THEME 2 IN VIVO EXPERIMENTAL MODELS

P18 MESENCHYMAL STEM CELL TRANSPLANTATION IN AN ALS MOUSE MODEL

WATANABE Y¹, MORITA ERI¹, NAKANO T¹, KITAYAMA M¹, YASUI K¹, DOI K¹, KARUNARATNE A², MURRELL W², MACKAY-SIM A², NAKASHIMA K¹

¹Tottori University, Yonago, Japan, ²Griffith University, Brisbane, Australia

E-mail address for correspondence: [email protected]

Keywords: transplantation, mesenchymal stem cell, transgenic mice

Background: ALS selectively affects motor neurons throughout the central nervous system (CNS) from the cerebral cortex to the lumbar spinal cord. This wide distribution of motor neurons is an obstacle in applying cell transplantation therapy for the treatment of ALS. We developed a transplantation protocol that enables transplanted cells to broadly diffuse throughout the CNS. We applied this method to the treatment of an ALS mouse model.

Objectives: To examine whether the transplantation of mesenchymal stem cells (MSCs) into the CNS shows any beneficial effect on clinical or pathological aspects of disease. The use of MSCs is advantageous since they have the capability to differentiate into several specific cell types including neural cells; furthermore, they can be relatively easily obtained by biopsy in humans, thus making autologous transplantation possible.

Methods: MSCs were derived from 5-week-old rats that ubiquitously express green fluorescent protein (GFP). As the host we used SOD1^Leu126delTT transgenic mice which express a mutated form of SOD1 with a 2-bp deletion at codon 126. The MSCs were transplanted into 100-day old SOD1^Leu126delTT transgenic mice via the fourth ventricle and a control group of mice received a sham operation. From one week prior to the transplantation until death occurred, clinical evaluations of body weight, hind limb extension reflex score, and footprint analysis were performed once a week. The MSCs in the host spinal cord were pathologically observed by immunofluorescent staining of frozen sections.

Results: There were beneficial tendencies in age of onset, death, and disease duration in MSC-transplanted mice compared to control mice, but these did not reach statistical significance. However, when we compared only MSC-treated female mice with sham-operated female mice, we could find a statistically significant difference for disease duration between the two groups (MSC; 21.6±5.4 days vs. control; 10.0±2.2 days). On pathological examination, we confirmed the existence of transplanted MSCs in the anterior horns of transgenic SOD1^Leu126delTT animals. Although there was a trend that the MSC group's motor neurons out numbered those of the sham group, the remaining motor neuron count immediately before death was similar between MSC transplanted animals and sham operated animals.

Conclusions: MSCs modestly prolong disease duration in female transgenic SOD1^Leu126delTT animals. Therefore, our transplantation method could be a promising way to deliver donor cells to the CNS. Further experiments to optimize timing and frequency of the cell transplantation are necessary. Additionally, it might prove effective to use genetically modified cells expressing a growth factor, such as the glial cell line-derived neurotrophic factor or insulin-like growth factor.

P19 DIGESTION OF THE EXTRACELLULAR CHONDROITIN SULFATE PROMOTES AN INTRINSIC REGENERATIVE PROCESS IN THE SPINAL CORD OF ALS TRANSGENIC RATS

WARITA H, MIZUNO H, AOKI M, ITOYAMA Y

¹Tohoku University Graduate School of Medicine, Sendai, ²Tohoku University Hospital ALS Center, Sendai, Japan

E-mail address for correspondence: [email protected]

Keywords: SOD transgenic microenvironment

Background: Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease characterized by adult-onset selective motor neuronal loss. Approximately 2% of all ALS cases are linked to mutations in the Cu/Zn superoxide dismutase (SOD1) gene. Upregulation of extracellular chondroitin sulfate proteoglycans (CSPGs) under various types of CNS insults is known to restrict intrinsic regenerative processes such as neurite outgrowth, synaptogenesis and cell migration.

Objectives: In order to clarify the role extracellular CSPGs under neurodegenerative condition such as ALS, we tested the effect of enzymatic degradation of CSPGs in the spinal cord of a transgenic rat model of ALS.

Methods: We examined the expression of CSPGs in the lumbar spinal cord of His46Arg mutant SOD1 transgenic (SOD1^H46R Tg) rats at presymptomatic, early symptomatic, and late symptomatic stages with their age-matched non-transgenic (non-Tg) littermates. Based on the results, a bacterial enzyme chondroitinase ABC (ChABC) was continuously infused into the subarachnoid space of mid-symptomatic SOD1^H46R Tg rats for 7 days. Over the same period, we also infused a thymidine analogue bromodeoxyuridine subcutaneously to label the newborn cells in vivo. After the infusion, we performed immunohistochemical analysis in the lumbar spinal cord to confirm the effect of ChABC. In addition, we compared the neuropathology, proliferative neural progenitors, and synaptogenesis in the spinal ventral horns between ChABC-treated and vehicle-treated SOD1^H46R Tg rats.

Results: In contrast to non-Tg rats, the SOD1^H46R Tg rats showed significant and progressive upregulation of CSPGs in the ventral spinal cord even from presymptomatic stage. In the spinal cord parenchyma of ChABC-treated SOD1^H46R Tg rats, immunohistochemistry revealed significant reduction of the chondroitin sulfate deposition as compared with vehicle-treated rats. In addition, bromodeoxyuridine-incorporated newborn cells including immature phenotype, synaptophysin-positive structures, and neurons with immature phenotype were significantly increased in the ventral horns of ChABC-treated rats. On the other hand, neuropathology such as loss of ventral horn neurons, ubiquitinated protein aggregates, and phosphorylated neurofilament accumulation was not exacerbated in the ChABC-treated rats.

Discussion and Conclusions: In the present study, we found significant and progressive deposition of CSPGs in the spinal cord of transgenic rat model of ALS. Although a possible neuroprotective role of CSPGs remains to be argued, the present results suggest that the excessive deposition of CSPGs may play an inhibitory role against intrinsic regenerative processes under the ALS-like disease in the present model. Therefore, regulation of extracellular molecules in the microenvironment surrounding motor neurons could be an important strategy to develop cell-restorative therapy in ALS.

P20 EFFECTS OF EDARAVONE, A FREE RADICAL SCAVENGER APPROVED IN JAPAN FOR INDICATIONS OF ACUTE ISCHEMIC STROKE, IN A TRANSGENIC RAT MODEL OF AMYOTROPHIC LATERAL SCLEROSIS

AOKI M¹, WARITA H¹, MIZUNO H¹, YUKI S², TAKAHASHI I³, ITOYAMA Y¹

¹Department of Neurology, Tohoku University School of Medicine, Sendai, Japan, ²Mitsubishi Tanabe Pharma Coporation, Yokohama, Japan, ³Mitsubishi Chemical Safety Institute LTD, Kumamoto, Japan

E-mail address for correspondence: [email protected]

Keywords: H46R SOD1 transgenic rat, free radical, Edaravone

Background: The toxic gain-of-function of mutant SOD1 enhances oxidative activity, produces hydroxyl radical or peroxynitrite has previously been hypothesised. Therefore free radicals are presumed to contribute to the pathogenesis of ALS.

Objectives: We investigated the efficacy of edaravone, a free radical scavenger previously approved for acute ischemic stroke in Japan, using a rat model of ALS expressing a human SOD1 transgene with ALS associated mutation: histidine to arginine at position 46 (H46R).

Methods: H46R SOD1 transgenic rats of both sexes were used in this experiment.

Edaravone (1.5 and 3.0 mg/kg/h, male: n = 8, female: n = 8, in each group) and saline (male: n = 8, female: n = 8) were administered intravenously to rats by continuous infusion (1 hour infusion per day). Two days of administration was followed by 2 days of drug holiday. This series commenced from when the rats were 18 weeks old and was repeated until the day of loss of righting reflex.

The lifetime and duration of illness were evaluated, and besides these, motor function was assessed using expansion of limbs by tail suspension (suspension test), width between hind limb when landing from 30 cm height (landing test), rota rod test and inclined plate test.

The number of motor neurons of the third lumber spinal cord in the vehicle group and edaravone 3.0 mg/kg/h group were also examined at the time point of 24-weeks old.

Results: We defined the onset of disease as the day of the observation of postural change in one limb when the rat was lifted by the tail (suspension test). The death of animal was defined as the day when the righting reflex disappeared.

Lifetime of saline treated, edaravone low dose and edaravone high dose group was 196±3, 200±1 and 202±4 days respectively, in males and 212±4, 214±5 and 224±10 days, respectively in females. Duration of illness of saline treated, edaravone low dose and edaravone high dose group was 52.0±3.2, 54.3±2.1 and 57.6±5.2, respectively, in males and 67.6±4.6, 72.1±4.6 and 78.7±10.5 respectively, in females. Edaravone elongated both lifetime and duration of illness according to the dose escalation, however, did not reach significance.

Evaluation point of motor function was set up on the day that half of the saline-treated animals died. In suspension test, rota rod and inclined plate test, high dose of edaravone tended to improve both sexes, especially in landing test where a high dose of edaravone treated male group showed a significant improvement.

No difference was observed in number of motor neurons between the vehicle group and the edaravone treated group at the time point of 24 weeks old.

Discussion and Conclusions: In this study, we demonstrated the possibility of edaravone as a therapeutic agent for ALS using H46R SOD1 transgenic rat. Further studies are needed to confirm the efficacy of edaravone in the near future.

P21 MELATONIN IS DETRIMENTAL TO SURVIVAL IN A TRANSGENIC MOUSE MODEL OF FAMILIAL ALS

KYRIAKIDES T¹, RIZKI G¹, MARQUEZ B¹, VONTA I², MALAS S¹, FELDMAN M¹, HADJISAVVAS A¹, KYRIAKOU K¹

¹Cyprus Institute of Neurology and Genetics, Nicosia, ²University of Cyprus, Nicosia, Cyprus

E-mail address for correspondence: [email protected]

Keywords: Melatonin,G93A, mice

Background: There is experimental and human data that reactive oxygen species (ROS) and apoptosis are involved in the pathogenesis of amyotrophic lateral sclerosis (ALS). In familial ALS due to mutations of the Cu/Zn superoxide dismutase (SOD1) it is postulated that the mutated SOD1 exhibits a novel gain of function and acts as a supplier rather than a scavenger of ROS. In G93A SOD1 transgenic mice there is an enhanced production of ROS in the spinal cord, which precedes motor neuron degeneration. Melatonin is an endogenous neurohormone with well recognized anti-oxidant properties, both as a direct ROS scavenger as well as by up regulating anti-oxidant enzymes. Melatonin has also been shown to protect neurons from kainate induced glutamate-receptor excitotoxicity and to attenuate lipid peroxidation and microglial activation.

Objective: To assess the effect of melatonin on disease phenotype in the G93A-SOD1 transgenic mouse model of ALS.

Methods: Four groups of 10 mice each, carrying the human G93A-SOD1 mutation (BS6JL-TgN SOD1-G93A), were injected with 0 mg/kg, 0.5 mg/kg, 2.5 mg/kg and 50 mg/kg of melatonin intraperitoneally from the age of 40 days. The end points were disease onset, survival, rotarod performance and morphometry of lumbar cord motor neurons. SOD1 expression was studied with histochemistry, Western blot and Real Time PCR.

Results: There was no statistically significant difference in disease onset between the four groups although there was a tendency for the 2.5mg/kg dose to delay onset. Survival was significantly reduced (P < 0.05, two tailed t-test) with the 0.5mg/kg and 50mg/kg doses and tended to be reduced with the 2.5mg/kg dose. Histochemistry against SOD1 revealed higher expression of the protein in motor neurons of mice treated with melatonin. Western blot and Real Time PCR revealed a dose dependent up regulation of SOD1 in melatonin treated animals.

Conclusion and Discussion: Melatonin, contrary to expectations, reduces survival in the G93A mouse model of ALS despite its known anti-oxidant and anti-apoptotic properties. This phenomenon is probably accounted for by the up regulation of anti oxidative enzyme gene expression including, in the G93A transgenic mice, the mutated and toxic human SOD1. This action of melatonin probably overrides any of its beneficial effects. This study reinforces the notion that, at least in SOD1 associated ALS, ROS play a pathogenic role.

P22 THE CHICK EMBRYO AS A TOXICITY SCREEN FOR GENES IN AMYOTROPHIC LATERAL SCLEROSIS

TRIPATHI VB¹, SREEDHARAN J¹, GUTHRIE S², SHAW C¹, AL-CHALABI A¹

¹Institute of Psychiatry, King's College London, ²Department of Developmental Neurobiology, Guy's Campus, King's College London, United Kingdom

E-mail address for correspondence: [email protected]

Keywords: Chick, SOD1, TDP43

Background: Genetic technology has advanced to the point that large numbers of people are being tested for gene variants contributing to ALS. The chick embryo provides a model system in which effects on intact tissues can be examined quickly and simply in a cell and stage-specific way. We have used the chick embryo model in order to study the effects of gene mutations found by linkage studies.

Methods: We electroporated SOD1^G93A-dsRed, SOD1^WT-dsRed and HA and Myc tagged TDP-43^WT, TDP-43^Q331K and TDP-43^M337V DNA into the spinal cords of HH stage 14 chick embryos using in ovo electroporation. After 24 hours the embryos were processed for frozen sectioning and then apoptotic cell death was observed using TUNEL assay. Immunohistochemistry of neuronal markers was carried out in order to investigate the types of neurons affected.

Results: SOD1^G93A-dsRed transfected neurons showed cell death as demonstrated by TUNEL staining (no. of embryos analysed = 36/36). We also observed a decrease in the population of motor neurons (stained for Islet1/2) on the transfected side of the embryo as compared to the non-transfected control side.

Embryos expressing mutant TDP43 (n = 49) showed a dramatic reduction in maturation as observed by a failure to develop normal limb and tail buds. The percentage of mature embryos electroporated with TDP-43^Q331K was 11.66% and for TDP-43^M337V was 15% as compared to 97.66% for embryos electroporated with TDP-43^WT. TUNEL staining demonstrated a significant increase in the number of apoptotic nuclei in embryos expressing either mutant TDP-43 when compared to TDP-43^WT (TDP-43^Q331K n = 29.5; TDP-43^M337V n = 21.6, TDP-43^WT n = 5.6, where n is mean of TUNEL positive cells/section for 5 sections analysed from 3 embryos).

Conclusions: The chick model offers a relatively quick and cost-effective means of validating genetic discoveries and will provide valuable insights into disease mechanisms.

P23 NEUROTOXIC SPECIES OF MISFOLDED MUTANT SOD1 DISPLAYING CROSS-REACTIVITY WITH ANTIBODIES AGAINST P2X4 SUBUNIT OF THE ATP RECEPTOR ACCUMULATE IN DEGENERATING NEURONS BUT NOT IN GLIAL CELLS OF TRANSGENIC SOD1^G93A RODENTS

HERNÁNDEZ S, CASANOVAS A, PIEDRAFITA L, ESQUERDA JE

Universitat de Lleida, Spain

E-mail address for correspondence: [email protected]

Keywords: SOD1, P2X4, neurodegeneration, microglia

Background: We have recently described that degenerating motor neurons (MNs) in spinal cord, brainstem and cerebral cortex of SOD1^G93A rats or mice, exhibit an intense immunoreactivity to P2X4 antibodies [1]. Neurons with strong P2X4-like immunoreactivity (P2X4-LI) are often associated with microglial cells displaying neuronophagic activity and do not show apoptotic phenotype when dying. Degenerating neurons with strong P2X4-LI and microglial cell recruitment were also observed in the cerebellar cortex and in noradrenergic or serotoninergic systems of the brainstem. After Western blot analysis it was found that P2X4 antibodies recognize, in addition to P2X4-membrane associated protein, an unidentified low MW band seen in cytosolic extracts from SOD1^G93A but not in wild type (WT) samples.

Objectives: To identify of the molecular counterpart of the strong P2X4-LI observed in association of neuronal degeneration and death in SOD1^G93A animals.

Methods: Soluble spinal cord extracts were subjected to 2D electrophoresis and P2X4 immunoreactive spots were analyzed by proteomics. Further studies include: immunoprecipitation, multiple fluorescent labelling of tissue sections and confocal microscopy.

Results: On Western blots of spinal cord extracts, we have found that the antibody recognizes a protein band in membrane fraction with the expected size of P2X4 receptor protein (40–60 kDa) that is upregulated in end-stages of SOD1 G93A rats. After MALDI-TOF it was found that the low MW, P2X4-immunoreactive (P2X4-IR) protein band seen in the cytosolic fraction of transgenic animals was SOD1. This result was further confirmed by immunoprecipitation experiments showing that anti-P2X4 antibody is able to immunoprecipitate SOD1and that anti SOD1 antibody immunoprecipitates a P2X4-IR band migrating at the level of SOD1. By double labelling immunohistochemistry it was found that P2X4-LI in spinal cord tissue sections partially overlaps with SOD1 in spinal cord from mSOD1^G93A mice. Whereas in mSOD1^G93A samples SOD1 immunostaining was present in both neurons and glia, P2X4-LI was only detected in degenerating neurons, being absent in spinal cord tissue sections from mice overexpressing human WT SOD1. P2X4 antibodies were able to immunoprecipitate erythrocyte-derived unfolded SOD1 more efficiently than the folded one. Affinity chromatography-isolated P2X4-LI SOD1 species were injected intracerebrally in mice in order to examine its pro-inflammatory properties; preliminary results indicate that microglial activation by P2X4-LIR-SOD1^G93A was higher than that by human WT SOD1, used as a control. Studies on P2X4-LI in NSC34 motor neuron cell line expressing mutant SOD1 are in progress.

Discussion and Conclusions: From these results we deduce that P2X4 antibody recognizes a form of misfolded mutant SOD1 which is expressed only in neuronal cells undergoing degeneration. It seems that the cross-reactivity could be due to the abnormal exposure of hydrophobic regions of SOD1 sharing structural homology with the P2X4 immunizing peptide that was used for raising the antibody.

P24 SOLUBLE MISFOLDED SOD1 IN TRANSGENIC MICE IS MOSTLY OLIGOMERIC

ZETTERSTRÖM P¹, SIXTENSDOTTER-GRAFFMO K², BRÄNNSTRÖM T², MARKLUND SL¹

¹Umeå University, Medical Biosciences, Clinical Chemistry, Umeå, Sweden, ²Umeå University, Medical Biosciences, Pathology, Umeå, Sweden

E-mail address for correspondence: [email protected]

Keywords: Misfolded SOD1, Oligomers, Disulphide reduction

Background: 140 different mutations in SOD1 have been found in ALS cases and all should share a common neurotoxic mechanism. A common denominator of transgenic mice expressing various mutant SOD1s, is enrichment of soluble misfolded forms of the protein in the vulnerable spinal cord [1].

Aims: To determine the molecular structure of misfolded SOD1 in murine spinal cords. Antibodies specific for misfolded forms of SOD1 were used for the purpose.

Methods: Sandwich ELISAs (misELISAs) were developed using antibodies specific for misfolded SOD1 and directed at different parts of the molecule. Tissue extracts from transgenic mice were subjected to gel chromatography. Eluting fractions were analyzed for total SOD1 with Western immunoblots. Misfolded SOD1 in the fractions was analyzed with the misELISAs. The disulphide status of SOD1 was determined with non-reducing Western immunoblots [2].

Results: In G93A transgenic mice two peaks were seen in the chromatography when analysed for total SOD1. One large peak of 32 kDa composed of dimeric SOD1 and a smaller of 16 kDa with monomeric SOD1. The misELISAs showed two peaks, one at 140 kDa and one at 45 kDa. This would correspond to SOD1 nonamers and trimers. Less than 2% of the total SOD1 was found to react in the misELISAs. Analysis of D90A and G85R mice with the misELISAs show similar nonamer-trimer patterns with an additional small monomeric peak. The oligomeric SOD1 was not disulphide-coupled and lacked the disulphide bond.

Discussion: In many neurological diseases large protein aggregates are found late in the disease process. If the aggregates are toxic or if precursor molecular forms exert the toxicity is not known. We show here that misfolded SOD1 in tissues adopts several conformations including monomeric, tri and nonameric forms. Whether these are toxic remains to be shown.

P25 BIOCHEMICAL CHARACTERIZATION OF HUMAN G93A SOD1 MUTANT PURIFIED FROM TRANSGENIC RAT SPINAL CORD

BHOGARAJU V, REED R, CROW J

University of Arkansas Medical School, Little Rock, AR, United States

E-mail address for correspondence: [email protected]

Keywords: copper, zinc, cysteine

Background: Mutations to SOD1 remain the only proven cause of ALS, but the mechanism whereby mutants cause neurotoxicity remains unknown. In vitro studies with recombinant SOD1 mutants have been shrouded in controversy as to which properties are “real” and which are related to the necessarily artificial conditions used to express and purify them. To address this question, we sought a way to quickly and effectively purify SOD1 from transgenic rat spinal cord using very mild conditions, so that the true “as isolated" properties of a disease-causing mutant could be unequivocally established.

Objective: To compare properties of the human G93A mutant enzyme purified from spinal cord of both pre-symptomatic and endstage transgenic rats with those of recombinant G93A obtained from E. coli. Determination of the properties of G93A “as isolated” from spinal cord may provide clues as the toxic mechanism of SOD1 mutants, and allow us to better assess the suitability of recombinant enzymes to test various toxic hypotheses.

Methods: Spinal cords were harvested from 65 day old and endstage transgenic rats and homogenized, clarified, and the supernatants injected into a polymeric, reversed-phase preparative HPLC column maintained and eluted at pH 7.4 (Tris with linear acetonitrile gradient). The fraction corresponding to native homodimer was collected and overloaded onto an SDS-PAGE gel to assess purity. For each sample (n = 3 for each condition), Zn, Cu, and reduced thiol content were determined, and normalized to enzyme subunit based on rigorous determination of protein concentration. Specific dismutase activity was also determined.

Results: SDS-PAGE gels revealed that the one-step HPLC purification procedure yielded > 99% pure G93A from rat cord within 20 minutes of tissue homogenization. Analysis of pure enzyme revealed ∼ 50% of the maximal content of Zn and Cu. Reduced thiol (-SH) content of native enzyme was ∼ 0.2/subunit, and increased to ∼ 1.2 upon denaturation (the same was seen with recombinant G93A). Specific activity of rat cord G93A was ∼ 4,000 units/mg, however, because the Cu content was only 50% of maximal, the true specific activity is arguably ∼ 8,000 units/mg of copper-saturated enzyme.

Discussion: The results are consistent with the surface accessible cysteine (Cys111) existing in the oxidized state, whereas Cys6 is reduced. The ratios of Zn and Cu per subunit are consistent with one metal ion per dimer. Limiting amounts of CCS and Cu could explain the low Cu levels (when SOD1 is grossly overexpressed), but this would not explain the low levels of Zn. With the exception of somewhat lower Zn and Cu content, G93A from rat cord was essentially identical to enzyme obtained from E. coli, suggesting that recombinant enzymes are, indeed, suitable for examining toxic properties, particularly when wild-type enzyme is prepared using identical conditions.

P26 TRANSCRIPTIONAL ANALYSIS BY SUPPRESSION SUBTRACTIVE HYBRIDIZATION OF TEMPORAL AND CONDITIONAL CHANGES IN SPINAL CORD OF THE SOD1 G93A MOUSE MODEL OF FAMILIAL AMYOTROPHIC LATERAL SCLEROSIS

YAN Z, LATHIA K, CLAPSHAW P

Solomon Park Research Institute, Kirkland, Washington, United States

E-mail address for correspondence: [email protected]

Keywords: subtractive hybridization, SOD1 mutation, differential gene expression

Background: Mice transfected with the human SOD1 mutation G93A develop amyotrophic lateral sclerosis (ALS) like symptoms at approximately 100 days of age. Biochemical changes in motor areas in these mice will provide clues as to the cause and possible treatment of this condition in human patients. Using the procedure of suppression subtractive hybridization (SSH) the transcription events before and after the onset of symptoms (temporal comparison) and between normal and SOD1 mice at the two ages tested (conditional comparison) provide valuable information as to the general metabolic events taking place in the spinal cord during this process.

Objectives: To compare normal (C57Bl6J) and hemizygous (SOD1-G93A) mouse spinal cords at 60 (pre-symptomatic) and 120 (post-symptomatic) days of development for changes in transcribed expressed sequence tags (ESTs). Direct comparisons are between SOD1 and normal mice at 60 and 120 days (conditional) and between normal mice at 60 and 120 days and SOD1 mice at 60 and 120 days (temporal) resulting in eight reciprocal comparisons.

Methods: Total and messenger RNA was isolated from spinal cords dissected from mice and immediately frozen on dry ice. Direct comparisons of cDNA derived from these preparations were compared by SSH in reciprocal subtractions. Subtraction efficiencies between 5 and 10 cycles only were used for comparisons and clones derived from these sequences were further selected by dot blot assay. Representative differentially expressed genes were ultimately verified by northern analysis.

Results: Over 600 sequences were compared between these eight conditions. No differentially expressed sequences were observed in any reciprocal subtractions indicating complete separation during these subtractions. A control subtraction of SOD1 60 day animals against normal 60 day animals without spiking out human SOD1 sequences, showed in excess of 75 percent of these sequences, indicating the power of this procedure. A large percentage of the sequences were either ESTs of unknown function or represented by to-date unrecorded transcripts making analysis of any influence on metabolic pathways difficult (the majority of unknown sequences were in the temporal comparisons rather than in the conditional comparisons). However, there is an obvious predominance of myelin sequences in younger SOD1 mice when subtracted against younger normal mice and a similar preponderance of these sequences in younger normal mice. Animals carrying the G93A SOD1 gene also showed a preponderance of structural and respiratory associated sequences when compared to normal controls.

Discussion and Conclusions: Overall patterns of expression indicate that SOD1 mice compensate for disturbances in oxidative processes in the pre- symptomatic animals. Normal and SOD1 animals both show a preponderance of structural and myelin sequences at earlier ages and a dominance of unknown and unrecorded ESTs at later ages. Results of this study are compared and contrasted to other studies using similar specimens and comparative transcription procedures.

P27 GENETIC MODIFIER AND PHENOTYPIC ANALYSIS OF Tg hSOD1 G93A CONGENIC MOUSE STRAINS

SHER R, WOOLEY C, SEBURN K, COX G

The Jackson Laboratory, Bar Harbor, Maine, United States

E-mail address for correspondence: [email protected]

Keywords: hSOD1 G93A, modifier genes, QTL

Background: Mutations in the gene encoding superoxide dismutase 1 (SOD1) cause amyotrophic lateral sclerosis (ALS) in humans. Studies of SOD1-induced ALS frequently employ a transgenic mouse model that over-expresses the human mutant form of SOD1 that has a glycine replaced by an alanine at position 93 (hSOD1-G93A). We investigated the influence of genetic background on hSOD1-G93A-induced ALS by breeding the transgene onto several inbred mouse strains to create fully congenic lines. Our results show that genetic background can significantly modify survival time. We selected two congenic SOD1 strains, C57BL/6J and ALR/LtJ for further investigation. B6.SOD1 mice live nearly two months longer than ALR.SOD1 mice. A mapping cross between these two strains identified a major B6-dominant modifier located on Chr. 17 and a slightly less significant B6-dominant modifier located on Chr. 4.

Objective: 1) To compare the phenotypic expression of the ALS-induced disease in the two congenic strains to identify differences that might contribute to the earlier disease onset. 2) To analyze candidate genes in the QTL regions that have previously been associated with ALS in humans.

Methods: Mice from each strain were age-matched relative to their 50% strain survival time and characterized at 3 time points. Phenotypic measures included innervation status, peripheral axon loss, gait analysis, muscle contractile force, and immunoblot analysis of SOD1 positive aggregates in the spinal cord. Initial examination of candidate genes was performed using RNA and cDNA from spinal cord tissue. Nucleotide sequence of candidate genes was examined for polymorphisms between B6.SOD1 and ALR.SOD1 and quantitative PCR was used to look for expression differences between candidate genes.

Results: To date phenotypic characterization has revealed no significant differences between B6.SOD1 and ALR.SOD1 mice age-matched relative to their 50% survival. The two strains show similar: 1) presymptomatic modification of gait, 2) loss of innervation at the neuromuscular junction, 3) decreases in axon counts of muscle nerves and 4) loss of contractile force. Data collection for immunoblot analysis and data from additional time points for other measures remain to be completed.

There are no major alterations in splice forms, sequence polymorphisms, or expression levels for Tardbp between B6 and ALR. Preliminary results indicate that there are several sequence polymorphisms in DAXX between the strains, and these are being further characterized. A further candidate gene, Mapk14, is also being analyzed.

Conclusions: Phenotypic characterization completed to date has not revealed any differences that predict the accelerated onset of the hSODG93A induced ALS between the strains. Candidate genes in the intervals are presently being further tested for sequence differences. Gene chip expression analysis is being initiated to discover other potential candidate genes.

P28 TIME-COURSE OF BULBAR V. SPINAL MOTOR DEFICITS IN MALE AND FEMALE SOD1-G93A MUTANT RATS

STANFORD J, SMITTKAMP S, SPALDING H, BROWN J

University of Kansas Medical Center, Kansas City, KS, United States

E-mail address for correspondence: [email protected]

Keywords: bulbar, rats, tongue

Background: Amyotrophic lateral sclerosis (ALS) is a progressive degenerative disease of upper and lower motor neurons with an invariably fatal outcome. Symptom onset may occur in the muscles of the limbs (spinal onset) or those of the head and neck (bulbar onset). Although the pathogenesis of sporadic ALS is unknown, mutations in the Cu/Zn superoxide dismutase-1 (SOD1) gene are associated with the familial form of ALS (fALS). Several transgenic rodent models of fALS have been developed, with the SOD1-G93A being the most widely studied. These animals develop an ALS-like phenotype that includes muscle weakness and atrophy, and degeneration of motor neurons in the cortex, brainstem and spinal cord. Spinal involvement has been the focus of most preclinical studies to date. Recently, we documented behavioural bulbar involvement in SOD1-G93A mice [1] and wished to extend this characterization to SOD1-G93A rats.

Objectives: The purpose of this study was to determine the relative onset of bulbar v. spinal motor signs in male v. female SOD1-G93A rats.

Methods: In the clinic, tests of bulbar function include evaluation of tongue strength and motility. Here, we measured orolingual motor function by placing water-restricted rats in a customized operant chamber that allows for the measurement of tongue force (g) and licking rhythm (licks/s) as animals lick water from an isometric disc. Rats were tested 3 times week for over 200 days, spanning the pre-symptomatic, symptomatic, and end stage segments of the disease. Peak tongue force and licking rhythm were recorded during the 6-minute test sessions. Body weight and fore- and hindlimb grip strength were also monitored concomitantly. Rats were killed within 7 days of onset of overt paralysis in at least 1 limb.

Results: Training-related increases in tongue motility were attenuated in affected rats from ∼40–90 days of age. Licking rhythm continued to be attenuated in the SOD1-G93A group and primarily in females throughout the experiment. Tongue force was not affected. In female SOD1-G93A rats, body weight gain diminished compared to wildtypes beginning ∼140 days of age, while weight gain in affected males diverged from healthy males at 180 days. Hindlimb grip strength deficits emerged at 6 months of age in both males and females.

Conclusions: To our knowledge, these findings are the first to detail the onset of bulbar v. spinal motor signs in this preclinical model of ALS. It is apparent that our rats were low-expressing SOD1-G93A transgenics. Our results suggest that bulbar symptoms emerge primarily in female SOD1-G93A rats and prior to spinal symptoms. Our results also suggest that water restriction may reveal sexually dimorphic effects in the rat SOD1-G93A model of ALS. A detailed characterization of bulbar involvement will be important for future evaluation of therapeutic agents.

P29 GENDER DIFFERENCES IN SURVIVAL OF G93A SOD1 MICE IS NOT DUE TO DIFFERENCES IN MUTANT SOD1 EXPRESSION

SANDHU P, ALEXANDER G, DEITCH J, MYERSON M, HEIMAN-PATTERSON T

Drexel University College of Medicine, Philadelphia, Pa, United States

E-mail address for correspondence: [email protected]

Keywords: SOD1,G93A SOD1 Transgenic mouse,phenotype

Background: The G93A SOD1 transgenic (Tg) mouse model of ALS remains the best pathologic model for amyotrophic lateral sclerosis. The phenotype is dependent on transgene copy number, levels of mutant SOD1, genetic background, and gender. In this regard higher transgene numbers, higher levels of SOD1, SJL background, and male gender all are associated with shortened survival.

Objectives: To determine if differences in survival between male and female G93A SOD1 Tg mice on the SJL-B6 background are due to differences in the expression of human mutant SOD1(hSOD1).

Methods: All B6SJL-TgN(SOD1-G93A)1Gur mice were genotyped by quantitative PCR to verify transgene copy number. Spinal cords and brains were examined from 5 male and 5 female each of control and mutant Tg mice from four age groups: 30, 60, 90, and 120 days of age. Cerebrum and lumbar spinal cord were homogenized in either TBS (for ELISA and activity assay) or SDS lysis buffer (for Western blot) and centrifuged. Protein content of the supernatant was determined by the BCA method. Total SOD1 activity (both human and mouse) of the supernatant was determined spectrophotometrically using a kit by Oxis International. Western blots of samples were performed antibodies to human SOD1. For quantitative purposes, each gel included three lanes loaded with 50, 100 and 200 ng of human SOD1 protein. In order to detect human SOD1, ELISAs of sample extracts (1:10000 dilution for mutant animals and 1:20 for control animals) were performed in wells coated with anti-human SOD1 antibody, then reacted with secondary antibodies conjugated with HRP.

Results: Immunoreactive hSOD1 protein undergoes a rapid upregulation in the brain and spinal cord of mutant animals at about 30 days of age that continues until about 90 days of age before leveling off. This upregulation is more pronounced in the spinal cord compared to the brain. There is no statistical difference in immunoreactive hSOD1 protein levels between males and females in the brain or spinal cord at any age. hSOD1 protein levels in control mice are undetectable. SOD1 enzymatic activity levels in mutant animals mirrored hSOD1 protein levels with a steady increase after 30 days of age with greater activity in spinal cord tissue compared to brain in both males and females. SOD1 activity levels in control animals were also greater in the spinal cord compared to the brain, slowly increasing with age but at a much slower rate; at about 8 fold lower levels than mutant animals.

Conclusion: Given that there is no difference between hSOD1 protein levels in the spinal cord of male and female Tg mice it is unlikely that variations in hSOD1 levels are responsible for the gender based differences in survival previously reported.

P30 A PREFERENTIAL LOSS OF THE LARGEST MOTOR UNITS IN FAST-TWITCH MUSCLES OF THE G93A TRANSGENIC MODEL OF ALS IS ATTENUATED BY ACTIVITY-DEPENDENT MOTOR UNIT AND MUSCLE FIBER TYPE CONVERSION

GORDON T, PUTMAN T, TYREMAN N, HEGEDUS J

University of Alberta, Edmonton, Canada

E-mail address for correspondence: [email protected]

Keywords: Functional motor units, saving motor units, model of hyperactivity

Transgenic G93A mice expressing mutant human superoxide dismutase (SOD1) with a glycine to alanine conversion at the 93^rd codon develop a stereotypic syndrome with pathology and symptoms resembling human ALS [1]. In both patients and G93A mice, a small proportion of motor units (MUs) remain intact in affected limbs even at end-stage disease [2],[3]. To explore the cellular properties that confer protection on selective motoneurones and the muscle fibres they innervate we characterised the surviving MUs throughout the lifespan of the G93A mouse. We reported an early and selective decline in the number of functional MUs in the fast-twitch tibialis anterior (TA), medial gastrocnemius (MG), and the extensor digitorum longus (EDL) muscles by 40 days of age and reduction in numbers in the slow-twitch soleus muscle (SOL) 50 days later at the onset of symptomatic disease [4]. The most forceful MUs were lost first, even though the number of muscle fibres innervated by each motoneurone (innervation ratio, IR) remained unchanged. Decline in MU force, but not IR, occurred due to an increase in the proportion of smaller, less forceful type IIA and IID/X muscle fibres likely due to 1) preferential die-back of the motoneurones innervating the most forceful type IIB muscle fibres and 2) activity-dependent conversion of the remaining innervated fibres to type IIA and IID/X types. If indeed activity-dependent conversion of MUs to the more fatigue resistant type increased their resilience and hence survival, we hypothesized that experimental increase in MU activity in muscles of the G93A mouse should “save” functional MUs. To test this hypothesis, we partially denervated hindlimb muscles in G93A mice and mice expressing wild type SOD1 by avulsion of either L4 or L5 spinal roots at 40 days of age. Isometric force recordings were made from the fast-twitch TA, MG, and EDL muscles and the slow-twitch SOL muscle. Using MUNE [5] to count functional MUs at 90 days of age, we found that the rapid age-dependent decline in numbers of functional MUs in fast-twitch muscles of the G93A mice was significantly reduced by the functional hyperactivity. The muscles comprised a significantly higher component of type IIA and type IID/X fibres. We conclude that the vulnerable MUs in the G93A mouse model of ALS, can be saved by increasing their neuromuscular activity and consequently, converting them to slower, less forceful, fatigue resistant MUs.

Acknowledgements: Grateful thanks to NRP Canada and AHFMR and CIHR for financial support.

P31 GENE EXPRESSION IN SPINAL CORD, BLOOD, AND MUSCLE REVEALS GOOD BIOMARKERS IN BLOOD FOR ALS MOUSE MODEL

SARIS CGJ¹, VAN VUGHT PWJ¹, GROEN E¹, VELDINK J¹, WOKKE JHJ¹, OPHOFF RA^2,3, VAN DEN BERG LH¹

¹Department of Neurology, ²Department of Medical Genetics, Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Netherlands ³Department of Human Genetics and Semel Institute of Neuroscience and Human Behavioral, School of Medicine, University of California Los Angeles, United States

E-mail address for correspondence: [email protected]

Keywords: Biomarkers, Gene expression, mouse model

The SOD1 G93A transgenic murine model is a well-studied model for amyotrophic lateral sclerosis. Previous studies using microarray techniques have focused on spinal cord transcriptome only and revealed changes in transcription and translation, inflammation, mitochondrial preprotein translocation and respiratory chain function.

We combined transcriptional profiles of spinal cord, blood and muscle to find overlapping changes in these tissues. Information on overlapping genes will give insight into 1) common pathways in ALS and 2) the value of blood and muscle as biomarker tissue.

Whole genome gene expression profiles of spinal cord, blood and muscle from 16 transgenic mice were compared to their wild type littermates. In the spinal cord we detected a dramatic down regulation of genes involved in mitochondrial metabolism. Thirty-five percent of the significant differential expressed genes in blood of the transgenic mice were also found to be significant in the spinal cord. Only 0.4% of muscle significant probes were found significant in spinal cord. At the presymptomatic age (70 days postnatal) down regulation in mitochondrial proteins was seen in all three tissues in the transgenic mice. Up regulation of genes involved in immune response was seen in spinal cord and blood. At symptomatic age (day 100) more genes involved in immune response were up regulated in both the spinal cord and the blood.

This study provides evidence for mitochondrial dysfunction presymptomatic in spinal cord, blood which is followed by a large immune response. Overlap between changes in spinal cord and blood was large. In mice, blood could therefore serve as biomarker tissue.

P32 NEURONAL DELETION OF C-JUN IN AMYOTROPHIC LATERAL SCLEROSIS (ALS) CU/ZN SUPEROXIDE DISMUTASE 1 SOD1 G93A TRANSGENIC MICE REDUCES NEURONAL CELL DEATH, GLIOSIS AND PROLONGS LIFESPAN

ACOSTA-SALTOS A, MAKWANA M, GILCHRIST A, HRISTOVA M, DA COSTA C, BEHRENS A, RAIVICH G

¹Perinatal Brain Repair Group, University College London, ²Mammalian Genetics Laboratory, Cancer Research, London, United Kingdom

E-mail address for correspondence: [email protected]

Keywords: SOD1 G93A, c-jun, deletion

Background: Mutations in the SOD1 gene are a cause of familial ALS. Mice carrying SODG93A mutations develop an adult onset form of lethal motorneuron disease, implicating programmed cell death. c-Jun is involved in neuronal survival and regeneration and its presence at increased levels in surviving spinal motoneurons of ALS patients suggests it may play a role in neurodegenerative disease.

Objectives: To investigate the effect the deletion of transcription factor c-jun has on the survival and pathology observed in transgenic SODG93A mice.

Methods: We crossed conditional mouse mutants lacking c-jun in the CNS (c-jun-deltaN) with transgenic SODG93A. Average survival times were measured using the onset of severe hind limb paralysis at the endpoint. H&E stain, Toluine-blue stains and immunohistochemistry were used to study the morphology of motoneuron, axons and glial cells.

Results: SODG93A; c-jun-deltaN survived significantly longer (115±3days), compared with SOD1 G93A mice with c-jun (104±24, p,5%). At day-70, before onset of ALS-like-symptoms, no morphological difference was seen between motoneuron in the L5 spinal cord segment (755±299) and axonal number in the L5 ventral root (329±250) in SOD1 G93A; c-jun-deltaN mice, compared with motoneuron (817±230, p = 58%) and axonal number (402±254, p = 40%) with SOD1 G93A mice with normal c-jun function. However, at the end stage of the disease, motoneuron (446±251) and axonal number (253±231) were higher in SOD1 G93A; c-jun-deltaN mice, compared with motoneuron (311±235, p = 5%) and axonal number (197±215, p = 16%) in control SOD1 G93A mice. In addition, both astrogliosis and microgliosis are considerably reduced in SOD1 G93A;c-jundeltaN animals compared with controls at day-70 and endstage time points.

Discussion and Conclusion: CNS-specific c-jun deletion is neuroprotective and prolongs lifespan in SOD1 G93A model of familial ALS. The SOD1 G93A;cjun-deltaN mice die with a higher number of residual motoneurons and axons, suggesting lack in adaptive mechanisms compensating for neurodegenerative loss. Furthermore, animals died despite an almost complete abrogation of microgliosis and astrocytosis suggesting that decrease in gliosis in SOD1 G93A;c-jun-deltaN mice is beneficial but not essential for their survival.

P33 LEMUR TYROSINE KINASE 2 (LMTK2) REGULATES PHOSPHORYLATION OF KLC2 AND ITS BINDING OF CARGO

GUILLOT F, MANSER C, DAVIES J, MILLER C

King's College London, United Kingdom

E-mail address for correspondence: [email protected]

Keywords: LMTK2, KLC2, GSK3beta

Background: Defective axonal transport is one of the earliest pathological features in several transgenic mouse models of amyotrophic lateral sclerosis (ALS). Cyclin-dependent kinase 5 (cdk5) influences transport of axonal vesicles via a pathway involving phosphorylation of kinesin light chain-2 (KLC2) by glycogen synthase kinase 3beta (GSK3beta). KLC2 is a subunit of the molecular motor kinesin-1 and KLC2 phosphorylation causes detachment of kinesin-1 from vesicles, thus halting their axonal transport [1]. We previously identified Lemur tyrosine kinase-2 (LMTK2) as a binding partner of the cdk5 activator p35 and showed that cdk5/p35 regulates LMTK2 activity [2]. Furthermore LMTK2 binds protein phosphatase 1Calpha (PP1Calpha) and the latter is implicated in the regulation of GSK3beta. Hence, LMTK2 may be involved in the regulation of axonal transport.

Objectives: The objective of this study was to investigate the role of LMTK2 in the regulation of KCL2 phosphorylation and cargo binding.

Results and Methods: To assess if LMTK2 influences KLC2 phosphorylation, we co-expressed LMTK2, or a mutant LMTK2 incapable of binding PP1Calpha (LMTK2-deltaPP1) with KLC2 in HEK293 cells and determined KLC2 phosphorylation by 2-dimensional gel electrophoresis and Pro-Q staining. LMTK2 decreased KLC2 phosphorylation whereas LMTK2-deltaPP1 had no effect. In the same cells phospho-specific Western blots showed that LMTK2 but not LMTK2-deltaPP1 significantly increased the inhibitory phosphorylation of PP1Calpha (on Thr-320) and GSK3beta (on Ser-9). Together these results indicate that LMTK2 regulates KLC2 phosphorylation by controlling the activity of GSK3beta in a PP1Cα-dependent way. Finally, to investigate if LMTK2-induced dephosphorylation of KLC2 affected KLC2-cargo interaction we co-immunoprecipitated KLC2 and SMAD2, a specific KLC2 cargo in the presence of LMTK2. LMTK2 increased binding of KLC2 to SMAD2, showing the functional relevance of LMTK2-dependent regulation of KLC2 phosphorylation.

Conclusions: We described a novel function of LMTK2 in the regulation of KLC2-cargo interactions: LMTK2 phosphorylates PP1Calpha and hence inhibits PP1Calpha activity, which consequently causes a reduction in GSK3beta activity, and ultimately reduced KLC2 phosphorylation. Consequently, LMTK2 promotes the binding of KLC2 to cargo. LMTK2 may therefore play an important role in kinesin-1 mediated axonal transport, and thus prove a potential therapeutic target to restore normal axonal transport in ALS.

This work was supported by grants from the UK MND Association and the Wellcome Trust.

P34 ALS2 MRNA SPLICING VARIANTS DETECTED IN KO MICE RESCUE SEVERE MOTOR DYSFUNCTION PHENOTYPE IN ALS2 KNOCK-DOWN ZEBRAFISH

GROS-LOUIS F², KRIZ J², KABASHI E¹, MCDEARMID J³, MILLECAMPS S², URUSHITANI M², LI L³, DION P¹, ZHU Q⁴, DRAPEAU P³, JULIEN JP², ROULEAU G¹

¹Center for Excellence in Neuromics, CHUM Research Center, Montreal, Quebec, Canada, ²Research Centre of CHUL (CHUQ), Quebec, ³Department of Pathology and Cell Biology and Groupe de recherche sur le système nerveux central, Montreal, ⁴Montreal Institute for Clinical Research, Montreal, Quebec, Canada

E-mail address for correspondence: [email protected]

Keywords: ALS2, knock-out mouse, knock-down zebrafish

Background: Recessive ALS2 mutations are linked to three related but slightly different neurodegenerative disorders: amyotrophic lateral sclerosis, hereditary spastic paraplegia and primary lateral sclerosis.

Objectives: To investigate the molecular function of alsin.

Methods: To investigate the function of the ALS2 encoded protein, we generated Als2 knock-out mice and zAls2 knock-down zebrafish. A number of polyclonal antibodies specifically recognising alsin in different organisms, including human, mouse and zebrafish, have been also generated.

Results: The Als2-deficient mice lacking exon 2 and part of exon 3 developed mild signs of neurodegeneration compatible with axonal transport deficiency. In contrast, zAls2 knock-down zebrafish had severe developmental abnormalities, swimming deficits and motor neuron perturbation. We identified, by RT-PCR, Northern and Western blotting novel Als2 transcripts in mouse central nervous system. These Als2 transcripts were present in Als2 null mice as well as in wild type littermates and some rescued the zebrafish phenotype.

Discussion: Thus, we speculate that the newly identified Als2 mRNA species prevent the Als2-KO mice from developing severe neurodegenerative disease and might also regulate the severity of the motor neurons phenotype observed in ALS2 patients.

P35 MITOCHONDRIAL ALTERATIONS IN TRANSGENIC MICE WITH AN H46R MUTANT SOD1 GENE

SASAKI S¹, NAGAI M², AOKI M³, ITOYAMA Y³

¹Tokyo Women's Medical University, Japan, ²Okayama University, Okayama, Japan, ³Tohoku University, Sendai, Japan

E-mail address for correspondence: [email protected]

Keywords: mitochondria, ultrastructure, H46R mutant SOD1 transgenic mice

Background: H46R mutant Cu/Zn superoxide dismutase (SOD1) transgenic (Tg) mice are a model for human familial ALS, characterized by initial muscle weakness and atrophy in the legs, a very long-term clinical course (approximately 15 years) and pathologically, widespread changes of the spinal cord that extend beyond the motor system, including many aggregates lacking vacuoles. No report on morphological changes of mitochondria has yet been published in these transgenic mice.

Objectives: To examine ultrastructural changes of mitochondria in H46R mutant SOD1 mice, and to clarify whether the mitochondrial alterations affect the pathomechanism of the Tg mice.

Methods: We electron-microscopically examined the spinal cord of Tg mice expressing an H46R mutant human SOD1 gene at early pre-symptomatic (12 w, n = 2), late presymptomatic (16 w, n = 2), early symptomatic (20 w, n = 2), and end (24 w, n = 2) stages, respectively. Age-matched non-Tg mice served as controls in each group. We also studied immunoelectron-microscopically the spinal cord of Tg mice at late pre-symptomatic (n = 1), early symptomatic (n = 1), and end (n = 1) stages, respectively, using post-embedding method.

Results: Non-Tg littermates: Small vacuolar changes and disorganization of the inner compartment of mitochondria were only occasionally seen.

Tg mice: In 12-week-old mice, largely swollen mitochondria with an increased number of cristae and mitochondria with disorganized inner compartment were only occasionally observed predominantly in the axon of the anterior column and, to a lesser degree, dendrites and somata of anterior horn neurons. In 16-week-old mice, in addition to those observed in 12-week-old mice, the alterations of the cristae such as electron-dense changes, multi-folded membranous structures and electron-dense amorphous membranous structures were only occasionally observed in proximal axons. In 20-week-old mice, alterations of the cristae observed in 16-week-old mice were more frequently observed in the proximal axons. Inner and outer membranes focally projected outward slightly on both the transverse and longitudinal sections. In 24-week-old mice, above-described alterations of the cristae became prominent and were frequently observed not only in the proximal axons, but also in the somata, dendrites, and presynaptic terminals. The whole interior of mitochondria frequently changed into electron-dense membranous or amorphous structure. Electron-dense membranous structures of inner and outer membranes prominently projected outward. Immunoelectron-microscopically, neither normal-appearing nor degenerated mitochondria exhibited SOD1- or ubiquitin-immunogold labeling.

Conclusions: Mitochondrial alterations may be involved in the pathomechanism of motor neuron degeneration in H46R mutant SOD1 Tg mice, probably by disrupting axonal transport of substrates needed for neuronal viability and thus causing motor neuron death.

P36 SLOW DEATH OF MOTOR NEURONS IN SPORADIC ALS MOUSE MODEL BY CONDITIONAL TARGETING OF RNA EDITING ENZYME ADAR2

HIDEYAMA T¹, YAMASHITA T¹, MISAWA H², SUZUKI T², TSUJI S¹, TAKAHASHI R³, SHIN K¹

¹University of Tokyo, ²Kyoritsu University of Pharmacy, Tokyo, ³University of Kyoto, Japan

E-mail address for correspondence: [email protected]

Keywords: ADAR2, RNA editing, knockout mouse

Background: Deficient RNA editing of GluR2 mRNA at the Q/R site occurs specifically in motor neurons of sporadic ALS patients regardless of disease phenotype; i.e. classical ALS, progressive bulbar palsy and ALS-dementia. On the contrary, this molecular change does not occur in dying motor neurons of other motor neuron diseases including familial ALS associated with mutated SOD1 (ALS1). Because adenosine deaminase acting on RNA type 2 (ADAR2) specifically catalyzes GluR2 Q/R site-editing, it is likely that ADAR2 activity is not sufficient to edit this site in motor neurons of sporadic ALS. Because these molecular changes occur in a disease and motor neuron-specific manner, we have proposed that GluR2 Q/R site-underediting due to ADAR2 underactivity is a neuronal death-inducing cause in sporadic ALS. Systemically ADAR2 knockout mice die young from status epilepticus, suggesting the role of ADAR2-mediated RNA editing in neuronal survival.

Objectives: To investigate whether deficient ADAR2 activity induces slow neuronal death as seen in motor neurons of sporadic ALS.

Methods: We have generated genetically modified mice in that the ADAR2 gene is conditionally knocked out in motor neurons using Cre-loxP system. After confirming that ADAR2 was deficient in a subset of motor neurons, we investigated the characteristics of phenotype of the mutant mice by means of weekly measurement of stay time on Rotarod and grip strength, analysis of survival time, and of histological observation of central nervous system.

Results: The mutant mice displayed significantly shorter mean survival than control mice and progressive motor-selective deficit in behaviour with delayed loss of spinal motor neurons. RNA editing at the GluR2 Q/R site was completely abolished in ADAR2-lacking motor neurons. The extent of editing at the CYFIP2 K/E site (a recently identified A-to-I editing position) was markedly decreased in motor neurons deficient for ADAR2 and GluR2 Q/R site-editing.

Discussion and Conclusions: The mutant mice exhibited an ALS-like phenotype without developing fatal epilepsy. Our results indicate that motor neurons would undergo slow progressive death if ADAR2 were deficient. Thus it is likely that a reduction of GluR2 Q/R site-editing in motor neurons of sporadic ALS is caused by ADAR2-underactivity and a direct cause of neuronal death. Our model would be useful in searching the aetiology and therapy of sporadic ALS.[12–15]

P37 CHANGES OF NOGO-A AND RECEPTOR NGR IN THE LUMBAR SPINAL CORD OF ALS MODEL MICE

MIYAZAKI K, NAGAI M, OHTA Y, MORIMOTO N, KURATA T, TAKEHISA Y, IKEDA Y, KAMIYA T, ABE K

Department of Neurology, Graduate School of Medicine and Dentistry, Okayama University, Japan

E-mail address for correspondence: [email protected]

Keywords: Nogo-A, NgR, spinal cord

Background: Nogo-A is a myelin-associated neuronal growth inhibitor protein, and has been identified as the target of a neutralizing antibody for a myelin protein. Nogo-A delivers it's inhibitory signals through it's receptor, Nogo receptor (NgR). High expressions of Nogo-A and NgR expression patterns were reported in trauma, stroke, and some neurodegenerative diseases; Nogo-A and NgR have been considered as possible therapeutic targets. Changes in the expression pattern of Nogo-A have also been reported in ALS ([1], [2]).

Objectives: Detailed assessment of Nogo-A and its receptor NgR at the spinal cord of the ALS mouse models or patients has not been reported. Therefore, we examined the expression and distribution pattern of Nogo-A and NgR in an ALS mouse model to determine whether this molecule plays a role in this disease.

Methods: We used 10, 15, and 18 weeks-old (W) G93A (Tg) mice and age-matched non-Tg wild-type littermates (WT) as a control. Ten W Tg mice were considered as the pre-symptomatic stage, 15 W as early-symptomatic stage, and 18 W as end stage of the disease. After dissecting lumbar spinal cord from each mouse, we performed Western Blot and immunohistochemistry analysis.

Results: As compared with WT mice, Tg mice showed that the expression levels of Nogo-A transiently increased in motor neurons at an age of 10 W, while it progressively decreased from 15 to 18 W. NgR expression in motor neurons of the Tg mice increased at 10 W, then progressively decreased from 15 to 18 W. In contrast, there was no significant change in the dorsal lumbar cord or the cerebellum of Tg mice throughout the progression of ALS.

Discussion and Conclusions: This study suggests that the function of Nogo-A may alter under certain conditions and locations, and thus transient overexpression of Nogo-A and NgR in motor neurons of this ALS mouse model at 10 W may represent a survival reaction of these cells under stressful conditions. These proteins then gradually decrease with disease progression probably due to depletion of energy in motor neurons, where mitochondrial disfunction was reported [3].

P38 EXPRESSION OF NOGO IN THE SPINAL CORD AND MUSCLE DURING DISEASE PROGRESSION IN THE G93A SOD1 MOUSE MODEL OF ALS/MND

ATHANASIOU D, VOUKALI E, PULLEN A

Institute of Neurology, University College London, United Kingdom

E-mail address for correspondence: [email protected]

Keywords: NOGO, Spinal cord, muscle

Background: Compensatory synaptic re-organisation in motor pathways in ALS patients and murine models [1], may be restricted by release of neurite outgrowth inhibitors like NOGO. Raised NOGO expression has been reported in muscle homogenates from autopsied ALS patients and end-stage SOD1 mice [2]. Irrespective of whether this is disease specific [3], NOGO limits neurite outgrowth in the injured spinal cord [4], but it's expression in the spinal cord in ALS has not been reported, and it's time-course of expression in disease development remains uncertain.

Objective: To explore NOGO in spinal cords and muscles of SOD1^G93A mice during disease progression.

Methods: Transgenic (Tg) SOD1G93A mice, and wild-type littermates (WT) were examined at pre-symptomatic (6 &13 wks) and symptomatic stage (18 wks). Cryostat sections of spinal cords were immunostained for NOGO-A (MAb clone 6D5, GSK 2), and tibialis anterior (TA) muscles were immunostained for NOGO-A (Abcam), NOGO-B & NOGO–C (Chemicon). SABC methods detected tissue-bound antigen. Image analysis measured cell size and staining levels under standardised conditions. Immunoblots confirmed both NOGO-A antibodies bind to NOGO-receptor NgR1(protein from R&D Systems), and validated immunostaining. Histochemistry typed muscle fibres.

Results: a) Spinal cord; NOGO-A localised to MN and oligodendrocytes (OD) in both Tg and WT at all ages. Relative to age-matched WT mice, MN in Tg mice showed raised levels of NOGO by 13 wks (p < 0.03), but lower levels at 18 wks reflecting profound MN damage. Significantly higher numbers of NOGO-A immunostained OD occurred in the anterior horn and ventral white matter than in other regions of the cord in Tg mice. Increases first occurred at 13 wks (p < 0.01), progressing further in symptomatic mice (p < 0.007). b) TA muscle; In WT mice of all ages, superficial regions of the TA contained mostly larger fibres with ‘low’ NOGO immunostaining. Their size and number resembled the histochemical profile of Type 2B fibres which dominate the superficial TA. Deeper regions contained smaller fibres with ‘high’ NOGO immunostaining, resembling Type 2A fibres. This pattern occurred for all NOGO-isoforms. Relative to WT mice, i) Tg mice at 6 wks, showed no differences in the regional pattern of immunostaining. Larger unstained fibres dominated superficial regions, with higher numbers of stained fibres in deeper regions in all isoforms. ii) By 13 wks, Tg mice showed mild fibre atrophy with significant increases in staining in the superficial TA for all isoforms (p < 0.0001), but variable changes in deeper regions (NOGO-A & -C, p > 0.1 to p < 0.001; NOGO-B, p < 0.03) iii) In symptomatic mice (18 wks), the major alteration was to the superficial region which showed severe muscle fibre atrophy (p < 0.0001), and significantly increased staining for NOGO- A, B & C (p < 0.0001).

Conclusions: NOGO up-regulation begins presymptomatically in spinal cord and muscle, and progresses until end-stage. In the spinal cord, the response targets the ventral motor areas. In the TA. the response shows preferential targeting of Type 2B containing regions, implying impaired capacity for reinnervation. On-going work using strategies targeting NOGO aim to improve compensatory reinnervation.

Supported by the MND Association and Action Medical Research.

P39 IDENTIFICATION OF AXON-MEDIATED NEURONAL SIGNALING TO ASTROCYTES. THE ENDOGENEOUS REGULATION ASTROGLIAL SPECIFIC SYNAPTIC PROTEINS IS DISRUPTED IN ALS

YANG Y, ROTHSTEIN J

Johns Hopkins University, Baltimore, MD, United States

E-mail address for correspondence: [email protected]

Keywords: astroglia, synapse, glt1

Background: Astrocytes play an essential role in the regulation of synaptic function and neuronal metabolism. The communication for this regulation between neuron/axon and astroglial remains unknown. In disease settings, severe disruption of this neuron-glial network, as reflected by a loss of synaptic astroglial glutamate transporter EAAT2/GLT1 and MCT1 has been repeatedly documented. Disruption of this astroglial function is a potently toxic mechanism that acts to promote disease progression in neurodegeneration associated with amyotrophic lateral sclerosis.

Objectives: To investigate various in vitro and in vivo models of the up- and down-stream signaling from neuron to astrocytes that regulate GLT1 expression.

Methods: We first developed a novel neuron astrocyte co-culture system by using a microfluidic platform that only allows axons to approximate and/or contact individual astroglia. By using this novel system, we showed that axons are sufficient and necessary to induce GLT1 expression through transcriptional activation in astrocytes, by both membrane contact and secretion of soluble factors. Axon mediated transcriptional activation of GLT1 in astrocytes is dependent on synaptic transmission from neuron to astrocyte as tetrodotoxin (TTX) and antagonists of both ionotropic and metabotropic glutamate receptors block neuronal induction of GLT1 in slice and astrocyte cultures. Furthermore, in vivo denervation of neuron to astrocyte signaling by acutely or chronically inducing neuron degeneration with Ricin or overexpression of mutant SOD1 also reduces both the GLT1 genomic promoter activity and GLT1 mRNA levels in lumbar spinal cord astrocytes. Subsequent computational and mutagenesis analysis of EAAT2/GLT1 promoter further identified a cis-element that is essential to the EAAT2/GLT1 promoter activity. By affinity purification and LC/MS, a kappa B-motif binding phosphoprotein was identified that specifically binds to this essential cis-element. The expression of kappa B-motif binding phosphoprotein is highly correlated to the expression of GLT1 in astrocytes under different physiological and pathological conditions. Moreover, silencing of kappa B-motif binding phosphoprotein results in reduced GLT1 expression in cultured astrocytes and in vivo, suggesting that kappa B-motif binding phosphoprotein is essential for GLT1 activation.

Conclusions: Taken together, we demonstrated that synaptic transmission from neuron to astrocyte is sufficient and essential for astroglial synaptic protein GLT1 induction. Axonal injury /dysfunction associated with diseases alters a key astroglial transcriptional regulatory pathway, kappa B-motif binding phosphoprotein and regulation of this essential synaptic protein EAAT2/GLT1, ultimately contributes to further neurodegeneration.

P40 ALS CAUSES ALTERED PROLIFERATION AND CYTOPLASMIC INJURY TO NG+ CELLS IN ADULT SPINAL CORD

CARMAN J, DIBIASE L, BERGLES D, ROTHSTEIN J

Johns Hopkins University, Baltimore, United States

E-mail address for correspondence: [email protected]

Keywords: NG2, stem cell, differentiation

Background: Amyotrophic lateral sclerosis (ALS) is characterized by the specific loss of motor neurons in the lumbar spinal cord as well as focal abnormalities in astroglia. Evidence from human patients as well as from the animal model for ALS suggests that non-neuronal cells significantly contribute to the death of motor neurons and the progression of disease. NG2⁺ cells are prevalent in the adult CNS and are thought to represent a population of progenitor cells capable of generating new oligodendrocytes and astrocytes in the injured adult CNS. Additionally, NG2⁺ cells have been shown to express neurotransmitter receptors that are activated at neuro-glia synapses.

Objectives: To study the biology of endogenous NG2 population in rodent models and human ALS and to follow the differentiation and reaction of these cells in ALS and what role they play in generation of new astroglia.

Methods: We employed NG2 BAC promoter DsRed reporter mice to study the developmental population of endogenous MG2 cells in wild type and ALS mice in multiple CNS regions and time points. We examined the differentiation of the cells into new astroglia and oligodendroglia in ALS mice. In addition we examined the population of NG2 cells in human ALS and non-ALS spinal cord.

Conclusions: We have found that there are significantly more NG2⁺ cells in the spinal cord of diseased mSOD1 mice compared to pre-symptomatic mice. Additionally, we have found that NG2⁺ cells in the diseased spinal cord display signs of pathologic injury. Most importantly we show a significant population of new astrocytes in disease psial cord are derived from differentiating NG2 adult progenitors. These findings suggest that NG2⁺ cells not only respond to mSOD1-mediated damage, but may also become diseased. Furthermore, we are studying how the normal activity of these cells is altered in the disease setting. These studies offer tantalizing information regarding the function of this distinct class of glial cells, as well as the ability of the adult CNS to generate new oligodendrocytes and astrocytes in disease.

P41 COMPLEMENT THERAPY PROMOTES PERIPHERAL NERVE REGENERATION

BAAS F¹, RAMAGLIA V¹, STA M¹, MORGAN P², KING R³

¹Academic Medical Center, Amsterdam, Netherlands, ²University of Cardiff, Cardiff, United Kingdom, ³Royal Free and University College, London, United Kingdom

E-mail address for correspondence: [email protected]

Keywords: regeneration, complement, animalmodel

Background: The complement system (C), highly expressed in the healthy peripheral nerve, is activated during Wallerian degeneration (WD). Using a C6 deficient rat strain that is unable to form the membrane attack complex ((MAC, C5b-9), the terminal activation product of the C cascade) we showed that MAC is required for rapid WD after peripheral nerve injury. Pharmacological inhibition of C activation with soluble C receptor 1 (sCR1) protected the nerve from axonal and myelin breakdown at 3 days post-injury; it inhibited macrophage infiltration by five fold and prevented their activation.

Objectives: To determine the effect of C inhibition on nerve regeneration after acute trauma.

Methods: We performed a crush injury of the sciatic nerve in C6 deficient rats and rats treated systemically for 1 week with sCR1. Recovery of function was assessed by footflick test every week up to 5 weeks post-injury. Axonal regeneration was monitored by retrograde tracing of sensory neurons at 1 week post-injury and pathology of the tibial nerve at 5 weeks post-injury.

Results: Both, genetic and pharmacological inhibition of C activation resulted in faster recovery of sensory function than PBS-treated animals whereas reconstitution with purified human C6 protein re-established the wildtype phenotype. The number of sensory neurons that could be labeled by retrograde tracing in C6 deficient rats was significantly higher than wildtypes. Neuropathological analysis of the sciatic nerve after 5 weeks recovery showed that C6 deficiency and sCR1 treatment improve regeneration as judged by axon diameter, myelin thickness and number of regenerative clusters.

Conclusion: We conclude that complement inhibition improves regeneration of the injured peripheral nerve.

Discussion: We propose that an early and destructive complement-mediated event during WD hampers the subsequent regenerative process. Interfering with the terminal C cascade offers an interesting therapeutic approach for many injury and non-injury related disorders in which activation of the C system has been implicated.

Amyotrophic lateral sclerosis (ALS) is an example of such diseases. We observed several components of the complement cascade, including C1q, C3c and C3d in active microglia and reactive astrocytes in spinal cord and motor cortex of both sporadic and familial ALS cases.

P42 ENHANCED NEUROMUSCULAR SYNAPTIC PROTECTION INDUCED BY MODIFIERS OF THE NEUROPROTECTIVE WLDS GENE

RIBCHESTER R¹, COLEMAN M², BLANCO G³, CONFORTI L², BEIROWSKI B², WONG F³, HARTLEY R¹, HYNES-ALLEN A¹, THOMSON D¹

¹University of Edinburgh, United Kingdom, ²The Babraham Institute, Babraham, United Kingdom, ³MRC Mammalian Genetics Unit, Harwell, United Kingdom

E-mail address for correspondence: [email protected]

Keywords: Neuromuscular junction, neuroprotection, WldS mouse

Background: Loss of neuromuscular junctions (NMJ's) precedes other signs of neurodegeneration in SOD1-dependent familial ALS. More generally, progressive muscle denervation with or without compensatory innervation by collateral sprouts from relatively healthy neurones is a feature of sporadic as well as familial ALS. Reducing the rate of neuromuscular synaptic loss and/or strengthening compensatory sprouts could therefore significantly preserve muscle utility, function and strength in ALS patients and thus both mitigate progression of disease and improve quality of life.

Objective: To seek generic mechanisms that protect motor nerve terminals from degeneration.

Methods: The principal model we used was the WldS mutant mouse, in which axotomy produces about a 10 fold delay of Wallerian degeneration after nerve injury. Axons are strongly protected by WldS protein expression but protection of neuromuscular synapses is much weaker; highly sensitive to the WldS gene dose; and declines progressively with age in native mutant WldS mice. We have adopted three basic strategies for identifying enhancers of neuromuscular synaptic protection using this mutation: first, systematic mutation of defined regions of interest in the WldS gene itself; second, random genomic point mutagenesis, using sensitized, heterozygous WldS as a background, with high-throughput screening for relevant progeny; and third, ‘environmental’ influences that alter neuromuscular activity. We assess the efficacy of the first strategy in transgenic lines using standard physiological and immunocytochemical techniques. For the second strategy, we screen progeny in anaesthetized F1 mice using confocal fibre-optic microendoscopy, 3 days after cutting the sciatic nerve. Candidate modifiers (phenodeviants) are recovered and used for inheritance testing of the neuromuscular protective phenotype. Thirdly, we either block activity chronically in peripheral nerves using tetrodotoxin infused from implanted microcapillaries; or enhance activity by giving mice access to exercise (running) wheels. Outcome is measured as in the first strategy.

Results: All three approaches have yielded evidence of modulated neuromuscular synaptic protection compared with native mutant WldS mice. From the first approach, we have found that targeted mutations in the chimeric WldS gene, followed by transgenic expression in mice, leads to significantly greater protection of NMJ's 3 days after axotomy compared with heterozygous WldS mice. In the second approach, we have identified a candidate ethylnitrosourea (ENU)-induced mutant in which there is also enhanced protection of some motor nerve terminals after axotomy. The inheritance of this phenotype is currently under test. From the third approach, preliminary data suggest that neuromuscular activity may condition the vulnerability of motor nerve terminals to axotomy-induced degeneration.

Conclusions: The data obtained thus far support a model in which neuromuscular synapses are an independent neurodegenerative compartment. Further study will include tests of the effects of the modifiers we have identified on SOD1 mouse models of ALS. This may help identify treatments that protect NMJ's from degeneration more effectively than WldS, an important step towards the goal of mitigating disease progression in either familial or sporadic ALS.

Supported by BBSRC, MRC, MND Association and SMNDA.

P43 LOSS OF FUNCTION OF FIG4 CAUSES AN ASYMMETRIC NEURONAL DEGENERATION WITH RAPID PROGRESSION

ZHANG X¹, CHOW C², SAHENK Z⁴, SHY M¹, MEISLER M², LI J³

¹Wayne State University, Detroit, Michigan, ²University of Michigan,, Ann Arbor, Michigan, ³John D. Dingell VA Medical Center, Detroit, Michigan, United States ⁴Columbus Children's Research Institute and the Ohio State University, Columbus, Ohio, United States

E-mail address for correspondence: [email protected]

Keywords: Fig4 gene, progression of neurodegeneration, intracellular organelle trafficking

Background: We have recently identified a novel recessive disorder Charcot-Marie-Tooth type-4J (CMT4J) in humans and in pale tremor mice (plt) that is caused by mutations in the Fig4 gene that encodes a PI3,5P₂. These homozygous mutant mice develop neurodegeneration with intracellular vacuoles in both the central and peripheral nervous system. However, a detailed analysis of the natural history of CMT4J remains to be described, and the mechanisms responsible for the effects of Fig4 mutations have not been identified.

Objectives: The present study was to investigate clinical, pathological, and electrophysiological abnormalities in CMT4J patients and in mice with loss of function of Fig4.

Methods: Two siblings are compound heterozygotes for mutations of Fig4, and were evaluated with serial neurological examinations, electrophysiological, pathological studies and time-lapse imaging. Eleven mice at six weeks of age and two mice at 3 weeks of age were also examined with similar techniques.

Results: We describe the 9-year clinical course of CMT4J, including asymmetric, rapidly progressive paralysis, in two siblings. Progression in both was preceded by trauma. Sensory symptoms were absent despite reduced numbers of sensory axons. Thus, the phenotypic presentation of CMT4J clinically resembles motor neuron disease. Time-lapse imaging of fibroblasts from CMT4J patients demonstrates impaired trafficking of intracellular organelles because of obstruction by vacuoles. Further characterization of plt mice identified axonal degeneration in motor and sensory neurons, limited segmental demyelination, lack of TUNEL staining and lack of accumulation of ubiquitinated protein in neuronal vacuoles.

Discussion and Conclusion: This study represents the first documentation of the natural history of CMT4J. The progression pattern of this disease suggests that CMT4J and plt mice may be a unique model for investigating the progression of motor neuron degeneration. Physical obstruction of organelle trafficking by vacuoles is a potential novel cellular mechanism of neurodegeneration.