebellar granule cells (CGCs) was observed. On the other hand, Purkinje cells, which are in very close proximity to CGCs, were spared (Hunter and Russell, 1954). These observations were later confirmed when a group of researchers studied changes in the cerebella of rats exposed to MeHg (Syversen et al., 1981). Young (30 days old) and adult rats were given a single injection of 10 mg/kg of MeHg-chloride and examined for changes in the cerebella. Swelling was observed in the majority of CGCs from young rats after one day of exposure. Interestingly, the swelling was followed by shrinkage, 6 three and seven days after MeHg treatment (Syversen et al., 1981). The main difference between the effects of MeHg in young versus adult rats was that changes in CGCs of young rats seemed to be reversible to a certain extent. This suggests that acute exposure of low MeHg doses in young individuals could be reversible perhaps due to a higher tolerance of young neurons to chemical insults. Aside from irregular branching of the endoplasmic reticulum, no significant difference in Purkinje cell morphology was detected in young or adult MeHg-treated rats (Syversen et al., 1981). As mentioned earlier, MeHg has high affinity for thiol groups found ubiquitously in proteins present in many different types of cells. Therefore, the finding that Purkinje cells are more resistant to MeHg-induced cell death came as a surprise. To date, research is ongoing to find the factors that render CGCs more susceptible to MeHg compared to Purkinje cells. It is 2+ hypothesized that voltage-gated Ca channels (VGCCs) have a role in the mechanism of MeHg-induced cytotoxicity because of the similarities between Hg 2+ 2+ and Ca and their location at the surface of neurons. A very important conclusion from the human and animal model studies is that all modes (acute or chronic; dietary intake or skin contact, etc.) of MeHg exposure result in the same neurological dysfunction and characteristic delay of onset. Also, even though there are differences in the susceptibility between developmental stages (embryo, childhood, and adulthood), MeHg causes the same pattern of neurotoxicity. 7 2+ B. Voltage-Gated Ca Channels Excerpt from: Marrero-Rosado, B., Fox, S.M., Hannon, H.E., Atchison, W.D. (2013) Effects of mercury and lead on voltage-gated calcium channel function. Encyclopedia of Metalloproteins. Springer Reference (In press) Ca 2+ plays a crucial role in the regulation of gene expression and neurotransmitter release, synaptic plasticity, and growth cone elongation in neurons. VGCCs mediate influx of Ca 2+ into the cell as a result of a depolarization of the plasma membrane. Their specific role in a neuron is dictated by the properties that characterize them. In general, two classes of VGCCs exist: the high voltage-activated (HVA) and low voltage-activated (LVA) channels; they are grouped according to the degree of depolarization needed for their activation. While the pharmacological properties of the channel are imparted by the α1 pore-forming subunit, the β, γ, and α2δ subunits modulate the voltage-dependence of activation/inactivation and kinetics of the protein (Benarroch, 2010) (Fig. 1.1). The HVA category, composed of channels that require a high degree of depolarization, is subdivided into L-, N-, P/Q-, and R-type (CaV1.x, CaV2.x) according to their pharmacological and electrophysiological characteristics (Table 1). The LVA category is only composed of the T-type channels (CaV3.x). Due to their low-threshold of activation, T-type VGCCs are responsible for providing neurons Ca 2+ spikes (transient low- threshold currents) at resting membrane potentials (Zamponi, 2005). Many roles for these Ca 2+ 2+ spikes have been suggested, including pacemaking activity, modulation of Ca -dependent ion channels, and neuronal development (Zamponi, 2005). 8 The process of discriminating among extracellular cations is due in part to a selectivity filter present at the mouth of the VGCC pore. A region of four negatively-charged glutamate residues projects into the lumen outside the pore. The ability of these residues to bind Ca 2+ is believed to be crucial for ion discrimination (Atchison, 2003; Zamponi, 2005). This selectivity is so efficient that, even though extracellular sodium ions are found in greater concentrations, 2+ + VGCCs are still more permeable to Ca . However, cations of similar atomic radius (i.e. Na , + + + 2+ 2+ Li , K , Cs , Sr , and Ba ) also can pass through VGCCs when extracellular Ca 2+ is replaced (Zamponi, 2005). These cations show a higher single-channel conductance due to their inability to bind efficiently to the binding sites present in the channel’s vestibule, a characteristic that has been advantageous in studying VGCC function. Toxic metals also interact with and disrupt the function of VGCCs. In particular, MeHg falls into the category of a non-physiological, environmentally-relevant organic metal whose neurotoxic effects on VGCCs have been well documented (Atchison, 2003). 2+ B.1. Effects of methylmercury on voltage-gated Ca channel function 2+ The first experiments to study the effects of inorganic mercuric ions (Hg ) on overall neuronal function were performed in neuromuscular junction (NMJ) preparations because of the extensive characterization of the process of acetylcholine (ACh) release from motor nerve terminals. Moreover, there was a high incidence of myasthenia gravis-like symptoms in the Iraqi MeHg exposure event (Rustam et al., 1975). Treatment of frog sciatic nerve-sartorius muscle preparations with low concentrations (< 1 µM) of HgCl2 showed an initial 720% increase in the 9 amplitude of end-plate potentials (EPPs), a measure of evoked neurotransmitter release (Manalis and Cooper, 1975). However, this increase gradually decreased to a complete block of ACh release within minutes. Similar experiments were performed in rat phrenic nerve-hemidiaphragm preparations continuously exposed to 4, 20, and 100 µM MeHg in the bath (Haines and Dietrichs, 2012); these concentrations were within range of those found in plasma of individuals poisoned with MeHg. Two main findings came out of these experiments. First, for concentrations of 20 and 100 µM MeHg, there was an initial increase in the frequency of miniature end-plate potentials (MEPPs; spontaneous ACh release) followed by a decrease to block. This biphasic event occurred faster at the 100 µM MeHg concentration, suggesting a time- and concentrationdependent effect. Second, the amplitude of the end-plate potentials (EPPs; evoked ACh release) was significantly reduced after 20 and 100 µM MeHg treatment, but not 4 µM. Application of ACh (without stimulation of the phrenic nerve) before and after MeHg treatment did not show a decrease in the amplitude of end-plate depolarizations, suggesting that MeHg may not act on the postsynaptic terminal by affecting the sensitivity of ACh receptors (Haines and Dietrichs, 2012). These experiments also studied the possibility that the increase in MEPP frequency was due to a depolarization of the nerve terminal caused by MeHg. Pretreatment of hemidiaphragms with tetrodotoxin, a neuronal voltage-gated Na + channel antagonist, did not prevent the MeHg- induced increase in MEPP frequency. This suggested that MeHg does not open neuronal voltage+ gated Na channels to cause an increase in spontaneous ACh release due to a presynaptic depolarization. It was also evident that the block of neurotransmitter release was irreversible when the hemidiaphragms were rinsed with MeHg-free solutions. Based on the increase in 10 MEPP frequency and decrease in EPP amplitude, the authors concluded that block of evoked neurotransmitter release involved a decrease in the readily releasable pool (RRP) of neurotransmitter vesicles. The nerve terminal contains two pools: the RRP, the first to be released as a result of Ca 2+ entry, and the reserve pool, which is transported to the sites of release as the RRP is depleted. Since the process of synchronous vesicle release associated with the EPP is dependent on entry of Ca 2+ to the nerve terminal, the decrease in evoked neurotransmitter release suggested that MeHg interfered with normal function of VGCCs. The effects of MeHg on VGCC function were subsequently studied in more detail. Because impaling NMJ nerve terminals with microelectrodes to measure Ca 2+ influx would + compromise their viability, synaptosomes were used instead to study the effects of MeHg on K induced 45 Ca 2+ uptake. Synaptosomes are pinched off nerve terminals from brain homogenates that retain ion channels and organelles involved in neurotransmitter release. They retain functional properties of nerve terminals including the ability to (1) release neurotransmitter, (2) generate relevant and appropriate ion influxes in response to depolarization, and (3) reuptake neurotransmitter or their precursors (Yuan et al., 2005). Upon treatment with 200 and 500 µM MeHg for very short periods of time (1 or 10 sec), a significant reduction (50% maximal block) of 45 Ca 2+ uptake was observed in synaptosomes isolated from adult rat forebrains (Atchison et al., 1986). It was also found that increasing concentrations of extracellular Ca prevented the reduction in 45 2+ Ca 2+ only partially uptake caused by MeHg. This suggested a non-competitive + blocking action on VGCCs. When synaptosomes are exposed to a depolarizing K solution, 11 Ca 2+ uptake is characterized by a slow and fast component. The fast component is believed to + involve the opening of fast-activating (1-2 sec after K depolarization) VGCCs involved in + neurotransmitter release; the slow component may be due to the activity of a Na /Ca exchanger that remains active 20-90 sec after depolarization and/or a slowly-inactivating Ca channel (Yuan et al., 2005). The authors found that 45 2+ Ca 2+ 2+ uptake by the fast component, was reduced faster than the slow component, and it occurred at lower concentrations of MeHg. However, there was a more pronounced decrease in 45 Ca2+ uptake by the slow component (Atchison et al., 1986). Experiments using electrophysiological techniques on a rat adrenal medulla pheochromocytoma cell line, PC12, were performed to study the effect of MeHg on Ba 2+ currents through L-type and N-type VGCCs (Shafer and Atchison, 1991). PC12 cells, when cultured in neuronal growth factor (NGF), endogenously express N- and L-type VGCCs. It is common in electrophysiological experiments that measure ion conductance through VGCCs to 2+ use Ba 2+ 2+ as charge carrier, in lieu of Ca . Ba does not bind sites present in the channel’s 2+ vestibule and is unable to induce Ca -dependent inactivation of VGCCs (Zamponi, 2005). This property allows VGCCs to have a higher single-channel conductance that can be easier to visualize (Zamponi, 2005). Similar to previous studies in rat brain synaptosomes, MeHg caused a concentration-dependent reduction in overall Ba 2+ influx. Of pivotal importance was the additional finding that increasing the frequency of VGCC stimulation reduced the time it took for MeHg to block 70% of the overall Ba 2+ current (Shafer and Atchison, 1991). This suggested that 12 VGCC block by MeHg can be facilitated by the frequent opening of these channels. However, depolarization of PC12 cells before MeHg exposure does not have an effect on the block of 45 Ca 2+ uptake, suggesting that it is not dependent on channel configuration (i.e. open, resting, or inactivated) (Shafer et al., 1990). Altogether, these findings strengthened the idea that VGCCs play a role in the MeHg-induced block of Ca 2+ influx and, therefore, evoked neurotransmitter release. Since chronic and acute MeHg poisonings cause cerebellar ataxia, the focus shifted to studies of the effects of MeHg on VGCC of CGCs isolated from animal models, better known as primary CGCs. CGCs express all subtypes of VGCC. Electrophysiological experiments have dissected the contribution of each subtype of VGCC to the overall Ca 2+ influx in mature CGCs. The P/Q-type VGCCs account for the majority (46%) of the overall Ca 2+ current; they are followed by L-type and N-type VGCCs with a contribution of 15 and 20%, respectively (Randall and Tsien, 1995). Submicromolar and micromolar concentrations (0.25-1 µM) of MeHg produced a concentration-dependent decrease in overall Ba 2+ conductance that was not reversible. In accordance with previous studies, increasing the stimulation frequency from 0.1 Hz 2+ to 0.2 Hz increased the degree of block of Ba current caused by 0.25 and 0.5 µM MeHg (Sirois and Atchison, 2000). It has been hypothesized that MeHg uses VGCCs as a pathway to reach the 2+ intracellular space; the enhancement of MeHg-induced Ba current block by an increase in stimulation frequency could, therefore, be attributed to an increase in access to these pathways. The specificity toward a particular subtype of VGCC was also examined in these studies. Rat 13 CGCs were treated with ω-conotoxin GVIA (N-type or Cav2.2 antagonist), ω-agatoxin IVA (Ptype or Cav2.1 antagonist at low concentrations), ω-conotoxin MVIIC (N- and P/Q-type antagonist), nimodipine (L-type or Cav1.2 antagonist) or calcicludine (L-, N-, and P-type antagonist) prior to MeHg exposure (Sirois and Atchison, 2000). None of the antagonists were able to prevent MeHg from further reducing overall Ba 2+ current, suggesting that all VGCC subtypes are equally blocked by MeHg. In the in vitro experiments discussed so far, the authors used concentrations of MeHg in the micromolar ranges, which are still below those found in plasma of poisoned individuals. However, the effects of long-term exposure of MeHg concentrations in the nanomolar range on Ca 2+ currents have also been examined (Fox et al., 1987). Differentiated/primed (NGF-treated) and differentiating (MeHg present during NGF treatment) PC12 cells were exposed to 30 nM MeHg for 24 hours; overall Ca 2+ + and Na currents were compared, along with cell morphology. The main purposes of these experiments were to (1) study in vitro the effect of MeHg on changes in morphology of developing neurons, and (2) determine if there is a link to VGCC function. Both types of PC12 cells showed a significant decrease in overall Ca 2+ current after 30 nM (primed cells) and 10 nM (differentiating cells) MeHg exposures. However, only PC12 cells + given the MeHg while differentiating showed a decrease in overall Na current as well. Cell morphology was only disrupted in primed PC12 cells, and was characterized by the shrinkage of axons. Even though chronic exposure of low concentrations of MeHg did not prevent neurite 14 growth in developing PC12, it is noteworthy that these concentrations were still able to decrease overall Ca 2+ current through N-type and L-type VGCCs. To study the effect of MeHg on VGCCs in isolation, each subtype can be heterologously expressed in human embryonic kidney 293 cells (HEK 293), which do not endogenously express HVA VGCCs. HEK 293 cells transfected with the human α1C, β3a, and α2δ (composing a recombinant L-type VGCC) subunits were exposed to 0.125-5 µM MeHg. MeHg caused the expected concentration- and time-dependent decrease in overall Ba 2+ current. The L-type VGCC blocker, nimodipine, was able to diminish completely the overall Ba 2+ current in the same magnitude reported for other cell types. However, the maximum concentration of MeHg being tested was not able to completely block overall Ba 2+ current; 20-25% of current remained (Johnson et al., 2011). Similar experiments were also performed on HEK 293 cells transiently expressing the α1B (N-type) or α1E (R-type or Cav2.3) subunits, along with β3a and α2δ. In contrast to recombinant L-type VGCCs, MeHg was able to completely block overall Ba 2+ current equally through both N-type and R-type recombinant VGCCs (Cicale et al., 2002). Thus, even though Ca 2+ current through all HVA VGCC subtypes is significantly reduced by MeHg, some subtypes show differences in sensitivity to MeHg. Finally, if MeHg-induced neurotoxicity is dependent on block of Ca 2+ current through VGCCs, then in vivo treatment of animal models with VGCC antagonists prior to MeHg exposure should be protective. Sakamoto, et al. examined this hypothesis by orally administering L-type VGCC blockers (flunarizine, verapamil, nicardipine, and nifedipine) 30 min before MeHg 15 treatment (5 mg/kg/day) for 12 days (Sakamoto et al., 1996). The authors assessed the toxicity of MeHg based on physiological and behavioral markers including change in body weight, hind leg paralysis, crossed and/or retracted hind legs when held by tail, and piloerection. All antagonists were able to reduce the incidence of MeHg-induced neurological dysfunction, reduce weight loss, and prolong survival. In the case of flunarizine, the protective effect was seen without it affecting MeHg absorption by the tissues, which indicates that it conferred its protection due to its pharmacodynamic properties. These protective effects were also confirmed in vitro (Sakamoto et al., 1996). CGCs treated in vitro with flunarizine prior to MeHg exposure showed a significant decrease in cell mortality. Altogether, the data strongly argue that VGCCs are involved in the mechanism of MeHg-induced CGC cytotoxicity. C. The cerebellum The cerebellum is the region of the brain responsible for motor coordination and spatial memory (Haines and Dietrichs, 2012; Hashimoto and Hibi, 2012). Structurally, the cerebellum of higher vertebrates consists of 10 lobules internally composed of three layers where different types of cells reside: the internal granular layer (IGL), Purkinje cell layer (PCL), and molecular layer (ML) (Fig. 1.2.A). At the core of the cerebellum is white matter, which contains the axons of afferent and efferent neurons. The IGL is packed with approximately billions of CGCs (6-9 µm in diameter) which receive one of the two main afferent inputs by means of the mossy fibers. The mossy fibers carry information into the cerebellum from peripheral nerves, the spinal cord, and the brain stem (Ito, 2006). They make excitatory connections in the IGL with dendrites of CGCs to form cerebellar glomeruli (labeled as MF rosettes in Fig. 1.2.A); this structure also contains the synaptic connections between CGCs and Golgi cells (11-24 µm in diameter), which 16 secrete the inhibitory neurotransmitter γ-aminobutyric acid (GABA) (Haines and Dietrichs, 2012; Hashimoto and Hibi, 2012). Whereas one CGC can receive information from 3-4 mossy fibers, one mossy fiber can make excitatory synapses with 400-600 CGCs or even more if it innervates more than one lobule (Ito, 2006). CGCs extend a long axon that splits into two parallel fibers in the outermost layer of the cerebellum, the ML, to make excitatory synapses with the dendrites of Purkinje cells (30-70 µm in diameter). It is important to note that Purkinje cells, whose cell bodies are located in the PCL, are the only neurons that convey information away from the cerebellum. Therefore, CGCs are important intercortical neurons that relay information from the efferent mossy fibers to the Purkinje cells. It has been reported that one parallel fiber can make excitatory synapses with at least 300 Purkinje cells; the large dendritic trees of Purkinje cells, conversely, can interact with as many as 180,000 parallel fibers in humans or 60,000-175,000 in rats. However, some “silent synapses” do not elicit an EPSP in the postsynaptic Purkinje cells; only approximately 3% of all parallel fiber input to a Purkinje cell is effective (Ito, 2006). Purkinje cells also receive excitatory information from the other main afferent input: the climbing fibers from the inferior olive (IO) nucleus. These connections are made in the proximal branchlets of the dendritic tree, as opposed to the parallel fiber-Purkinje cell connections which are located at the distal branchlets (Haines and Dietrichs, 2012) (Fig. 1.2.B). The key to smooth motor coordination lies in part on the ability of Purkinje cells to integrate both excitatory signals, from CGCs and IO neurons, into one efferent signal. Stimulation of Purkinje cells by climbing fibers has been shown to produce long-term depression (LTD) of the parallel fiber synapse. In this scenario, LTD is a reduction of the amplitude of response of Purkinje cells to stimulation from parallel fibers due to a reduction in sensitivity to glutamate (excitatory neurotransmitter) 17 (Ito, 2006). It is long-lasting because endocytosis of glutamate receptors at the synapse occurs. However, this form of synaptic plasticity occurs bidirectionally. Long-term potentiation (LTP), the enhancement of signal transduction in a synapse as a result of increased input from parallel fibers, can remove LTD imposed by climbing fiber input. Purkinje cells also receive inhibitory input from two interneurons present in the ML: the basket cells (BC) and stellate cells (SC). These cells receive excitatory input from the CGCs. When stimulation of Purkinje cells occurs via the parallel fiber-Purkinje cell synapse, the result is the inhibition of cerebellar nuclear and/or vestibular nuclei neurons that are postsynaptic to Purkinje cells. On the other hand, if CGCs stimulate basket/stellate cells, they cause the inhibition of Purkinje cells, which releases the inhibitory effect on the cerebellar nuclear and/or vestibular nuclei neurons (Haines and Dietrichs, 2012). What happens first is determined by the type of afferent signal that arrives at the cerebellar cortex. C.1. Development of cerebellar granule cells The development of CGCs, the most abundant neurons in the central nervous system, entails their proliferation, migration and maturation, all of which occur during the first three postnatal weeks in the mouse (Nakanishi and Okazawa, 2006; Hashimoto and Hibi, 2012; Xu et al., 2013). The developing cerebellum contains an external germinal layer (EGL), aside from the three previously discussed cerebellar layers present in adult cerebellum (Fig. 1.3). Proliferation of precursor CGCs takes place in the EGL up to PND 9. At this stage, CGCs are in a relatively depolarized state with a resting membrane potential of -24 mV (Rossi et al., 1998). Extensive research shows that the highly depolarized state allows the increased entry of Ca 2+ through activated N-methyl-D-aspartic acid (NMDA) receptor ion channels and VGCCs, particularly N18 2+ type (Takayama and Inoue, 2006; Okazawa et al., 2009). Consequently, Ca -dependent 2+ enzymes, such as the Ca /calmodulin-dependent protein kinase (CaMK) and calcineurin phosphatase (CaN), are activated and play a role in the regulation of developmental gene expression (Barbado et al., 2009; Okazawa et al., 2009). From PND 10-15, CGCs migrate (with the aid of a scaffold provided by Bergmann glia cells) towards the innermost layer of the cerebellum, the IGL. During this migration period, CGCs undergo a decrease in resting membrane potential to a final -70 mV (Okazawa et al., 2009). This developmental change in resting membrane potential can be mimicked in vitro by changing the concentration of KCl in the growth media from 25 mM to 5 mM (Sato et al., 2005; Nakanishi and Okazawa, 2006; Okazawa et al., 2009). Interestingly, data from mouse cerebellar slices dissected at PND 6 and cultured under depolarizing conditions show that migration of CGCs still occurs, but their dendrites fail to form the “claw-like” morphology characteristic of mature CGCs (Okazawa et al., 2009). This suggests that, while some genes are temporally expressed regardless of resting membrane potential, other proteins are only present in mature CGCs because of the change in resting membrane potential. One such protein is the α6 subunit of the inhibitory GABAA receptor, which has been shown to only be expressed in vitro when CGCs are cultured under low + K conditions (Mellor et al., 1998), and in vivo after CGCs establish functional synapses in the IGL (Okazawa et al., 2009). Since CGCs do not form functional inhibitory synapses until they are mature and located in the IGL, it makes sense for the expression of the α6 subunit of GABAA receptors to be regulated by the shift in membrane potential. 19 Ca 2+ influx through N-type VGCCs is crucial for the migration of CGCs during development (Zhang et al., 1993). The observed increase in the incidence of congenital cerebellar ataxia in populations exposed to MeHg could then be explained by an impaired migration of CGCs due to the block of Ca 2+ current. Scientists have been able to replicate the congenital ataxia in mouse and rat models treated with MeHg during their gestational period. Acute oral doses of 10 mg/kg/day and 5 mg/kg/day from days 6 to 17 of gestation killed the dams before they gave birth and caused stillbirths, respectively. Dams given a 1 mg/kg/day dose for the same period gave birth to pups that did not show any signs of neurological dysfunction for up to PND 28. Upon microscopic examination of the cerebellum at PND 2, the loss of CGCs on the EGL was evident. These cells also showed reduced staining of a variety of mitochondrial enzymes (Brustovetsky et al., 2002). Chronic fetal and lactational exposure to MeHg in mice also showed the characteristic changes in the cerebellum. Mouse dams were treated with 4 mg/kg/day from gestational day 1 to PND 30. The authors observed shrinkage of the ML and increased cell death in the IGL, without changes in the density of the Purkinje cells. They also found an increase in the number of cells outside of the IGL, and suspected they were granule cells that were “stranded” during the process of migration (Markowski et al., 1998). Altogether, these data confirm the increased susceptibility of embryos to low levels of MeHg exposure, and explain the etiology of neurological dysfunction in children exposed to MeHg in utero. C.2. Effects of methylmercury on cerebellar granule cells: As previously mentioned, chronic and acute MeHg exposure causes massive loss of CGCs in humans and animal models. To understand the factors responsible for their heightened 20 susceptibility to MeHg, research has been focused on elucidating the events that precede their 2+ death. Single-cell microfluorimetry assays have used the ratiometric fluorescent Ca -binding 2+ compound fura-2 to examine changes in [Ca ]i in CGCs during acute MeHg treatment. The ester form of fura-2, fura-2 acetoxymethyl ester (fura-2AM), is cell-membrane permeable. Once inside the cell, endogenous esterases remove the acetoxymethyl ester group from fura-2AM, converting it into fura-2 which is non-permeable. Hence, fura-2 becomes trapped inside the cell. Inside the cell, fura-2 exists in two forms: a free anionic form with an excitation λmax of 380 nm, 2+ and a Ca -bound with an excitation λmax of 340 nm; they both show a Ca 2+ concentration2+ dependent emission at 530 nm λmax (Grynkiewicz et al., 1985). Upon binding of Ca , the 2+ amount of Ca -bound fura-2 increases as the free anionic form decreases; this causes a change in the ratio of fluorescence at 340 nm excitation λmax over fluorescence at 380 nm excitation λmax (F340/F380; Fig. 1.4). Factors including cell size and diffusion rate can play a role in the 2+ amount of fura-2AM that each cell uptakes. Therefore, we analyze the ratio of the Ca -bound fura-2 (F340) signal over the unbound fura-2 (F380) signal to essentially eliminate the variable of differential uptake of the compound between CGCs. Our lab has demonstrated that rat primary 2+ CGCs undergo a time- and concentration-dependent biphasic increase in intracellular Ca 2+ concentration ([Ca ]i) following exposure to MeHg (Marty and Atchison, 1997) (Fig. 1.4). Removing extracellular Ca 2+ during acute MeHg treatment in vitro eliminates the onset of the 21 second phase, but the first phase appears to be independent of extracellular Ca 2+ (Marty and Atchison, 1997; Limke and Atchison, 2002). In other words, phase 1 occurs regardless of an 2+ onset of phase 2. These data suggest that while the first phase of [Ca ]i increase in CGCs is due to the release of Ca extracellular Ca 2+ 2+ from intracellular stores, the second phase corresponds to a robust entry of (Fig. 1.4). 2+ In neurons, there are two organelles believed to be involved in [Ca ]i buffering: the mitochondrion and the smooth endoplasmic reticulum (SER). When the SER is depleted of its 2+ 2+ Ca , there is a delay in the onset of the first phase of MeHg-induced [Ca ]i increase (Limke et 2+ al., 2004). Since Ca from the SER can be released by activation of the ryanodine receptors or inositol-1,4,5-trisphosphate (IP3) receptors, experiments were performed to determine if both were involved. Interestingly, the first phase was only delayed when IP3 receptors were desensitized prior to MeHg treatment. This effect was also reproduced when M3 muscarinic ACh receptors (mAChR), located at the surface of the cell membrane, were desensitized and down regulated, by 24 hr treatment with bethanecol, prior to treatment with MeHg (Limke et al., 2+ 2004). It is believed that one of the events preceding the increase in [Ca ]i is that MeHg increases the production of IP3 (Sarafian, 1993), perhaps via activation of M3 muscarinic receptors, which results in release of Ca 2+ from the SER (Hare and Atchison, 1995) (Fig. 1.5). The role of mitochondria in these experiments could not be ruled out. Pretreatment of CGCs with cyclosporine A, an immunosuppressant that delays the opening of mitochondrial transition pore 22 2+ (MTP), also delays the onset of the first phase of [Ca ]i increase caused by 0.2 and 0.5 µM MeHg (Limke and Atchison, 2002). This suggests that mitochondria, through opening of MTP, 2+ are also involved in the first wave of [Ca ]i increase. Interestingly, prevention of MTP opening also delays the onset of the second phase, suggesting a possible link between both events. It is noteworthy that the effect of MeHg on mitochondria in vitro has also been observed in vivo. CGCs of mice that were given low to moderate (1.0 to 5.0 mg/kg) doses of MeHg over a period of 20 days showed increased levels of reactive oxygen species (ROS), reduced mitochondrial membrane potential (MMP) due to MTP opening, and increased levels of cytochrome c in the cytosol (Bellum et al., 2007). MeHg blocks Ca 2+ current through VGCCs within seconds (Atchison et al., 1986; Shafer and Atchison, 1989; Sirois and Atchison, 2000). Therefore, it has been hypothesized that these channels serve as a mode of entry for MeHg into the intracellular space. Consistent with this idea is the fact that pretreatment of CGCs with ω-conotoxin MVIIC (N- and P/Q-type VGCC blocker) 2+ delays the onset of both phases of MeHg-induced [Ca ]i increase. Nimodipine, an L-type VGCC antagonist, was only able to delay the onset of the second phase (Marty and Atchison, + 1997). In contrast, pretreatment with tetrodotoxin (voltage-gated Na channel antagonist) and various excitatory amino acid receptor antagonists did not produce a significant delay in the 2+ MeHg-induced [Ca ]i increase (Marty and Atchison, 1997). The MeHg-induced increase in 2+ [Ca ]i is sufficient to cause a delayed increase in CGC death in vitro. Moreover, our lab has 2+ demonstrated that faster time-to-onset of the biphasic [Ca ]i increase correlates with a higher 23 percentage of CGC death. When CGCs are pretreated with nimodipine or ω-conotoxin MVIIC, MeHg-induced cytotoxicity is significantly delayed for up to 24 hrs, albeit not prevented (Marty and Atchison, 1998). Taken together, these data strengthen the idea of a key role of VGCCs in the mechanism of MeHg-induced CGC cytotoxicity. Another process in CGCs that seems to be affected by MeHg is the GABA-mediated inhibitory transmission. GABA is the most important inhibitory neurotransmitter in the mammalian central nervous system (Obata et al., 1967). The GABAA ligand-gated ion channel is believed to be heteropentameric, composed of multiple combinations of seven subunits (α1-6, β1-3, γ1-3, δ, ε, π, and θ); only a few combinations are found naturally (Takayama, 2005). When - activated by GABA, GABAA receptors increase their permeability to Cl , resulting in an inhibitory post-synaptic current (IPSC) that reduces the amplitude of the excitatory post-synaptic postsynaptic current (EPSC) in mature CGCs. Therefore, the excitability of CGCs is regulated by GABA receptors. However, depending on the location of the GABAA receptor being activated, the result will be one of two types of inhibition. “Phasic” inhibition is seen as an IPSC of short duration that is mediated by GABAA receptors located post-synaptically to GABAergic neurons (Farrant and Nusser, 2005). This type of inhibition is involved in cell-to-cell signaling of normal synaptic inhibition. On the other hand, “tonic” inhibition involves GABAA receptors that are located extrasynaptically and has the main role of modulating the size and duration of an action potential (Farrant and Nusser, 2005). Our lab has demonstrated that MeHg blocks IPSCs through GABAA receptors in both CA1 hippocampal neurons and CGCs (Yuan and Atchison, 1997; 24 Yuan and Atchison, 2003). In rat cerebellar slices, MeHg initially increased and subsequently reduced the amplitude and frequency of spontaneous IPSCs in Purkinje cells and CGCs (Yuan and Atchison, 2003). The time to block spontaneous IPSCs in CGCs was much longer than in Purkinje cells, a key difference between both types of cells. It is noteworthy that in hippocampal CA1 neurons MeHg first increases excitatory neurotransmission before suppressing it. Pretreatment with bicuculline, a GABAA receptor antagonist, eliminated the characteristic increase in amplitude and frequency of excitatory postsynaptic potentials (EPSPs) caused by MeHg (Yuan and Atchison, 1997). This suggests that suppression of the GABAA-mediated neuronal inhibitory pathway might be responsible for the initial MeHg-induced increase in excitability. D. The Tottering Mouse The tottering (tg) mouse is the result of a natural point mutation in the Cacna1a gene, which codes for the pore-forming α1A subunit of the P/Q-type VGCCs (Fletcher et al., 1996) (Fig. 1.6). It is a non-lethal mutation, exhibiting an autosomal recessive mode of inheritance, which causes a proline to leucine amino acid change in the S5-S6 linker region of repeat domain II of α1A, in close proximity to the pore. (Fig. 1.6) (Fletcher et al., 1996; Sawada et al., 2000). Their phenotype is characterized by absence epilepsy, motor seizures, and ataxia. In tg Purkinje cells, this mutation causes a ~60% reduction of whole-cell current density compared to WT cells (Wakamori et al., 1998). However, since single-channel conductance and voltage-dependence of activation is not affected by the point mutation, we are able to rule out the possibility that the 25 amino acid change results in hindrance of ion conductance through the pore (Wakamori et al., 1998). It has also been reported that the missense mutation causes an 85% reduction of α 1A protein expression (Campbell and Hess, 1999; Leenders et al., 2002) in tg forebrain homogenates, even though mRNA levels are comparable to WT (Fletcher et al., 1996). To partially compensate for the lack of normal numbers of P/Q-type VGCCs, which normally play a main role in neurotransmitter release, Purkinje cells in the tg mouse increase the expression of Ltype channels (Campbell and Hess, 1999), while CGCs increase the expression and contribution of N-type channels (Zhou et al., 2003). Even though L-type channels show higher Ca 2+ conductance and slower inactivation rate than P/Q-type channels (Moran et al., 1991; Hammond, 2001), the fact that they activate at more positive membrane potentials than P/Q-type channels could play a role in the smaller overall Ca 2+ current density. As a result of this impaired Ca 2+ homeostasis, some Purkinje cell shrinkage is observed in the tg cerebellum (Wakamori et al., 1998). In the tg CGCs, the dysfunction of P/Q-type channels allows N-type VGCCs to take over the main role of neurotransmitter release (Leenders et al., 2002; Zhou et al., 2003). Even though N- and P/Q-type channels have similar degrees of Ca 2+ conductance, N-type channels inactivate at more hyperpolarizing membrane potentials than P/Q-type channels; whereas at -60 mV 50% of the N-type population is already inactivated, it takes a depolarization of -45 mV to produce the same effect in the P/Q-type population (Hammond, 2001). Nonetheless, extensive evidence from tg brain synaptosomes (Leenders et al., 2002), NMJs (Pardo et al., 2006), and parallel fiber-Purkinje cell synapses (Zhou et al., 2003) shows that compensation by N-type or other subtypes of VGCCs does not affect the release of neurotransmitter. 26 Parallel fiber-Purkinje cell synapses are normally modulated on the CGC extrasynaptic end by GABAB receptors. This type of GABA receptor activates G proteins that can reduce Ca 2+ influx through VGCCs. In addition to opening at more positive membrane potentials, N- type VGCCs are more susceptible to inhibitory modulation by GABAB receptors (Zhou et al., 2003). Consistent with this idea are data from experiments in which baclofen, a GABAB receptor agonist, was applied prior to stimulation of parallel fibers and recording of field EPSPs. The IC50 for baclofen in WT transverse cerebellar slices was 1.71 µM, whereas tg preparations showed an IC50 of 0.58 µM (Zhou et al., 2003). This suggests that because of the increase in the contribution of N-type VGCCs CGC parallel fibers might be more susceptible to GABAB receptor modulation. Since CGCs receive GABAergic input from inhibitory neurons, their ability to excite Purkinje cells might be negatively affected. Very little information on the effect of the tg mutation on Ca 2+ influx in CGCs is available. The reason for this is that Purkinje cell dysfunction has been identified as the main cause of the motor dysfunction (Campbell and Hess, 1999; Erickson et al., 2007), and research focused on the tg mouse as a model for studying human epilepsy. However, there are three main observations that suggest a compensation mechanism in the CGCs that is different from that reported in Purkinje cells: (1) neurotransmitter release at tg parallel fiber-Purkinje cell synapses is highly dependent on Ca 2+ influx through N-type VGCCs (Zhou et al., 2003); (2) these synapses show an enhancement in the modulation of presynaptic activity by GABAB receptor27 mediated inhibition (N-type VGCCs are more sensitive than P/Q-type VGCCs to GABA2+ mediated inhibition) (Zhou et al., 2003); and (3) tg CGCs in culture show a lower basal [Ca ]i compared to WT CGCs (Bawa and Abbott, 2008). Another crucial change observed in tg CGCs is a significant decrease in the number of GABAA receptors composed of the α6βxγ2 subunits (Kaja et al., 2007). The α6 subunit is normally only found in mature CGCs, and the α6βxγ2 GABAA receptor is specifically found at the synapses where CGCs receive inhibitory input {Essrich, 1998 #167}. As discussed previously, Purkinje cells receive input from two excitatory 2+ neurons that compete for synaptic strength. A possible lower overall influx of Ca , enhanced 2+ inhibition of the synapse by GABAB receptors, and lower basal [Ca ]i could favor LTD by climbing fiber input and impair neurotransmission in the parallel fiber-Purkinje cell synapse. E. Specific Aims Pretreatment of rat primary CGC cultures with VGCC antagonists has demonstrated that these cation channels play a crucial role in MeHg-induced cell death. With that in mind, the aim of the present research was to determine the effect of MeHg on CGCs with impaired function of the main VGCC involved in neurotransmitter release, the P/Q-type. We sought to test our hypothesis that a natural reduction in Ca 2+ influx caused by a mutation in this highly expressed VGCC would have a protective effect on CGCs in a similar manner as the pharmacological 28 block. We used primary CGC cultures and ex vivo (cerebellar organotypic slices) model systems to answer the following questions:  2+ Does acute exposure to MeHg alter [Ca ]i homeostasis in a biphasic manner in tg CGCs? If yes, is there a difference in time of onset of the MeHg-induced changes between WT, +/tg, and WT CGCs?  Are there any differences in VGCC subtype composition in our CGC cultures? If so, is there a correlation with the response to MeHg?  Does the tg mutation offer some protection against the cytotoxic effects of MeHg by reducing percent of CGC death?  Are there differences in response to MeHg as a result of developmental changes in vitro?  Does the tg mutation alter MeHg response of CGCs in an ex vivo model? Of great importance is the finding that tg-like mutations in the human α1A subunit gene, CACNA1A, have been linked to human disorders including episodic ataxia type 2, spinocerebellar ataxia type 6, and familial hemiplegic migraine (Pietrobon, 2002) (Fig. 1.6). Therefore, I will also discuss how we can potentially correlate our findings to susceptibility of particular human populations to MeHg exposure. 29 2+ Figure 1.1: Schematic representation of voltage-gated Ca Voltage-gated Ca 2+ channel quaternary structure. channels (VGCCs) are composed of an α1 pore-forming subunit which confers the pharmacological and voltage-sensing signatures that allow us to distinguish among them (the HVA Cav1.1-1.4 and Cav2.1-2.3, and the LVA Cav3.1-3.3). Four auxiliary subunits (β, α2, δ, and γ) are also associated with the α1 subunit to modify the kinetic properties of the VGCC. The α1 subunit has 4 transmembrane domains (I-IV), each consisting of 6 transmembrane segments (S1-S6). The S4 segments (seen marked with plus (+) signs) contain the voltage-sensing part of the channel, and the loops between S5 and S6 (indicated by red P) form the channel’s pore. The α2 and δ subunits are bound together by a disulfide bond (S-S) 2+ Regulatory proteins including Ca -binding calmodulin (CaM), protein kinase C (PKC), and activated G-proteins (Gβγ) can bind to intracellular sites of the protein to modify the activity of 2+ the channel. One such site is the Ca -binding domain (CBD) in the intracellular C-terminus of the α1 protein. Modified figure from Benarroch, 2010. 30 Figure 1.1 (cont’d) 2+ S-S Ca α1 γ α2 III IV δ β Cav1.1 L I Cav1.2 II Cav1.3 SSS S S 1 23 5 P6 Cav1.4 P/Q Cav2.2 R Cav2.3 SSS S S S SS S S 1 2 3 5 P6 1 2 3 5 P 6 Cav2.1 N SSS S S 123 5 P6 N PKC 2+ Gβγ C C B D Ca /CaM Cav3.1 T Cav3.2 Cav3.3 For interpretation of the reference to color in this and all other figures, the reader is referred to the electronic version of this dissertation. 31 Table 1: Characteristics of Voltage-Gated Calcium Channels (From Marrero-Rosado, B., et al., 2013) Type Channel Subtype Antagonist Electric Properties (conductance/τ) α1Coding Gene Cell/Tissue-Specific Expression Skeletal muscle α1S (Cav1.1) L Phenylalkylamines, dihydropyridines, benzothiazapines, and ω-agatoxin IIIA α1C (Cav1.2) 25 pS/20-50 ms α1D (Cav1.3) P/Q N ω-conotoxin GVIA 9-20 pS/>100 ms for Ptype; <100 ms for Qtype. 13 pS/50-110 ms (500-800 ms in sympathetic neurons) Plasticity/cardiac muscle contraction Sensory transduction Neurotransmitter release Pacemaker activity Retina and Tlymphocytes Visual function/un-known function in Tlymphocytes Presynaptic terminals Neurotransmitter release in neuromuscular junctions and CNS Purkinje neurons and cerebellar granule cells Neurotransmitter release Thalamus Involved in depolarization of thalamic neurons Only expressed in neurons Neurotransmitter release Nociceptive dorsal root ganglion neurons Depolarization (Cav1.4) ω-agatoxin IVA (P-type), ω-conotoxin MVIIC (Q-type) Excitation-contraction coupling CNS (dendrites and cell bodies)/cardiac muscle Cochlea Striatal medium spiny neurons Substantia nigra pars compacta α1F HVA Function α1A (Cav2.1) α1B (Cav2.2) 32 Table 1 (cont’d) Antagonist Electric Properties (conductance/τ) α1Coding Gene Cell/Tissue-Specific Expression Function Neurotransmitter release SNX 482 ? α1E Presynaptic terminals R (Cav2.3) Hippocampus Depolarization T Kurtoxin, mibefradil, amiloride Dendrites and cell bodies of some CNS neurons Rhythmic burst firing and pacemaker activity Type Channel Subtype HVA LVA α1G, α1H, 5-11 pS/? α1I (Cav3.13.3) 33 Figure 1.2. The layers and neuronal circuitry of the adult mouse cerebellar cortex. (A) Each lobule of the adult cerebellum is composed of three layers: the molecular layer (ML), Purkinje cell layer (PCL; labeled Pur in drawing), and the internal granular cell layer (IGL; labeled Gr in drawing). CGC cell bodies are located in the IGL. (B) They extend their axonal projections into the ML and make excitatory connections with the Purkinje cells (parallel fiber). Purkinje cells also receive excitatory input from climbing fibers of afferent inferior olivary neurons. Figures A and B from Haines and Dietrichs (2012) (A) Purkinje cell Golgi cell Purkinje cell Mol Pur Golgi cell Gr Transverse Sagittal Parallel fiber Basket cell Granule cell axon Golgi cell Golgi cell axon Granule cell Dendritic digits MF rosettes Mossy fiber (MF) 34 Figure 1.2 (cont’d) (B) Parallel fiber Climbing fiber Parallel fiber Purkinje cell dendrites and spines Climbing fiber 35 Figure 1.3: Cerebellar granule cell migration in the mouse cerebellum During the first two postnatal weeks of the mouse, CGC (labeled in image as GC) development consists of: proliferation in the EGL starting up to PND 10, followed by migration from the EGL to the IGL. Bergmann glia cells (BG) provide a scaffold during migration. The EGL becomes the ML, while the CGCs and Purkinje cells (PCs) make functional connections in the adult mouse. Figure from Xu, H., et.al. (2013) Pia ML EGL PCL IGL P3 GC BG 36 P15 Figure 1.4. Representative tracings of fura-2 signals during acute MeHg treatment: F340, F380, and ratio (F340/F380). The tracings shown are representative of the changes in both fura-2 signals of a CGC during acute treatment with 0.5 µM MeHg. Before start of MeHg treatment, + CGCs are depolarized with 40 mM K HBS to assess their viability (shown as first peak in all traces). This depolarization causes the opening of VGCCs, causing the influx of Ca 2+ and the 2+ binding of intracellular fura-2 to Ca ; this is seen as a simultaneous decrease in F380 and increase in F340 signal. CGCs that have a steady (non-declining) baseline and are able to return to original baseline after depolarization are considered healthy. Only healthy cells are analyzed at the end of the recording. MeHg treatment is started when cells return to baseline after depolarization (as marked in tracings), and continues until phase 2 reaches a plateau. Onset of the phases is determined based on changes in slope of the ratio (F340/F380) tracing. As shown below, the changes in slope on the ratio tracing are the attributed to simultaneously opposing changes in F340 and F380. 37 Figure 1.4 (cont’d) F340 Fluorescence Intensity (arbitrary units) F380 MeHg 70 Phase 1 Phase 2 63 55 43 41 39 37 0 12 24 36 48 60 Time (min) MeHg 1.00 Ratio (F340/F380) Phase 1 Phase 2 0.75 0.50 0 12 24 36 Time (min) 38 48 60 2+ Figure 1.5: Schematic drawing of MeHg-induced dysregulation of [Ca ]i homeostasis. 2+ Multiple events contribute to the MeHg-induced dysregulation of [Ca ]i that is sufficient to cause CGC death. (1) It is hypothesized that MeHg utilizes the VGCCs to reach the intracellular space. MeHg causes a block of Ca 2+ current through VGCCs within seconds. (2) MeHg also 2+ increases IP3 production, which releases Ca from the SER into the cytoplasm through IP3 receptors; (3) MeHg impairs mitochondrial respiration, causing loss of mitochondrial membrane potential and opening of MTP to release more Ca 2+ and cytochrome c (involved in apoptosis) into the cytoplasm. Mitochondrion VGCC 2+ Ca 2+ 2+ MTP [Ca ]i Ca , MeHg (?) IP3 R IP3 MeHg Ca M3 mACh R SER 39 2+ Figure 1.6: Location of the tottering mutation in the α1A subunit of the P/Q-type voltage2+ gated Ca channel. The α1A subunit of the P/Q-type is composed of four transmembrane domains, each consisting of 6 transmembrane segments (S1-S6). The location of the tg mutation is indicated by the arrow; it is located near one of the loops that forms the pore of the channel. Also shown are the locations of mutations that have been linked to episodic ataxia type-2 (circle), familial hemiplegic migraine (triangle), and spinocerebellar ataxia type-6 (square). Figure from Sawada, et.al., 2000 tg extracellular SSSS S 1234 5 S 6 S SSSS 12345 S 6 SS SSS 12 345 S 6 SSSSS 12345 S 6 intracellular H3N COOH episodic ataxia type-2 familial hemiplegic migraine spinocerebellar ataxia type-6 40 CHAPTER TWO EFFECT OF METHYLMERCURY ON INTRACELLULAR Ca 2+ HOMEOSTASIS OF TOTTERING CEREBELLAR GRANULE CELLS 41 A. Abstract Acute MeHg exposure causes a biphasic time- and concentration-dependent increase in 2+ 2+ [Ca ]i in CGCs in vitro. Pretreatment with VGCC antagonists delays the onset of the [Ca ]i increase, but does not prevent it. Nonetheless, this suggests that VGCCs play an important role in 2+ the mechanism of MeHg-induced [Ca ]i dysregulation. CGCs express different subtypes of VGCCs that are characterized by different kinetic and pharmacological properties. The main VGCC subtype involved in the process of neurotransmitter release, the P/Q-type (Cav2.1), is a major contributor of Ca 2+ influx at the synapses. The following experiments study the effects of 2+ MeHg on [Ca ]i homeostasis in primary CGCs of the Cav2.1 channelopathy mouse model, the tg mouse. We hypothesized that the decrease in P/Q-type VGCC density at the plasma membrane caused by the tg mutation would hinder MeHg from gaining access to the intracellular Ca 2+ 2+ 2+ stores and disrupting [Ca ]i homeostasis. The Ca -binding fluophore, fura-2, was used 2+ for analysis of [Ca ]i changes in primary cultures of tg CGCs caused by acute exposure to 0.5, 2+ 1, and 3 µM MeHg. [Ca ]i homeostasis was examined at different days in vitro (DIV) to account for changes in developmental changes in VGCC expression. +/tg and tg/tg CGCs 2+ showed a significant delay in the onset of both phases of [Ca ]i increase after exposure to 0.5 and 1 µM, at DIV 4. +/tg and tg/tg also spend more time in phase 1, and this lag between the onset of the first and second phase is inversely proportional to the concentration of MeHg. Interestingly, a slow increase in the time-to-onset of both phases, over the time investigated (DIV 4, 6, 8, and 10), was observed for WT CGCs; this gradual change resulted in an eventual loss of 42 differences between genotypes by DIV 10. These effects did not correlate with differences in expression of the α1 subunits of VGCC subtypes between WT, +/tg, and tg/tg, but do not rule out the possibility of an effect of the tg mutation in the expression of intracellular modulators of the 2+ MeHg-induced [Ca ]i dysregulation pathway. The data presented in this chapter suggest that 2+ +/tg and tg/tg CGCs are more resistant to the effects of MeHg on [Ca ]i dysregulation at earlier stages of development in vitro. 43 B. Introduction A significant loss of CGCs is observed in humans and animal models chronically and acutely exposed to MeHg. In vitro, acute exposure to micromolar and submicromolar concentrations of MeHg causes a biphasic, time- and concentration-dependent increase in 2+ 2+ [Ca ]i in CGCs (Marty and Atchison, 1997). The first phase of [Ca ]i increase occurs as a result of MeHg causing an increase in the production of IP3 via activation of M3 mACh receptors, and subsequent opening of SER IP3 receptors (Sarafian, 1993; Hare and Atchison, 1995). There is also evidence suggesting the release of Ca 2+ from mitochondria by opening of the MTP during this phase (Limke and Atchison, 2002). The second phase is believed to involve the robust entry of extracellular Ca 2+ through an unknown pathway that may not necessarily involve VGCCs. Even though overall Ca 2+ current through VGCCs in CGCs is reduced within seconds after acute MeHg exposure (Sirois and Atchison, 2000), these ion channels are believed 2+ to have an important role in the mechanism of MeHg-induced [Ca ]i increase. This is demonstrated by the delay in onset of both phases when CGCs are pretreated with VGCC antagonists (Marty and Atchison, 1997). HVA VGCC subtypes are characterized by different pharmacological (imparted by the α1 subunit) and kinetic properties (due to different β subunit composition) (see Table 1). Mature CGCs express all subtypes of HVA VGCCs in vitro, albeit they do not all contribute equally to overall Ca 2+ influx (Randall and Tsien, 1995) and neurotransmitter release (Varming et al., 1997). Even though MeHg reduces Ca 2+ current through all subtypes of VGCCs (Sirois and 44 Atchison, 2000), they are differentially expressed in CGCs. Also, since block of Ca 2+ current is influenced by rate of stimulation (VGCC opening) (Sirois and Atchison, 2000), the different kinetic properties (i.e. voltage-dependence of activation/inactivation, rate of inactivation, etc.) of 2+ VGCC subtypes could also influence the mechanism of [Ca ]i increase. Evidence supporting these hypotheses has been reported. Pretreatment with ω-conotoxin MVIIC (N- and P/Q-type antagonist) was able to reduce significantly CGC death in vitro after acute exposure to 0.5 and 1 µM MeHg. On the other hand, nimodipine (L-type VGCC antagonist) was only able to provide significant protection to CGCs against the lower MeHg concentration (Marty and Atchison, 1998). The present experiments were designed to study the hypothesis that a deleterious mutation in the main VGCC involved in neurotransmitter release in mature CGCs, the P/Q-type, 2+ would have an effect on the time-to-onset of the MeHg-induced [Ca ]i increase. The Cacna1a gene codes for the α1A pore-forming subunit of the P/Q-type VGCC. A leucine to proline mutation in Cacna1a causes a decrease in the number of P/Q-type VGCCs and increase in the expression of the N-type VGCC in mature CGCs of the tg mouse (Campbell and Hess, 1999). The difference in VGCC subtype composition between tg and WT CGCs may alter the time of 2+ 2+ 2+ onset of the MeHg-induced [Ca ]i increase. To monitor changes in [Ca ]i, a ratiometric Ca binding fluorophore, fura-2AM, was used. Following start of MeHg treatment, the time-to-onset 2+ of phase 1 and phase 2 of the [Ca ]i increase were measured and subsequently compared between genotypes. Since CGCs show an increase in whole-cell Ba 2+ current through VGCCs over time (Randall and Tsien, 1995), all measurements were performed at DIV 4, 6, 8, and 10. 45 C. Materials and methods tg Mice. A colony was established from B6.D2-Cacna1a /J breeding pairs obtained from The Jackson Laboratory (Bar Harbor, Maine). Homozygous tg mice showed a very pronounced phenotype (i.e. seizures, wobbly gait, and ataxia) at weaning age and were, therefore, not tail clipped. Littermates that did not show the aforementioned phenotype were genotyped at PND 21. PND 7-8 pups used for CGC isolations were tail clipped post-mortem. All experiments were performed following Michigan State University Laboratory Animal Resources and National Institutes of Health guidelines, and were approved by the Michigan State University Animal Use and Care Committee. Cerebellar granule cell isolation. PND 7-8 Cacna1a tg pups of either gender were decapitated, followed by removal of cerebella in ice-cold Krebs-Ringer buffer (KRB; 120 mM NaCl, 4.8 mM KCl, 1.21 mM KH2PO4, 14.17 mM dextrose, 25.3 mM NaHCO3, and 30 μM Phenol Red) supplemented with 4.5 mM MgSO4 and 3% w/v bovine serum albumin. All enzymes used in the isolation process are reconstituted in this KRB solution. Since the genotype of pups is not known prior to dissection, each cerebellum is worked with separately for the subsequent steps. Cerebella were minced with two scalpel blades prior to digestion with trypsin (0.028% w/v) in a 37ºC shaking water bath for 15 min. Cell suspensions are transferred to a 15 mL centrifuge tube; a 0.008% w/v trypsin inhibitor and 0.008% w/v DNAse solution is added, and cells are centrifuged at 134 x g for 3 min. The supernatants are discarded and replaced by a 0.052% w/v trypsin inhibitor and 0.001% w/v DNAse solution, followed by mechanical trituration (~25 times) with a 46 fire-polished Pasteur pipette. The dissociated cells were centrifuged one final time at 67 x g for 7 min. and cell pellets were resuspended in 6-7 mL of high glucose Dulbecco’s Modified Eagle medium (DMEM) with pyruvate (Invitrogen, New York) supplemented with 10% fetal bovine 6 serum, 1% w/v antibiotic-antimycotic (Invitrogen, New York), and 25 mM KCl. 2-3 x 10 CGCs were plated on coverslips previously coated with poly-D-lysine (0.1 mg/mL; Sigma-Aldrich, St. Louis, MO), inside 35 mm petri dishes, and kept in a 37ºC/5% CO2 incubator. The following day (DIV 1), half of the medium was replaced with non-antibiotic/antimycotic DMEM containing arabinosylcytosine (Ara-C; 5 µM final concentration) to prevent proliferation of glial cells. The antibiotic/antimycotic supplement was omitted from the medium after DIV 1 because previous research has shown that certain antibiotics can reduce VGCC function (Atchison et al., 1988; Suarez-Kurtz, 1989). Half of the medium was replaced with non-antibiotic/antimycotic DMEM (no Ara-C) on DIV 6 and 8. Each mouse yielded at least five plates; one plate was used on each DIV for a particular MeHg concentration. 2+ Single-cell measurements of changes in [Ca ]i. At the appropriate DIV, cells were washed twice with HEPES-buffered saline (HBS; 150 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl2, 0.8 mM MgSO4, 20 mM dextrose, and 20 HEPES at pH 7.3) buffer warmed to 37ºC. Cells were incubated for 45 min with a 5-7 µM (depending on cell density) solution of fura-2AM (Invitrogen, New York) in HBS. Using an IonOptix system (IonOptix, Milton, MA), sequential images were taken at 340 and 380 nm λmax, and the ratio of the 530 nm emission intensity at the 2+ two λmaxs was used to calculate the time of onset of changes in [Ca ]i after MeHg exposure. A 47 range of 5-7 cells per plate were selected for monitoring, and a mean time-to-onset of MeHg2+ induced [Ca ]i increase for each plate was calculated. Prior to start of recordings, excess fura2AM not taken up by cells was washed away for 10 min. Before the start of MeHg exposure, cells were superfused with a brief pulse of 40 mM K+ HBS to assess their viability. Cells were considered healthy if they reacted to the depolarizing solution by allowing influx of Ca 2+ (seen as a decrease in the signal at 380 nm λmax and increase at 340 nm λmax), and returned to original + baseline when superfused with normal K (5.4 mM) HBS. A 10 mM stock solution of MeHgchloride was diluted in HBS to the final concentrations being tested (0.5, 1.0, and 3.0 µM). Experiments were performed at different DIV (4, 6, 8, and 10) to account for changes in VGCC expression and overall Ca 2+ influx as a result of the in vitro environment (Randall and Tsien, 2+ 1995). MeHg was continuously superfused until both phases of [Ca ]i increase (Marty and Atchison, 1997) were observed. Between each plate, the superfusion tubing was rinsed with 0.5 2+ mM EDTA (Hg chelator) for 2 min, followed by rinsing with deionized water. This was done to prevent the subsequent plates from being exposed to trace amounts of MeHg before the start of treatment. Immunocytochemistry. Some CGCs that were used for fura-2 experiments were also used for immunocytochemistry assays. Cells were plated in 96-well plates. At each DIV, cells were rinsed once with ice-cold phosphate buffered saline (PBS) solution composed of 137 mM NaCl, 2.7 mM KCl, 1.4 mM NaH2PO4, 4.3 mM Na2HPO4, pH 7.4. The PBS buffer was always used ice48 cold. 4% w/v paraformaldehyde was added for 15 min as a fixative, followed by two washes with PBS. Since all antibodies used bind to an intracellular epitope, cells were permeabilized with acetone for 10 min. CGCs were then treated for at least 30 min with a blocking solution of 10% v/v normal goat serum and 0.1% v/v Tween-20 in PBS. All antibodies were diluted in this solution. Antibodies raised in rabbit against the α1A (P/Q-type), α1B (N-type), α1C (L-type), and α1E (R-type) subunits (Alomone Labs, Jerusalem, Israel) were used at a 1:200 dilution and incubated with cells overnight at 4ºC. Since all antibodies were raised in rabbit, single immunostaining had to be performed; each antibody was added to different wells containing CGCs. After primary antibody staining, cells were washed with PBS, three times, each for 5 min. Cells were then treated for 1 hr with a 1:200 dilution of the secondary goat anti-rabbit antibody, conjugated with fluorescein isothiocyanate (FITC). After secondary antibody staining, cells were washed for 5 min. with PBS, three times. Images were taken using a fluorescent microscope at 400X magnification. All quantifications were performed using the ImageJ software (National Institutes of Health, USA). A range of 5-7 cells were selected for measurement of pixel intensity. The following formula was used to normalize whole-cell pixel quantity for background signal (Corrected Total Fluorescence): Integrated density – (area of selected cell X mean fluorescence of background pixels) Integrated density refers to the product of the area of cell and mean pixel value. This formula has been previously used in peer-reviewed publications (Burgess et al., 2010; Gavet and Pines, 2010). 49 Statistical analysis. The Statistical Product and Service Solutions (SPSS) program (IBM, New York) was used for all statistical analyses. Differences in means between genotypes were analyzed using a multivariate analysis of variance (MANOVA) for microfluorimetry and immunocytochemistry data. For analysis of duration of phase 1 data, a univariate ANOVA was used. Statistically significant differences are set at p < 0.05 for all analyses. Measurements are expressed as mean ± standard error of the mean (S.E.M.). n ≥ 3 for means that were statistically analyzed. For immunocytochemistry data analysis, the statistical power for all VGCC subtypes at all DIV was < 0.8 due to the low sample size; at DIV 10, there is an n = 2 for each genotype. For microfluorimetry data, statistical power for detection of differences between genotypes is only > 0.8 for the DIV 4 means; there is an n = 1 for tg/tg time-to-onset of phase 1 and phase 2 after 3 µM MeHg treatment and, therefore, statistical analysis was not performed for this concentration. 50 D. Results VGCC subtype composition at each DIV in WT, +/tg, and tg/tg CGCs cultured in 25 mM + K . We studied the possibility that the differences in time-to-onset of the phases seen so far between genotypes could be due to differences in the composition of VGCC subtypes. CGCs + cultured in 25 mM K were immunostained with antibodies against the pore-forming subunits of each VGCC subtype: α1A, α1B, α1C, and α1E. Mean pixel values corrected for background signal were compared between genotypes at DIV 4 (Fig. 6A), 6 (Fig. 6B), 8 (Fig. 6C), and 10 (Fig. 6D). Overall, we observed an increase in total expression of VGCCs between DIV 4 and 8: 67.5% increase in WT, 33.7% in +/tg, and 38.0% in tg/tg. However, at DIV 10 this increase is reduced to an increase of just 11.6% in WT and 14.4% tg/tg compared to values at DIV 4. In CGCs from all genotypes, at all DIVs, we also observe a higher degree of immunostaining for α1B and α1C than for α1A and α1E. At DIV 4, CGCs from all genotypes show the same VGCC subtype makeup; there is no statistically significant difference in α1 subunit composition between genotypes. Immunostaining for α1B and α1C subunits starts to increase on DIV 6, until at DIV 8 we observe a significant difference in the quantity of α1C between WT and +/tg CGCs. The +/tg CGCs show a decreased staining for α1C compared to WT. Even though tg/tg appear to also show a decrease in α1C compared to WT, the measurement is representative of an n = 2 and statistical analysis was not performed. At DIV 10, there is only an n = 2 for each staining of α1 subunits and, therefore, I 51 was not able to perform a statistical analysis. It is important to note that the statistical power for analysis of mean immunostaining between genotypes was > 0.8 due to low sample size, which could affect the sensitivity of the test and prevent us from detecting differences in VGCC subtype composition between genotypes. 2+ Effect of MeHg on [Ca ]i homeostasis in WT, +/tg, and tg/tg CGCs at DIV 4, 6, 8, and 10. 2+ Fura-2AM, a ratiometric Ca -binding fluorophore, was used to visualize changes in 2+ [Ca ]i after MeHg treatment. At DIV 4 (Fig. 4A) and 10 (Fig. 4D) we observe the characteristic 2+ time- and concentration-dependent (p < 0.05) increase in [Ca ]i after 0.3, 1.0, and 3.0 µM. However, statistically significant differences between genotypes in the time to onset of both phases were only observed at DIV 4. On this DIV, tg/tg CGCs showed a slower onset of the first 2+ and second phase (45.7 ± 3.9 min and 77.4 ± 5.8 min, respectively) of [Ca ]i increase after treatment with 0.5 µM MeHg, compared to WT CGCs (11.2 ± 1.5 min for phase 1; 39.2 ± 12.7 min for phase 2). There was also a statistically significant delay in the onset of phase 1 and phase 2 between +/tg (44.5 ± 5.6 min and 78.3 ± 6.1 min, respectively) and WT CGCs, after 0.5 µM 2+ MeHg. There was no significant difference in the onset of [Ca ]i increase between +/tg and tg/tg. At DIV 4, we also observe a significant difference between phase 2 onset of WT (18.3 ± 0.9 min) and +/tg (35.4 ± 3.6 min) after 1µM MeHg treatment. DIV 4 is the only time point in which the statistical power of our analysis of means in time-to-onset was > 0.8. There are no differences between genotypes after any concentration of MeHg at DIV 6, 8, and 10. 52 2+ Effect of the tg mutation on the duration (min) of the first phase of MeHg-induced [Ca ]i increase. 2+ The biphasic increase in [Ca ]i caused by acute treatment with MeHg represents the outcome of two events: Ca 2+ release from intracellular Ca 2+ stores and robust influx of 2+ extracellular Ca . Delays in time-to-onset of phase 1 have been shown to cause subsequent delays of equal magnitude in the onset of phase 2 (Marty and Atchison, 1997; Limke and Atchison, 2002). This suggests that the two phases may be linked in some yet unknown way. The duration (min) of phase 1 was calculated for each DIV and MeHg concentration (Fig. 2.3). At DIV 4 we observe that +/tg (33.7 ± 2.4 min) and tg/tg (31.7 ± 6.9 min) CGCs spend a significantly longer amount of time on phase 1 than WT CGCs (14.7 ± 0.9 min) when acutely exposed to 0.5 µM MeHg. There are no other statistically significant differences between genotypes. However, we can observe a concentration-dependent pattern in +/tg and tg/tg CGCs that is not shown by WT CGCs. From DIV 4-8, WT CGCs spend roughly the same amount of time in phase 1 regardless of DIV or MeHg concentration. On the other hand, the time spent in phase 1 seems to be inversely proportional to MeHg concentration in +/tg and tg/tg CGCs, for all DIVs. We also observe a gradual increase of duration of phase 1 in WT CGCs overtime in vitro; this correlates with the same gradual shift we observe in phase 1 time-to-onset from DIV 4-10. 53 E. Discussion The results presented in this study indicate that: (1) all CGCs, regardless of genotype, 2+ show the characteristic time- and concentration-dependent increase in [Ca ]i after MeHg treatment; (2) at DIV4, +/tg and tg/tg CGCs show a significant delay in the onset of the first and 2+ second phase of [Ca ]i increase after acute exposure to 0.5 and 1 µM MeHg; (2) this difference is lost in subsequent DIVs, although subtle differences remain between WT and tg/tg CGCs on DIV 8 after 1 µM MeHg acute treatment; (3) tg/tg and +/tg CGCs spend more time in the first 2+ phase of [Ca ]i increase on DIV 4; (4) the significant differences in phase 1 time-to-onset observed at DIV4 between WT, +/tg, and tg/tg CGCs may not be explained by a difference in VGCC subtype composition, but an increase in sample size is needed to confirm this result; and (5) +/tg CGCs show a significant decrease in α1C immunostaining compared to WT CGCs at DIV 8. CGCs from WT rats express in culture all subtypes of HVA VGCCs, and the contribution of each on overall Ca 2+ current (Randall and Tsien, 1995) and neurotransmitter release (Varming et al., 1997) has been studied. It is important to note that the two aforementioned reports on the contribution of VGCC subtypes examined CGCs cultured under different conditions. Whereas Randall and Tsien (1995) recorded overall Ca 2+ influx from cells cultured + in 5 mM K , Varming, et.al., (1997) studied contribution of VGCCs in neurotransmitter release + in cells cultured in 25 mM K . As previously discussed, changes in resting membrane potential are responsible in vivo for the expression of genes only present in mature CGCs. CGCs can, 54 + therefore, be kept from fully developing into mature neurons by culturing them under high-K conditions (Sato et al., 2005; Nakanishi and Okazawa, 2006; Okazawa et al., 2009). These previous studies have reported the differences in the contribution of Ca 2+ influx through HVA + VGCC subtypes between high- and low-K cells in vitro. In rat primary CGCs cultured in low+ K (5 mM; mimicking mature cells in vivo), the biggest contributor of Ca 2+ current is the P/Q- type (46%), followed by N-type (20%), R-type (19%), and L-type (15%) (Randall and Tsien, + 1995). On the other hand, mouse CGCs cultured in high-K (25 mM) media showed a global Ca 2+ current also mostly dependent on P/Q-type (42%), but with a higher contribution of L-type (35%) than that reported by Randall, et.al.; R-type and N-type were reported to compose 13 and 10% of the Ca 2+ current, respectively (Varming et al., 1997). The tg mutation is located in the pore-forming subunit of the P/Q-type VGCC, α1A (Fletcher et al., 1996). Without reportedly causing changes in single-channel ion conductance, there is a decrease in contribution of P/Qtype VGCCs in overall Ca 2+ influx that is believed to be the result of a reduced presence at the cell membrane (Wakamori et al., 1998; Leenders et al., 2002). However, research in the tg mouse cerebellum has focused on the effects of the mutation on Purkinje cells, which are the only efferent neurons of the cerebellum and believed to be behind the cause of the ataxic and epileptic behavior (Wakamori et al., 1998; Campbell and Hess, 1999; Erickson et al., 2007). There is no information in the current literature about the effect of the tg mutation on overall 2+ Ca /Ba 2+ current in CGCs. The two types of cells do not compensate in the same manner for the decrease in P/Q-type VGCC content. Whereas in Purkinje cells there is an increased role of 55 L-type VGCCs in neurotransmitter release (Campbell and Hess, 1999; Erickson et al., 2007), neurotransmitter release in parallel fiber-Purkinje cell synapses is mostly dependent on N-type VGCC function (Zhou et al., 2003). Our immunocytochemistry data show a higher level of staining for the pore-forming subunits of N- and L-type VGCCs, α1B and α1C, in the cell bodies of CGCs of all genotypes. This result contrasts with functional studies of VGCC subtype contribution previously reported in mouse primary CGCs. In WT mouse CGCs cultured in 25 + mM K , P/Q-type VGCCs contribute to 42% of the overall Ba 2+ current (Varming et al., 1997), whereas WT CGCs in our assay showed that α1A is ~7% of the VGCC subtype levels on all DIV. Conversely, Varming, et al. (1997) also demonstrated that the N-type VGCC contribution on overall Ba 2+ in these cells is only 10%, whereas our assay showed that α1B is ~45% of all VGCC subtype levels on all DIV. We cannot imply that protein levels correlate with VGCC contribution to Ca2+ influx. For example, in forebrain synaptosomes of adult tg/tg mice there is an increase in the contribution of N-type VGCC on glutamate release without an increase in protein levels of the α1B subunit (Leenders et al., 2002). Recordings of overall Ca 2+ influx in the cell bodies of tg/tg CGCs would have to be performed to really assess the contribution of each VGCC subtype. 2+ MeHg causes a biphasic time- and concentration-dependent increase in [Ca ]i (Fig. 1.4) that can be delayed with pretreatment of VGCC antagonists (Marty and Atchison, 1997). This and other data (Marty and Atchison, 1998) have suggested a crucial role of VGCCs in the 2+ mechanism of MeHg-induced [Ca ]i dysregulation. Therefore, the central aim of the experiments in this chapter was to study the effect of the tg mutation on the MeHg-induced 56 2+ 2+ 2+ [Ca ]i increase. Fura-2, a Ca -binding fluorophore, was used to visualize changes in [Ca ]i in WT, +/tg, and tg/tg CGCs after acute treatment with 0.5, 1, and 3 µM MeHg. The concentrations of MeHg chosen to conduct the microfluorimentry experiments are below the reported body burden (200-312 mg of Hg) for Iraq individuals acutely exposed to MeHg and showing signs of extremity weakness and ataxia (Bakir et al., 1973). A body burden of 200-312 mg is approximately equivalent to 19.5 µM MeHg in blood (Atchison, 1986). Microfluorimetry experiments were performed on a wide range of DIVs because the overall Ca 2+ influx in CGCs has been shown to increase with time in vitro (Randall and Tsien, 1995). The only significant difference between genotypes was observed at DIV 4. tg/tg CGCs 2+ show a delay in the onset of the first and second phase of [Ca ]i increase after acute exposure to 0.5 µM MeHg (Fig. 2.2.A). Even though there are no data on global Ca 2+ influx in tg CGCs, an 2+ effect of the mutation on resting basal [Ca ]i has been reported. In a study by Bawa and Abbott 2+ (2008) fura-2 was used to compare the resting (no depolarization) concentrations of [Ca ]i 2+ between WT and tg/tg CGCs. They found that resting [Ca ]i was significantly reduced in tg CGCs (20% reduction) compared to WT. There was no difference in mitochondrial membrane 2+ potential, which could indicate that the role of this organelle in [Ca ]i homeostasis is not affected in tg/tg CGCs (Bawa and Abbott, 2008). There are many other key players in regulating 2+ 2+ + [Ca ]i homeostasis including the SER, Ca -ATPase, Na /Ca function of these proteins has not been examined in tg CGCs. 57 2+ exchanger, among others. The We also determined the time spent in phase 1 to study the hypothesis that intracellular factors may be affecting the response of tg CGCs to acute MeHg exposure. For 0.5 µM MeHg at DIV 4, +/tg and tg/tg CGCs spend more time in phase 1 than WT CGCs (Fig. 6). Even though no other significant difference between genotypes was found, we can observe that in general (up to DIV 8) +/tg and tg/tg CGCs show a duration of phase 1 that is inversely dependent on the concentration of MeHg. In other words, the lower the concentration of MeHg, the longer the time lapse between the onsets of phase 1 and 2. This pattern is only observed for WT CGCs at DIV 10. This difference in behavior of the CGCs could indicate a difference between genotypes in the accumulation of MeHg inside the cell. Another explanation for the longer duration of phase 1 2+ tg/tg could involve their lower basal [Ca ]i (Bawa and Abbott, 2008). It is not known what 2+ triggers the onset of phase 2, only that it is dependent on the presence of extracellular Ca . If 2+ 2+ the onset of phase 2 depends on [Ca ]i to reach a certain threshold, then lower basal [Ca ]i in tg/tg CGCs could prolong the time spent in phase 1. Alternatively, because MeHg has high affinity for thiol groups, the difference in duration of phase 1 could be due to unreported differences in intracellular cysteine-containing proteins between tg/tg, +/tg, and WT CGCs. Data from the immunocytochemistry assay performed in this chapter do not show a difference in VGCC subtype expression between tg/tg, +/tg and WT CGCs on DIV 4 (Fig. 2.1.A), even though on this DIV we observe a significant delay in the onset of phase 1 and 2 on +/tg and tg/tg CGC cultures (Fig. 2.2.A). However, not only does the immunocytochemistry assay not provide information on the function of the α1 subtype detected, but it also does not indicate if the protein is located in the cytosol or at the cell membrane. It is also important to 58 note that the VGCC immunostaining data presented in this chapter represent fluorescence intensity in cell cultures from only three mice per genotype. This is a very low sample size and, therefore, the statistical power was < 0.8 for the analysis performed on the immunocytochemistry data. It has been reported that the tg mutation causes a decrease in overall Ba 2+ current in mature Purkinje cells without causing changes in voltage of activation/inactivation and single-channel conductance (Wakamori et al., 1998). It is suspected that the point mutation results in a posttranslational modification of the α1A subunit that results in an 85% decrease of P/Q-type VGCCs at the cell membrane (Leenders et al., 2002). It is very interesting that, even though the phenotype of +/tg mice is undistinguishable from WT, +/tg CGCs do not respond in the same manner to MeHg. In fact, most of the current research literature on tg mice and primary cell cultures only compare data between WT and tg/tg. Some researchers even use +/tg and WT mice interchangeably as normal controls (Kaja et al., 2007). The experiments presented in this and subsequent chapters are unique in that they present and compare data from all genotypes, and do not make the assumption that +/tg CGCs are equivalent to WT CGCs. We do not know the full effect of the tg mutation in +/tg mice and primary cell cultures. The only report found in the current literature providing electrophysiological data from +/tg pertains to studies at the NMJ. It was found that NMJs from +/tg showed a significant increase (~40% of WT) in MEPP frequency, a measurement of spontaneous neurotransmitter release; tg/tg NMJs recorded an increase in MEPP frequency of ~100% (Plomp et al., 2000), however, this was not recapitulated in studies performed by Pardo, et al. (2006). Even though we cannot say that the effects seen in NMJs occur in CGCs in vitro or in vivo, we could still argue the possibility that +/tg and WT CGCs 59 present differences below the threshold for a detectable change in phenotype, but enough to result in a difference in MeHg susceptibility. Data presented in this chapter show that the tg missense mutation in the pore-forming 2+ subunit of P/Q-type VGCCs can delay the effect of MeHg on [Ca ]i homeostasis, at early DIV + in cells cultured in depolarized conditions. CGCs cultured in 25 mM K have been shown to maintain electrophysiological characteristics and gene expression patterns of immature CGCs in vivo (Sato et al., 2005; Okazawa et al., 2009). Therefore, we could argue that +/tg and tg/tg 2+ CGCs could be more resistant to MeHg-induced [Ca ]i disruption during early development in the mouse. Also, because +/tg CGCs respond to MeHg in a similar manner to that of tg/tg CGCs, having one mutant allele of the Cacna1a gene appears to be sufficient to protect the cells from 2+ MeHg-induced [Ca ]i disruption. 60 2+ Figure 2.1. Voltage-gated Ca channel subtype composition in tottering CGCs cultured in + 25 mM K . Immunostaining of WT, +/tg, and tg/tg CGCs was performed at DIV 4 (A), 6 (B), 8 (C), and 10 (D) using antibodies against the α1 pore-forming subunits of HVA VGCCs. Values represent mean corrected fluorescence ± S.E.M; n=6 for WT, n=4 for +/tg, and n=3 for tg/tg (except for DIV 8). (A) DIV 4 P/Q-type L-type N-type R-type 48 (Thousands) Corrected Total Fluorescence 60 36 24 12 0 WT +/tg 61 tg/tg Figure 2.1 (cont’d) (B) DIV 6 P/Q-type N-type R-type L-type 48 (Thousands) Corrected Total Fluorescence 60 36 24 12 0 WT +/tg 62 tg/tg Figure 2.1 (cont’d) (C) DIV 8 P/Q-type L-type N-type R-type 48 (Thousands) Corrected Total Fluorescence 60 36 n=2 24 12 0 WT +/tg 63 tg/tg Figure 2.1 (cont’d) (D) DIV 10 P/Q-type L-type N-type R-type 48 (Thousands) Corrected Total Fluorescence 60 36 24 12 0 WT +/tg 64 tg/tg Figure 2.2. Mean time-to-onset (min) ± SEM of first and second phase of MeHg-induced 2+ + [Ca ]i increase in tottering cerebellar granule cells cultured in 25 mM K , at DIV 4, 6, 8, 2+ and 10. Single-cell fluorimetric analysis of changes in [Ca ]i after MeHg treatment was 2+ performed. We observed a biphasic, concentration- and time- dependent increase in [Ca ]i caused by MeHg (representative image of phases in Fig. 1.4). 5-7 cells were monitored in each plate to calculate an average of time-to-onset of the phases, and the averages were compared among genotypes. Each value represents the mean ± S.E.M of time-to-onset of phase 1 (left) and phase 2 (right) for DIV 4 (A), 6 (B), 8 (C), and 10 (D). The asterisk (*) represents a statistically significant difference between two genotypes. n = the number of mouse pups from which cultures were made. (A) DIV 4 tg/tg +/tg WT Phase 1 100 Phase 2 Time (min) 80 60 40 20 0 0.5 1 0.5 3 MeHg concentration (µM) 65 1 3 Figure 2.2 (cont’) (B) DIV 6 tg/tg +/tg WT Phase 1 50 Phase 2 Time (min) 40 30 20 10 0 1 3 1 MeHg concentration (MeHg) 66 3 Figure 2.2 (cont’d) (C) DIV 8 tg/tg +/tg WT Phase 2 Phase 1 50 Time (min) 40 30 20 10 0 1 1 3 MeHg concentration (µM) 67 3 Figure 2.2 (cont’d) (D) DIV 10 tg/tg +/tg WT Phase 2 Phase 1 100 Time (min) 80 60 40 20 0 0.5 1 0.5 3 MeHg concentration (µM) 68 1 3 2+ Figure 2.3. Mean duration of phase 1 (min.) ± SEM of MeHg-induced [Ca ]i increase in + tottering cerebellar granule cells cultured in 25 mM K . Duration of phase 1 was calculated by subtracting the average time of onset of phase 1 from the average time of onset of phase 2, in each plate. This was calculated from data collected at DIV 4, 6, 8, and 10. Each value represents the mean duration of phase 1 (± S.E.M). The asterisks (*) represent statistically significant difference between genotypes (p < 0.05). tg/tg +/tg WT Duration of Phase 1 (min) 45 36 27 18 9 0 0.5 1 DIV 4 3 0.5 1 3 DIV 6 0.5 1 DIV 8 MeHg concentration (µM) 69 3 0.5 1 DIV 10 3 CHAPTER THREE EFFECT OF METHYLMERCURY ON TOTTERING CEREBELLAR GRANULE CELL VIABILITY IN VITRO 70 A. Abstract 2+ Dysregulation of [Ca ]i homeostasis is one of the hallmark events that occurs in CGCs 2+ in vitro as a result of acute exposure to MeHg. Prolonged elevated [Ca ]i can cause impairment 2+ of mitochondrial respiration as a result of the organelle’s increased buffering of [Ca ]i, and the 2+ activation of Ca -dependent kinases involved in the intrinsic apoptotic pathway. Evidence of 2+ the role of [Ca ]i dysregulation in MeHg-induced cell mortality comes from data that show that 2+ exposing CGCs in vitro to MeHg for the amount of time required to increase [Ca ]i is sufficient to cause a significant increase in cell mortality. The aim of the present study was to determine + the effect of MeHg on cell viability of tg primary CGCs cultured at 25 and 5 mM K . Acute treatment with 0.5, 1, and 3 µM MeHg under depolarizing conditions (mimicking developing CGC in vivo) causes an overall increase in cell mortality. The effect of MeHg on the viability of WT, +/tg, tg/tg CGCs becomes more pronounced as DIV increases. tg/tg CGCs are more susceptible to MeHg-induced cytotoxicity than are +/tg and WT at DIV 6 and 8, respectively. At DIV 4 we did not observe any differences between genotypes. As discussed in the previous 2+ chapter, at DIV 4 there is a significant delay in +/tg and tg/tg CGCs of the [Ca ]i increase + caused by MeHg compared to WT. When extracellular K is decreased to 5 mM to resemble more the conditions in which fura-2 microfluorimetry was performed, there seems to be a reduction in cell mortality of tg/tg preparations. The data presented in this chapter indicate that the tg mutation causes increased susceptibility to acute MeHg exposure under in vitro conditions 71 that mimic early postnatal development. These findings suggest that MeHg may be more cytotoxic to CGCs in the developing tg mouse. 72 B. Introduction Previous experiments by Marty and Atchison (1998) have demonstrated that the increase 2+ in [Ca ]i caused by acute exposure to MeHg is sufficient to cause cell death. Rat CGCs exposed in vitro to MeHg for just the amount of time it takes for them to reach phase 2 (Fig. 1.3), show an increase in cell mortality (as early as 3.5 hrs after treatment) compared to preparations not exposed to MeHg. This increase in cell death can be significantly delayed when CGCs are pretreated with VGCC antagonists (Marty and Atchison, 1998). This pivotal piece of data 2+ conclusively linked [Ca ]i dysregulation and VGCCs as two key players in MeHg-induced CGC cytotoxicity. Other events that occur at CGCs as a result of MeHg exposure are an increase in ROS, decrease in glutathione (a cysteine-rich antioxidant), and an overall loss of mitochondrial function (Sarafian and Verity, 1991; Limke and Atchison, 2002). Experiments discussed in the previous chapter showed that +/tg and tg/tg CGCs present a 2+ significant delay in the onset of MeHg-induced [Ca ]i increase compared to WT CGCs (Fig. 2.2.A). Moreover, we observed that +/tg and tg/tg CGCs spend more time in phase 1 at lower concentrations of MeHg. Therefore, the current experiments were designed to study the effect of the tg mutation on MeHg-induced cell death. We wanted to examine the hypothesis that the delay in the onset of phase 1 observed at DIV 4 would translate into reduced cytotoxicity, in the same manner that pretreatment with VGCC antagonists delays cell death. Others have reported differences in resting membrane potential and electrophysiological properties between immature and mature CGCs (Nakanishi and Okazawa, 2006; Okazawa et al., 2009). To closely resemble the resting membrane potential of CGCs in early development in vivo, CGCs were exposed to + MeHg in the culture medium in a depolarizing (25 mM K ) environment; an extracellular 73 concentration of 25 mM K + is calculated to generate a resting membrane potential of approximately -35 mV in CGCs, which is close to the average potential previously reported for pre-migratory cells in the EGL (Rossi et al., 1998; Okazawa et al., 2009). Since an extracellular + concentration of 5 mM K is calculated to generate a resting membrane potential of approximately -50 mV, CGCs were exposed to MeHg in this condition to mimic the resting membrane after CGCs have migrated to the IGL (Rossi et al., 1998; Okazawa et al., 2009). Aside from examining MeHg effects on increasing DIV, the cytotoxicity experiments were performed at 4, 8 and 24 hrs post-MeHg treatment. MeHg causes maximum cytotoxicity 24 hrs after onset of MeHg exposure (Marty and Atchison, 1998); however, we also assayed at earlier time points to account for potential differences in susceptibility between genotypes that may occur before the 24-hr mark. 74 C. Materials and Methods Cerebellar granule cell isolation. The protocols for isolation and maintenance of primary CGCs are described in Chapter 2. However, 24-well plates, instead of coverslips, were treated with a 0.1 mg/mL solution of poly-d-lysine for CGC plating. For experiments in which CGCs in low + + K were used, 5 mM K DMEM was used as cell culture medium. Each mouse was considered to be an n = 1. MeHg treatment. Each mouse was able to yield sufficient CGCs for a total of 12 wells. A 24well plate contained 4 wells of CGCs from each mouse; one well per MeHg concentration was used (0, 0.3, 1, and 3 µM). There were a total of three 24-well plates, one for each time point assayed after MeHg treatment was given (4, 8, or 24 hr). At a particular DIV (4, 6, 8 or 10), half of the medium was removed and replaced with MeHg-containing DMEM (either 5 or 25 mM K + without antibiotics). Cells were exposed to MeHg for 2 hrs, at which time all of the culture medium was removed and replaced with MeHg-free DMEM. The duration of the MeHg treatment was chosen based on the slowest time-to-onset of phase 2 seen in the fura-2 experiments (Chapter 2). Cytotoxicity assays were then performed at 4, 6, and 24 hrs after onset of MeHg exposure. Calcein-AM and ethidium homodimer-1 cytotoxicity assay. At each time point assayed, DMEM was removed and CGCs were washed once with HBS. We used the Live/Dead® viability/cytotoxicity kit for mammalian cells from Invitrogen (Carlsbad, CA) to assess the percentage of cell death in our MeHg-treated cultures. Cells were treated for 20 min. in the dark 75 with a 0.2 µM solution of calcein-AM and ethidium homodimer-1 (EthD-1). Then cells were rinsed twice (5 min. each) with HBS. Calcein-AM is cell-membrane permeable, but is trapped inside of healthy cells when endogenous enzymes cleave the ester group. On the other hand, EthD-1 only permeates damaged nuclear membranes; once inside, it binds to DNA to increase its fluorescence. 40X images were taken under a fluorescent microscope with a fluorescein optical filter (485 ± 10 nm) for calcein-AM, and a rhodamine optical filter (530 ± 12.5 nm). Automated cell quantification of CGCs with ImageJ. Each image for cells stained with calcein-AM and EthD-1 contained throusands of cells. Therefore, an automated quantification method with ImageJ (NIH) was used. Background signal was minimized using the “rolling ball radius” method provided by the software, set at 15. The brightness/contrast of the picture was also automatically corrected. The images were then converted into binary form, and the “watershed” feature used to digitally separate rare cell clumps. Cells were chosen for counting 2 based on two parameters: a 30-120 µm size (CGC bodies are reported to be between 6-10 µm in diameter) and a 0.50-1.00 circularity (CGCs are very spherical compared to glial cells). Even though glial cell contamination is minimal (~5%), these parameters allowed me to be confident that I was counting only CGCs. The percentage of dead cells was calculated using the following formula: % dead cells = # EthD-1-stained cells X 100 # Calcein-AM + # EthD-1 cells The automated quantification method allowed an unbiased analysis because all images were processed in the same manner and the same parameters were used for the quantification. 76 Also, this was a blind study because genotypes of the mice were not known until after all images were analyzed. Statistical analysis. The SPSS program (IBM, New York) was used for all statistical analyses. Data were analyzed using a mixed design of two-way ANOVA with repeated measures: one between-factor (genotype) and one within factor (concentration). Pairwise comparisons between genotypes at different concentrations of MeHg were performed using a Bonferroni adjustment. Statistically significant differences were set at p < 0.05 for all analyses. Measurements are expressed as mean ± S.E.M.; n ≥ 3 for all means. All analyses had a statistical power < 0.8, except the analysis of data from 24 hrs after MeHg treatment on DIV 6. 77 D. Results Effect of acute exposure to MeHg on the viability of WT, +/tg, and tg/tg CGCs cultured in + 25 mM K . The present studies were designed to investigate the effect of acute exposure to MeHg on + cell viability of tg/tg CGCs that were essentially kept in an immature state with high K . Fig. 3.1 shows the mean percent of cells stained with EthD-1 for each genotype, at different time points after MeHg exposure, on DIV 4 (A), 6 (B), and 8 (C). Overall, we observed that the effect of MeHg on cell viability is more pronounced as DIV increases. At DIV4, neither genotype nor concentration of MeHg has a significant effect on percent of cell death at either time point. On DIV 6, we report a significant concentration-dependent response after MeHg treatment at all time points. A significant difference in percent of cell death between +/tg (15.6 ± 3.0%) and tg/tg (26.6 ± 1.1%) was observed 8 hrs after 0.3 and 1 µM (16.9 ± 0.3% for +/tg; 29.5 ± 3.0% for tg/tg) MeHg treatment. On that same DIV, at the 24 hr time point, cell death in negative control tg/tg (27.0 ± 3.7%) CGC cultures was significantly higher than WT (13.9 ± 3.8%) and +/tg (14.6 ± 1.6%; Fig. 3.1.B). Even though there are also differences between genotypes for the 0.3 and 1 µM treatments after 24 hrs, we cannot conclude that this is due to differences in genotype because of the differences observed in negative controls. At DIV 8, a significant effect of MeHg concentration on cell death was only observed 24 hrs after MeHg treatment. tg/tg (56.0 ± 3.9%) CGC cultures show a higher percent of cell mortality than WT (16.6 ± 1.4%) and +/tg (32.1 ± 9.0%) CGCs at the 8 hr time point after 1 µM MeHg. The difference in percent of cell death between WT (17.0 ± 1.6%) and tg/tg (42.4 ± 8.9) CGCs 24 hrs after 0.3 µM MeHg barely missed 78 the cutoff point set for statistical significance (p = 0.052). These experiments were not performed at DIV 10. We also present preliminary data on the effects of MeHg on cell viability of CGCs + cultured in low K (Fig. 3.2). Even though the mean percentage of cell death from WT CGCs represents an n = 2 and statistical analysis was not performed, there may be differences in + response to MeHg between cells exposed to MeHg in depolarizing (25 mM K ) and non+ + depolarizing (5 mM K ) conditions. Additional repetitions of acute MeHg exposures in low-K conditions would have to be conducted to explore this hypothesis. 79 E. Discussion Data from experiments described in this chapter indicate that: (1) as DIV increases, susceptibility of cells from all genotypes increases; (2) the constant exposure to high + concentrations of K (25 mM) causes an increase in cell mortality of tg/tg cell not exposed to MeHg; (3) tg/tg CGCs are more susceptible than +/tg CGCs 8 hours after MeHg exposure on DIV 6; (4) tg/tg CGCs are more susceptible to MeHg toxicity than WT CGCs at DIV 8; and (5) + CGCs may respond differently to MeHg when MeHg is given in low-K conditions. It is of interest that the overall increase in MeHg-induced cell death observed from DIV 4-8 occurs as the expression of the VGCCs increases (Fig. 2.1 A-D). This would be in agreement with previous research that demonstrates the role of VGCCs in cell mortality after acute MeHg 2+ exposure. Pretreatment with the [Ca ]i chelator 1,2-bis(o-aminophenoxy)ethane- N,N,N',N'tetraacetic acid (BAPTA) is able to delay cell death, but not prevent it (Marty and Atchison, 2+ 1998). This suggests that [Ca ]i is involved in the mechanism of cytotoxicity after acute exposure to MeHg, but is not the only player involved. In fact, a decrease in glutathione and an increase in production of ROS have been implicated in the mechanism of MeHg-induced cell death (Sarafian and Verity, 1991). The observation of a higher percent of cell death in negative + controls of tg/tg CGCs cultured in high K may stem from the fact that we are assuming that these mutant cells undergo the same developmental changes in resting membrane potential as WT cells. During the migration of CGCs in mouse and rat models, CGCs are reported to change their resting membrane potential from a relatively depolarized to a more negative membrane potential (Rossi et al., 1998; Nakanishi and Okazawa, 2006; Okazawa et al., 2009). There are 80 conflicting reports on the resting membrane potential of immature and mature CGCs. Rossi, et al. (1998) reported that CGCs in the EGL (immature) of WT mouse cerebellar slices showed resting membrane potentials between -22 and -22.9 mV, whereas Okazawa, et al. (2009) reported it to be -47.2 mV. Mature WT mouse CGCs that have migrated and settled in the IGL have a resting membrane potential between -57.5 (Rossi et al., 1998) and -68.6 mV (Okazawa et + al., 2009). By culturing CGCs in 25 mM K , we calculate the resting membrane potential to be + at -35 mV; CGCs in 5 mM K should have a resting membrane potential of -50 mV. Whatever the true resting membrane potentials are in WT CGCs in vivo, we do not know if any of these occur in tg/tg CGCs. Sustained elevated resting membrane potentials caused by culture in 25 + 2+ mM K are toxic to tg/tg CGCs, most likely due to an abnormally increased influx of Ca . It is important to note that the present model of exposure to MeHg is very different from the method used in previous published work by Marty and Atchison (1998). The exposure of + MeHg in the previous work was applied by superfusion in a normal-K HBS buffer; this was done to resemble the fura-2 microfluorimetry experiments on which the cytotoxicity assays were based. On the other hand, the experiments described in this chapter involved the exposure of MeHg in the culture medium. An advantage of this latter method is that we could mimic better an in vivo paradigm in which cells are exposed to neurotrophic factors and depolarized resting membrane potentials. One caveat, however, is the presence of cysteine-containing proteins in the culture medium supplements that may bind to MeHg and essentially reduce the concentration of MeHg to what the cells are exposed. This may be the reason why, even though we observe a 2+ significant delay in the phases of [Ca ]i increase at DIV 4, results from the cytotoxicity assay 81 do not show a difference between genotypes on the same DIV. Because the methods of MeHg + exposure are not the same, we cannot correlate the cytotoxicity assays performed at high-K with data from the fura-2 microfluorimetry assay. Nonetheless, the data demonstrate some differences between genotypes that may be relevant for future studies of susceptibility of the tg/tg CGCs to the effects of MeHg on neuronal migration. CGCs of the early-postnatal developing tg/tg cerebellum might be more susceptible to the cytotoxicity caused by MeHg exposure. In vivo MeHg exposure experiments would have to be performed to study this hypothesis. 82 + Figure 3.1. Tottering cerebellar granule cell (cultured in 25 mM K ) viability after acute exposure to MeHg. CGCs maintained in depolarized conditions were exposed to 0.3, 1, and 3 µM MeHg for two hours. Calcein-AM and EthD-1 were used at a final concentration of 0.2 µM to assess the viability of cells after 4, 8, and 24 hrs of onset of MeHg exposure. Mean percent of cells stained with EthD-1 (dead) was calculated for each genotype on DIV 4 (A), 6 (B), and 8 (C) an indication of dead cells. Data represent the mean ± S.E.M. (n ≥ 3). Asterisk (*) represents a statistically significant difference between genotypes; pound sign (#) represents a statistically significant effect of MeHg concentration on percent of cell mortality (concentrationdependence). (A) DIV 4 tg/tg +/tg WT 65 % EthD-1-stained CGCs 52 39 26 13 0 0 0.3 1 3 0 0.3 1 3 8 hrs 4 hrs MeHg concentration (µM) 83 0 0.3 1 24 hrs 3 Figure 3.1 (cont’d) (B) DIV 6 tg/tg +/tg WT 65 % EthD-1-stained CGCs 52 39 26 13 0 0 0.3 1 3 0 0.3 1 3 8 hrs 4 hrs MeHg concentration (µM) 84 0 0.3 1 24 hrs 3 Figure 3.1 (cont’d) (C) DIV 8 tg/tg +/tg WT 65 p = 0.052 % EthD-1-stained CGCs 52 39 26 13 0 0 0.3 1 3 0 0.3 1 3 8 hrs 4 hrs MeHg concentration (µM) 85 0 0.3 1 24 hrs 3 + Figure 3.2. Tottering cerebellar granule cell (cultured in 5 mM K ) viability after acute + exposure to MeHg. CGCs cultured in conditions of low K were exposed to 0.3, 1, and 3 µM MeHg for 2 hrs on DIV 4. Calcein-AM and EthD-1 were used at a final concentration of 0.2 µM to assess the viability of cells after 4, 8, and 24 hrs of onset of MeHg exposure. Mean percent of cells stained with EthD-1 (dead) was calculated for each genotype as an indication of cell death. Data represent the mean ± S.E.M. Statistical analysis was not performed because WT data represents the average of n = 2; tg/tg data represents average of n = 4. DIV 4 tg/tg WT 65 % EthD-1-stained CGCs 52 39 26 13 0 0 0.3 1 3 0 0.3 4 hrs 1 3 8 hrs MeHg concentration (µM) 86 0 0.3 1 24 hrs 3 CHAPTER FOUR EFFECT OF METHYLMERCURY ON CEREBELLAR GRANULE CELL VIABILITY IN CEREBELLAR ORGANOTYPIC SLICES 87 A. Abstract CGCs in culture provide the opportunity of studying the effects of a neurotoxicant in isolation. However, the contribution of other neuronal and non-neuronal components of the cerebellar circuitry should not be ignored. CGCs in the adult mouse, aside from receiving excitatory input from afferent neurons, receive tonic and phasic inhibitory signals from local GABAergic neurons that can modify CGC firing rate and, consequently, the activity of VGCCs. The inhibitory input that CGCs receive is of particular interest because N-type VGCCs, which in tg CGCs have a higher contribution to whole-cell Ca 2+ current than in WT CGCs, are more susceptible to GABA-dependent modulation than are P/Q-type VGCCs. Therefore, we hypothesized that the reported increased GABA-dependent modulation of Ca 2+ influx in tg CGCs would decrease their susceptibility to MeHg and reduce the percentage of CGC death in the tg cerebellum. To investigate this hypothesis, we made use of cerebellar organotypic slices, an ex vivo model to study CGC cytotoxicity after MeHg exposure. Organotypic slices maintain local functional neuronal circuitry in vitro for long periods of time and have been used extensively as an alternative to in vivo models. Data presented in this chapter show that, while there is no statistically significant difference in CGC mortality between WT and tg/tg cerebellar organotypic slices, +/tg showed a higher percent of cell death when exposed to 10 and 30 μM MeHg for 24 hrs. Moreover, pretreatment with ω-conotoxin GVIA reduced considerably the percentage of CGC death in +/tg slices, suggesting that their increased susceptibility is dependent upon N-type VGCC activity. The data presented in this chapter suggests that CGCs of the adult +/tg mouse could be more susceptible to in vivo MeHg exposure, possibly due to increased activity of N-type VGCCs. 88 B. Introduction Much research has been performed over the years on the effects of MeHg in isolated CGC cultures. The advantage of this approach is being able to focus on studying cell-specific changes without contamination of other cell types. This advantage is also the main limitation of using primary CGC cultures for MeHg cytotoxicity studies. For CGCs it is especially the case because in vivo they are part of a more elaborate circuit of neurons that can regulate each other’s activity. CGC cultures are 90-95% CGC, with some glial contamination. However, in these cultures we lack the input of inhibitory cells that can modulate the firing frequency and, therefore, VGCC activity. An alternative to studying the cytotoxic effects of MeHg in vivo, is the use of an ex vivo model, the cerebellar organotypic slice. Organotypic slices are a modification of acute brain slice models with the benefit of longterm culture. These slices preserve their local circuitry intact, remain functional, and more closely mimic in vivo models than do primary cell cultures (Xiang et al., 2000; Lonchamp et al., 2006; Cho et al., 2007). These types of slices are usually dissected from mice during their first postnatal week because of the ability of younger neurons to adapt better to changes in their environment (i.e. from the animal to a petri dish). A higher percentage of cell viability is observed in organotypic slices from younger animals (PND 3-9) than young adults (Cho et al., 2007). For the experiments discussed in this chapter, the main caveat of using slices from PND 3-9 is that, at this stage of development, tg/tg mice do not exhibit the characteristic motor dysfunction that closely resembles ataxia in humans. The ataxic behavior of tg/tg mice presents during the third postnatal week. Therefore, the current experiments were performed in cerebellar organotypic slices of PND 23-25 mice. 89 Data from experiments discussed in Chapter Three showed that at DIV 6 and 8, tg/tg CGC cultures have a higher degree of cell death than WT 8 hrs after acute exposure to 0.3 µM MeHg. However, those experiments were performed under depolarizing conditions, which are reportedly different than those found in mature CGCs in vivo (Rossi et al., 1994). We suspect, + based on preliminary results from low K CGC cultures (Fig. 3.2), that mature CGCs will show a reduced susceptibility to MeHg in an ex vivo model. The present studies were designed to study MeHg-induced cytotoxicity of CGCs in their own environment composed of various neuronal interactions. 90 C. Materials and Methods Organotypic slices. Tg litters were genotyped at PND 21. Genotyping took two days to complete and was performed before the start of dissections. All instruments and materials used for the surgical dissection of the cerebella were sterilized to prevent contamination of the brain slice cultures. At PND 23-25, mice were anesthetized with isoflurane and decapitated. Whole brains were removed and placed in a petri dish with a sterile filter paper soaked in the culture medium. Culture medium was kept in ice and consisted of Neurobasal-A, supplemented with B27 (1X), N-4 (1X), and 1 mM glutamine (all from Invitrogen, Carlsbad, CA). The culture medium also contained antibiotic-antimycotic (final concentration of 1X; 100 units/mL penicillin, 100 units/mL streptomycin, and 0.25 µg/mL of Fungizone®; Gibco/Invitrogen, Carlsbad, CA) during the dissection and the first two hrs after slices were plated. Cerebella were separated from brain stem and forebrain with a razor. 250 µm sagittal slices were cut using a manual tissue slicer that had been previously sterilized with 80% ethanol and 1 hr UV light exposure. After each cut, cerebellar slices were carefully picked up and placed in a 35 mm petri dish that contained 2 mL of the culture medium. Cut slices were not kept on ice because cell viability is improved if the temperature is not excessively low (data not shown). Cerebellar slices were inspected under a light microscope for signs of excessive mechanical damage; those which appeared damaged were discarded (very rare occasions). Immediately after all slices were collected, they were plated on Transwell®-COL collagen-coated membrane inserts (Corning, Tewksbury MA) placed inside 6-well plates. One membrane insert was used for each treatment and it contained two slices per mouse (4 slices total). In other words, for each mouse, data from a particular treatment represented the average of two slices. 91 MeHg exposure of organotypic slices. Cerebellar slices were allowed to attach to the collagencoated membrane inserts for 2 hrs before start of MeHg treatment. During this time, they were kept in medium supplemented with 1X antibiotic-antimycotic (Gibco/Invitrogen, Carlsbad, CA). This medium was removed completely, followed by addition of 3, 10, or 30 µM MeHg in nonantibiotic Neurobasal-A medium. These concentrations are 10X higher than those used in previous in vitro experiments because of the higher content of lipids present in tissue compared to isolated CGC cultures. The organotypic slices were exposed to MeHg for 24 hrs at 37ºC in a 5% CO2/95% O2 incubator. ω-conotoxin GVIA treatment. For some experiments, organotypic slices were pretreated with ω-conotoxin GVIA prior to exposure to MeHg. After allowing organotypic slices to attach for 2 hrs post-dissection, a 1 hr treatment with 500 nM ω-conotoxin GVIA was given in non-antibiotic Neurobasal-A medium. The concentration of ω-conotoxin GVIA was chosen based on a previous electrophysiology study performed in CGC cultures which show that 0.3-1 µM concentrations of this toxin provide the same magnitude of N-type VGCC block (Randall and Tsien, 1995). The toxin was not removed during MeHg treatment. Instead, MeHg was added directly to the toxincontaining medium. The slices were kept in the MeHg medium for 24 hrs at 37ºC in a 5% CO2/95% O2 incubator before the cytotoxicity assay was performed. Calcein-AM and ethidium homodimer-1 cytotoxicity assay. Some modifications were made to the cytotoxicity protocol described in Chapter Three. Organotypic slices were kept on the collagen-membrane inserts throughout the procedure. MeHg medium was removed and slices 92 were washed once with warm (37ºC) HBS. A treatment with a solution of 2 µM calcein-AM and 4 µM EthD-1 (Mancini et al., 2009) was applied for 1 hr. Slices were then washed once before imaging. Confocal microscopy and image analysis. Z-series images of fluorescent organotypic slices were obtained using a 40X water objective fitted to a Leica DM LFSA (Leica Optics, Bannockburn, IL) confocal microscope. For each treatment, there were two slices per mouse. Organotypic slices were removed from the collagen-coated membrane with a tissue puncher and transferred to a chamber with a glass bottom for imaging. An Excitation Beam Splitter FW: TD 488/543/633 was used, with a frame average of 2 and line average of 2 (excitation λ of 485 nm for calcein-AM and 530 for EthD-1). The photomultiplier tube (PMT) gain was set at 380 volts for calcein, and at 657 volts for EthD-1. An automated quantification of fluorescent cells was performed with ImageJ (NIH, Bethesda, MD) using the maximum projection of each Z-stack. Compared to images from cytotoxicity assays in primary CGC cultures (Chapter Three), background signal in images from organotypic slices was significantly higher. Therefore, a threshold depending on the background signal was set before automated quantification was performed. After setting of fluorescence intensity threshold, images were transformed to binary and the “watershed” function was performed. The same size and circularity parameters as described in Chapter Three were used for the automated quantification analysis. Statistical analysis. Mean percent of dead cells represents an n ≥ 6 for MeHg-only experiments; ω-conotoxin GVIA pretreatment experiments represent an n = 3 for each genotype. Cerebellar slices from each mouse received all treatments. Therefore, a repeated measures (for MeHg 93 concentration) two-way ANOVA was performed for data analysis. Pairwise comparisons of genotype and treatments were performed with a Bonferroni correction for multiple comparisons. Statistical significance was set at p < 0.05 for all measurements. 94 D. Results CGC viability in WT, +/tg, and tg/tg organotypic slices after 24-hr exposure to MeHg. Cerebellar organotypic slices from WT, +/tg, and tg/tg mice at PND 23-25 were exposed to micromolar concentrations of MeHg for a period of 24 hrs and cytotoxicity was measured. Maximal projection of 100X Z-stack images was used to assess the overall morphology of the cerebellar slices (Fig. 4.1.A); Representative images of the maximal projection of 400X Z-stack images are presented in Fig. 4.1.B. Data of percentages of EthD-1-stained cell represent the mean of n ≥ 6 for each genotype. We observed an overall significant (p < 0.05) MeHg concentration-dependent response. There are no statistically significant differences between genotypes when slices are exposed to 3µM (57.7 ± 3.9% for WT, 53.2 ± 6.5% for +/tg, and 56.2 ± 5.6% for tg/tg). After 10 µM MeHg treatment, +/tg (77.2 ± 2.7%) slices show a higher percent of CGC death than WT (58.4 ± 5.4%) and tg/tg (62.9 ± 3.9%). At this concentration there is no difference in cell mortality between WT and tg/tg. After 30 µM MeHg there is a higher percent of dead cells in +/tg (93.9 ± 2.0%) slices than in tg/tg (77.6 ± 3.4%) slices. There is no difference in CGC mortality between WT (85.6 ± 4.2%) and +/tg after treatment with 30 µM MeHg. CGC viability in WT, +/tg, and tg/tg organotypic slices treated with ω-conotoxin GVIA prior to 24-hr exposure to MeHg. Treatment with ω-conotoxin GVIA prior to MeHg exposure was performed to assess the contribution of N-type VGCCs to the increased percentage of CGC death observed in +/tg organotypic slices. Mean percent of cell death represents the mean of n = 3 for each genotype, at each concentration. Whereas treatment of MeHg by itself caused a significantly higher degree of cytotoxicity in +/tg organotypic slices (Fig. 4.2), pretreatment with ω-conotoxin GVIA was able 95 to reduce the mortality seen in these slices; this reduction was enough to eliminate the previously seen differences between +/tg and WT, and +/tg and tg/tg at 10 and 30 μM MeHg (Fig. 4.2). 96 E. Discussion Data from experiments described in this chapter indicate that: (1) there is no difference in MeHg-induced CGC mortality between PND 23-25 WT and tg/tg cerebellar organotypic slices; (2) After 10 and 30 μM MeHg exposure, +/tg slices show a significantly higher CGC mortality than WT and tg/tg slices; and (3) higher +/tg susceptibility to MeHg can be partially rescued at all concentrations by pretreatment with ω-conotoxin GVIA. One caveat of studying MeHg-induced cytotoxicity in primary CGC cultures and correlating the results to in vivo responses is that CGCs exist in the animal as part of a circuit (discussed in Chapter One). The cerebellum is not only composed of CGCs, but also other neurons that make connections with CGCs and have the potential of influencing their response to MeHg in vivo. For example, CGCs receive inhibitory input from Golgi cells. Primary CGC cultures from PND 7-8 mice do not contain these inhibitory neurons and are composed of ~9095% CGCs. More importantly, these primary cultures are derived from mice during the first PNDs, when the neurons are still in their early developmental stages. The phenotype of a tg/tg mouse is indistinguishable from that of a WT or +/tg mouse during the first 3 postnatal weeks; the characteristic impaired motor function of the tg/tg mice is manifested from weaning age to adulthood (Fletcher et al., 1996). We opted to not do in vivo studies because tg mice show a severe ataxic behavior that would hinder the collection of useful data from performance tests, including rotarod and balance beam. Although an in vivo model of treatment would have been ideal, we would not have been able to discern motor dysfunction caused by MeHg from the phenotype due to the tg mutation. In lieu of in vivo treatment with MeHg, cerebellar organotypic slices (ex vivo model) were used to examine CGC viability as a result of acute exposure to MeHg. 97 Organotypic slices from young adult animal models have been successfully used by other groups for studies of the hippocampal (Xiang et al., 2000; Wise-Faberowski et al., 2009) and cerebellar circuitry (Lonchamp et al., 2006). Our studies used cerebellar organotypic slices from PND 23-25 mice, after the onset of their ataxic behavior. We observed that after 24 hr exposure period with 10 and 30 μM MeHg there is a significantly higher percentage of CGC mortality in the IGL of +/tg cerebellar slices than in WT and +/tg slices (Fig. 4.2). Very few research studies have examined the effects of the mutation in the +/tg mouse. Electrophysiological recordings at +/tg NMJs revealed an increase in frequency of spontaneous release (Plomp et al., 2000), that was not recapitulated in similar studies performed by Pardo et al. (2006). In WT CGCs P/Q-type and N-type VGCCs contribute 49% and 32% of the Ca 2+ influx required for glutamate release, respectively, while tg/tg CGCs have a higher contribution from N-type (53%) than from P/Qtype (13%) (Zhou et al., 2003). N-type VGCCs, aside from exhibiting a faster inactivation rate than P/Q-type VGCCs, display a greater extent of G-protein-mediated inhibition by GABAB receptors (Currie and Fox, 1997). GABA-mediated G-protein inhibition is characterized by a decrease in Ca 2+ current amplitude and slower rate of voltage-dependent activation of the N-type VGCC; this inhibition can be removed by high depolarizations steps (Barrett and Rittenhouse, 2000). G-protein modulation is voltage-dependent: some of the G-protein-dependent inhibition of N-type VGCCs is eliminated under depolarized conditions (Dolphin, 1998). The increase in N-type VGCC activity in tg/tg CGCs is hypothesized to be the cause of their enhanced GABAmediated inhibition (Zhou et al., 2003), which could play a role in the susceptibility of these cells to MeHg exposure. The subtle differences in CGC mortality after MeHg treatment between WT and tg/tg cerebellar slices did not reach statistical significance. However, it is important to 98 note that the MeHg-induced cytotoxicity assay was performed at only one time point (24 hrs after onset of MeHg treatment). It would be interesting to explore possible differences in susceptibility between the genotypes at earlier time points. Tg/tg mature CGCs in vivo show certain characteristics that are different from WT and could explain possible differences in susceptibility. Overall Ca 2+ influx and neurotransmitter release is dependent mostly on N-type VGCCs, as opposed to P/Q-type VGCCs in WT CGCs (Leenders et al., 2002; Zhou et al., 2003). Also, tg/tg CGCs show a reduction in the expression - of the α6βxγ2 GABAA receptor (Kaja et al., 2007). This ionotropic, ligand-gated Cl channel is involved in inhibition of CGC firing as a result of input from GABAergic neurons involved in inhibition of the CGC; interestingly, the α6 subunit has been hypothesized to confer CGCs their characteristic susceptibility because Purkinje cells do not express this subunit (Herden et al., 2008). As previously mentioned, Purkinje cells are relatively spared in cases of acute and chronic exposure to MeHg. It has been hypothesized that the decreased expression of α6-GABAA receptors could cause a reduced degree of tonic inhibition and increased excitability of tg/tg 2+ CGCs (Kaja et al., 2007). However, this hypothesis contrasts with the reduced basal [Ca ]i levels in these cells (Bawa and Abbott, 2008); it also would not explain the lower degree of cell mortality after MeHg exposure in tg/tg cerebellar organotypic slices. To date, the studies performed by Zhou et al. (2003) are the only ones that have reported functional differences between +/tg and WT CGCs. N-type VGCC contribution to glutamate release in +/tg CGCs is similar to that in WT CGCs at 33%; however, the activity of P/Q-type VGCCs in +/tg CGCs is reduced to approximately half of the percent in WT CGCs. Data from 99 experiments described in this chapter show that CGCs in +/tg cerebellar organotypic slices are more susceptible to MeHg-induced toxicity. Because pretreatment with ω-conotoxin GVIA was able to eliminate the differences in MeHg-induced cell mortality between the genotypes (Fig. 4.3), the susceptibility of +/tg CGCs seems to be influenced by the activity of N-type VGCCs. It is possible that effects of MeHg on cell viability might be influenced by the ratio of P/Q-type and N-type contribution to Ca 2+ contribution. In +/tg CGCs, there is an almost equal contribution of P/Q- and N-type VGCCs to overall Ca 2+ influx, which may interfere with the enhanced GABA- mediated inhibition normally found in tg/tg CGCs. There is no information on the degree of GABA-mediated inhibition in +/tg CGCs and, therefore, more research needs to be performed to study this hypothesis. 100 Figure 4.1. Representative images of sagittal cerebellar organotypic slices stained with calcein-AM and Ethd-1. Cerebella from WT, +/tg, and tg/tg mice were sagittally sliced (250 µm thick) and exposed to various MeHg concentrations for 24 hrs. After staining with calcein-AM and EthD-1, a Z-stack fluorescence confocal image analysis of the cerebellar lobules was performed. (A) 100X maximum projection image of a z-stack from a negative control (no MeHg treatment). The cerebellar layers in a lobule are labeled in the 100X image as such: molecular layer, ML; internal granular layer, IGL; Purkinje cell layer, PCL. 100X z-stack images were captured to visualize the overall morphology of a lobule; maximum projection of each 400X zstack images was used for the quantification of EthD-1-stained cells. (B, C, D) Shown are the representative maximum projections of calcein and EthD-1 staining in 400X z-stack images from WT, +/tg, and tg/tg cerebellar organotypic slices exposed to different treatments (as labeled). 101 Figure 4.1 (cont’d) (A) PCL Calcein IGL (live cells) ML 100 µm EthD-1 (dead cells) 100 µm 102 Figure 4.1 (cont’d) (B) No ω-conotoxin GVIA pretreatment 0 µM MeHg 500 nM ω-conotoxin GVIA pretreatment 30 µM MeHg 0 µM MeHg 30 µM MeHg Calcein 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm WT EthD-1 103 Figure 4.1 (cont’d) (C) No ω-conotoxin GVIA pretreatment 0 µM MeHg 500 nM ω-conotoxin GVIA pretreatment 30 µM MeHg 0 µM MeHg 30 µM MeHg Calcein 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm +/tg EthD-1 104 Figure 4.1 (cont’d) (D) No ω-conotoxin GVIA pretreatment 0 µM MeHg 500 nM ω-conotoxin GVIA pretreatment 30 µM MeHg 0 µM MeHg 30 µM MeHg Calcein 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm 40 µm tg/tg EthD-1 105 Figure 4.2. CGC viability in cerebellar organotypic slices exposed in vitro to MeHg. Cytotoxicity assays were performed on organotypic cerebellar slices from WT, +/tg, and tg/tg after a 24-hour treatment with 3, 10, and 30 µM MeHg. Data are presented as mean percentage of EthD-1-stained (dead) cells ± S.E.M. n ≥ 6 for each genotype at all MeHg concentrations tested. An overall concentration-dependent response is observed (p < 0.05; indicated by # sign). After 3 µM MeHg treatment, no statistically significant difference between genotypes is observed. However, after 10 µM MeHg +/tg slices show a significantly higher percentage (77.2 ± 2.7%) of CGC death compared to WT (58.4 ± 5.4%) and tg/tg (62.9 ± 3.9%). After 30 µM MeHg, we can still observe a significantly higher percentage of CGC death in +/tg (93.9 ± 2.0%) compared to tg/tg (77.6 ± 3.4%). The differences between WT and tg/tg never reached statistical significance. 106 Figure 4.2 (cont’d) +/tg WT tg/tg 100 % EthD-1-stained cells 89 78 67 56 45 35 0 0 3 10 MeHg concentration (µM) 107 30 Figure 4.3. CGC viability in cerebellar organotypic slices exposed in vitro to MeHg after pretreatment with 500 nM ω-conotoxin GVIA. Cytotoxicity assays were performed on organotypic cerebellar slices from WT, +/tg, and tg/tg pretreated with 500 nM ω-conotoxin GVIA followed by a 24-hour treatment with 3, 10, and 30 µM MeHg. Data are presented as mean percentage of EthD-1-stained (dead) cells ± S.E.M. n = 3 for each genotype at all MeHg concentrations tested. An overall concentration-dependent response is observed (p < 0.05). No statistically significant difference between genotypes is observed. WT tg/tg +/tg 100 % EthD-1-stained cells 89 78 67 56 45 35 0 Neg. control 0 µM MeHg 3 µM MeHg 10 µM MeHg 500 nM ω-conotoxin GVIA 108 30 µM MeHg CHAPTER FIVE SUMMARY AND FUTURE DIRECTIONS 109 Summary and future directions MeHg is an environmental neurotoxicant that causes a plethora of intracellular events that lead to neuronal cell death. Despite this, there are both very clear cell-specific neurotoxicity and more sensitive targets (Yuan et al., 2005). Cases of acute and chronic MeHg human poisonings have revealed that affected individuals present with similar signs, regardless of mode of exposure; these signs include sensory disturbances, motor dysfunction, and death (severe MeHg poisoning) (Nuttall, 2006; Ceccatelli et al., 2010). The motor dysfunction caused by MeHg exposure can be explained by a robust loss of CGCs, which have a crucial role in the relay of information from afferent neurons to the only efferent neurons of the cerebellum, the Purkinje 2+ cells (Hunter and Russell, 1954). MeHg causes CGC death by (1) disrupting [Ca ]i homeostasis (Marty and Atchison, 1996), and (2) increasing ROS levels, which leads to oxidative stress (Sarafian, 1991; Yee 1994). Interestingly, CGCs are particularly susceptible to MeHg compared to other neurons present in the cerebellum, even though MeHg seems to interact with all proteins that contain amino acids with thiol groups. Our lab and others have demonstrated the important influence of VGCC function in the 2+ pathway of [Ca ]i dysregulation (and subsequent death) in CGCs exposed acutely and chronically to MeHg (Marty and Atchison, 1996; Marty and Atchison, 1997). MeHg causes the release of Ca 2+ from the SER and mitochondria, creating a cytosolic environment of sustained 2+ elevated levels of [Ca ]i that is sufficient to cause CGC cell death (Marty and Atchison, 1997; Marty and Atchison, 1998; Limke and Atchison, 2002; Limke et al., 2004). The MeHg-induced 2+ 2+ increase in [Ca ]i in turn activates Ca -dependent proteases, known as calpains, that are 110 involved in the intrinsic apoptotic signaling pathway (Sakaue, 2005). In this caspase-independent apoptotic pathway, calpains trigger the release of the apoptosis inducing factor (AIF) from the mitochondria, which translocates to the nucleus and causes the condensation and fragmentation 2+ of DNA (Fig. 5.1; Sakaue, 2005). However, because pretreatment with an [Ca ]i chelator does not prevent MeHg-induced CGC death (Marty and Atchison, 1998), it is believed that other apoptotic pathways that do not involve increased levels of Ca 2+ are occurring simultaneously. CGCs of mice acutely exposed to MeHg show increased ROS and cytochrome c levels (Bellum 2+ et al., 2007). Opening of the MTP, aside from contributing to the increased [Ca ]i levels (Limke and Atchison, 2002), releases cytochrome c (Brustovetsky et al., 2002) which interacts with the apoptotic protease activating factor 1 (Apaf-1) to trigger the activation of the caspasedependent apoptotic pathway (Fig. 5.1, Bratton, 2000). MTP opening also causes the increase of ROS due to the loss of MMP; since lysosomes are vulnerable to oxidative stress, these can rupture and release enzymes (cathepsins) that activate pro-apoptotic proteins (Bax/Bak) involved in the mitochondrial apoptotic pathway (Ceccatelli et al., 2010). The experiments described in this dissertation were designed to explore the hypothesis that a mutation that causes impaired function of a highly important VGCC in CGCs, alters their susceptibility to MeHg. To test this hypothesis we made use of a Cav2.1 human channelopathy animal model, the tg mouse. The tg mouse is extensively used in the research community as an animal model to study the causes and therapeutic treatments of epilepsy. Tg mice present an ataxic phenotype that is also characterized by the onset of absence seizures that closely resemble human petit mal epilepsy (Fletcher et al., 1996). The data presented in this study demonstrated that a deleterious mutation in the main VGCC subtype expressed in CGCs can have an effect on 111 the response of these neurons to MeHg exposure, in vivo and ex vivo. Data from fura-2 singlecell microfluorimetry experiments show that +/tg and tg/tg CGCs acutely exposed to MeHg in + 2+ normal-K HBS present a delayed onset of the biphasic [Ca ]i increase after acute MeHg exposure. +/Tg and tg/tg CGCs also show a longer period of time spent in the first phase of 2+ MeHg-induced [Ca ]i increase (release of Ca findings that suggest that overall Ca 2+ 2+ from SER and mitochondria). There are several influx from VGCCs could be decreased in tg/tg CGCs, 2+ which may explain the delay in the onset of [Ca ]i increase after acute MeHg exposure in vitro: (1) Purkinje cells from tg mice show a decrease in Ba 2+ current (Wakamori et al., 1998), (2) 2+ basal [Ca ]i levels are significantly reduced in tg/tg CGCs (Bawa and Abbott, 2008), and (3) 2+ the number of α6 GABAA receptors, whose expression is dependent on Ca /calmodulinactivated CaN, is reduced in the cerebellum (Kaja et al., 2007). A possible reduction in overall Ca 2+ 2+ influx in tg/tg CGCs could explain the delay in the increase of [Ca ]i after acute MeHg exposure observed on DIV 4. This delay loses statistical significance at DIV 6-10, even though 2+ +/tg and tg/tg CGCs still appear to spend more time in the first phase of [Ca ]i increase. Presently, there is no information available in the literature on Ca 2+ influx through VGCCs in +/tg and tg/tg CGCs, in vivo or in vitro; additional experiments need to be performed to determine if the restoration of normal levels of Ca their similar behavior as WT CGCs. 112 2+ influx at latter the DIV is responsible for + CGCs cultured under high K conditions show an increased influx of Ca 2+ compared to + cells that are cultured under low K media (Gallo et al., 1987). This reported increase in Ca 2+ is believed to trigger intracellular signaling mechanisms that promote proliferation and migration of CGCs during early development. However, complete maturation of CGCs requires a shift toward more negative resting membrane potentials, demonstrated by the lack of claw-like morphology of CGCs in the IGL of mouse cerebellar organotypic slices cultured in high K + + (Okazawa et al., 2009). Therefore, CGCs cultured in high K media are believed to remain in a virtually immature state. Our data from acute treatment of CGCs in vitro with MeHg in high K + media show that tg/tg CGCs are more susceptible to MeHg than WT CGCs. A 25 mM + extracellular K concentration shifts the resting membrane potential of CGCs to approximately -35 mV. N-type VGCCs show significant activation between -30 and -10 mV in chick dorsal root ganglion neurons (Fox et al., 1987); P/Q-type VGCCs activate at membrane potentials that are positive to -40 mV in (Zhang et al., 1993; Wakamori et al., 1998). Therefore, at a -35 mV resting membrane potential, CGCs are closer to the activation threshold for N- and P/Q-type VGCCs and the conditions could increase Ca 2+ influx through both subtypes. The tg mutation also 2+ causes a reduction in the expression of calretinin, an intracellular Ca -binding protein, in CGCs (Cicale et al., 2002). Under depolarized conditions, MeHg may be more toxic to tg/tg CGCs 2+ because of their inability to cope with the increase in [Ca ]i caused by the toxicant. The + observation that tg/tg CGC cultures show higher percentage of cell death in high K medium without MeHg could be evidence of the reduced ability of these cells to buffer the 113 depolarization-induced increase in Ca 2+ influx. The level of calretinin in +/tg CGCs is comparable to that found in WT CGCs (Cicale et al., 2002), which is consistent with +/tg CGCs + not being as susceptible to MeHg as tg/tg CGCs in vitro under high-K conditions. Since CGC + cultured in high K media mimic closely their state during early development (Gallo et al., 1987; Mellor et al., 1998; Okazawa et al., 2009), it would be of interest to study the effect of MeHg on the migration of tg CGCs during early development. In humans and murine models, MeHg causes a halt of CGC migration during development. Based on the data presented in this work, we hypothesize that tg/tg CGCs are more susceptible to MeHg during the first three postnatal weeks than WT CGCS; we expect that in vivo MeHg exposure during the early postnatal development of CGCs will result in higher cell death in tg/tg pups than in WT pups. MeHg exposure of cerebellar organotypic slices, an ex vivo model of cytotoxicity, provided a glimpse into how mature CGCs would respond to in vivo MeHg exposure. We observed a decreased percent of cell death in tg/tg cerebellar organotypic slices exposed to MeHg, compared to +/tg CGCs. Moreover, the relatively high susceptibility of +/tg CGCs in cerebellar slices to MeHg could be partially rescued by pretreatment with ω-conotoxin GVIA (N-type antagonist); this could suggest a decreased degree of N-type VGCC inhibition in +/tg CGCs. Since organotypic slices closely resemble in vivo circuitry (Xiang et al., 2000; Lonchamp et al., 2006; Cho et al., 2007), we expect in vivo MeHg exposure of young adult tg/tg, +/tg, and WT mice to yield similar results as those observed in our ex vivo MeHg paradigm. However, one potential confounding factor to keep in mind is that afferent neurons that normally make excitatory connections with CGCs are severed during isolation of the cerebellar organotypic slices, which could reduce the activity of CGCs and, therefore, VGCCs. 114 The data from experiments presented in this dissertation provide additional evidence of a 2+ role for VGCCs in the mechanism of MeHg-induced disruption of [Ca ]i homeostasis and death of CGCs. Previous studies had made use of VGCC antagonists to explore the effects of MeHg on CGCs with decreased activity of specific VGCC subtypes (Sakamoto et al., 1996; Marty and Atchison, 1997; Marty and Atchison, 1998). This appears to be is one of the few studies in which gene-environment interactions pertaining to MeHg exposure have been studied in model organisms. We have previously reported that low oral doses (1-3 ppm/day) of MeHg accelerated the onset of rotarod failure of G93A superoxide dismutase 1 (SOD1) mutant mice, a model of amyotrophic lateral sclerosis (ALS) (Johnson et al., 2011). SOD1 is an intracellular enzyme that prevents the accumulation of ROS. The G93A SOD1 mouse model has “knocked-in” multiple copies of the human SOD1 gene with a point mutation that is believed to cause a detrimental gain-of-function result. Even though CGC function is not dysfunctional in the G93A SOD1 mouse, it presents evidence of a mutation altering the function of a protein and altering the outcome of in vivo MeHg exposure. Such in vivo MeHg exposure experiments have yet to be performed on CaV2.1 channelopathy mouse models; these experiments could potentially lead to identifying human populations exposed to MeHg that are at higher risk for neurotoxicity. Some tg-like mutations in the CACNA1A gene have been linked in the human population to disorders including spinocerebellar ataxia type-6, familial hemiplegic migraine, and episodic ataxia type 2 (Pietrobon, 2002; Rajakulendran et al., 2012) (Fig. 1.6). Even though not all mutations in CACNA1A produce the same deleterious effect, some of these mutations are believed to impair Ca 2+ influx into neurons in a similar manner as the tg mutation in mice (Pietrobon, 2002). For example, the signs of episodic ataxia type-2 (rare autosomal dominant pattern of inheritance) 115 closely resemble those presented in tg mice: ataxia, wobbly gait, loss of motor coordination, all of which can be induced by stress. Moreover, the fact that some of these disorders have a latency of onset suggests that, similar to what happens in the tg mouse, there might be compensatory mechanisms by other VGCC subtypes. However, we cannot say that the same compensatory mechanisms (i.e. increase in expression and contribution of N-type VGCC subtype, decreased expression of α6 subunit of GABAA receptors, etc.) occur in human subjects with tg-like mutations. It would be interesting to study in isolation, by heterologous expression in HEK cells, the effect of MeHg on cells expressing VGCCs containing the exact mutations linked to human disorders. Not only would these experiments aid in identifying human populations at risk, but also further aid in the understanding of the mechanism of MeHg-induced cytotoxicity. In conclusion, even though the specific role of VGCCs in MeHg-induced CGC death is still a matter of debate, this dissertation provides evidence of an altered MeHg cytotoxicity as a result of a point mutation in the pore-forming subunit of the most abundant VGCC subtype in CGCs, the P/Q-type. The data presented in the previous chapters argue that changes in the function or composition of VGCC subtypes in CGCs can alter the susceptibility of these neurons to MeHg exposure. Since numerous mutations in CACNA1A with different effects on P/Q-type VGCC function have been reported in the human population, the idea of individuals exhibiting different susceptibilities to MeHg exposure is a reality that should be explored. 116 Figure 5.1. MeHg-induced intracellular cell death pathways. MeHg causes an increase in 2+ [Ca ]i that is sufficient to cause cell death via caspase-dependent and -independent pathways. The caspase-dependent pathway involves release of cytochrome c from mitochondria and subsequent binding to Apaf-1 (forming the apoptosome) to cause the cleavage and activation of caspases involved in apoptosis (caspase 9 and 3). MTP opening also causes the increase of ROS due to the loss of MMP; this causes the rupture of lysosomes and release of enzymes (cathepsins) that activate pro-apoptotic proteins (Bax/Bak) involved in the mitochondrial apoptotic pathway. 2+ The prolonged high levels of [Ca ]i caused by MeHg can also activate a caspase-independent 2+ pathway via the activation of Ca -dependent calpains. Calpains in turn can result in the release of AIF, which translocates to the nucleus and causes the condensation and fragmentation of DNA, leading to apoptosis. Image from Ceccatelli, et al. (2010). 117 Figure 5.1 (cont’d) ROS Cell Membrane Cyt c Mito AIF Bax/Bak Lysosomes Cathepsins Pro-caspase-9 Caspase-9 tBid Calpains Pro-caspase-3 Bid 2+ Caspase-3 [Ca ]i Proteolysis Apoptosome Apaf-1 ER Nucleus 118 APPENDIX EFFECT OF METHYLMERCURY ON LETHARGIC CEREBELLAR GRANULE CELL VIABILITY IN VITRO 119 A. Abstract To date, VGCCs are known to be composed of five subunits: α1, β1-4, α2δ, and occasionally a γ subunit. The function of the β4 subunit is to aid in the transport of α1A and α1B to the plasma membrane and modify the kinetic properties of the P/Q- and N-type VGCCs, respectively. We previously presented data from in vivo and ex vivo MeHg exposure experiments of CGCs with a deleterious point mutation in α1A. In the present experiments, we explore the effect of an essentially-null mutation in Cacnb4, the gene that codes for the β4 subunit of VGCCs, on MeHg-induced cytotoxicity in vitro. The lethargic (lh) mouse contains a frameshift mutation in Cacnb4 that results in a β4 subunit protein lacking 60% of its C-terminus, rendering it unable to bind to α1A and some of the α1B subunits. Consequently, other β subunits (mainly β1 and β3) associate with α1A and α1B, altering the kinetics of P/Q- and N-type VGCCs at the plasma membrane of CGCs. We hypothesized that, in a similar manner as the tg mutation, the lh mutation would alter the degree of susceptibility of CGCs to MeHg exposure. WT, +/lh and lh/lh CGCs were acutely exposed in vitro to 0.5, 1, and 3 µM MeHg under depolarizing conditions (mimicking the resting membrane potential of CGCs during development in vivo) at DIV 6 and 10. Similar to tg/tg CGCs, there was an increased percentage of cell death in lh/lh CGC preparations that were not treated with MeHg, which could indicate an increased susceptibility to + the high extracellular K conditions. On DIV 6, +/lh and lh/lh CGC cultures show a higher percentage of cell death than WT CGC preparations. However, at DIV 10, lh/lh CGCs show higher resistance to the cytotoxic effects of acute MeHg exposure. The data presented in this 120 chapter suggest that changes in the β subunit associated with the pore-forming subunits of the most abundant VGCC subtypes in CGCs can alter the susceptibility of these neurons to MeHg exposure. Moreover, we provide additional evidence of the crucial role that VGCCs play in the mechanism of MeHg-induced cytotoxicity. 121 B. Introduction The lh mouse, another Cav2.1 channelopathy model of human epilepsy, is the result of a 4bp insertion in an intron of Cacnb4 which codes for the β4 subunit of P/Q-type VGCCs (Fig. 1.1). The β4 subunit is believed to have a main role in directing the α1A subunit of P/Q-type VGCCs to the plasma membrane, where it modulates the channel’s voltage-dependence of activation and inactivation (De Waard and Campbell, 1995; Buraei and Yang, 2012). However, the lh frameshift mutation causes abnormal splicing of the β4 mRNA and results in a truncated protein lacking more than 60% of its C-terminus and the motif that binds to the α1A subunit (Burgess et al., 1997). Interestingly, the expression of α1A and its presence in the plasma membrane are unaffected in the cerebellum of lh mice, suggesting that other β subunits interact with α1A to maintain some P/Q-type VGCC function (McEnery et al., 1998; Burgess et al., 1999). This β subunit reshuffling is only able to compensate partially for the loss of β4, but is sufficient to allow some of the mice that are homozygotes for the lh mutation to reach adulthood. The lh mouse presents a similar phenotype as the tg mouse (i.e. ataxia and spontaneous focal motor seizures), but with an earlier onset at approximately PND 15 (Burgess et al., 1997). Given that the lh mutation potentially causes changes in the kinetic properties of the P/Qand N-type VGCCs, the present experiments were performed to study in vitro the effects of acute MeHg exposure on CGCs of the lh mouse. CGCs were treated for 2 hrs with MeHg on DIV 6 and 10. Cytotoxicity was assessed by a calcein-AM/EthD-1 assay at various time points after MeHg treatment. 122 C. Materials and methods Cerebellar granule cell isolation. The protocols for isolation and maintenance of primary cultures of CGCs are described in Chapter Two. However, for these experiments, CGCs were plated in 24-well plates, instead of coverslips, treated with a 0.1 mg/mL solution of poly-d+ lysine. Cells were cultured in 25 mM K DMEM (supplemented as described in Chapter Two). MeHg exposure. The protocol for MeHg treatment is described in Chapter Three. However, MeHg treatments were only performed at DIV 6 and 10. Calcein-AM and ethidium homodimer-1 cytotoxicity assay. The cytotoxicity protocol was performed as described in Chapter Three. Automated quantification of CGCs with ImageJ. The percent of dead CGCs was calculated as described in Chapter Three. Statistical analysis. The SPSS program (IBM, NY) was used for all statistical analyses. Data were analyzed using a mixed design of two-way ANOVA with repeated measures: one betweenfactor (genotype), one within factor (concentration). Pairwise comparisons between genotypes at different concentrations of MeHg were performed using a Bonferroni adjustment. Statistically significant differences were set at p < 0.05 for all analyses. Measurements are expressed as mean ± S.E.M.; n ≥ 3 for all means, where “n” refers to a mouse. The statistical power was > 0.8 for the 4-hr mark at DIV 6 and DIV 10. 123 D. Results Effect of acute exposure to MeHg on the viability of WT, +/lh, and lh/lh CGCs cultured in + 25 mM K . The present studies were designed to investigate the effect of acute exposure to MeHg on + cell viability of lh/lh CGCs cultured in high K . At DIV 6, we observe a concentrationdependent response 8 and 24 hrs after MeHg treatment (Fig. A.2.A). +/lh (70.8 ± 2.1%) and lh/lh (45.7 ± 7.4%) cultures show higher percentage of CGC death than WT (10.2 ± 0.56%) at the earliest time point assayed after 3 µM MeHg. However, even though it did not reach statistical significance, we cannot ignore the fact that at this time point +/lh (35.0 ± 7.2%) and lh/lh (31.1 ± 5.3%) cultures show elevated levels of cell death without MeHg treatment compared to WT (10.8 ± 1.3%). By 24 hrs after 1 and 3 µM MeHg, +/lh cultures still show higher percentage of CGC death compared to WT and +/lh. At this time point, the differences in MeHg-induced cell mortality between lh/lh and WT did not reach statistical significance. At DIV 10, we observe a significant concentration-dependent response at all time points assayed. Both +/lh (24.6 ± 1.9%) and lh/lh (29.4 ± 2.4%) CGC cultures show higher cytotoxicity without MeHg treatment compared to WT (15.3 ± 2.3%) (Fig. A.2.B). However, this difference is only statistically significant at the 4 hr time point. Therefore, all significant differences observed at this time point are due to the difference in cell mortality observed without MeHg treatment. Interestingly, whereas at DIV 6 +/lh CGCs are more susceptible to MeHg, at DIV 10 this effect is reversed so that 24 hrs after a 1 µM MeHg treatment lh/lh (28.0 ± 0.8%) cultures show significantly less percentage of cell death than WT (45.8 ± 3.5%). 124 E. Discussion Data from the experiments described in this chapter show that at DIV 6, under high K+ conditions, +/lh cultures show higher percentage of CGC death than WT or lh/lh (Fig. A.2.A). It is noteworthy that +/lh and lh/lh CGC cultures show elevated levels of cell death in untreated controls in a similar manner as do tg/tg cultures (Fig. 3.1). We also observe this effect in untreated control cultures at DIV 10 (Fig. A.2.B). WT CGCs in vivo present a highly depolarized resting membrane potential in their immature state (first postnatal week) (Okazawa et al., 2009). However, there is no information available on the resting membrane potential of tg and lh CGCs during their early stages of development. Since both mutations cause an increase in cell mortality without MeHg treatment, it is plausible that sustained depolarized conditions (~ -55 mV) are toxic to these neurons, albeit more in lh than in tg CGCs. The effect of MeHg on lh cell viability changes from higher susceptibility at DIV 6 to higher resistance at DIV 10, compared to WT CGCs. This could be explained by a developmental change in either the β subunits associated with P/Q- and N-type VGCCs, or VGCC composition over time. Immunocytochemistry on the CGC cultures would have to be performed to study these hypotheses. However, it has been suggested that during the first and second postnatal weeks the N-type VGCC is predominantly responsible for Ca 2+ influx (reference needed). The α1B pore-forming subunit of N-type VGCCs also normally associates with the β4 subunit. Therefore, it is to be expected that either expression, function, or both for this VGCC subtype is also altered in lh mice. Indeed, in the lh cerebellum α1B also shows an increased association with β1B and β3 (Burgess et al., 1999). This combination could be responsible for the increased cell mortality observed in lh/lh compared to WT cultures after MeHg treatment. On the other hand, 125 the decreased susceptibility to MeHg observed in lh/lh at DIV 10 could be explained by the reported decrease in expression of N-type VGCCs in forebrain and cerebellum lysates of adult lh mice (McEnery et al., 1998). It is around this time point that there is normally an increase in the contribution of P/Q-type VGCCs on Ca 2+ influx in vivo. However, WT and lh P/Q-type VGCCs differ in their subunit composition. There is an increased association of α1A with β3 and β1B in whole brain lysates of adult lh mice (Burgess et al., 1999). Heterologous coexpression of α1A and β3 in Xenopus laevis oocytes has revealed that this combination requires a larger depolarization step to open than does an α1Aβ4 combination. In addition, the α1Aβ3 combination has a faster inactivation rate than α1Aβ4 (De Waard and Campbell, 1995). There are contradicting reports on the effects of the lh mutation on overall Ca brain, and no information of its effect in Ca depolarization-induced 45 Ca 2+ 2+ 2+ current density in the influx in CGCs. Lin, et al. reported a decrease in uptake in isolated cerebral cortex synaptosomes of 8-week old lh mice (Lin et al., 1999). This decrease in Ca 2+ uptake may be the cause behind the decrease in glutamatergic synaptic transmission observed in neurons of the thalamus of lh mice (Caddick et al., 1999). Qian, et al., on the other hand, found that presynaptic Ca 2+ influx is not affected in lh CA3-CA1 synapses of the hippocampus (Qian and Noebels, 2000). In Purkinje cells, the lh mutation and compensation by β1B and β3 does not cause a significant change in whole-cell Ptype current nor the channel’s voltage-dependence of activation and inactivation (Burgess et al., 126 1999). Regardless of the effect of the mutation on overall Ca 2+ influx, the presence of an α1Aβ3 VGCC combination could explain the resistance to MeHg observed in lh CGCs. Unfortunately, there is no information available in the current literature about the effect of lh-like mutations in humans. One report links the R482X frameshift mutation in CACNB4 to a case of juvenile myoclonic epilepsy (Escayg et al., 2000). However, this mutation only eliminates 10% of the β4 protein at the C-terminus and does not prevent it from binding to the α1A subunit. A point mutation located closer to the site of the lh mutation has been linked in humans to idiopathic generalized epilepsy and seizures (Escayg et al., 2000). But, again, this mutation does not prevent β4 subunit from associating with α1A. Nonetheless, the data shown in the present experiments provides further evidence of an important role of VGCCs in the mechanism of MeHg-induced cytotoxicity, and that changes in their kinetic properties can alter the effect of MeHg on CGCs. 127 Figure A.1: Location of the lethargic mutation in the β4 subunit of P/Q- and N-type VGCCs. The β4 subunit normally interacts directly with the α1A and α1B subunits of P/Q- and N-type VGCCs. The lh mutation (location in β4 subunit is indicated in the figure by an arrow) causes the loss of 60% of the C-terminus, which includes the domain that binds to the α1 subunit. Image from Burgess, et al. (1997). II I α1 subunit 12345 6 IV III 12345 6 1234 5 6 12345 6 NH2 COOH β subunit NH2 COOH 128 + Figure A.2: Lethargic cerebellar granule cell (cultured in 25 mM K ) viability after acute exposure to MeHg. CGCs maintained in depolarized conditions were exposed to 0.3, 1, and 3 µM MeHg for two hours. Calcein-AM and EthD-1 were used at a final concentration of 0.2 µM to assess the viability of cells after 4, 8, and 24 hrs of onset of MeHg exposure. Mean percent of cells stained with EthD-1 was calculated for each genotype on DIV 6 (A) and 10 (B) as a representative of dead cells. Data represent the mean ± S.E.M. (n ≥ 3). Asterisks (*) represent statistically significant difference (p < 0.05) between genotypes; pound signs (#) represent a significant concentration-dependent response. (A) DIV 6 +/lh WT lh/lh 100 % EthD-1-stained CGCs 80 60 40 20 0 0 0.3 1 3 0 0.3 1 3 8 hrs 4 hrs MeHg concentration (µM) 129 0 0.3 1 24 hrs 3 Figure A.2. (cont’d) (B) DIV 10 WT lh/lh +/lh 100 % EthD-1-stained CGCs 80 60 40 20 0 0 0.3 1 3 0 0.3 1 3 8 hrs 4 hrs MeHg concentration (µM) 130 0 0.3 1 24 hrs 3 BIBLIOGRAPHY 131 BIBLIOGRAPHY Atchison, W.D., 1986. Extracellular calcium-dependent and -independent effects of methylmercury on spontaneous and potassium-evoked release of acetylcholine at the neuromuscular junction. J Pharmacol Exp Ther 237, 672-680. Atchison, W.D., 2003. Effects of toxic environmental contaminants on voltage-gated calcium channel function: from past to present. J Bioenerg Biomembr 35, 507-532. Atchison, W.D., Adgate, L., Beaman, C.M., 1988. Effects of antibiotics on uptake of calcium into isolated nerve terminals. J Pharmacol Exp Ther 245, 394-401. Atchison, W.D., Joshi, U., Thornburg, J.E., 1986. Irreversible suppression of calcium entry into nerve terminals by methylmercury. J Pharmacol Exp Ther 238, 618-624. Bakir, F., Damluji, S.F., Amin-Zaki, L., Murtadha, M., Khalidi, A., al-Rawi, N.Y., Tikriti, S., Dahahir, H.I., Clarkson, T.W., Smith, J.C., Doherty, R.A., 1973. Methylmercury poisoning in Iraq. Science 181, 230-241. Barbado, M., Fablet, K., Ronjat, M., De Waard, M., 2009. Gene regulation by voltage-dependent calcium channels. Biochim Biophys Acta 1793, 1096-1104. Barrett, C.F., Rittenhouse, A.R., 2000. Modulation of N-type calcium channel activity by Gproteins and protein kinase C. J Gen Physiol 115, 277-286. Bawa, B., Abbott, L.C., 2008. Analysis of calcium ion homeostasis and mitochondrial function in cerebellar granule cells of adult CaV2.1 calcium ion channel mutant mice. Neurotox Res 13, 1-18. Bellum, S., Bawa, B., Thuett, K.A., Stoica, G., Abbott, L.C., 2007. Changes in biochemical processes in cerebellar granule cells of mice exposed to methylmercury. Int J Toxicol 26, 261269. Benarroch, E.E., 2010. Neuronal voltage-gated calcium channels: brief overview of their function and clinical implications in neurology. Neurology 74, 1310-1315. 132 Bridges, C.C., Zalups, R.K., 2010. Transport of inorganic mercury and methylmercury in target tissues and organs. J Toxicol Environ Health B Crit Rev 13, 385-410. Brustovetsky, N., Brustovetsky, T., Jemmerson, R., Dubinsky, J.M., 2002. Calcium-induced cytochrome c release from CNS mitochondria is associated with the permeability transition and rupture of the outer membrane. J Neurochem 80, 207-218. 2+ Buraei, Z., Yang, J., 2012. Structure and function of the beta subunit of voltage-gated Ca channels. Biochim Biophys Acta. Burgess, A., Vigneron, S., Brioudes, E., Labbe, J.C., Lorca, T., Castro, A., 2010. Loss of human Greatwall results in G2 arrest and multiple mitotic defects due to deregulation of the cyclin BCdc2/PP2A balance. Proc Natl Acad Sci U S A 107, 12564-12569. Burgess, D.L., Biddlecome, G.H., McDonough, S.I., Diaz, M.E., Zilinski, C.A., Bean, B.P., 2+ Campbell, K.P., Noebels, J.L., 1999. beta subunit reshuffling modifies N- and P/Q-type Ca channel subunit compositions in lethargic mouse brain. Mol Cell Neurosci 13, 293-311. 2+ Burgess, D.L., Jones, J.M., Meisler, M.H., Noebels, J.L., 1997. Mutation of the Ca channel beta subunit gene Cchb4 is associated with ataxia and seizures in the lethargic (lh) mouse. Cell 88, 385-392. Caddick, S.J., Wang, C., Fletcher, C.F., Jenkins, N.A., Copeland, N.G., Hosford, D.A., 1999. lh Excitatory but not inhibitory synaptic transmission is reduced in lethargic (Cacnb4 ) and tg tottering (Cacna1a ) mouse thalami. J Neurophysiol 81, 2066-2074. Campbell, D.B., Hess, E.J., 1999. L-type calcium channels contribute to the tottering mouse dystonic episodes. Mol Pharmacol 55, 23-31. Ceccatelli, S., Dare, E., Moors, M., 2010. Methylmercury-induced neurotoxicity and apoptosis. Chem Biol Interact 188, 301-308. Cho, S., Wood, A., Bowlby, M.R., 2007. Brain slices as models for neurodegenerative disease and screening platforms to identify novel therapeutics. Curr Neuropharmacol 5, 19-33. 133 Cicale, M., Ambesi-Impiombato, A., Cimini, V., Fiore, G., Muscettola, G., Abbott, L.C., de Bartolomeis, A., 2002. Decreased gene expression of calretinin and ryanodine receptor type 1 in tottering mice. Brain Res Bull 59, 53-58. Clarkson, T.W., 1995. Environmental contaminants in the food chain. Am J Clin Nutr 61, 682S686S. Clarkson, T.W., Strain, J.J., 2003. Nutritional factors may modify the toxic action of methyl mercury in fish-eating populations. J Nutr 133, 1539S-1543S. Clarkson, T.W., Vyas, J.B., Ballatori, N., 2007. Mechanisms of mercury disposition in the body. Am J Ind Med 50, 757-764. Currie, K.P., Fox, A.P., 1997. Comparison of N- and P/Q-type voltage-gated calcium channel current inhibition. J Neurosci 17, 4570-4579. De Waard, M., Campbell, K.P., 1995. Subunit regulation of the neuronal alpha 1A Ca expressed in Xenopus oocytes. J Physiol 485 ( Pt 3), 619-634. 2+ channel Dolphin, A.C., 1998. Mechanisms of modulation of voltage-dependent calcium channels by G proteins. J Physiol 506 ( Pt 1), 3-11. Erickson, M.A., Haburcak, M., Smukler, L., Dunlap, K., 2007. Altered functional expression of Purkinje cell calcium channels precedes motor dysfunction in tottering mice. Neuroscience 150, 547-555. Escayg, A., De Waard, M., Lee, D.D., Bichet, D., Wolf, P., Mayer, T., Johnston, J., Baloh, R., Sander, T., Meisler, M.H., 2000. Coding and noncoding variation of the human calcium-channel β4-subunit gene CACNB4 in patients with idiopathic generalized epilepsy and episodic ataxia. Am J Hum Genet 66, 1531-1539. Farrant, M., Nusser, Z., 2005. Variations on an inhibitory theme: phasic and tonic activation of GABAA receptors. Nat Rev Neurosci 6, 215-229. Fletcher, C.F., Lutz, C.M., O'Sullivan, T.N., Shaughnessy, J.D., Jr., Hawkes, R., Frankel, W.N., Copeland, N.G., Jenkins, N.A., 1996. Absence epilepsy in tottering mutant mice is associated with calcium channel defects. Cell 87, 607-617. 134 Fox, A.P., Nowycky, M.C., Tsien, R.W., 1987. Single-channel recordings of three types of calcium channels in chick sensory neurones. J Physiol 394, 173-200. Gallo, V., Kingsbury, A., Balazs, R., Jorgensen, O.S., 1987. The role of depolarization in the survival and differentiation of cerebellar granule cells in culture. J Neurosci 7, 2203-2213. Gavet, O., Pines, J., 2010. Progressive activation of CyclinB1-Cdk1 coordinates entry to mitosis. Dev Cell 18, 533-543. Grynkiewicz, G., Poenie, M., Tsien, R.Y., 1985. A new generation of Ca greatly improved fluorescence properties. J Biol Chem 260, 3440-3450. 2+ indicators with Haines, D.E., Dietrichs, E., 2012. The cerebellum - structure and connections. Handbook of clinical neurology 103, 3-36. Hammond, C., 2001. Cellular and molecular neurobiology. Academic Press, San Diego. Harada, M., 1995. Minamata disease: methylmercury poisoning in Japan caused by environmental pollution. Crit Rev Toxicol 25, 1-24. Hashimoto, M., Hibi, M., 2012. Development and evolution of cerebellar neural circuits. Dev Growth Differ 54, 373-389. Herden, C.J., Pardo, N.E., Hajela, R.K., Yuan, Y., Atchison, W.D., 2008. Differential effects of methylmercury on gamma-aminobutyric acid type A receptor currents in rat cerebellar granule and cerebral cortical neurons in culture. J Pharmacol Exp Ther 324, 517-528. Hunter, D., Russell, D.S., 1954. Focal cerebellar and cerebellar atrophy in a human subject due to organic mercury compounds. J Neurol Neurosurg Psychiatry 17, 235-241. Ito, M., 2006. Cerebellar circuitry as a neuronal machine. Prog Neurobiol 78, 272-303. Jensen, S., Jernelov, A., 1969. Biological methylation of mercury in aquatic organisms. Nature 223, 753-754. 135 Johnson, F.O., Yuan, Y., Hajela, R.K., Chitrakar, A., Parsell, D.M., Atchison, W.D., 2011. Exposure to an environmental neurotoxicant hastens the onset of amyotrophic lateral sclerosis2+ 2+ like phenotype in human Cu /Zn superoxide dismutase 1 G93A mice: glutamate-mediated excitotoxicity. J Pharmacol Exp Ther 338, 518-527. Kaja, S., Hann, V., Payne, H.L., Thompson, C.L., 2007. Aberrant cerebellar granule cell-specific GABAA receptor expression in the epileptic and ataxic mouse mutant, Tottering. Neuroscience 148, 115-125. Leenders, A.G., van den Maagdenberg, A.M., Lopes da Silva, F.H., Sheng, Z.H., Molenaar, P.C., Ghijsen, W.E., 2002. Neurotransmitter release from tottering mice nerve terminals with reduced 2+ expression of mutated P- and Q-type Ca -channels. Eur J Neurosci 15, 13-18. Limke, T.L., Atchison, W.D., 2002. Acute exposure to methylmercury opens the mitochondrial permeability transition pore in rat cerebellar granule cells. Toxicol Appl Pharmacol 178, 52-61. 2+ Limke, T.L., Bearss, J.J., Atchison, W.D., 2004. Acute exposure to methylmercury causes Ca dysregulation and neuronal death in rat cerebellar granule cells through an M3 muscarinic receptor-linked pathway. Toxicol Sci 80, 60-68. Lin, F., Barun, S., Lutz, C.M., Wang, Y., Hosford, D.A., 1999. Decreased 45 Ca 2+ uptake in P/Q- lh type calcium channels in homozygous lethargic (Cacnb4 ) mice is associated with increased β3 and decreased β4 calcium channel subunit mRNA expression. Brain Res Mol Brain Res 71, 1-10. Lonchamp, E., Dupont, J.L., Beekenkamp, H., Poulain, B., Bossu, J.L., 2006. The mouse cerebellar cortex in organotypic slice cultures: an in vitro model to analyze the consequences of mutations and pathologies on neuronal survival, development, and function. Crit Rev Neurobiol 18, 179-186. Manalis, R.S., Cooper, G.P., 1975. Evoked transmitter release increased by inorganic mercury at frog neuromuscular junction. Nature 257, 690-691. Mancini, J.D., Autio, D.M., Atchison, W.D., 2009. Continuous exposure to low concentrations of methylmercury impairs cerebellar granule cell migration in organotypic slice culture. Neurotoxicology 30, 203-208. 136 Markowski, V.P., Flaugher, C.B., Baggs, R.B., Rawleigh, R.C., Cox, C., Weiss, B., 1998. Prenatal and lactational exposure to methylmercury affects select parameters of mouse cerebellar development. Neurotoxicology 19, 879-892. Marrero-Rosado B., Fox S.M., Hannon H.E., Atchison W.D., 2013. Effects of mercury and lead on voltage-gated calcium channel function. In: Kretsinger R., Uversky V., Permyakov E. (Ed.) Encyclopedia of Metalloproteins: SpringerReference. Springer-Verlag Berlin Heidelberg. 2+ Marty, M.S., Atchison, W.D., 1997. Pathways mediating Ca entry in rat cerebellar granule cells following in vitro exposure to methyl mercury. Toxicol Appl Pharmacol 147, 319-330. 2+ Marty, M.S., Atchison, W.D., 1998. Elevations of intracellular Ca as a probable contributor to decreased viability in cerebellar granule cells following acute exposure to methylmercury. Toxicol Appl Pharmacol 150, 98-105. McEnery, M.W., Copeland, T.D., Vance, C.L., 1998. Altered expression and assembly of N-type calcium channel α1B and β subunits in epileptic lethargic (lh/lh) mouse. J Biol Chem 273, 21435-21438. Mellor, J.R., Merlo, D., Jones, A., Wisden, W., Randall, A.D., 1998. Mouse cerebellar granule cell differentiation: electrical activity regulates the GABAA receptor α6 subunit gene. J Neurosci 18, 2822-2833. Moran, O., Lin, F., Zegarra-Moran, O., Sciancalepore, M., 1991. Voltage dependent calcium channels in cerebellar granule cell primary cultures. Eur Biophys J 20, 157-164. 2+ Nakanishi, S., Okazawa, M., 2006. Membrane potential-regulated Ca signalling in development and maturation of mammalian cerebellar granule cells. J Physiol 575, 389-395. Nierenberg, D.W., Nordgren, R.E., Chang, M.B., Siegler, R.W., Blayney, M.B., Hochberg, F., Toribara, T.Y., Cernichiari, E., Clarkson, T., 1998. Delayed cerebellar disease and death after accidental exposure to dimethylmercury. N Engl J Med 338, 1672-1676. Nuttall, K.L., 2006. Interpreting hair mercury levels in individual patients. Ann Clin Lab Sci 36, 248-261. 137 Obata, K., Ito, M., Ochi, R., Sato, N., 1967. Pharmacological properties of the postsynaptic inhibition by Purkinje cell axons and the action of gamma-aminobutyric acid on deiters neurones. Experimental brain research. Experimentelle Hirnforschung. Experimentation cerebrale 4, 43-57. Okazawa, M., Abe, H., Katsukawa, M., Iijima, K., Kiwada, T., Nakanishi, S., 2009. Role of calcineurin signaling in membrane potential-regulated maturation of cerebellar granule cells. J Neurosci 29, 2938-2947. Pardo, N.E., Hajela, R.K., Atchison, W.D., 2006. Acetylcholine release at neuromuscular junctions of adult tottering mice is controlled by N-(Cav2.2) and R-type (Cav2.3) but not L-type 2+ (Cav1.2) Ca channels. J Pharmacol Exp Ther 319, 1009-1020. Passos, C.J., Mergler, D., 2008. Human mercury exposure and adverse health effects in the Amazon: a review. Cad Saude Publica 24 Suppl 4, s503-520. Pietrobon, D., 2002. Calcium channels and channelopathies of the central nervous system. Mol Neurobiol 25, 31-50. Plomp, J.J., Vergouwe, M.N., Van den Maagdenberg, A.M., Ferrari, M.D., Frants, R.R., Molenaar, P.C., 2000. Abnormal transmitter release at neuromuscular junctions of mice carrying 2+ the tottering α1A Ca channel mutation. Brain 123 Pt 3, 463-471. 2+ Qian, J., Noebels, J.L., 2000. Presynaptic Ca influx at a mouse central synapse with Ca channel subunit mutations. J Neurosci 20, 163-170. 2+ Rajakulendran, S., Kaski, D., Hanna, M.G., 2012. Neuronal P/Q-type calcium channel dysfunction in inherited disorders of the CNS. Nat Rev Neurol 8, 86-96. Randall, A., Tsien, R.W., 1995. Pharmacological dissection of multiple types of Ca currents in rat cerebellar granule neurons. J Neurosci 15, 2995-3012. 2+ channel Rossi, P., D'Angelo, E., Magistretti, J., Toselli, M., Taglietti, V., 1994. Age-dependent expression of high-voltage activated calcium currents during cerebellar granule cell development in situ. Pflugers Arch 429, 107-116. 138 Rossi, P., De Filippi, G., Armano, S., Taglietti, V., D'Angelo, E., 1998. The weaver mutation causes a loss of inward rectifier current regulation in premigratory granule cells of the mouse cerebellum. J Neurosci 18, 3537-3547. Sakamoto, M., Ikegami, N., Nakano, A., 1996. Protective effects of Ca against methyl mercury toxicity. Pharmacology & toxicology 78, 193-199. 2+ channel blockers Sanfeliu, C., Sebastia, J., Cristofol, R., Rodriguez-Farre, E., 2003. Neurotoxicity of organomercurial compounds. Neurotox Res 5, 283-305. Sarafian, T., Verity, M.A., 1991. Oxidative mechanisms underlying methyl mercury neurotoxicity. Int J Dev Neurosci 9, 147-153. Sato, M., Suzuki, K., Yamazaki, H., Nakanishi, S., 2005. A pivotal role of calcineurin signaling in development and maturation of postnatal cerebellar granule cells. Proc Natl Acad Sci U S A 102, 5874-5879. 2+ Sawada, K., Haga, H., Fukui, Y., 2000. Ataxic mutant mice with defects in Ca channel α1A subunit gene: morphological and functional abnormalities in cerebellar cortical neurons. Congenital Anomalies 40, 99-107. Shafer, T.J., Atchison, W.D., 1989. Block of 2+ 45 Ca uptake into synaptosomes by methylmercury: + Ca - and Na -dependence. J Pharmacol Exp Ther 248, 696-702. 2+ Shafer, T.J., Atchison, W.D., 1991. Methylmercury blocks N- and L-type Ca channels in nerve growth factor-differentiated pheochromocytoma (PC12) cells. J Pharmacol Exp Ther 258, 149157. Shafer, T.J., Contreras, M.L., Atchison, W.D., 1990. Characterization of interactions of 2+ methylmercury with Ca channels in synaptosomes and pheochromocytoma cells: radiotracer flux and binding studies. Mol Pharmacol 38, 102-113. Sirois, J.E., Atchison, W.D., 2000. Methylmercury affects multiple subtypes of calcium channels in rat cerebellar granule cells. Toxicol Appl Pharmacol 167, 1-11. 139 Suarez-Kurtz, G., 1989. Effects of aminoglycoside antibiotics on calcium action potentials and calcium channel currents. Acta Physiol Pharmacol Latinoam 39, 407-418. Syversen, T.L., Totland, G., Flood, P.R., 1981. Early morphological changes in rat cerebellum caused by a single dose of methylmercury. Arch Toxicol 47, 101-111. Takayama, C., 2005. Formation of GABAergic synapses in the cerebellum. Cerebellum 4, 171177. Takayama, C., Inoue, Y., 2006. Developmental localization of potassium chloride co-transporter 2 in granule cells of the early postnatal mouse cerebellum with special reference to the synapse formation. Neuroscience 143, 757-767. Tsuda, T., Yorifuji, T., Takao, S., Miyai, M., Babazono, A., 2009. Minamata disease: catastrophic poisoning due to a failed public health response. J Public Health Policy 30, 54-67. Varming, T., Christopherson, P., Schousboe, A., Drejer, J., 1997. Pharmacological characterisation of voltage-sensitive calcium channels and neurotransmitter release from mouse cerebellar granule cells in culture. Journal of neuroscience research 48, 43-52. Wakamori, M., Yamazaki, K., Matsunodaira, H., Teramoto, T., Tanaka, I., Niidome, T., Sawada, K., Nishizawa, Y., Sekiguchi, N., Mori, E., Mori, Y., Imoto, K., 1998. Single tottering mutations responsible for the neuropathic phenotype of the P-type calcium channel. J Biol Chem 273, 34857-34867. Wise-Faberowski, L., Robinson, P.N., Rich, S., Warner, D.S., 2009. Oxygen and glucose deprivation in an organotypic hippocampal slice model of the developing rat brain: the effects on N-methyl-D-aspartate subunit composition. Anesth Analg 109, 205-210. Xiang, Z., Hrabetova, S., Moskowitz, S.I., Casaccia-Bonnefil, P., Young, S.R., Nimmrich, V.C., Tiedge, H., Einheber, S., Karnup, S., Bianchi, R., Bergold, P.J., 2000. Long-term maintenance of mature hippocampal slices in vitro. J Neurosci Methods 98, 145-154. Xu, H., Yang, Y., Tang, X., Zhao, M., Liang, F., Xu, P., Hou, B., Xing, Y., Bao, X., Fan, X., 2013. Bergmann glia function in granule cell migration during cerebellum development. Mol Neurobiol 47, 833-844. 140 Yuan, Y., Atchison, W.D., 1997. Action of methylmercury on GABAA receptor-mediated inhibitory synaptic transmission is primarily responsible for its early stimulatory effects on hippocampal CA1 excitatory synaptic transmission. J Pharmacol Exp Ther 282, 64-73. Yuan, Y., Atchison, W.D., 2003. Methylmercury differentially affects GABAA receptormediated spontaneous IPSCs in Purkinje and granule cells of rat cerebellar slices. J Physiol 550, 191-204. Yuan, Y., Otero-Montanez, J.K., Yao, A., Herden, C.J., Sirois, J.E., Atchison, W.D., 2005. Inwardly rectifying and voltage-gated outward potassium channels exhibit low sensitivity to methylmercury. Neurotoxicology 26, 439-454. Zamponi, G.W., 2005. Voltage-gated calcium channels. Landes Bioscience/Eurekah.com ; Kluwer Academic/Plenum Publishers, Georgetown, Tex. New York, N.Y. Zhang, J.F., Randall, A.D., Ellinor, P.T., Horne, W.A., Sather, W.A., Tanabe, T., Schwarz, T.L., 2+ Tsien, R.W., 1993. Distinctive pharmacology and kinetics of cloned neuronal Ca channels and their possible counterparts in mammalian CNS neurons. Neuropharmacology 32, 1075-1088. Zhou, Y.D., Turner, T.J., Dunlap, K., 2003. Enhanced G protein-dependent modulation of 2+ excitatory synaptic transmission in the cerebellum of the Ca channel-mutant mouse, tottering. J Physiol 547, 497-507. 141