THEME 10 IN VITRO EXPERIMENTAL MODELS

P253 BIOPHYSICAL AND BIOLOGICAL CHARACTERISATION OF INCLUSION BODIES CONTAINING TDP-43

Capitini C

Conti S

Cascella R

Cecchi C

Chiti F

Section of Biochemistry, Department of Biomedical, Experimental and Clinical Sciences, Università di Firenze, Firenze, Italy

Email address for correspondence: [email protected]

Keywords: amyloid, TDP-43, toxicity

Background: At present, it is not clear if TDP-43 intracellular inclusions accumulating in the cortex or spinal cord of patients are amyloid-like fibrils or another type of protein aggregates. Indeed, reports describing a structural characterisation of the TDP-43 inclusions accumulating in the cortex or spinal cord of patients or formed in vitro indicate contradicting reports (1–4).

Objectives: The aim of the present study is to clarify the structural and morphological nature of the TDP-43 aggregates and assess their toxic activity.

Methods: We have over-expressed TDP-43 in bacterial Escherichia. coli cells and found that the expressed protein spontaneously aggregates as inclusions bodies (IBs). Since bacterial IBs contain both native-like proteins and amyloid fibrils, depending on the intrinsic propensity of the expressed protein, IBs represent a suitable system to investigate the type of protein aggregates formed spontaneously by TDP-43. We have thus purified IBs containing TDP-43, and control IBs devoid of TDP-43 and analysed them with a number of biophysical techniques.

Results: We found that the TDP-43 component of IBs does not increase thioflavin T fluorescence or cause a red-shift of the Congo Red optical absorption. The TDP-43 component of IBs was found to possess a random-coil secondary structure, as detected with far-UV circular dichroism and infrared spectroscopy. Proteolytic digestion with proteinase K shows that TDP-43 aggregates are digested rapidly, revealing a non- compact, non-amyloid structure. Finally, the analysis of the IBs with atomic force microscopy shows that TDP-43-containing IBs have a more irregular structure than control IBs.

We have also tested the toxicity of TDP-43 and control IBs to neuroblastoma SH-SY5Y cells, monitoring the decrease in MTT reduction, the accumulation of intracellular ROS and the increase of caspase-3 activity as indicators of cellular toxicity. The results indicate that TDP-43 IBs, unlike the control IBs, are toxic to SH-SY5Y cells when they are both added to the extracellular medium and brought inside the cells using a transfection method based on a cationic amphiphilic molecule.

Discussion: Overall, these results show that TDP-43 aggregates accumulating into IBs appear to be structurally disordered and morphologically amorphous, thus showing a non-amyloid nature. In this form, the aggregates appear toxic highlighting that protein aggregate toxicity is coupled to non-amyloid aggregation for TDP-43.

Acknowledgements:

We thank ARiSLA for financial support.

References:

• Kerman et al. Acta Neuropathol 2010;119:335–344.
   PubMed Web of Science ®Google Scholar
• Bigio et al. Acta Neuropathol. 2013;125:463–465.
   PubMed Web of Science ®Google Scholar
• Johnson et al. J. Biol. Chem. 2009;284:20329–20339.
   PubMed Web of Science ®Google Scholar
• Guo et al. Nature Struct. Mol. Biol. 2011;18:822–831.
   PubMed Web of Science ®Google Scholar

P254 MOTONEURON AND MUSCLE-SELECTIVE REMOVAL OF ALS-RELATED MISFOLDED PROTEINS

Crippa V¹

Galbiati M¹

Boncoraglio A¹

Rusmini P¹

Onesto E²

Zito A¹

Giorgetti E¹

Cristofani R¹

Pennuto M³

Carra S⁴

Poletti A¹

^aDipartimento di Scienze Farmacologiche e Biomolecolari (DiSFeB), “Centro di Eccellenza per lo studio delle Malattie Neurodegenerative” (CEND), Universita’ degli Studi di Milano, Milano, Italy

^bDipartimento di Neurologia, IRCCS Istituto Auxologico Italiano, Cusano Milanino, Italy

^cDipartimento di Neuroscienze e Tecnologie del Cervello, Istituto Italiano di Tecnologia, Genova, Italy

^dDipartimento di Scienze Biomediche, Metaboliche e Neuroscienze, Universita’ degli Studi di Modena e Reggio Emilia, Modena, Italy

Email address for correspondence: [email protected]

Keywords: protein misfolding, muscle, autophagy

Background: Amyotrophic lateral sclerosis (ALS) occurs in clinically indistinguishable sporadic (sALS) or familial (fALS) forms. Most of the fALS-related mutant proteins identified so far, such as mutant SOD1, TDP-43, FUS, are prone to misfold; also the product of the mutant C9ORF72 gene aberrantly codes for small highly hydrophobic dipeptides. Both misfolded proteins and hydrophobic peptides accumulate into insoluble proteinaceous material inside motor neurons. This material must be cleared away from cells with the assistance of the molecular chaperones. Chaperones may act on aberrant proteins either by assisting their refolding or by directing them to degradation through the proteasome (UPS) or the autophagic system.

Results: Motor neurons are very sensitive to misfolded protein toxicity, but other cell types, such as astrocytes, oligodendrocytes, and muscle cells could also be affected by their presence. Notably, muscle-restricted expression of mutant SOD1 (mutSOD1), responsible for some fALS, induces muscle atrophy and motor neuron death. We found that several genes are altered in the skeletal muscle of mutSOD1 mice. In fact, we observed upregulation of specific muscle genes, such as MyoD, myogenin, and also of several components of cell response to proteotoxicity (atrogin-1, HspB8, Bag1, and Bag3). Similar changes were found to occur in cultured ALS myoblasts.

We then compared the potential mutSOD1 toxicity in motor neuron (NSC34) and muscle (C2C12) cells. Initially, we found that muscle ALS models possess much higher chymotryptic proteasome activity and autophagy power than motor neuron ALS models. The mutSOD1 molecular behaviour was also very different. MutSOD1 clearance was much higher in muscle than in motor neurons, and the misfolded protein formed aggregates and impaired proteasome only in motor neurons. The motor neuronal cells were also more sensitive to superoxide-induced oxidative stress. In muscle cells, mutSOD1 remained soluble even after proteasome inhibition, possibly because of high mutSOD1 autophagic clearance. Finally, N-terminal TDP-43 fragment accumulated in NSC34, but not in C2C12 cells. In the case of TDP-43, proteasome inhibition resulted in a large accumulation of both wt and N-terminal fragment of TDP-43.

Discussion: Therefore, our results suggest that muscle cells differentially manage misfolded mutSOD1 and TDP-43 and their toxicity in muscle may not directly depend on aggregation.

Acknowledgements:

Italian Ministry of Health (ConvMondino/UNIMI); Università di Milano; RegioneLombardia; Fondation Thierry Latran, France; AFM-France.

P255 ACTIVATION OF TRANSFORMING GROWTH FACTOR-BETA/SMAD SIGNALING REDUCES AGGREGATE FORMATION OF MISLOCALIZED TAR DNA-BINDING PRTOTEIN-43

Nakamura M¹

Kaneko S¹

Ito H²

Fujisawa J¹

Kusaka H¹

^eKansai Medical University, Osaka, Japan

^fWakayama Medical University, Wakayama, Japan

Email address for correspondence: [email protected]

Keywords: transforming growth factor-beta, phosphorylated Smad2, TDP-43

Background: We have previously reported that phosphorylated Smad2/3 (pSmad2/3), the major intracellular mediators of transforming growth factor (TGF)-beta signaling, abnormally accumulated in TDP-43 inclusions of anterior horn cells (AHCs) in amyotrophic lateral sclerosis (ALS). The nuclear pSmad2/3 signal is decreased in AHCs bearing TDP-43 inclusions in comparison with those without these inclusions. It suggested that neuroprotective effect of TGF-beta might be disturbed in ALS. TGF-beta is multifunctional cytokine that regulates cell growth, differentiation, and apoptosis. Furthermore, TGF-beta activates autophagy. In cultured cells, the nuclear localization signal (NLS) deletion mutant of TDP-43 (ΔNLS-TDP-43) has been reported to aggregate in the cytoplasm in the presence of a proteasome inhibitor.

Objective: To investigate the possible pathophysiological linkage between TDP-43-positive inclusions and TGF-beta/Smad signaling system.

Methods: We constructed a plasmid of the nuclear localization signal deletion mutant of TDP-43, and transfected it into HEK293T cells. TDP-43-positive inclusions were formed in the cytoplasm of the HEK293T cell under proteasome inhibition. Plasmids of Smad2 protein and constitutive active or dominant negative forms of TGF-type1 receptor were also transfected into the HEK293T cells. The direct effects of TGF-beta/Smad signaling on the aggregation formation were analyzed using immunocytochemistry and Western blotting.

Results: The aggregates formed in the cytoplasm of the HEK293T cells contained ubiquitinated, phosphorylated, and fragmented TDP-43, consistent with the essential features of the human pathology. Moreover, the aggregates were co-localized with phosphorylated Smad2 under continuous TGF-beta stimulation. These aggregates were immunopositive for p62 and LC3. Overexpression of Smad2 reduced the amount of cytoplasmic aggregates in HEK293T cells. TGF-beta stimulation augmented this reduction effect in a dose-dependent manner.

Discussion: As phosphorylated Smad2 was sequestered within cytoplasmic aggregates, TGF-beta/Smad signaling might be disturbed in the aggregation cells. The co-localization of TDP43 aggregates with p62, and LC3 suggests that autophagy may take part in the degradation of the aggregates. Stimulation with TGF-beta/Smad signaling may result in overcoming this sequestration of pSmad2 by TDP-43 aggregates, and autophagy could be induced under proteasome inhibition.

Conclusions: Our data indicate that activation of TGF-beta/Smad signaling system is protective against aggregate formation of cytoplasmically mislocalized TDP-43, and may be a potential therapeutic approach to delay the progression of ALS.

Acknowledgements:

This study was supported by the Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

References:

• Nakamura M et al. Activation of Transforming Growth Factor-beta/Smad Signaling Reduces Aggregate Formation of Mislocalized TAR DNA binding prtotein-43. Neurodegenerative Disease. Neurodegener Dis. 2013;11:182–93.
   PubMed Web of Science ®Google Scholar
• Nakamura M et al. Phosphorylated Smad2/3 immunoreactivity in sporadic and familial amyotrophic lateral sclerosis and its mouse model. Acta Neuropathol. 2008;115: 327–34.
   PubMed Web of Science ®Google Scholar

P256 FUNCTIONAL EFFECTS OF TDP-43 MUTATIONS IN HUMAN IPSC-DERIVED MOTOR NEURONS AND GENOMIC DNA MODELS

Mutihac R¹

Alegre-Abarrategui J^1,2

Yamasaki-Mann M^1,3

Vowles J³

Cowley S^2,3

Talbot K^2,4

Wade-Martins R^1,2

^gDepartment of Physiology, Anatomy and Genetics

^hOxford Parkinson's Disease Centre

ⁱOxford Stem Cell Institute, Sir William Dunn School of Pathology

^jNuffield Department of Clinical Neurosciences; University of Oxford, Oxford, UK

Email address for correspondence: [email protected]

Keywords: TDP-43, calcium signaling, iPSC derived motor neurons

Background: The transactive-response DNA-binding protein (TDP-43) is a major protein component of the characteristic ALS neuronal cytoplasmic inclusions. Mutations in the gene encoding TDP-43 have been identified in sporadic and familial ALS cases, but molecular mechanisms of disease associated with TDP-43 pathogenic mutations remain to be elucidated.

Objectives: The aim of our work is to investigate pathological and physiological phenotypes associated with TDP-43 pathogenic mutations in human motor neurons, using improved genomic DNA expression models and induced pluripotent stem cell (iPSC)-derived motor neurons.

Methods: To study the molecular mechanisms of TDP-43 mutations, we generated novel in vitro cellular models by site-specific bacterial artificial chromosome (BAC) integration in HEK293 of a fluorescently tagged full-length human genomic TARDBP locus carrying one of two ALS-associated mutations (A382T or M337V). Induced pluripotent stem cells were generated from healthy volunteers and differentiated to mature motor neurons, which were characterized by immunostaining, electrophysiology and calcium imaging. TDP-43 BACs were transduced into iPSC-derived motor neurons by Herpes simplex virus type 1 (HSV-1), and functional investigations were carried out.

Results: While both exogenous and endogenous TDP-43 were mainly localized to the nucleus in HEK293 cells, a higher frequency of cytoplasmic TDP-M337V was observed compared to wild-type TDP-Ypet. Mis-localization was exacerbated by the presence of oxidative stress. In iPSC-derived motor neurons transduced with TDP-43 BACs, cytoplasmic translocation was observed at high levels of oxidative stress, without significant differences between TDP-43 mutant and wild-type TDP-Ypet. Increased susceptibility to cell death was detected in HEK293 cells carrying M337V mutations, with 38% of cells positive for cleaved caspase 3, compared to only 18% of cells carrying wild-type TDP-Ypet, while no significant differences in survival assays were observed in iPSC-derived motor neurons carrying TDP-43 mutations. Calcium signaling from the endoplasmic reticulum (ER) was found to be impaired in HEK293 clonal cell lines carrying TDP-43 mutations, which showed 50% reduction in ER Ca²⁺ concentration and decreased Ca²⁺ amplitude compared to TDP-Ypet. Calcium dysregulation correlated with high levels of Bcl-2 in cells carrying M337V and A382T mutations. Knockdown of Bcl-2 restored amplitude of ER Ca²⁺ release in TDP-M337V similar to the levels of ER Ca²⁺ release amplitude detected in TDP-Ypet.

Discussion and conclusion: Using a novel full-length genomic TDP-43 BAC construct tagged with the fluorescent protein Ypet, our study shows that two TDP-43 ALS-specific mutations, A382T and M337V, increase the probability of cytoplasmic mis-localization in HEK293 cell lines and correlate with ER Ca²⁺ dysregulation, as well as with cell death. High levels of Bcl-2 were associated with TDP-43 mutations, and its knockdown restored ER Ca²⁺ dynamics, suggesting a novel potential mechanistic link between TDP-43 and Bcl-2 in the regulation of ER Ca²⁺ release.

Acknowledgements:

Lady Edith Wolfson Studentship, MND Association.

P257 TDP-43 PROTEINOPATHY: LOSS- AND GAIN-OF-FUNCTION DISEASE CELL MODELS

Onesto E¹

Colombrita C^1,2

Buratti E³

Baralle FE³

Silani V^1,2

Ratti A^1,2

^kIRCCS Istituto Auxologico Italiano, Milano, Italy

^lUniversità degli Studi di Milano, Milano, Italy

^mInternational Centre for Genetic Engineering and Biotechnology, Trieste, Italy

Email address for correspondence: [email protected]

Keywords: TDP-43, progranulin, disease cell model

Background: TDP-43 protein forms pathological inclusions in affected tissues of ALS and FTLD patients. In disease conditions, TDP-43 mislocalizes from the nucleus to the cytoplasm and it is not well-established yet if TDP-43- associated toxicity is due to a nuclear loss- or a cytoplasmic gain-of-function mechanism. The loss-of-function hypothesis is based on the observed clearance of TDP-43 from the nucleus together with its sequestration into aggregates that may affect its main pre-mRNA splicing activity and its cytoplasmic activities, such as mRNA transport, stability and translation. In the gain-of-function hypothesis, the presence of cytoplasmic TDP-43 aggregates is supposed to trigger cell death by affecting important and still unknown cellular processes.

Objectives: To better investigate this issue, we studied how and whether the expression and protein content of one of TDP-43 mRNA targets, the neurotrophic factor Progranulin (PGRN), is altered in TDP-43 proteinopathy cell models.

Methods: The loss-of-function cell models were obtained by knocking down TDP-43 in human neuroblastoma SKNKBE and murine motoneuron-like NSC34 cells, while the gain-of-function model by over-expressing the TDP-43 C-terminal Gln/Asn-rich region (12X-Gln/Asn), recently shown to induce aggregation.

Results: Since we previously reported that TDP-43 post-transcriptionally regulates Pgrn mRNA stability, influencing also its protein level in NSC34 cells, we confirmed that TDP-43 depletion determined a significant increase in PGRN protein levels also in the human SKNBE cells. The increased PGRN content was observed also for the secreted protein in the medium, but specifically in the mouse NSC34 cells and not in SKNKBE cells. Conversely, the presence of TDP-43-positive aggregates in the gain-of-function cell models did not seem to change the content of both the endogenous and the secreted PGRN protein in the human and mouse cell lines.

Discussion and conclusion: PGRN levels need to be maintained within certain physiological ranges because a reduced content is pathogenetic in FTLD patients carrying non-sense mutations in Pgrn gene and an increased level is observed in ALS-affected tissues. Our data show that TDP-43 depletion rather than its sequestration into aggregates is able to influence the protein amount of its target Pgrn, suggesting that reduced levels of TDP-43 are likely to alter the neuronal and/or the neighbouring cell metabolism. Our disease cell models provide a useful tool to better investigate the molecular mechanisms and the cellular pathways involved in TDP-43 proteinopathy in vitro. However, the differences observed in the post-transcriptional regulation of Pgrn in human and in the mouse cell lines should be carefully taken into account when studying TDP-43 and its dysfunction in different disease models.

Acknowledgements:

Financial support was received by AriSLA.

P258 ALTERNATIVE SPLICING OR POLYADENYLATION, WHICH IS THE MAJOR MECHANISM FOR AUTO-REGULATION OF TDP-43?

Koyama A¹

Sugai A²

Kato T²

Konnno T²

Ishihara T²

Nishizawa M²

Onodera O³

ⁿCenter for Transdisciplinary Research, Niigata University, Niigata, Japan

^oDepartments of Neurology, Clinical Neuroscience Branch, Niigata University, Niigata, Japan

^pDepartment of Molecular Neuroscience, Resource Branch for Brain Disease Research, Center for Bioresource-based Research, Niigata University, Niigata, Japan

Email address for correspondence: [email protected]

Keywords: TDP-43, nonsense-mediated mRNA decay, polyadenylation,

Background: TAR DNA-binding protein-43 (TDP-43), a predominantly nuclear protein, plays a key role in the pathogenesis of amyotrophic lateral sclerosis. Accumulating evidence suggests that TDP-43 levels in the nucleus should be strictly regulated especially in central nervous system. TDP-43 is auto-regulated via binding to its own 3’UTR. Mechanisms following auto-regulation remain unclear.

Objective: To investigate the mechanism of TDP-43 auto-regulation.

Materials and methods: Flp-In 293 cell lines stably expressing myc-tagged wild-type TDP-43 cDNA was used to investigate elevated TDP-43 effects on endogenous TDP-43 mRNA. Minigene-containing wild-type exon-6 of TDP-43, which includes TDP-43-binding region, was used to investigate depleted TDP-43 effects on the transcripts. In situ hybridization, northern blotting of polyA (+) RNA in cytoplasmic and nuclear extraction, and 3’-end qRT-PCR were performed to evaluate distribution and expression levels of each isoform of the transcripts.

Results: Northern blot analysis and qRT-PCR revealed that TDP-43 mRNA was alternatively polyadenylated. The levels of transcripts using distal polyadenylation sites were increased with increasing TDP-43 levels. In situ hybridization and northern blot analysis of RNA extracted from nucleus or cytoplasm showed that substantial amounts of TDP-43 mRNA with distal polyadenylation were located in the nucleus. The increasing level of TDP-43 reduced its own mRNA from cytoplasm, whereas the TDP-43 mRNAs in the nucleus with distal polyadenylation sites were unchanged. The mRNA with different polyadenylation sites, however, did not alter its stability. Northern blot analysis, upon cycloheximide treatment showed shorter isoforms, excited 2 or 3 introns in exon-6. These isoforms fulfilled the criteria for nonsense-mediated mRNA decay (NMD).

Discussion and conclusion: Increasing TDP-43 undergoes excision intra-exonic introns, resulting in the reduction of its own mRNA via NMD pathway. In addition, the amounts of TDP-43 altered the polyadenylation sites in combination with alternative splicing, which regulates intracellular distribution (nuclear or cytoplasm) of the mRNA. These mechanisms are collaboratively involved in the auto-regulation of TDP-43.

References:

• Ayala YM, De Conti L, Avendaño-Vázquez SE et al. The EMBO journal 2011;30:277–288.
   PubMed Web of Science ®Google Scholar
• Egawa N, Kitaoka S, Tsukita K et al. Science Translational Medicine 2012;4:145ra104.
   Google Scholar
• Gitcho MA, Bigio EH, Mishra M et al. Acta neuropathologica 2009;118:633–645.
   PubMed Web of Science ®Google Scholar
• Jenal M, Elkon R, Loayza-Puch F et al. Cell 2012;149: 538–553.
   PubMed Web of Science ®Google Scholar
• Kraemer BC, Schuck T, Wheeler JM et al. Acta neuropathologica 2010;119:409–419.
   PubMed Web of Science ®Google Scholar
• Lagier-Tourenne C, Polymenidou M, Cleveland DW Human molecular genetics 2010;19:R46–64.
   PubMed Web of Science ®Google Scholar
• Polymenidou M, Lagier-tourenne C, Hutt KR, et al. Nature neuroscience 2011;14:459–468.
   PubMed Web of Science ®Google Scholar
• Swarup V, Phaneuf D, Dupré N et al. The Journal of experimental medicine 2011;208:2429–47.
   PubMed Web of Science ®Google Scholar
• Tollervey JR, Curk T, Rogelj B et al. Nature neuroscience 2011;14:452–8.
   PubMed Web of Science ®Google Scholar
• Tsai K-J, Yang C-H, Fang Y-H et al. The Journal of experimental medicine 2010;207:1661–1673.
   PubMed Web of Science ®Google Scholar

P259 DISTINCT SPLICING PATTERNS FOR TDP-43 AND FUS RNA-BINDING PROTEINS IN NEURAL-LIKE CELLS

Colombrita C^1,2

Onesto E¹

Buratti E³

Silani V^1,2

Baralle FE³

Ratti A^1,2

^qIRCCS Istituto Auxologico Italiano, Milano, Italy

^rUniversità degli Studi di Milano, Milano, Italy

^sInternational Centre for Genetic Engineering and Biotechnology, Trieste, Italy

Email address for correspondence: [email protected]

Keywords: splicing, TDP-43, FUS

Background: The two RNA-binding proteins TDP-43 and FUS are implicated in the pathogenesis of ALS and FTD, and they both share similar features and functional activities in post-transcriptional regulation of gene expression. Although they similarly regulate pre-mRNA splicing in the nuclear compartment and mRNA transport into neurites, to what extent their function is overlapping biologically is not established yet. As TDP-43 and FUS lose their main nuclear localization in ALS-affected tissues, RNA metabolism and alternative splicing mediated by these proteins may be defective in disease conditions and trigger neurodegeneration.

Objectives: Our aim was to define the effects of the nuclear depletion of TDP-43 and FUS proteins on the alternative splicing patterns and gene expression profiles in neural-like cells.

Methods: TDP-43 and FUS were knocked down in human neuroblastoma SKNBE cells and alternative splicing, and gene expression analyses were performed using the Affymetrix Human Exon 1.0ST GeneChips. EASANA analysis tool was used for data visualization (GenoSplice). Q-PCR and RT-PCR were employed to validate gene expression and splicing changes, respectively.

Results: We found 265 and 64 genes differentially spliced (splicing index, ≥ 2) in TDP-43 and FUS knocked-down SKNBE cells, respectively, 25 of which were commonly mis-spliced. Moreover, silencing of TDP-43 and FUS resulted in 386 and 143 genes differentially expressed, respectively, with 56 common genes. In particular, almost 60% of the differentially expressed genes for both TDP-43 and FUS were down-regulated.

Functional annotation analyses revealed that genes alternatively spliced in condition of TDP-43 and FUS depletion belong to different GO categories enriched in GTPase regulator activity and apoptosis for TDP-43 and in neuron differentiation for FUS. On the contrary, differentially expressed genes in TDP-43- and FUS-silenced cells showed overlapping GO categories, including apoptosis, cell migration and response to oxygen levels.

Importantly, our microarray data for TDP-43 have confirmed some of the alternatively spliced transcripts recently identified by other groups, including TNIK and POLIDIP3 genes. However, to further confirm our experimental results, we validated by RT-PCR some alternatively spliced transcripts selected on the basis of their reported involvement in neuronal cell metabolism. In parallel, changes in the expression of selected genes (fold change > 2) were also confirmed by quantitative real-time PCR.

Discussion and conclusions: Our findings indicate that TDP-43 regulates the alternative splicing and expression pattern of specific targets compared to FUS, although common targets were identified, too. The recent literature data also seem to support the idea that TDP-43 and FUS proteins behave differently in regulating alternative splicing and are not associated in the same ribonucleoprotein complexes. The altered transcriptomes emerging from our study will help give further insights into the potential cellular pathways disrupted by the loss of TDP-43 and FUS activity in neurodegenerative diseases.

Acknowledgements:

Financial support was received by AriSLA.

P260 ALS-ASSOCIATED FUS MUTANTS RETAIN SPLICEOSOMAL SNRNPS IN THE CYTOPLASM

Gerbino V^1,2

Rossi S^1,2

Mirra A⁴

Carrì MT^1,2

Cozzolino M³

Achsel T⁵

^tDepartment. Biology, University of Tor Vergata, Rome, Italy

^uFondazione Santa Lucia IRCCS, Rome, Italy

^vInstitute for Translational Pharmacology, CNR, Rome, Italy

^wUniversity La Sapienza, Rome, Italy

^xVIB Center for the Biology of Disease and Department of Human Genetics, KU Leuven, Leuven, Belgium

Email address for correspondence: [email protected]

Keywords: FUS, SMN, alternative splicing

Background: Motor neuron diseases such as amyotrophic lateral sclerosis (ALS) and spinal muscular atrophy (SMA) are associated with defects in proteins involved in RNA metabolism (TDP43 and FUS, and SMN, respectively). SMN, the causative factor in SMA, is crucial for the biogenesis of the spliceosomal snRNPs. FUS forms cytoplasmic aggregates, as a consequence of disturbed nuclear import due to disease-causing mutations. It is extremely likely that the cytoplasmic aggregates are cytotoxic because they trap important factors; the nature of these factors, however, remains to be elucidated.

Objectives: We investigated whether mutant FUS and reduced SMN might disturb the same pathway, namely the biogenesis of snRNPs, looking for mutual interactions and disturbed expression of these factors.

Methods: Mouse motoneuronal NSC34 cells transfected with wild-type or mutant FUS were used as cellular models of ALS. qPCR coupled to immunoprecipitation, as well as to immunofluorescence analysis and fluorescence in situ hybridisation (FISH), was used to assess expression, FUS binding, and subcellular distribution of the snRNPs. A splice reporter plasmid including the exon 7 of human SMN2 was used to monitor alternative splicing variations by FUS.

Results: To test the hypothesis that FUS might be involved in snRNP biosynthesis, we checked for a physical association with SMN, the catalyst of snRNP assembly, and with the snRNAs themselves. We found that FUS and SMN associated with each other, and FUS bound to Sm-snRNPs. Mutations in FUS did not affect association with the snRNPs, but caused their retention in the cytoplasm. Since the total snRNP concentration did not change, this reduced the availability of functional snRNPs in the nucleus. As a result, alterations in the alternative splicing of a reporter plasmid were observed.

Discussion and conclusion: Our results suggest that aggregated FUS may indeed be toxic because they sequester spliceosomal snRNPs in the cytoplasm, lowering their availability in the nucleus and thus leading to changes in alternative splicing patterns. In this sense, the FUS mutations and genetic depletion of SMN interfere with the same pathway, which might represent a unifying theme in the FUS-related ALS and SMA.

Acknowledgments:

This work was supported by ARiSLA (MTC, MC) SMA Europe (TA), and CARIPLO foundation (MTC and TA).

P261 THE ALS-ASSOCIATED PROTEIN FUS/TLS IS A COMPONENT OF THE CELLULAR RESPONSE TO DNA DAMAGE

Rulten S¹

Rotheray A^1,2

Moore D¹

Green R²

Caldecott K¹

Hafezparast M²

^yGenome Damage and Stability Centre, Brighton, UK

^zSchool of Life Sciences, University of Sussex, Brighton, UK

Email address for correspondence: [email protected]

Keywords: FUS/TLS, DNA damage response, RNA processing

Background: Fused in sarcoma/translocated in sarcoma (FUS/TLS) is a member of the hnRNP family of RNA-processing factors that bind thousands of pre-mRNAs and can regulate their splicing, raising the possibility that ALS associated with FUS mutation may involve loss-of-function defects in RNA processing. However, how defects in RNA processing might lead to neurodegeneration is unclear. One possibility is that FUS might be important to maintain the expression of genes that have acquired DNA damage, to prevent inappropriate mRNA splicing and/or polyadenylation. DNA suffers from an intrinsic level of instability and from attack by reactive oxygen species and, as a result, accrues DNA lesions that can block transcription. Since the cumulative number of DNA lesions encountered by a gene is dependent on cell age, DNA damage poses the greatest threat to gene transcription in long-lived cells such as neurons.

Objectives: To examine whether FUS is a component of the response machinery to DNA damage in mammalian cells.

Methods: KU58948, or KU55933, was used to inhibit PARP1 or ATM, respectively. GFP-positive cells were irradiated with a 351-nm UVA laser. UVA (0.44 J/m²) was introduced to an area of ˜12 μm x 0.1 μm, and images were captured at 15-s intervals. Pre-sensitisation was carried out as above, and individual cells were irradiated with 4.4 J/m² UVA.

Results: We observed that both mouse and human GFP-FUS were rapidly recruited to sites of UVA-induced oxidative DNA damage, which include both single- and double-strand breaks. Importantly, endogenous FUS similarly accumulated at these sites as measured by two separate anti-FUS antibodies, ruling out that this response was an effect of overexpressing GFP-tagged protein. Moreover, the recruitment of FUS to the DNA lesions was dependent on the activation of poly (ADP-ribose)-polymerase-1 (PARP-1), a critical protein sensor of chromosomal DNA strand breaks. Intriguingly, the ALS-associated R521G mutation in the C-terminal domain of FUS significantly reduced FUS accumulation at sites of UVA-induced DNA damage in both A549 and HeLa cells. In addition, treatment of mammalian cell lines A549 and HeLa Cells expressing GFP-FUS with camptothecin (CPT; a DNA single-strand-break inducing toxin) led to rapid accumulation of GFP-FUS in the nucleoli. We observed similar results for endogenous FUS in U20S cells. Interestingly, this accumulation was reversible following CPT removal, consistent with the rapid repair of Top1-induced single strand breaks.

Conclusions: These data suggest that FUS relocalises in response to DNA damage, including damage that induces transcriptional stress by blocking the progression of RNA polymerases. Our data highlight a new aspect of FUS function, and supports a model in which FUS is an important component of the mechanism/s by which cells respond to transcriptional stress.

Acknowledgements:

We thank Drs Acevedo-Arozena and Joyce for providing FUS plasmid constructs.

P262 AUTOSOMAL DOMINANT INHERITANCE OF RAPIDLY PROGRESSIVE JUVENILE-TYPE ALS DUE TO A FUS MUTATION

Kent L¹

Vizard T¹

Smith B²

Topp S²

Shaw CE²

Talbot K¹

^aaNuffield Department of Clinical Neurosciences, Oxford University, Oxford, UK

^abDepartment of Clinical Neurosciences, Kings College London, London, UK

Email address for correspondence: [email protected]

Keywords: FUS, familial, truncation

Background: Approximately 5% of ALS cases are familial. Mutations in the gene encoding FUS/TLS (fused in sarcoma/translocated in liposarcoma) are responsible for about 4% of familial and < 1% of sporadic disease, and a high frequency of cases have a young onset and an aggressive clinical course. The FUS protein contains a nuclear localisation signal (NLS) at the extreme C terminal, a common site for mutations. In vitro studies suggest that mutations can cause FUS mislocalisation from the nucleus to the cytoplasm, and subsequent aggregation in stress granules.

Objectives: Identification of a disease-causing mutation in a family with an unusually severe form of autosomal dominant ALS, and initial functional studies using cell culture.

Methods: We undertook whole exome sequencing in a family with an unusually severe ALS phenotype with onset in adolescence or early adult life. We then characterised the cellular effects of an identified FUS mutation using site-directed mutagenesis to create an expression construct which was expressed in cell lines and primary neurons. The effect of this mutation on FUS distribution was compared to that on wild-type and P525L mutant FUS.

Results: DNA from members of a family with an aggressive form of lower motor neuron predominant ALS with onset from 15 to 30 years was analysed using whole exome sequencing, demonstrating a heterozygous AG deletion in exon 14 of FUS (c.1509_1510delAG). This frameshift results in a truncated protein lacking the nuclear localisation signal. The mutation was expressed in neuronal cell lines and primary neurons. The effect of this mutation on FUS distribution and stress granule formation, compared to overexpressing wild-type and P525L mutant FUS, demonstrated a severe cellular phenotype, in keeping with previously described truncation mutations.

Discussion: This is the first reported family with several members bearing a severe truncating mutation in FUS. Truncation mutations tolerated long enough to allow vertical transmission have not previously been described, presumably due to the severity of the cellular phenotype. These studies will further examine the hypothesis that mutations causing clinically aggressive disease are paralleled by greater in vitro changes. This supports the argument that failure of nuclear localisation is central to the toxicity of mutant FUS.

Conclusions: Severe truncating mutations can rarely be transmitted prior to disease onset. Further work will examine the effect of this mutation on cellular homeostasis in primary motor neurons.

Funding: Patrick Berthoud Trust and MND Association.

P263 EVALUATION OF THE ROLE OF SMN (SURVIVAL OF MOTOR NEURON) PROTEIN IN PATHOLOGICAL FUS STRESS GRANULES IN PRIMARY NEURONS

Vizard T¹

Kent L¹

Oliver P²

Bäumer D³

Ansorge O³

Talbot K¹

^acNuffield Department of Clinical Neurosciences

^adDepartment of Physiology, Anatomy and Genetics, University of Oxford, Oxford, UK

^aeDepartment of Neuropathology, John Radcliffe Hospital, Oxford, UK

Email address for correspondence: [email protected]

Keywords: FUS, SMN, stress

Background: Mutations in the gene encoding the RNA-binding protein FUS (Fused in sarcoma) cause a subtype of ALS characterised pathologically by cytoplasmic insoluble FUS inclusions in neurons and glia. Stress granules are storage and sorting sites for stalled translation complexes, and incorporation of mutant FUS into cytoplasmic stress granules has been proposed to be a precursor of insoluble FUS inclusions. Despite the important advances in our knowledge of FUS pathological deposition, the mechanism of toxicity of mutant FUS is less clear. Recent studies have suggested that wild-type FUS interacts with nuclear SMN, the spliceosome, and is involved in nuclear Gem formation, all of which are perturbed by pathogenic FUS mutations. Evidence of SMN incorporation into cytoplasmic stress granules, and pathological evidence of SMN immunoreactivity in FUS-positive inclusions at autopsy, supports the need for a further evaluation of SMN in FUS-mediated toxicity.

Objectives: To examine the involvement of SMN in pathological FUS stress granules in primary cortical and motor neurons.

Methods: Primary motor neuron and cortical cultures were established from embryonic E13.5 and E16.5 mice, and were transfected by magnetotransfection with pcDNA-based constructs expressing wild-type, R521C or P525L FUS. NSC-34 cells were transfected with the same constructs using Lipofectamine. Neurons and NSC-34 cells were stressed by treatment with 0.5mM sodium arsenite, fixed and stained for FUS, SMN and the stress granule marker PABP-1. Stringent analyses were performed to clarify SMN:FUS co-localisation in cytoplasmic stress granules.

Results: We found R512C and P525L FUS incorporated readily into PABP-1-positive stress granules in both cortical and motor neurons, consistent with previous reports. Here, we provide further clarification of the extent of SMN incorporation into FUS-containing stress granules in both motor and cortical neurons.

Discussion and conclusion: Previous reports, principally using neuronal or non-neuronal transformed cell lines, have shown the ability of SMN to form cytoplasmic stress granules when exposed to cellular stress. We have used two different populations of primary neurons and a motor neuron-like cell line to study further the role of SMN in native and mutant FUS-containing stress granules. Although further studies are required, the current findings suggest that results from transformed cell lines may not be applicable to primary neurons and further clarify the role of SMN in cytoplasmic stress granules.

Funding: MND Association and the Patrick Berthoud Trust.

P264 UBIQUILIN 2 MUTATIONS INDUCE ENDOPLASMIC RETICULUM STRESS

Halloran M¹

Soo KY²

Yang S³

Blair I³

Atkin J²

^afDepartment of Neuroscience, Psychological Science

^agDepartment of Biochemistry, La Trobe University, Melbourne, Bundoora, VIC, Australia

^ahANZAC Research Institute, Concord Hospital, University of Sydney, Sydney, NSW, Australia

Email address for correspondence: [email protected]

Keywords: Calreticulin, CHOP, ER stress, UBQLN2, XBP-1

Background: Mutations in ubiquilin 2, an ubiquitin-like protein which regulates protein degradation, have been shown to cause dominant X-linked ALS and ALS/dementia. Previous studies have shown that endoplasmic reticulum (ER) stress is a common pathogenic mechanism shared in both sporadic and familial ALS cases. Ubiquilin1 gene (UBQLN1) variants have been shown to be associated with an increased risk of Alzheimer's disease, and additionally ubiquilin1 has been demonstrated to interact with ER-associated proteins, presenilin 1 and presenilin 2, as well as with the molecular chaperone protein disulphide isomerase (PDI). Furthermore, ubiquilin1 has been shown to associate with the ER during hypoxia-induced ER-stress, and works co-operatively with PDI to attenuate C/EBP homology protein (CHOP) induction.

Objectives: To investigate the relationship between the ER and ubiquilin 2.

Methods: Neuro2A and COS-7 cells were transiently transfected with wild-type (WT) and mutant ubiquilin 2 constructs (P497H and P506T) and fixed with 4% PFA or 100% methanol at 48 h. Immunocytochemistry using antibodies against CHOP and X-box binding protein 1 (XBP-1) were used to detect ER stress, and ER marker calreticulin was used to observe changes in ER morphology. Neuro2A cells were also transfected with ubiquilin 2 constructs and harvested at 24 h post-transfection. Immunoblotting was performed with antibodies against CHOP, XBP-1, IRE1, Bip and P-PERK.

Results: Immunocytochemistry using Neuro2A and COS-7 cells transfected with ubiquilin 2 and stained with calreticulin revealed no perturbations in ER morphology. Ubiquilin 2 ALS mutants (P497H and P506T) demonstrated a significant increase in nuclear CHOP immunoreactivity, demonstrating activation of ER stress 48 h post-transfection. Western blot analysis at 24 h post-transfection demonstrated an upregulation of CHOP and Bip for ubiquilin 2 mutants (P497H, P506T).

Discussion and conclusion: The results obtained in this study demonstrate that ubiquilin 2 mutations induce ER stress, adding ubiquilin 2 to a growing list of proteins linked to ALS that demonstrate ER dysfunction.

P265 YEAST MODEL EXPRESSING ALS-LINKED P56S-VAPB EXHIBITS INCREASED SENSITIVITY TO OXIDATIVE AND ENDOPLASMIC RETICULUM-INDUCED STRESSES

Palma F¹

Mitne-Neto M^1,2

Gomes F¹

Demasi M³

Zatz M^1,2

Netto L¹

^aiUniversity of São Paulo, São Paulo, Brazil

^ajHuman Genome Research Center, São Paulo, Brazil

^akButantan Institute, São Paulo, Brazil

Email address for correspondence: [email protected]

Keywords: VAPB, oxidative stress, endoplasmic reticulum stress

Background: Some mutations in the gene encoding VAPB are responsible for familial amyotrophic lateral sclerosis type 8 (FALS8). Despite its function not being fully established in mammals, its homologous protein in Saccharomyces cerevisiae, SCS2, is involved in different aspects of cell biology of the endoplasmic reticulum, lipid metabolism and in unfolded protein response (UPR) which can result in apoptosis.

Objectives: Since mutations in the VAPB gene result in ALS and the same mutation may show very different clinical phenotypes, we seek to study the susceptibility to oxidative stress and endoplasmic reticulum stress as factors underlying this clinical heterogeneity. Therefore, we performed an integrated analysis in order to understand the molecular mechanisms involved in this disease using S. cerevisiae as a model.

Results: BY4741 strain was transformed with expression vectors containing genes of wild or mutant (P56S) human VAPB and a control strain with an empty expression vector. Growth curves and serial dilutions were made to assess the viability and sensitivity of S. cerevisiae under treatment with different concentrations of hydrogen peroxide (an oxidative stress inducer) and DTT (an endoplasmic reticulum stress inducer). As a parameter of redox state of the cells, the ratio of reduction to oxidized glutathione by HPLC with electrochemical detection was determined by enzymatic assays in normal conditions and under treatment with hydrogen peroxide and DTT.

Discussion and conclusion: Our results show lower growth and viability of cells carrying the mutant gene (VAPB^P56S) when compared to those carrying the wild-type gene and the control strain both under treatment with stress inducers and under normal conditions. Interestingly the expression of wild-type VAPB was also toxic suggesting that high levels of VAPB can modify some metabolic pathways in yeast cells. The ratio of reduced to oxidized glutathione in normal conditions was two-fold in the control strain than in the VAPB^P56S mutant strain. Under conditions of induced oxidative stress, VAPB^P56S mutants also displayed a lower ratio of reduction to oxidized glutathione and, comparatively to those data obtained in normal conditions, this ratio was even smaller. Similar results were obtained in the treatment with DTT, suggesting a link between oxidative and endoplasmic reticulum stresses.

P266 THE ROLE OF RBM45 IN ANTIOXIDANT RESPONSES IN ALS

Bakkar N

Bowser R

Barrow Neurological Institute/St Joseph Hospital and Medical Center, Phoenix, AZ, USA

Email address for correspondence: [email protected]

Keywords: RNA binding proteins, oxidative stress, NRF2

Background: Numerous signaling pathways have been proposed to contribute to ALS disease pathogenesis, including oxidative stress/injury (1). RNA-binding proteins FUS and TDP-43 have been implicated in the disease etiology, although their exact role in pathogenic mechanisms remains unclear. Our group has recently described the involvement of a new RNA-binding protein, RBM45 in ALS (2). Increased levels of RBM45 were detected in the cerebrospinal fluid of ALS patients, and the protein was localized to cytoplasmic inclusions that often co-localized with TDP43- and ubiquitin-positive aggregates.

Objectives: In this study, we further characterize RBM45 function and subcellular distribution within cultured cells and primary neurons using various molecular and biochemical approaches. We generate various tools to overexpress or knockdown full-length RBM45 as well as various truncations mutants to examine their effect on cellular function and viability.

Results: We found RBM45 to bind and stabilize KEAP, the inhibitor of the antioxidant response transcription factor NRF2. Overexpression of RBM45 increases KEAP levels, inhibiting NRF2 and the antioxidative response element signaling pathway, thus increasing cellular death in response to oxidant insult. We further mapped the functional region of the protein responsible for such effects.

Discussion and conclusion: Our findings define a novel role of RBM45 in the regulation of the oxidative status of the cell. Specifically, we show a detrimental effect of RBM45 on cellular response to oxidative injury. Given that oxidative damage plays a major role in neuronal death, these results provide the first link between an RNA-binding protein that can form cytoplasmic inclusions and the KEAP/NRF2/ antioxidant response element signaling pathway in ALS.

Acknowledgements:

Funding support by NS061867 and NS068179 to R Bowser.

References:

• Rossi L et al. Int J Cell Biol. 2012.
   Google Scholar
• Collins M et al. ActaNeuropathol. 2012;124:717–32.
   Google Scholar

P267 EVALUATION OF EXPRESSION AND LOCALIZATION OF ELAV PROTEINS IN ALS

Milani P¹

Dell’Orco M^1,2

Amadio M³

Gagliardi S¹

Laforenza U⁴

Diamanti L^2,5

Ceroni M^2,5

Cereda C¹

^alLaboratory of Experimental Neurobiology

^amDivision of General Neurology, “C. Mondino” National Institute of Neurology Foundation, IRCCS, Pavia, Italy

^anDepartment of Brain and Behavioural Sciences

^aoDepartment of Drug Sciences, Pharmacology Section

^apDepartment of Molecular Medicine, University of Pavia, Pavia, Italy

Email address for correspondence: [email protected]

Keywords: RBPs, neurodegeneration

Background: Altered RNA metabolism contributes to the pathogenesis of neurodegeneration and several RNA-binding proteins (RBPs) have been implicated in ALS pathogenesis (1). Indeed post-transcriptional regulatory mechanisms mediated by RBPs are fundamental for the development of the nervous system and its maintenance. ELAV proteins are RBPs which exert a pivotal role in modulating several aspects of RNA metabolism.

HuB, HuC and HuD represent the neuron-specific members of ELAV family (nELAVproteins), while HuR is ubiquitously expressed. In accordance with their function, nELAVs are mainly cytoplasmic, while HuR is characterized by a nucleus-cytoplasm shuttling ability. ELAV genes are spatially and temporally regulated, showing different expression patterns within the developing and adult nervous system.

Objectives: In this study, we aimed to explore potential variations in ELAVs expression in an in vitro cellular model of neurodegeneration and in samples from sporadic ALS (SALS) patients and controls. We focused on HuR and HuD since their involvement in neurodegenerative processes has been largely reported (2).

Methods: Human neuroblastoma SH-SY5Y cells treated with 1mM H₂O₂ for 30 and 60 min have been used. ELAV protein expression and subcellular localization were evaluated using immunocytofluorescence and Western botting (WB), while mRNA levels were quantified by real-time PCR. HuR phosphorylation in peripheral blood mononuclear cells (PBMCs) from SALS patients and healthy controls were assessed using immunoprecipitation coupled with WB. Immunohistochemistry and real-time PCR experiments were carried out in cerebral motor cortex tissues from SALS patients and controls to evaluate HuR and HuD protein expression/localization and mRNA levels, respectively.

Results: In our in vitro cellular model, we did not observe significant variations in HuD intracellular distribution, while we detected an increased HuR expression in the cytoplasm under oxidative stress condition. No significant variations were reported in HuD and HuR mRNA levels after the treatment. In PBMCs from SALS patients compared to controls, we could not appreciate changes in HuR expression and distribution; nevertheless, we found a specific increase in HuR phosphorylation, suggesting that this protein is more activated in the pathology. Finally, in line with the findings reported in our cellular model, we observed that, although there were no significant variations in mRNA levels, HuR and HuD protein expression was increased in the cytoplasm and in the perinuclear area in cerebral motor cortex from SALS patients compared to healthy individuals.

Discussion and conclusion: These preliminary data suggest a possible involvement of ELAVs protein in neurodegeneration. Further studies will be needed to confirm our data on SALS patients and to identify the functional role of ELAVs protein in ALS.

References:

• Gagliardi S, Milani P, Sardone V et al. Neurol Res Int. 2012;2012:278725.
   Google Scholar
• Pascale A, Govoni S. Cell Mol Life Sci. 2012;69(4): 501–17.
   PubMed Web of Science ®Google Scholar

P268 CAUSES AND CONSEQUENCES OF MICRORNA MALFUNCTION IN ALS

Hornstein E¹

Emde A¹

Yardeni T¹

Reichenstein I¹

Skorovsky M¹

Moeller T²

Ravits J³

^aqweizmann Institute of Science, Rehovot, Israel

^arUniveristy of Washington, Seattle, USA

^asUniveristy of California, San Diego, USA

Email address for correspondence: [email protected]

Keywords: RNA, microRNA, dicer

Background: ALS-causing mutations, recently discovered in genes encoding for RNA-binding proteins, encourage exploration of RNA-related processes in ALS pathogenesis. microRNAs provide homology-based silencing that is essential for brain integrity.

Objective: We tested the causes and consequences of microRNA dysregulation in ALS.

Methods and results: We are employing molecular biology, mouse genetics and human molecular histopathology. Our study reveals global downregulation of microRNAs in ALS lower-motoneurons from the lumbar region of sporadic and familial cases, but not in RNA extracted from surrounding, neuron-depleted ventral horn tissue or from the neurons of Clarke's column in the same autopsies.

We sought molecular mechanistic insight for dysregulation of microRNAs in tissue culture studies. We show that the canonical microRNA bioprocessing pathway is disrupted at the level of Dicer1 activity, by ALS-causing mutants FUS495X, FUSR521G, TDP-43A315T and TDP-43M337V. Dicer impairments result in pre-microRNA precursor accumulation and mature microRNA downregulation (1). Accordingly, loss of Dicer1 and microRNAs activity in a Dicer1 conditional knockout in spinal motoneurons resulted in degeneration of spinal motoneurons and in denervation-dependent muscle atrophy (2). Our newest pieces of data provide mechanisms for Dicer dysregulation and suggest that cellular stress impact the activity of the Dicer complex. We will present unpublished data that dissect how stress signalling impairs Dicer activity in ALS.

Conclusions: microRNA plays a key role in brain integrity and is probably involved in several forms of ALS. Dysregulation of Dicer is related to stress signalling and suggests that potentiating of Dicer activity may be beneficial in ALS. Our data link post-transcriptional regulation to stress signalling in ALS.

References:

• Emde et al. unpublished results.
   Google Scholar
• Haramati S et al. Proc Natl Acad Sci USA 2010; 107:13111–6.
   PubMed Web of Science ®Google Scholar

P269 IN VITRO CHARACTERIZATION OF RBM45, A NEW RNA-BINDING PROTEIN IMPLICATED IN ALS AND FTLD

Li Y

Bowser R

Barrow Neurological Institute-St. Joseph’s Hospital, Phoenix, AZ, USA

Email address for correspondence: [email protected]

Keywords: RNA-binding protein, inclusion, CLIP-seq

Background: A major advancement to our understanding of the ALS and FTLD pathogenesis was the identification of mutations in the RNA-binding proteins TDP-43 and FUS in familial cases of ALS and FTLD, and the observation of cytoplasmic aggregates of these proteins in familial and sporadic ALS and FTLD. However, the motor neuron degeneration mechanism remains unknown. Our laboratory has recently identified a new RNA-binding protein, RBM45, with pathologic alterations in ALS and FTLD. RBM45 containing cytoplasmic inclusions were observed in both ALS and FTLD patients (1). RBM45 also co- localized with TDP-43 and ubiquitin inclusions in affected neurons. However, very little is known about the physiological functions and RNA-binding targets of RBM45.

Objective: We have used in vitro cultured cells to characterize the structural functions of RBM45. We have also identified the RNA-binding targets of RBM45 to study the RNA-mediated pathways of neurodegeneration.

Methods: We have used in vitro culture cells (HEK293, Neuro2A and SHSY5Y) as models. Biochemical, molecular and cellular biology techniques were used for this study.

Results: Domain analysis shows that RBM45 contains 3 RNA-recognition motifs, sharing similar structural elements with TDP-43 and FUS. Immunolocalization showed RBM45 is a nuclear protein, and the disruption of its nuclear localization signal (NLS) mis-localizes RBM45 to the cytoplasm. Reciprocal co-immunoprecipitation assays indicated physical interactions between RBM45 and TDP-43 and FUS. RBM45 domains responsible for these protein–protein interactions were determined using truncation analysis. We also discovered that RBM45 can self-aggregate and determined the domain required for self-aggregation. The sequestering effect of RBM45 aggregation was examined. To identify the RNAs bound and regulated by RBM45, we have developed a novel CLIP-seq (crosslinking and immunoprecipitation-coupled RNA-seq) approach. The CLIP-seq data and the genes jointly regulated by RBM45, TDP-43 and FUS will be presented.

Discussion: Our results demonstrate that RBM45 is a new RNA-binding protein implicated in ALS and FTLD. We propose that RBM45 is a nuclear protein that functions in RNA splicing and transport. Nuclear import defects and environmental stress mis-localize RBM45 to the cytoplasm, where cytoplasmic aggregation can occur. The accumulation of cytoplasmic RBM45 disrupts normal RNA processing, contributing to cell death. Future mechanistic studies of RBM45 are warranted to further define the roles of RBM45 inneurodegeneration, which will broaden therapeutic options for ALS and FTLD.

Acknowledgement:

Funding support by NS061867 and NS068179 to R Bowser.

References:

• Collins M, et al. The RNA-binding motif 45 (RBM45) protein accumulates in inclusion bodies in amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration with TDP-43 inclusions (FTLD-TDP) patients. Acta Neuropathol 2012;124(5): 717–32.
   PubMed Web of Science ®Google Scholar

P270 THE RNA-BINDING PROTEIN RBM45 ASSOCIATES WITH NUCLEAR STRESS BODIES DURING CELLULAR STRESS EVENTS

Collins M¹

Rice R²

Bowser R²

^atUniversity of Pittsburgh, Pittsburgh, PA, USA

^auBarrow Neurological Institute, Phoenix, AZ, USA

Email address for correspondence: [email protected]

Keywords: RBM45, RNA binding proteins, nuclear stress bodies

Background: The RNA-binding protein RBM45 is a component of the inclusion bodies found in neurons and glia in neurodegenerative disorders such as amyotrophic lateral sclerosis (ALS), frontotemporal lobar degeneration (FTLD), and Alzheimer's disease (AD). In neurons in human neurodegenerative disease post-mortem tissue, RBM45 also exhibits a speckled nuclear staining pattern that is independent of inclusion bodies. This staining pattern is suggestive of several subnuclear structures, including nuclear speckles, nuclear gems, Cajal bodies, and nuclear stress bodies. Which of these contain RBM45 is unclear.

Objectives: The purpose of this study was to investigate the potential association of RBM45 with a variety of subnuclear structures, and determine the biological factors governing the assembly of RBM45 into granules within the nucleus.

Methods: We examined subnuclear structures via immunocytochemistry using well-characterized markers of these structures, including nuclear speckles (SC35), Cajal bodies (coilin), nuclear gems (SMN), and nuclear stress bodies (HSF1 and SAM68). To investigate potential mechanisms of RBM45 inclusion formation, we also examined stress granules using TIA-1 and TIAR. The presence of absence of RBM45 within each subnuclear structure was determined using a combined digital deconvolution/quantitative immuno- colocalization approach. Automated image analysis was also used to determine the number and size of RBM45-positive granules within the nucleus of cells. These methods were supplemented with co-immunoprecipitation experiments to identify potential RBM45-interacting proteins.

Results: We found that RBM45 does not co-localize with markers of nuclear speckles (SC35), Cajal bodies (coilin), or nuclear gems (SMN) and likewise is not a component of cytoplasmic stress granules. In untreated cells, RBM45 was diffusely localized throughout the nucleus, with few granules present. In contrast, when cells were subjected to heat shock or genotoxic stress, the number and size of RBM45-positive granules significantly increased (p < 0.05). We identified significant colocalization of RBM45 and the nuclear stress body-associated proteins HSF1 and SAM68 in stressed cells (p < 0.05). RBM45 did not exhibit significant colocalization with any of the other markers used (p > 0.05). Co- immunoprecipitation experiments demonstrated that RBM45 could be pulled down using anti-HSF1 antibodies and that HSF1 could be pulled down using anti-RBM45 antibodies.

Discussion: Collectively, these results demonstrate that RBM45 is a new component of nuclear stress bodies and define a role for RBM45 in the cellular response to stress. The incorporation of RBM45 into these structures may have important implications for the process of RBM45 inclusion formation. The results likewise identify a novel biological process as a potential therapeutic target for neurodegenerative disorders such as ALS, FTLD, and AD.

P271 CHARACTERIZATION OF THE ROLE OF SIGMA RECEPTOR 1 (SIGMAR1) IN MOTONEURON FUNCTION AND DISEASE

Bernard-Marissal N

Azzedine H

Chrast R

UNIL, DGM, Lausanne, Switzerland

Email address for correspondence: [email protected]

Keywords: sigmar1, MAMs, motoneuron

Background: Mitochondria-associated membranes (MAMs) allow a bidirectional cross-talk between endoplasmic reticulum (ER) and mitochondria. MAMs have been shown to be involved in the control of lipid biosynthesis, mitochondrial division, calcium signaling and in the dynamics of the two organelles (1). Sigmar1, which is highly expressed in motoneurons and localized in large cholinergic postsynaptic densities (2), was previously shown to play a role in MAMs (3). Importantly, previous studies revealed that a recessive mutation in SIGMAR1 is associated with a juvenile form of amyotrophic lateral sclerosis (ALS) (4). Sigmar1 localization and potentially its function was also shown to be affected in ALS models (5) underscoring the importance of MAMs for motoneuron function. While these data strongly support an involvement of Sigmar1 in MND, the mechanisms linking a loss and/or inactivation of Sigmar1 to motoneuron dysfunction remain to be clarified.

Objectives: Our project aims at characterizing the role of Sigmar1 in the function of motoneurons both in vitro and in vivo, in order to understand whether and how its deficiency contributes to the development of neurodegenerative disorders.

Methods and results: To this end, we performed in vitro experiments using primary motoneuron cultures in which Sigmar1 was either pharmacologically blocked (via specific antagonists) or genetically deleted (Sigmar1Gt(OST422756)Lex mice). Using these paradigms, we showed that blockade of Sigmar1 was sufficient to induce death of a subpopulation of motoneurons. Interestingly we showed that motoneuron death was linked to dysregulation of calcium pathways associated with ER and mitochondria dysfunction. In addition, our preliminary in vivo histological data confirmed that Sigmar1-/- mice exhibit molecular and cellular defects in both motoneurons and muscles. Those defects may underlie previously described motor disabilities observed in this model (2).

Discussion and conclusion: Our results so far confirm the critical role of Sigmar1 in motoneuron function, and provide insight into the pathophysiological mechanisms involved in the development of ALS (3). We anticipate that further characterization of cellular pathways linked to Sigmar1 inhibition and/or loss will contribute to the identification of new therapeutical targets in the context of ALS.

References:

• Rowland AA. 2012.
   Google Scholar
• Mavlyutov TA. 2010.
   Google Scholar
• Hayashi T. 2007.
   Google Scholar
• Al-Saif A. 2011.
   Google Scholar
• Prause HMG. 2013.
   Google Scholar

P272 METABOLOMIC APPROACH ON AN IN VITRO MODEL OF AMYOTROPHIC LATERAL SCLEROSIS, A CO-CULTURE OF ASTROCYTES AND MOTOR NEURONS EXPOSED TO OXIDATIVE STRESS

Veyrat-Durebex C^1,2

Blasco H^1,2

Dangoumau A¹

Vourc’h P^1,2

Piver E⁴

Laumonnier F¹

Jonneaux A⁵

Devos D⁵

Marchetti P⁶

Garçon G⁷

Andres CR^1,2

Corcia P^1,3

^avUMR INSERM U930, Equipe 2 Neurogénétique et Neurométabolomique, Tours, France

^awCHRU de Tours, Laboratoire de Biochimie et de biologie moléculaire, Tours, France

^axCHRU de Tours, Service de Neurologie, Tours, France

^ayINSERM U966, Tours, France

^azEA 1046, Faculté de Médecine Lille 2, Lille, France

^baINSERM U837 Equipe 4 Faculté de Médecine Université de Lille 2, Lille, France

^bbEA 4483, Faculté de Médecine Lille 2, Lille, France

Email address for correspondence: [email protected]

Keywords: metabolomics, co-culture, oxidative stress

Background: Although genetic and environmental factors have been associated with amyotrophic lateral sclerosis (ALS) and despite evidences for implication of oxidative stress and excitotoxicity, pathophysiology of ALS remains unknown in the majority of cases (1). Preliminary results from our laboratory on cerebrospinal fluid components of ALS patients indicate a potential modification of energetic substrates consumed by the central nervous system (2).

Objectives: In order to improve the knowledge of the pathophysiology of this disease and associated metabolic alterations, we developed a metabolomic approach on a cellular model-reproducing genetic (in particular mutation in SOD1 gene) and “environmental” (oxidative stress) conditions of ALS.

Methods: A co-culture model of mouse astrocytes (C8-D1A cell line) and motor neurons (NSC-34 cell lines) expressing or not human SOD1 protein wild-type or G93C mutant were exposed to oxidative stress (hydrogen peroxide and menadione). Oxidative stress status was characterized by ROS measurement, 8-hydroxy-2’-deoxyguanosine and malondialdehyde measurement, oxidant and antioxidant status evaluation and measurement of enzymes activities such as glutathione peroxidase, glutathione reductase and superoxide dismutase and evaluation of the ratio of reduced glutathione on oxidized glutathione (GSSG/GSH).

The intra- and extracellular media were collected at 3 times post-stress (for 48h) and analyzed by a multimodal analytical approach (liquid chromatography coupled to high-resolution mass spectrometry, proton nuclear magnetic resonance, and gas chromatography coupled to mass spectrometry). The evolution of metabolites concentrations during oxidative stress was studied in parallel to assessment of viability.

Results: Once the optimal conditions for the co-culture and the sample pretreatment procedure were established, we identified various metabolites such as amino acids, organic acids, and sugars. Preliminary analysis of results has showed about ten metabolites whose concentrations varied significantly during exposure to oxidative stress. In addition to increased levels of antioxidant molecules (vitamin C and taurine) in stressed co-cultures, we noted a decrease in their global metabolism, including reduced consumption of glucose and many amino acids.

Discussion and conclusion: Preliminary results from this study revealed the feasibility of a metabolomic multimodal approach on a cellular model of ALS. These data could highlight metabolic pathways altered in ALS and could open perspective of functional, genomic, and transcriptomic approaches.

Acknowledgements:

We thank the Association ARSLA for its financial support.

References:

• Barber SC, Shaw PJ. Oxidative stress in ALS: key role in motor neuron injury and therapeutic target. Free Radic Biol Med 2010;48:629–641.
   PubMed Web of Science ®Google Scholar
• Blasco H, Corcia P, Moreau C et al. 1H-NMR-based metabolomic profiling of CSF in early amyotrophic lateral sclerosis. PLoS One 2010;5:e13223.
   PubMed Web of Science ®Google Scholar

P273 ALS ASTROCYTES KILL MOTOR NEURONS VIA LIGATION OF DEATH RECEPTOR 6 BY A FRAGMENT OF N-APP/APLP1

Re DB¹

Le Verche V¹

Ikiz B¹

Alvarez M¹

Politi K¹

Doulias P²

Papadimitriou D¹

Grecco T²

Than M³

Nikolaiev A⁴

Califano A¹

Ishiropoulos H²

Tessier-Lavigne M⁴

Przedborski S¹

^bcColumbia University, New York, NY, USA

^bdThe Children’s Hospital of Philadelphia, Philadelphia, PA, USA

^beLeibniz Institute for Age Research, Leipzig, Germany

^bfDivision of Research, Genentech, Inc, San Francisco, CA, USA

Email address for correspondence: [email protected]

Keywords: non-cell autonomous death, astrocytes, death receptor 6

Mutations in superoxide dismutase-1 (SOD1) cause a form of amyotrophic lateral sclerosis (ALS). Previously, we have shown that primary or embryonic stem cell (ES)-derived MNs are killed by mutant SOD1-expressing astrocytes or by their conditioned medium (CM). Here, we show that the deleterious effects of mutant astrocytes are due to a toxic activity and not to lack of beneficial effects on MNs. This toxic activity is mediated by a negatively charged protein of ˜5–30 kDa, which we surmised might be a ligand to a MN cell surface transduction protein.

Based on this premise, using liquid chromatography-mass spectrometry analysis, we found a list of 121 potential ligands selectively enriched in, or unique to, the anionic fraction of the toxic mutant astrocyte CM. To identify the extracellular ligand mediating the astrocyte dead signal within this list and its associated cell surface transducer, we decided to adopt a non-bias genome-wide approach. For this, purified ES-MNs were exposed to the ALS astrocyte CM for 72 h and then analyzed using RNA sequencing. From these gene expression profiling data, differential protein activity was inferred through a regulatory network-based approach that we have recently developed.

We found significant changes in the activity of 84 membrane proteins (FDR < 0.01). To narrow-down this list of candidate receptors, we combined the inferred activity with previous knowledge on protein–protein interaction catalogued in the STRING V9.1 database. Specifically, we selected those putative cell surface transduction proteins whose interaction with any of the 121 mutant SOD1 astrocyte-released proteins has been reported with high confidence (STRING score, > 850). We ended up with 5 putative receptor/ligand couples (FDR, < 0.2) that we have tested systematically in our co-culture models using a combination or immunological and genetic strategies.

We found that the neutralization or genetic ablation of the second most activated receptor, death receptor 6 (DR6, also known as TNFRSF21), in the MN compartment was fully protective. In agreement, in the astrocytes, the mirror neutralization or ablation of amyloid beta precursor protein (APP) or of amyloid precursor-like protein 1 (APLP1), two potential ligands of this orphan receptor, also completely reversed MN death. It was also noted that inhibitors of beta-secretase (BACE1) were protective, whereas a recombinant of the E1 domain of N-terminal APP was toxic to MNs in a DR6-dependent manner. Supporting the relevance of these findings to ALS is our observation that astrocytes from sporadic ALS patients also kill MNs by a DR6/APP/APLP1-dependent mechanism.

Thus, the present study not only reports on a new disease mechanism that rests on a deleterious molecular interaction between MNs and glial cells, but also opens new promising therapeutic avenues for this incurable disease.

Acknowledgements:

P2ALS, NIH/NINDS NS062180, NS064191-01A1, NS042269-05A2, NS072182-01, NS062055-01A1, NS078614-01A1, DOD W81XWH-08-1-0522, W81XWH-12-1-0431, NIEHS ES009089.

P274 UNRAVELING THE MOLECULAR DEATH CASCADE TAKING PLACE IN MOTOR NEURONS IN RESPONSE TO ALS-LINKED ASTROCYTE TOXICITY

Politi K

Ikiz B

Le Verche V

Re DB

Alvarez M

Califano A

Przedborski S

Columbia University, New York, USA

Email address for correspondence: [email protected]

Keywords: non-cell autonomous death, NFkB, RIPK1

Background: Mutations in superoxide dismutase-1 (SOD1) is a cause of amyotrophic lateral sclerosis (ALS), which is characterized by the death of large cholinergic motor neurons (MNs) in the spinal cord. Transgenic mice expressing mutant SOD1 develop a phenotype, which emulates the clinical and pathological hallmarks of ALS, including the early retraction of nerve terminals from the neuromuscular junction. However, the molecular mechanisms that lead to MN death remain unknown. We have reported that wild-type primary or embryonic stem cell-derived spinal MNs (ES-MNs) show significant death upon exposure to astrocytes expressing mutant SOD1 or their conditioned medium. We have also shown that the death response to mutant SOD1-expressing astrocytes is selective to MNs in a mixed culture model.

Objectives: Here, our objective is to elucidate the molecular death cascade which is specifically induced in MNs.

Results: We found that purified ES-MNs exposed to mutant astrocyte-conditioned medium (ACM) are a powerful model system to unravel the molecular mechanism(s) underlying this ALS-linked astrocyte neurotoxicity using microarray analysis. Using this model system, we have found 72 h of exposure to mutant ACM to be the point-of-no-return for MNs. Following this finding, we have completed gene array analyses on MNs after 12, 24, 48, and 72 h of exposure to mutant ACM to decipher early transcriptional alterations. In addition to common microarray analyses to assess genes that are differentially expressed, we have utilized an innovative reverse-engineering method. Here, gene expression data are integrated on a signaling interactome to infer the differential activity of signaling proteins. Using this approach, we have found NF-kappaB transcription factor activity at 12 and 24 h to be up-regulated in MNs exposed to mutant ACM. In parallel, by using cell death profiling and pharmacological approaches, we have found that RIP1, which is instrumental to the induction of necroptosis through its interaction with RIP3, is independently necessary for MN death induction. Indeed, inhibition of RIP1 kinase activity by the pharmacologic agent necrostatin-1 (Nec-1), or viral sh-RNA mediated knockdown of RIP1 transcript protects MNs from the death-mediated mutant SOD1 astrocyte-mediated MN death.

We are in the process of determining whether NFKB and RIP1, which are both independently necessary for MN death induction, function in the same pathway or distinct pathways.

Discussion: In further understanding the molecular pathways underlying MN death induction in our model, we anticipate that the generated information will be of critical importance for the development of effective neuroprotective therapies for familial ALS linked to mutant SOD1, as well as for sporadic ALS, as we and others are reporting that human sporadic ALS astrocytes are also toxic to MNs.

Acknowledgements:

Supported by Project-ALS, the ALS Association, P2ALS, NIH/NINDS Grants NS062180, NS064191-01A1, NS042269-05A2, NS072182-01, NS062055-01A1, NS078614-01A1, NIEHS ES009089, Philippe Foundation, TL1 Award TR000082-07 from NIH/NCATS.

P275 ABNORMAL GLUTAMATE RELEASE INDUCED BY GROUP I METABOTROPIC GLUTAMATE RECEPTORS IN EXPERIMENTAL ALS

Bonifacino T¹

Giribaldi F¹

Milanese M¹

Rossi PIA²

Pittaluga A^1,3

Puliti A²

Usai C⁴

Bonanno G^1,3

^bgDepartment of Pharmacy, Pharmacology and Toxicology Unit

^bhMolecular Genetic and Cytogenetic Unit, Gaslini Institute and Department of Pediatric Sciences

^biCenter of Excellence for Biomedical Research University of Genoa, Genoa, Italy

^bjInstitute of Biophysics, National Research Council, Genoa, Italy

Email address for correspondence: [email protected]

Keywords: excitotoxicity, glutamate release, presynaptic mGlu1/mGlu5 receptors

Background: The mechanisms of neuronal death in ALS are still largely obscure. It is well known that glutamate (Glu)-mediated excitotoxicity plays a major role in the degeneration of motor neurons. Although astrocyte uptake is thought to be a major cause (1), according to our previous studies, we have suggested that the high levels of synaptic Glu are also due to abnormal release of the amino acid (2).

Objectives: To investigate the modulation of Glu release by Group I metabotropic glutamate auto-receptors (mGluR1 and mGluR5) in spinal cord of SOD1^G93A mice.

Methods: SOD1^G93A mice expressing high-copy number of mutant human SOD1 with a Gly93Ala substitution and wild-type human SOD1-expressing control mice were used (3). Spinal cord synaptosomes were purified by homogenization and separation on discontinuous Percoll^® gradient and used for glutamate release, measured by labeling synaptosomes with [³H]D-Aspartate ([³H]D-Asp), and confocal microscopy experiments (4), Western blot protein determination (5), cytosolic IP3 quantification, was measured using a commercial assay kit.

Results: Exposure of spinal cord synaptosomes to increasing concentrations of 3,5-DHPG, a mGluR1/5 agonist, produced distinct effects in SOD1^G93A and control mice: concentration above 0.3 μM stimulated the basal release of [³H]D-Asp, both in control and in SOD1^G93A mice. At variance, concentrations of 3,5-DHPG equal to or lower than 0.3 μM increased [³H]D-Asp release in SOD1^G93A mice only. Experiments with selective mGluR1 or mGluR5 antagonists indicated that the 3,5-DHPG effects involved both mGluR1 and mGluR5 activation. According to release experiments, high 3,5-DHPG concentrations increased IP3 in both mouse strains, whereas low 3,5-DHPG induced IP3 formation in SOD1^G93A mice only. 3,5-DHPG elicited [³H]D-Asp exocytotic release involving intra-terminal Ca²⁺ release through IP3-sensitive channels. Confocal microscopy indicated the co-existence of both receptors in the same glutamatergic nerve terminal and Western blot analysis showed higher expression of mGluR5 in SOD1^G93A mice.

Discussion and conclusion: We can conclude that the activation of both mGluR1 and mGluR5, sited at spinal cord glutamatergic nerve terminals, produces abnormal Glu release in SOD1^G93Amice, suggesting their involvement in determining the high extracellular Glu levels present in ALS. These results would prompt for new pharmacological approaches.

Acknowledgements:

This work was supported by grants from Ministero dell’Università e Ricerca, by Compagnia San Paolo Torino and by Ministero della Salute.

References:

• Rothstein JD, Van Kammen M, LeveyAI et al. Ann. Neurol. 1995;38(1):73–84.
   PubMed Web of Science ®Google Scholar
• Milanese M, Zappettini S, Onofri F et al. J. Neurochem. 2011;116(6):1028–1042.
   PubMed Web of Science ®Google Scholar
• Gurney ME, Pu H, Chiu AY et al. Science 1994;264: 1772–1775.
   PubMed Web of Science ®Google Scholar
• Stigliani S, Zappettini S, Raiteri L et al. J. Neurochem. 2006;96:656–668.
   Google Scholar
• Laemmli UK Nature 1970;227(5259):680–685.
   PubMed Web of Science ®Google Scholar

P276 MOLECULAR MECHANISM UNDERLYING EXCESSIVE AND PRECOCIOUS GLUTAMATE RELEASE IN THE SPINAL CORD OF THE SOD1G93A MOUSE MODEL OF AMYOTROPHIC LATERAL SCLEROSIS

Milanese M¹

Bonifacino T¹

Giribaldi F¹

Musazzi L²

Treccani G²

Onofri F³

Usai C⁴

Popoli M²

Bonanno G^1,5

^bkDepartment of Pharmacy, Pharmacology and Toxicology Unit, Genova, Italy

^blDepartment of Pharmacological and Biomolecular Sciences, Laboratory of Neuropsychopharmacology and Functional Neurogenomics, Milano, Italy

^bmDepartment of Experimental Medicine, Section of Physiology, Genova, Italy

^bnInst. of Biophysics, National Research Council, Genova, Italy

^boCenter of Excellence for Biomedical Research, Genova, Italy

Email address for correspondence: [email protected]

Keywords: glutamate release, excitotoxicity, synaptic proteins

Background: Selective and progressive neurodegeneration of upper and lower motoneurons is a typical hallmark of amyotrophic lateral sclerosis (ALS). Motoneuron cell death in ALS has been ascribed to multiple causes, including glutamate (Glu)-mediated excitotoxicity. Impaired astrocytic Glu transport was suggested as the reason of high synaptic Glu levels (1); however, other causes may concur in elevating extracellular Glu in ALS. Our previous studies highlighted the presence of abnormal Glu release induced by activation of heterologous GABA and glycine transporters expressed at Glu-releasing nerve terminals in the spinal cord of SOD1^G93A mice (2–4).

Objectives: This study was aimed to verify the occurrence of excessive Glu exocytosis in experimental ALS as a possible cause of the disease.

Methods: Pre-symptomatic and symptomatic SOD1^G93A mice were used. Nerve terminals (synaptosomes) were obtained from the spinal cord of control and SOD1^G93A mice (5), and utilized for release experiments by exploiting the superfusion technique. Cytosolic calcium determination, confocal microscopy and Western blot experiments were performed to investigate the expression and the activation state of synaptic proteins.

Results: Both the spontaneous and the stimulus-evoked exocytotic Glu release was increased in SOD1^G93A symptomatic and pre-symptomatic mice, compared to that in controls. Increased pre-synaptic Ca²⁺ levels, over-activation of calcium/calmodulin-dependent kinase-II and ERK/MAP kinases, as well as hyper-phosphorylation of synapsin-I, were determined at both symptomatic and pre-symptomatic clinical stages. The expression of several synaptic proteins implicated in neurotransmission does not show differences, except for the over-expression of synaptotagmin in SOD1^G93A mice. Pre-synaptic hyper-phosphorylation at the inhibitory sites of GSK3 and accumulation of high molecular weight SNARE complexes was also measured. In line with these findings, release experiments suggested that the excessive Glu exocytosis involves the readily releasable pool of vesicles.

Discussion and conclusion: Our results indicate a dysregulation of glutamate exocytosis in the spinal cord of symptomatic and pre-symptomatic SOD1^G93A mice. This event was accompanied by marked changes in a number of pre-synaptic molecular mechanisms that lead to a significant augmentation of the readily releasable pool of vesicles and determine a higher probability of vesicles to fuse. Synaptic alterations, found to be present also in a pre-symptomatic stage, could represent a key feature in the early phase of ALS, thus playing a role in the etiopathogenesis of the disease.

References:

• Rothstein JD, Tsai G, Kuncl RW et al. Ann. Neurol. 1990;28:18–25.
   PubMed Web of Science ®Google Scholar
• Raiteri L, Paolucci E, Prisco S et al. Br. J. Pharmacol. 2003;138:1021–1025.
   Google Scholar
• Raiteri L, Stigliani S, Zappettini S et al. Neuropharmacology 2004;46:782–792.
   PubMed Web of Science ®Google Scholar
• Milanese M, Zappettini S, Jachetti E et al. J. Neurochem 2010;113:489–501.
   PubMed Web of Science ®Google Scholar
• Stigliani S, Zappettini S, Raiteri L et al. J. Neurochem. 2006;96:656–668.
   Google Scholar

P277 REDUCE GLUTAMATE UPTAKE IN ACTIVATED ASTROCYTES IN SOD1(G93A) MICE AND ITS IMPLICATIONS ON ALS PATHOGENESIS

Benkler C

Barhum B

Ben-Zur T

Offen D

Tel Aviv University, Petha Tikva, Israel

Email address for correspondence: [email protected]

Keywords: neurotrophic factors, glutamate, astrocyte activation

Background: Astrocyte activation occurs in response to central nervous system (CNS) insult and is considered a double-edged sword in many pathological conditions. We propose that reduced glutamatergic and trophic response of astrocytes to activation may, over time, lead to accumulative CNS damage, thus facilitating neurodegeneration.

Results: We found that astrocytes derived from the SOD1^G93A ALS mouse model exhibit a reduced glutamatergic and trophic response to specific activations compared to their wild-type counterparts. Wild-type astrocytes exhibited a robust response when activated with lipopolysaccharide (LPS), G5 or treated with ceftriaxone in many parameters evaluated. These parameters include increased expression of GLT-1 and GLAST the two major astrocytic glutamate transporters, accompanied by a marked increase in the astrocytic glutamate clearance and up-regulation of neurotrophic factor expression. However, not only do un-treated SOD1^G93A astrocytes take up glutamate less efficiently, but in response to activation they show no further increase in any of the glutamatergic parameters evaluated. Furthermore, activation of wild-type astrocytes, but not SOD1^G93A astrocytes, improved their ability to protect the motor neuron cell line NSC-34 from glutamate-induced excitotoxicity. All experiments were performed at least in three separate biological replications in triplicates.

Conclusion: Our data indicate that altered astrocyte activation may well be pivotal to the pathogenesis of ALS.

P278 E6-AP PROMOTES SOD1 PROTEIN DEGRADATION AND SUPPRESSES MUTANT SOD1 TOXICITY

Mishra A²

Maheshwari M⁴

Chhangani D²

Fujimori-Tonou N³

Endo F^1,3

Prakash Joshi A¹

Jana N²

Yamanaka K^1,3

^bpNagoya University, Nagoya, Japan

^bqIndian Institute of Technology Rajasthan, Jodhpur, India

^brRIKEN Brain Science Institute, Wako, Japan

^bsNational Brain Research Centre, Dehli, India

Email address for correspondence: [email protected]

Keywords: protein misfolding, SOD1, degradation

Background: Recent studies indicated that E6-AP, a homologous to E6-AP C terminus (HECT)-type E3 ubiquitin ligase, has a role in cellular protein quality control system and is implicated in the degradation of expanded polyglutamine proteins. However, the roles of E6-AP in motor neuron diseases have not been explored.

Objectives: The aim of this study is to examine the role of E6-AP in mutant SOD1-mediated ALS.

Methods: The expression levels and localization of E6-AP proteins were examined in the spinal cord of various mutant SOD1 mice. Using cultured cells, protein interaction between SOD1 and E6-AP and the ubiquitination and degradation of SOD1 protein in the presence of E6-AP were analyzed. Further, the role of E6-AP in ameliorating mutant SOD1-mediated cellular toxicity was analyzed in cultured cells with overexpression of molecular chaperone, Hsp70.

Results: Endogenous E6-AP was depleted from the motor neuron nuclei of SOD1-ALS mouse models prior to neurodegeneration. E6-AP co-immunoprecipitates with the SOD1 protein and is predominantly mislocalized in mutant SOD1-containing inclusion bodies. Overexpression of E6-AP increases the ubiquitination and facilitates degradation of SOD1 proteins. Finally, overexpression of E6-AP suppresses the aggregation and cell death mediated by mutated SOD1 proteins, and cellular protective effect is more prominent when E6-AP is overexpressed along with Hsp70.

Discussion and conclusion: These data suggest that enhancing the activity of E6-AP ubiquitin ligase might be a viable therapeutic strategy to eliminate mutant SOD1-mediated toxicity in ALS.

P279 CYSTATIN C PROTECTED NEURONAL CELLS AGAINST MUTANT COPPER-ZINC SUPEROXIDE DISMUTASE-MEDIATED TOXICITY IN VITRO

Watanabe S¹

Wakasugi K²

Yamanaka K^1,3

^btLaboratory for motor neuron disease, RIKEN Brain Science Institute, Wako-shi, Saitama, Japan

^buDepartment of Life sciences, Graduate school of Arts and Sciences, University of Tokyo, Meguro-ku, Tokyo, Japan

^bvResearch Institute of Environmental Medicine, Nagoya University, Nagoya-shi, Aichi, Japan

Email address for correspondence: [email protected]

Keywords: cystatin C, superoxide dismutase 1, autophagy

Background: Recent studies have demonstrated Cystatin C (CysC), an endogenous cysteine protease inhibitor, plays neuroprotective roles in Alzheimer’s and Parkinson’s diseases. In sporadic ALS, CysC is a major component of Bunina bodies and decreased in the patients’ cerebrospinal fluid. However, it remains unclear whether CysC has a protective role in ALS.

Objectives: The aim of this study is to examine the potential of CysC as a novel neuroprotective agent in ALS.

Methods: We added recombinant human CysC to neuro2a cells transiently expressing mutant SOD1. The cell viability was measured using MTS assay. We also examined the neuroprotective effect of CysC using the primary neuron-glia mix culture derived from Hb9:GFP/SOD1^G85R double transgenic mouse embryo. Furthermore, to investigate the neuroprotective mechanism of CysC, we examined the possible involvement of autophagy pathway.

Results: Exogenously added CysC protected neuro2a cells against mutant SOD1-mediated cytotoxicity in a dose- dependent manner. Intracellular aggregates of mutant SOD1 were remarkably decreased in the CysC-treated cells. CysC was also protective for primary cultured neurons. Moreover, the level of LC3-II was increased in the CysC-treated cells and 3-methyl adenine treatment clearly inhibited the protective effect of CysC, suggesting that the induction of autophagy was required for neuroprotection by CysC.

Discussion and conclusion: These findings suggest that CysC is a novel therapeutic candidate that can protect neurons against mutant SOD1 toxicity through induction of autophagy-lysosomal pathway.

P280 INCREASED AMPK ACTIVITY AND DOWNREGULATION OF HSP70 EXPRESSION DECREASE THE LIFESPAN OF SOD1G93A MOUSE MODEL OF AMYOTROPHIC LATERAL SCLEROSIS

Zhao Z

Sui Y

Gao W

Cai B

Fan D

Peking University Third Hospital, Beijing, China

Email address for correspondence: [email protected]

Keywords: AMPK, Hsp70, high-fat diet

Metabolic abnormalities are observed in amyotrophic lateral sclerosis (ALS). The molecular mechanisms of the abnormal metabolism remain unclear. AMP-activated protein kinase (AMPK) is essential for neuronal integrity and survival. AMPK is activated by metabolic stresses, including ischemia, hypoxia and caloric restriction (CR), while hyperglycaemia, feeding and a high-fat diet (HFD) inhibit AMPK activation. In this study, we found that AMPK activity increased and heat shock protein-70 (Hsp70) was downregulated in the spinal cord of SOD1^G93A transgenic mice compared to SOD1^WT controls. Furthermore, we showed that CR increased AMPK activity, downregulated Hsp70 expression, aggravated the loss of motor neurons, hastened disease onset and reduced the lifespan of SOD1^G93A transgenic mice compared to ad libitum (AL) SOD1^G93A controls. In contrast, HFD inhibited AMPK activity and motor neuron loss, upregulated Hsp70 expression, delayed disease onset and extended survival in these mice. These results suggest that AMPK activity plays a negative role in the survival of motor neurons, possibly through a novel mechanism involving Hsp70 downregulation, such that treatments causing inhibition of AMPK slow the progression of motor neuron disease in this mouse model.

Acknowledgments:

This study was supported by grants from the National Natural Sciences Foundation of China (81030019) and Doctoral Fund of Chinese Ministry of Education (20100001110084).

P281 PHENOTYPIC DISCOVERY AND CHARACTERIZATION OF NEUROPROTECTIVE COMPOUNDS RELEVANT TO ALS

Rudhard Y¹

Höing S²

Reinhardt P²

Glatza M²

Slack M¹

Schöler H R²

Sterneckert J²

^bwEvotec AG, Hamburg, Germany

^bxDepartment of Cell and Developmental Biology, Max Planck Institute for Molecular Biomedicine, Münster, Germany

Email address for correspondence: [email protected]

Keywords: drug discovery, high content screening, in vitro pharmacology

Disease modalities such as Parkinson’s, ALS and Alzheimer’s represent complex neurological disorders where certain neuronal populations decline via only partially understood mechanisms that frequently involve neuro-inflammation. In order to identify new and disease-relevant, neuroprotective compounds for ALS, we built a screenable phenotypic assay that re-builds aspects of the disease in a microtiter plate. We modelled the non-cell autonomous nature of ALS by integrating stem cell-derived motor neurons, astrocytes and activated microglia into the screening set up. As primary biomarker of ALS, we used degeneration of motor neurons and identified a small number of neuroprotective compounds in a screen of 11,000 compounds. Through diverse secondary assays, we showed that these hit compounds act through multiple mechanisms, including the inhibition of microglial activation, directed protection of neurons from nitric-oxide-induced degeneration, and glia-specific activation of genes controlled by Nrf2, a transcription factor previously shown to be effective in the SOD1^G93A mouse model.

The modular nature of the presented phenotypic assay allows for integration of both additional stress paradigms and cells relevant to specific types of neurodegenerative disease. Further integration of chemical biology, proteomics, and network biology tools will allow unravelling targets and molecular mechanisms of action. Thus, hit compounds identified and characterized in these ways may represent ideal starting points for the development of new drugs to treat various neurodegenerative diseases.

P282 VITAMIN D CONFERS PROTECTION TO MOTONEURONS AND IS A PROGNOSTIC FACTOR OF AMYOTROPHIC LATERAL SCLEROSIS

Camu W^1,2

Tremblier B²

Salasc C²

Scamps F²

Alphandery S¹

Pageot N¹

Juntas-Morales R¹

Raoul C²

^byALS center, Montpellier, France

^bzThe Neuroscience Institute of Montpellier, Inserm UMR1051, Montpellier, France

Email address for correspondence: [email protected]

Keywords: vitamin D, motoneuron culture, survival

Background: Vitamin D (VD) is a potent secosteroid hormone with diverse biological functions that include protection against neuronal damage. The detrimental consequences of VD dietary deficiency have been documented in several neurodegenerative diseases. We reported in 2012 that ALS patients with severe VD deficiency had a 6 times more rapid worsening than those with normal levels.

Objectives: We assessed the effect of VD on rat purified motoneurons in vitro. We also actualized data from our ALS cohort, for overall survival.

Methods: We studied the effect of 1,25(OH)₂ D₃ (the biologically active hormone),vitamin D2 (VD2) and vitamin D3 (VD3) on the electrophysiological properties and survival features of purified motoneurons (MNs) from E12.5 mice embryos. Purified MNs were obtained using iodixanol density gradient centrifugation and then plated on poly-ornithine/laminin-treated wells.

Results: The addition of 1,25(OH)₂D₃ in the culture medium increased intensity of nuclear staining of theVD receptor (VDR), while addition of VD₂ and VD₃ did not. In culture, 1,25(OH)₂D₃ significantly improved MN survival by 45% (100nM, p < 0.001), while VD2 and VD3 did not. Interestingly, the effect of 1,25(OH)₂D₃ was abolished when neurotrophic factors were removed from the culture medium. As accumulating evidence suggests that the MN-restricted Fas death pathway may contribute to the degenerative process in ALS, MNs were cultured for 24 h before being treated for 48 h with soluble Fas Ligand in combination of either VD₂, D₃ or 1,25(OH)₂D₃. After this delay, 1,25(OH)₂D₃ completely rescued MNs from Fas-induced death, while neither VD₂ nor D₃ influenced cell survival. In vitro, 1,25(OH)₂D₃, did not modify electrical properties of MNs.

In parallel, we determined the outcome of our initial group of 74 ALS patients with VD level measurement. These patients were included in a follow-up, until death, between 2010 and 2011. By April 2013, 100% of the patients with severe VD deficiency (< 25 nmol/l) were dead, compared to 62% of those belonging to the group with VD deficiency (>25 and < 75) and only 25% of the patients with normal levels. This corresponded to a median survival of 22, 31 and 49 months for the groups of severe deficiency, deficiency and normal levels, respectively (p < 0.001).

Discussion: In ALS patients, VD deficiency is associated with a worse outcome when compared to patients with normal levels. In vitro experiments give results that are consistent with clinical data, showing that the active hormone 1,25(OH)₂D₃ promotes MN survival and protects against the Fas-induced apoptotic pathway. These concordant results may pave the way towards new therapeutic approaches in ALS.

Acknowledgements:

We thank the INSERM, the AFM and the ARSLA for their financial support.

P283 SODIUM AND CALCIUM OVERLOAD INDUCED BY VERATRIDINE IN NSC-34 CELLS: A NOVEL IN VITRO MODEL OF ALS TO EXPLORE NEW NEUROPROTECTIVE COMPOUNDS

Cano-Abad MF^1,2

Moreno-Ortega AJ^1,3

Mouhid L^1,3

Ruiz-Nuño A^1,3

^caInstituto Teófilo Hernando, Madrid, Spain

^cbDepartamento de Farmacología y Terapéutica, Facultad de Medicina, Universidad Autónoma de Madrid, Madrid, Spain

^ccServicio de Farmacología Clínica, Instituto de Investigación Sanitaria, Hospital Universitario de la Princesa, Universidad Autónoma de Madrid, Madrid, Spain

Email address for correspondence: [email protected]

Keywords: NSC-34 cells, neuroprotection, TDP-43

Background: At present, therapeutic alternatives for amyotrophic lateral sclerosis (ALS) are scarce and controversial. Only riluzole delays death a few months. The aetiology of motorneuron vulnerability in this disease is being studied in NSC-34 cells that were created as hybrids of neuroblastoma cells and spinal motor neurons to find out the underlying mechanisms leading to the selective loss of motorneurons in ALS (1). One such mechanism is associated withTDP-43 (TARD DNA-binding protein 43) aggregates and Ca²⁺ buffering proteins driving more vulnerability of motoneurons degeneration.

Objectives: The main aim of this study proposes the overload of Na⁺ and Ca²⁺ induced by veratridine (VTD) in NSC-34 cell line as an in vitro model of ALS. As an additional aim, we assessed whether the cytotoxic damage evoked by VTD could be altering the expression of TDP-43 and Ca²⁺ buffering proteins.

Methods: Cell viability was measured by MTT and cytosolic concentration of Ca²⁺ ([Ca²⁺]_c) with the fluorescent probe Fura-2. Protein expression was measured using immunofluorescence.

Results: Cell viability was decreased by 79% (16 cells of 6 different experiments) after 48-h incubation with VTD (100 μM). Riluzole produced 13% neuroprotection at 1 μM and 9% at 3 μM against VTD after 6 h pre-treatment (9 cells of 3 different experiments). ITH33/IQM9.21, a novel neuroprotective compound (2, 3), did not elicit neuroprotection at any of the concentrations tested (1, 3 and 10 μM) after 2 or 6 h pre-treatment (9 cells of 3 different experiments). Furthermore, VTD induced oscillations of the cytosolic concentration of Ca²⁺ ([Ca²⁺]_c) in 46% of studied cells (141 cells of 7 different cultures). Riluzole abolished these oscillations of [Ca²⁺]_c in 93% of the cells (29 cells of 2 different cultures).

Discussion and conclusion: We propose VTD as a Na⁺ and Ca²⁺ overload in vitro model to reproduce ALS in order to test new neuroprotective compounds.

Acknowledgements:

This work was partly supported by FIS No. PI052124 to ARN and Fundación Teófilo Hernando. AJMO is granted by Ministry of Education/FPU Program Ref.: AP2009/0343.

References:

• Cashman NR, Durham HD, Blusztajn JK et al. Dev Dyn 1992;194:209–221.
   PubMed Web of Science ®Google Scholar
• Maroto M, de Diego AM, Albinana E et al. Cell Calcium 2011;50:359–369.
   PubMed Web of Science ®Google Scholar
• Lorrio S, Gómez-Rangel V, Negredo P et al. Neuropharmacology 2013;67:403–411.
   PubMed Web of Science ®Google Scholar

P284 IN VITRO EVIDENCE FOR THE THERAPEUTIC POTENTIAL OF MENSENCHYMAL STROMAL CELLS IN ALS

Sun H^1,3

Bernardais K^1,3

Stanslowsky N^1,3

Thau-Habermann N^1,3

Hensel N²

Huang D⁴

Claus P^2,3

Stangel M^1,3

Dengler R^1,3

Petri S^1,3

^cdHannover Medical School, Department of Neurology, Hannover, Germany

^ceHannover Medical School, Institute of Neuroanatomy, Hannover, Germany

^cfCenter for Systems Neuroscience, Hannover, Germany

^cgDepartment of Neurology, East Hospital, Tongji University, Shanghai, China

Email address for correspondence: [email protected]

Keywords: mesenchymal stromal cells, growth factors, inflammation

Background: Administration of mesenchymal stromal cells (MSC) has beneficial effects on motor function and survival in the SOD1G93A mouse model of ALS as well as in models of other neurological disorders.

Objectives: We intended to study the effect of the interaction between MSC and motor neurons and glial cells and to analyse the mechanisms underlying the protective effects of MSC in vitro.

Methods: MSC or MSC-conditioned medium (MSC CM) was added to embryonic primary motor neurons (derived from both non-transgenic and mutant SOD1^G93A transgenic mice), NSC-34 cells and glial cells (astrocytes and microglia, derived from both non-transgenic and mutant SOD1^G93A transgenic mice), and the effects against staurosporine-induced cell death were determined by immunocytochemistry and MTT assay. MSC CM-induced changes in mRNA expression of pro-and anti-inflammatory mediators in glial cells and on growth factor expression in motor neurons were quantified by real-time PCR.

Results: In primary motor neurons, NSC-34 cells and astrocytes, MSC CM attenuated staurosporine-induced apoptosis in a concentration-dependent manner. Studying MSC CM-induced expression of neurotrophic factors in astrocytes and NSC-34 cells, we found that glial cell line-derived neurotrophic factor (GDNF) and ciliary neurotrophic factor (CNTF) gene expression in astrocytes were significantly enhanced by MSC CM, with differential responses of non-transgenic and mutant astrocytes. Expression of vascular endothelial growth factor (VEGF) in NSC-34 cells was significantly upregulated upon MSC CM-treatment. MSC CM significantly reduced the expression of the cytokines TNFα and IL-6 and of iNOS in both transgenic and non-transgenic astrocytes. Gene expression of the neuroprotective chemokine Fractalkine (CX3CL1) was also upregulated in mutant SOD1^G93A transgenic astrocytes by MSC CM treatment. Correspondingly, MSC CM increased the respective receptor, CX3CR1, in mutant SOD1^G93A transgenic microglia.

Discussion and conclusion: Our data demonstrate that MSC modulate the motor neuronal and glial response to apoptosis and inflammation, and induce gene expression changes capable to contribute to neuroprotection. MSC therefore represent an interesting candidate for further preclinical and clinical evaluation in ALS.

P285 EMBRYONIC STEM CELL-DERIVED MOTONEURON/MUSCLE FIBER CO-CULTURES: A MODEL SYSTEM FOR STUDYING AMYOTROPHIC LATERAL SCLEROSIS PATHOPHYSIOLOGY

Shettar B

Rafuse V

Dalhousie University, Halifax, Canada

Email address for correspondence: [email protected]

Keywords: Motor neuron, Stem cells, co-culture

Background: Amyotrophic lateral sclerosis (ALS) is an adult-onset neurodegenerative disease characterized by motor neuron death. This devastating disease leads to paralysis of muscles of respiration and those of the limbs, eventually leading to death. Genetic defects in the copper/zinc superoxide dismutase (mSOD1) gene have been linked to familial ALS. One of the early features in the disease process is denervation of the neuromuscular junction (NMJ). It is not known whether denervation is simply a result of motor neuron (MN) death or whether denervation is one of the contributing variables causing the MNs to die. To address this question, we co-cultured wild-type mouse embryonic stem cell-derived motor neurons (ECSMNs) with chick myotubes to correlate NMJ formation, synaptic stability with the onset and time course of ECSMN cell death.

Methods: ESCMNs were derived from embryonic stem cells isolated from wild-type and transgenic mice expressing the G93A mutant form of humanSOD1. Both strains of mice expressed eGFP under the control of the motor neuron-specific Hb9 promoter. Stem cells were directed to differentiate into ESCMNs using retinoic acid and a smoothen agonist. Five days later, the ESCMNs were plated onto chick myotubes in a serum-free media that was supplemented with GDNF and CNTF. Growth factors were withdrawn 1 week later to promote neuromuscular synapse formation. ESCMS survival and NMJ morphology were quantified and correlated over the next 3–4 weeks.

Results and discussion: To date, our results indicate that there is no difference in the survival of WT and mSOD1 ESCMNs after the first week of co-culture in the presence of growth factors. After the withdrawal of growth factors, NMJs typically mature in WT ESCMN/myotube co-cultures over the next 2- to 3-week period. Maturation is characterized by an increase in size and complexity of the pre- and post synaptic structures at putative NMJs (ie, neurite, synaptic vesicles and acetylcholine receptors). Unlike WT ESCMN/myotube co-cultures, very few mature NMJs were observed in the mSOD1 ESCMNs co-cultures. The lack of mature NMJs was accompanied by a dramatic increase in mSOD1 ESCMN death compared to the WT ESCMN/myotube co-cultures. We are currently conducting detailed time course analysis to determine whether NMJs failed to form in the mSOD1 ESCMN/myotube co-cultures or whether they formed and later withdrew. If the latter is true, these results would support for the hypothesis that axonal withdrawal precedes, and may contribute to, MN cell death in ALS. Furthermore, it would indicate that the ESCMN/chick myotube co-culture is an ideal model system to study pathophysiologies associated with the NMJ.

P286 ALTERED NEURAL FATE OF EPENDYMAL STEM PROGENITOR CELLS IN ALS G93A-SOD1 MOUSE MODEL: THE MICRORNA REGULATION

Marcuzzo S

Kapetis D

Bonanno S

Cavalcante P

Barzago C

Bernasconi P

Mantegazza R

Fondazione Istituto Neurologico “Carlo Besta”, Milan, Italy

Email address for correspondence: [email protected]

Keywords: ependymal stem cells, microRNA

Background: Amyotrophic lateral sclerosis (ALS) is a fatal disorder characterized by motoneuron degeneration. In ALS SOD1^G93A mice, motoneuron degeneration is associated with proliferation of ependymal stem progenitor cells (epSPCs), usually quiescent in the spinal cord (1) which does not lead to restoration. MicroRNAs (miRNAs) regulate gene expression at post-transcriptional level and coordinate self-renewal and stem cell differentiation in the spinal cord (2); hence, their evaluation is crucial to understand ALS neurodegenerative mechanisms.

Objective: We investigated: a) the in vitro neurogenesis of epSPCs isolated from the spinal cord ofSOD1^G93A at asymptomatic and symptomatic phases of the disease; b) the expression of miRNAs specific for neural cell fate (miR-9 and miR-124a) and cell-cycle regulation (miR-19a and miR-19b) during neural differentiation of cultured epSPCs.

Methods: We characterized the capacity of SOD1^G93A and control epSPCs to proliferate and differentiate in vitro into the three neural cell lineages: neuron, astrocytes and oligodendrocytes. Neural cell lineages were identified by immunocytochemistry. The expression of miR-9, miR-124a, miR-19a, miR-19b, and their predicted mRNA targets were assessed using real-time PCR in undifferentiated and differentiated epSPCs.

Results: SOD1^G93A epSPCs produced neurospheres that differentiated into the three neural cell lineages: neurons were more numerous than astrocytes and smaller than those obtained from controls; astrocytes had an activated phenotype. miRNA analysis revealed that miR-9, miR-124a, miR-19a and miR-19b expression was altered during SOD1^G93A epSPC differentiation. A paired expression analysis of miRNAs and their predicted targets allowed the identification of a functional network composed of genes implicated in neural cell fate, cell maturation, cell migration and proliferation.

Discussion and conclusion: We demonstrated that SOD1^G93A epSPCs can differentiate into neural cells whose phenotypic alterations resemble those observed in ALS pathology. Analysis of the functional network suggested an involvement of miRNAs in epSPC neural fate, maturation, migration and proliferation through their effects on mRNA targets.

Our findings indicate that SOD1^G93A epSPCs are a valuable model recapitulating the pathogenetic steps of ALS and represent a new tool to investigate the mechanisms of neurodegeneration, their molecular control, and may serve to identify new targets for ALS therapy.

References:

• Guan YJ, Wang X, Wang HY et al. Increased stem cell proliferation in the spinal cord of adult amyotrophic lateral sclerosis transgenic mice. J Neurochem 2007;102: 1125–38.
   PubMed Web of Science ®Google Scholar
• Liu C, Zhao X. MicroRNAs in adult and embryonic neurogenesis. Neuromolecular Med 2009;11:141–52.
   Google Scholar