OUP user menu

Global gene expression profiling of somatic motor neuron populations with different vulnerability identify molecules and pathways of degeneration and protection

Eva Hedlund, Martin Karlsson, Teresia Osborn, Wesley Ludwig, Ole Isacson
DOI: http://dx.doi.org/10.1093/brain/awq167 2313-2330 First published online: 26 July 2010

Summary

Different somatic motor neuron subpopulations show a differential vulnerability to degeneration in diseases such as amyotrophic lateral sclerosis, spinal muscular atrophy and spinobulbar muscular atrophy. Studies in mutant superoxide dismutase 1 over-expressing amyotrophic lateral sclerosis model mice indicate that initiation of disease is intrinsic to motor neurons, while progression is promoted by astrocytes and microglia. Therefore, analysis of the normal transcriptional profile of motor neurons displaying differential vulnerability to degeneration in motor neuron disease could give important clues to the mechanisms of relative vulnerability. Global gene expression profiling of motor neurons isolated by laser capture microdissection from three anatomical nuclei of the normal rat, oculomotor/trochlear (cranial nerve 3/4), hypoglossal (cranial nerve 12) and lateral motor column of the cervical spinal cord, displaying differential vulnerability to degeneration in motor neuron disorders, identified enriched transcripts for each neuronal subpopulation. There were striking differences in the regulation of genes involved in endoplasmatic reticulum and mitochondrial function, ubiquitination, apoptosis regulation, nitrogen metabolism, calcium regulation, transport, growth and RNA processing; cellular pathways that have been implicated in motor neuron diseases. Confirmation of genes of immediate biological interest identified differential localization of insulin-like growth factor II, guanine deaminase, peripherin, early growth response 1, soluble guanylate cyclase 1A3 and placental growth factor protein. Furthermore, the cranial nerve 3/4-restricted genes insulin-like growth factor II and guanine deaminase protected spinal motor neurons from glutamate-induced toxicity (P < 0.001, ANOVA), indicating that our approach can identify factors that protect or make neurons more susceptible to degeneration.

  • motor neuron
  • SOD1G93A rat
  • microarray
  • hierarchical clustering
  • cranial nerves
  • cervical spinal cord
  • IGF-II

Introduction

Neurodegenerative diseases are characterized by the selective vulnerability of specific neuronal populations to toxic processes of genetic and/or environmental origin. Somatic motor neurons degenerate in diseases such as amyotrophic lateral sclerosis, spinal muscular atrophy and spinobulbar muscular atrophy. However, not all somatic motor neurons are equally affected by the events leading to degeneration. While ventral spinal motor neurons are affected in all three diseases, and motor neurons of the lower cranial nerves (CNs) [e.g. hypoglossal (CN12)] degenerate in amyotrophic lateral sclerosis and spinobulbar muscular atrophy, upper CNs [e.g. oculomotor/trochlear (CN3/4)] are generally spared in spinal muscular atrophy, amyotrophic lateral sclerosis and spinobulbar muscular atrophy (Leveille et al., 1982; Gizzi et al., 1992; Reiner et al., 1995; Sobue, 1995; Nimchinsky et al., 2000; Haenggeli and Kato, 2002). While spinal muscular atrophy is recessively inherited and caused by loss of functional survival of motor neuron (SMN1) protein (Bussaglia et al., 1995; Lefebvre et al., 1995), spinobulbar muscular atrophy is an X-linked disorder, caused by the expansion of CAG repeats in the androgen receptor gene (Kennedy et al., 1968; La Spada et al., 1991). The vast majority of amyotrophic lateral sclerosis cases appear sporadic (∼90%). However, amyotrophic lateral sclerosis can be inherited (familial) (∼10%) due to mutations in e.g. superoxide dismutase 1 (SOD1) (Rosen et al., 1993), angiogenin (Greenway et al., 2006) or the DNA/RNA-binding proteins transactivation response element DNA binding protein (TDP-43) (Kabashi et al., 2008; Sreedharan et al., 2008) or FUS (Kwiatkowski et al., 2009; Vance et al., 2009). Importantly, differential vulnerability among motor neurons appears independent of the cause of disease, since the pathology and pattern of selective motor neuron vulnerability is similar in familial and sporadic amyotrophic lateral sclerosis (Shaw et al., 1997).

Neurodegeneration in many diseases appears to involve cell-autonomous and non-cell-autonomous events (Garden et al., 2002; Boillee et al., 2006; Jaarsma et al., 2008; Chung et al., 2009). Familial amyotrophic lateral sclerosis model data indicate that factors intrinsic of motor neurons are crucial for initiation of degeneration, while non-cell-autonomous events are instrumental for disease progression (Beers et al., 2006; Boillee et al., 2006; Nagai et al., 2007; Di Giorgio et al., 2008; Hedlund and Isacson, 2008; Marchetto et al., 2008; Yamanaka et al., 2008). Furthermore, neuronal death in Huntington’s disease, a polyglutamine expansion disease like spinobulbar muscular atrophy, involves intrinsic and exogenous events (Ross, 2004; Gu et al., 2005), suggesting that spinobulbar muscular atrophy could be due to a combination of these two.

We therefore hypothesized that dissecting the intrinsic molecular code underlying the normal physiology of motor neurons that display differential vulnerability to disease could provide a basis for revealing why one motor neuron subpopulation is more vulnerable to degeneration than another. Previous analysis of the differential vulnerability of substantia nigra and ventral tegmental area dopamine neurons to degeneration in Parkinson’s disease showed that the differential gene expression pattern of neuronal populations in the normal animal can be used to elucidate mechanisms that can protect vulnerable neurons from disease (Chung et al., 2005). Here we used laser capture microscopy (LCM) to isolate motor neurons from CN3/4, CN12 and the lateral motor column of the cervical spinal cord of the normal rat and performed an analysis of the entire rat transcriptome. Differential expression of selected genes with implications for motor neuron vulnerability was confirmed by localization of the resulting proteins. Functional in vitro analysis revealed that the CN3/4-specific proteins insulin like growth factor (IGF)-II and guanine deaminase, could protect motor neurons from glutamate-induced toxicity.

We believe that this report provides insight into the intrinsic properties of different motor neuron subpopulations and gives important clues to mechanisms of relative vulnerability. Therefore, our extensive expression analysis could provide a basis for understanding why degeneration in amyotrophic lateral sclerosis, spinal muscular atrophy and spinobulbar muscular atrophy involve some, but not all, motor neuron populations and may hopefully be used to develop treatments for these diseases.

Materials and methods

Animal procedures

All animal procedures were performed in accordance with the National Institute of Health guidelines and were approved by the Animal Care and Use Committee at McLean Hospital, Harvard Medical School. Animals were housed according to standard conditions, with access to food and water ad libitum and a dark/light cycle of 12 h.

Analysis of disease onset in the SOD1G93A rat model of amyotrophic lateral sclerosis

Transgenic rats over-expressing mutant SOD1 (SOD1G93A) were used as a model of amyotrophic lateral sclerosis (Howland et al., 2002). Disease onset in these animals was carefully monitored (for details, see Supplementary material).

Tissue preparation

Presymptomatic, 60-day-old female SOD1G93A transgenic and wild-type litter-mates (Taconic) and symptomatic SOD1G93A rats and age-matched wild-type litter-mates were anaesthetized with sodium pentobarbital (150 mg/kg i.p.). For LCM and real-time polymerase chain reaction, tissues were removed, snap-frozen in 2-methylbutane (−60°C), sectioned (12 µM coronal sections), mounted onto LCM slides (Arcturus Engineering, Inc, Mountain View, CA) and stored at –70°C. For immunohistochemistry, animals were perfused intracardially with 100 ml heparinized saline and 200 ml 4% paraformaldehyde. Brains, brain stems and spinal cords were dissected, post-fixed for 6 h, cryoprotected in 20% sucrose, sectioned (30–40 µm) and stored at −70°C.

Quantification of differential motor neuron loss in the SOD1G93A rat

The number of motor neurons present in the CN3/4, trigeminal nucleus (CN5), facial nucleus (CN7), CN12 and in the lateral motor column across C2 and C3 segments (Kakinohana et al., 2004) in the cervical spinal cord from 60-day-old SOD1G93A rats, 60-day-old wild-type litter mates and symptomatic SOD1G93A rats and age-matched wild-type litter mates were quantified. Sections were incubated with blocking buffer (phosphate buffered saline, 10% normal donkey serum or normal goat serum and 0.1% Triton-X100) for 1 h. Sections were incubated overnight at 4°C with primary antibody against choline acetyltransferase (1:750, Millipore). Sections were washed in phosphate buffered saline and incubated with a biotinylated secondary antibody (1:300; Vector Laboratories, Burlingame, CA) for 1 h at room temperature, followed by incubation in streptavidin–biotin complex (Vectastain ABC kit Elite, Vector laboratories) for 1 h and visualized by incubation in 3,3′-diaminobenzidine solution (Vector Laboratories). The number of cranial and cervical spinal cord choline acetyltransferase positive motor neurons was quantified (for additional details see Supplementary material).

Quick immunostaining and dehydration of sections for laser capture microdissection

To visualize motor neurons for LCM, sections on slides were fixed in 75% ethanol for 1 min, washed in distilled water for 2 min, stained for 4 min in HistoGene staining (Arcturus) and washed again for 30 s in distilled water. The sections were dehydrated for 30 s in 75% ethanol, 2 min in 95% ethanol, 1 min in 100% ethanol and 5 min in xylene, air dried and placed into the Veritas LCM (Arcturus).

Laser capture microdissection of motor neurons

An Arcturus Veritas LCM System was used to isolate motor neurons from the CN3/4, CN12 and the lateral motor column of the cervical spinal cord (C2 and C3 segments) of normal rats onto CapSure Macro LCM caps (Arcturus). Motor neurons (500–1000) were isolated from each subpopulation and each individual animal (n = 4–5). Settings were optimized to capture nucleus and cytosol from the motor neurons, while minimizing inclusion of surrounding tissues.

RNA preparation, amplification and oligo-microarray analysis

RNA was purified from 250 to 500 motor neurons isolated from CN3/4, CN12 or cervical spinal cord (PicoPure RNA isolation kit, Arcturus), and amplified (aRNA) (RiboAmp RNA amplification kit, Arcturus). Amplified RNA quality was analysed (Agilent 2100 Bioanalyser, Agilent technologies) and hybridized to whole rat genome oligo-microarrays (Rat Genome 230 2.0 Array, Affymetrix, for raw microarray data, see Supplementary Table 4). The microarray study consisted of a comparison between motor neurons isolated from CN3/4, CN12 or cervical spinal cord. Each group constituted 4–5 replicates (arrays). The data set was analysed using Gene Pattern (http://www.broad.mit.edu/cancer/software/genepattern/). The MultiExperiment Viewer of TM4 (http://www.tm4.org) was used for correspondence analysis, utilizing the K-nearest neighbour algorithm (number of neighbours = 10). For the hierarchical clustering and gene distance matrix analysis based on Euclidian distance, genes with significant differential expression (1968 genes) were extracted by ANOVA, P ≤ 0.01 (based on F-distribution). Genes differentially expressed between CN3/4, CN12 and cervical spinal cord, P ≤ 0.05, were selected and visualized in heat maps, gene lists and annotations (Fig. 2, Table 1; Supplementary Fig. 3, Supplementary Tables 1 and 2). All gene lists were annotated using the DAVID Bioinformatics Database Gene Id Conversion Tool (http://david.abcc.ncifcrf.gov/conversion.jsp) and NCBI Entrez gene database and BLAST tool. The functional annotation chart tool in DAVID (http://david.abcc.ncifcrf.gov/) was utilized to detect differences in gene groups between motor neuron subpopulations (Huang da et al., 2009) (Fig. 2F–H, Supplementary Table 3), with classification stringency set using the medium parameter and the similarity threshold to 0.5.

View this table:
Table 1

Genes differentially expressed in motor neurons of the cervical spinal cord, hypoglossal nucleus (CN12) and the oculomotor/trochlear complex (CN3/4), selected based on fold change and function

GeneBank NoGene nameGene symbolFold changeP-valueGene function/categoryReferences
Genes upregulated in CN3/4 versus CN12 and cervical spinal cord with a P < 0.05, sorted by fold
NM_012560.1Forkhead-like transcription factor bf-1Fkhr48.50.002Transcription, development, anti-apoptosis, neuronal survival/degenerationYuan et al., 2008
AA818342Guanine deaminaseGda41.50.018Purine metabolism, CNS development, post synaptic protein sorting, dendritic branchingFirestein et al., 1999; Akum et al., 2004
NM_017110.1Cocaine and amphetamine regulated transcriptCart41.20.008Neurotransmitter/hormone, regulation of appetite and stressRogge et al., 2008
BG374268Orthodenticle homologue 2 (predicted)Otx233.80.010Transcription factor, CNS developmentBroccoli et al., 1999; Millet et al., 1999
BI285485DermatopontinDpt24.50.028Cell adhesionOkamoto and Fujiwara, 2006
AF322217.1Immunoglobulin superfamily, member 1Igsf123.70.002Signal transduction, transcription, cell recognitionMazzarella et al., 1998
AI412117Serine protease inhibitor, kunitz type 2Spint222.90.010TranscriptionKawaguchi et al., 1997
BI296460Laminin b1 subunit 1Lamb121.40.002Cell adhesion and migration, axon outgrowth and growth cone behaviourHopker et al., 1999
BM390227B-cell cll/lymphoma 11bBcl11b18.30.002Transcription, immune systemWakabayashi et al., 2003
BF386692Insulin receptor substrate 4Irs417.70.046Signal transduction, developmentBohni et al., 1999
AB028461.1Leucine-rich repeat, immunoglobulin-like and transmembrane domains 1Lrit117.60.008Retina specific receptor (morphogenesis), bindingGomi et al., 2000
AW920064Cathepsin CCtsc17.30.032Degradation of proteins, immune responseMcDonald et al., 1969
BI278379Reticulocalbin 3, EF-hand calcium binding domainRcn315.20.010Calcium binding protein involved in protein folding and sorting (endoplasmic reticulum)Honore, 2009
AA943808Cyclic AMP-regulated phosphoprotein, transcript variant 1Arpp-2114.90.002Regulation of calmodulin-dependent enzymes e.g. calcineurin in neuronsRakhilin et al., 2004
BE105492Forkhead box p2 (predicted)Foxp210.30.002Transcription, developmentLai et al., 2001
NM_017090.1Guanylate cyclase 1 soluble alpha 3Gucy1A39.60.002Nitric oxide-signalling, regulation of cGMP biosynthesisZabel et al., 1998
NM_016989.1Adenylate cyclase activating polypeptide 1Adcyap19.30.002Cell-cell signalling, cell cycle regulation, neuroprotectionMorio et al., 1996
AI059914Ets variant gene 1Etv18.20.002Transcription, formation of sensory-motor connectionsArber et al., 2000
M15481.1Insulin-like growth factor 1Igf16.70.004DNA-replication, anti-apoptosis, cell motion, signal transduction, development, motor neuron survival, axonal regenerationNachemson et al., 1990; Powell-Braxton et al., 1993; Kaspar et al., 2003
NM_012743.1Forkhead box a2Foxa26.60.008Transcription, development, neuronal survivalLee et al., 2005; Kittappa et al., 2007
NM_012551.1Early growth response 1Egr15.20.002Transcription, inhibition of Fas expressionDinkel et al., 1997
NM_031511.1Insulin-like growth factor IIIgf22.50.006Cell proliferation, development, motor neuron survival and axon regenerationCaroni and Grandes, 1990; DeChiara et al., 1991; Near et al., 1992
Genes upregulated in CN12 versus CN3/4 and cervical spinal cord with a P < 0.05 sorted by fold
AI511432Similar to tripartite motif-containing 58/olfactory receptor Olr1433Trim58/Olr143314.00.002Protein and metal ion binding/signal transduction
BI297651Dopamine receptor 1a/msh homeobox 2Drd1A/Msx29.10.002Signal transduction/transcription, cell proliferation, developmentLiu et al., 1992; Satokata et al., 2000
AI236118Phospholipid scramblase 4Plscr48.00.002Phospholipid scramblingWiedmer et al., 2000
BE101670Neurotensin (predicted)Nts7.50.002Signal transductionMai et al., 1987; Marondel et al., 1996
BF389738Similar to Homeobox B4 (M. musculus)Hoxb46.90.002Transcription, development, positional identity, motor neuron identity and target connectivityMcGinnis and Krumlauf, 1992; Dasen et al., 2005
BE095733Homeobox B5 (predicted)Hoxb56.70.002Transcription, development, positional identity, motor neuron identity and target connectivityKrumlauf, 1994; Dasen et al., 2005
NM_031598.1Phospholipase a2, group iia (platelets, synovial fluid)Pla2G2A5.00.042Phospholipid metabolism, Ca-dependentBirts et al., 2008
AW532697Myeloid ecotropic viral integration site 1 homologue (predicted)Meis14.90.002Transcription, developmentMoskow et al., 1995; Mercader et al., 1999
NM_031752.1Lutheran blood group (auberger b antigen includedBcam4.90.002Cell adhesion, signal transductionParsons et al., 1995
BE116745Wingless-type mmtv integration site 5aWnt5A4.70.010Signal transduction, cell–cell signalling, development, cell polarity, motor neuron columnar specificationWitze et al., 2008; Agalliu et al., 2009; Mosimann et al., 2009
AI030806Similar to Slit and Ntrk-like family, Member 3/butyrylcholinesteraseSlitrk3/ Bche4.70.002Neurite outgrowth/neurite outgrowth, axotomy-regulated, adhesionFlumerfelt and Lewis, 1975; Aruga and Mikoshiba, 2003; Paraoanu et al., 2006
AI549199Prostaglandin e receptor 4 (subtype ep4)Ptger44.60.002Immune response, signal transductionBastien et al., 1994
AI170441Homeobox A5Hoxa54.50.006Transcription, development, positional identity, motor neuron identity and target connectivityKrumlauf, 1994; Dasen et al., 2005
AI556803PleckstrinPlek4.10.024Intracellular signalling cascade, actin assemblyAbrams et al., 1995; Lian et al., 2009
AI598945Gastrulation brain homeobox 2Gbx24.00.002Transcription, development, cell proliferation, CNS developmentKowenz-Leutz et al., 1997; Waters and Lewandoski, 2006
BE108648Eph receptor a3Epha33.20.002Signal transduction, separation of axial motor and sensory pathwaysBoyd et al., 1992; Gallarda et al., 2008
BM385237Annexin A4Anxa42.90.002Ca-dependent phospholipid-bindingBarrow et al., 1994
NM_017238.1Vasoactive intestinal peptide receptor 2Vipr22.60.002Cell–cell signalling, circadian function, CNS developmentXia et al., 1996; Basille et al., 2000; Harmar et al., 2002
Genes upregulated in cervical spinal cord versus CN3/4 and CN12 with a P < 0.05, sorted by fold
AI235507Homeobox C8 (mapped)Hoxc871.90.002Transcription, development, positional identity, motor neuron identity and target connectivityDasen et al., 2005; McGinnis and Krumlauf, 1992
BE096332Crystallin, gamma NCrygn50.30.002Visual perception, cellular response to reactive oxygen speciesWistow et al., 2005
AI177143Similar to Homeobox D8/D4Hoxd8/d431.60.002Transcription, development, positional identity, motor neuron identity and target connectivityDasen et al., 2005; Krumlauf, 1994
AA956024Homeobox C5Hoxc526.50.002Transcription, development, positional identity, motor neuron identity and target connectivityDasen et al., 2005; Krumlauf, 1994
AA996507Homeobox A9Hoxa914.60.002Transcription, development, positional identity, motor neuron identity and target connectivityKrumlauf, 1994; Dasen, 2005, p. 140
AI501494Homeobox C6Hoxc611.60.016Transcription, development, positional identity, motor neuron identity and target connectivityDasen et al., 2005; Krumlauf, 1994
AW530378Potassium voltage-gated channel, shaker-related subfamily, member 1Kcna17.80.010Ion transport, synaptic transmissionAdelman et al., 1995; Gu et al., 2003; Raab-Graham et al., 2006
BE118557Supervillin (predicted)Svil6.60.004Cytoskeleton organization, skeletal muscle development, differentiationPestonjamasp et al., 1997
BE108523Solute carrier family 35, member D3Slc35d36.00.002Carbohydrate transportChintala et al., 2007
BF388562Similar to Homo sapiens seizure related 6 homologue (mouse)-like 2Sez6L26.00.032Signal transductionShimizu-Nishikawa et al., 1995
U36899.1Vomeronasal receptor 2Vnr25.90.046Olfactory sensory perceptionSpeca et al., 1999
AI146158Protein phosphatase 3, catalytic subunit, alpha isoformPpp3Ca5.70.028Protein amino acid dephosphorylationMuramatsu and Kincaid, 1993
BF555051Growth arrest specific 6Gas65.60.002Cell proliferation, signal transductionVarnum et al., 1995
AA859669Neuropilin 2Nrp24.80.002Axon guidance, CNS developmentKolodkin et al., 1997
AI060247N-terminal EF-hand calcium binding protein 3Necab34.00.002Protein metabolic process, regulation of amyloid precursor protein and beta-amyloid generationLee et al., 2000
AI177304Fibroblast growth factor 7Fgf73.70.038Cell proliferation, development, signal transduction, presynaptic organizationRubin et al., 1989; Umemori et al., 2004
NM_012633.1PeripherinPrph2.70.002Structural protein, axotomy-regulated, motor neuron degeneration, amyotrophic lateral sclerosis relatedBeaulieu et al., 1999; Robertson et al., 2003
BF281271Placental growth factorPgf1.90.006Cell proliferation, signal transduction, cell–cell signalling, angiogenesisMaglione et al., 1991
  • Cross comparisons of genes differentially expressed between motor neurons of the cervical spinal cord, CN12 and CN3/4. Displayed genes were selected based on their high differential expression between the motor neuron populations showing differential vulnerability to degeneration and their functions (two-sided t-test using 1000 permutations, P ≤ 0.05).

Preparation of primary spinal cord cultures and in vitro analysis of neuroprotection

Time-pregnant Sprague Dawley wild-type rats were anaesthetized, decapitated and embryonic day (E) 15.5 embryos collected, decapitated and spinal cords isolated in Hanks balanced salt solution (Invitrogen). Dissections of spinal cords were done carefully to avoid the inclusion of somites or other external tissues (Hedlund et al., 2004). Cells were dissociated by gentle trituration and incubation with papain (Worthington Biochemical Corporation). Cells were cultured for 6 days in Neurobasal media (Invitrogen) containing 10% foetal bovine serum (Fisher scientific), 1 × B27 supplement (Invitrogen), 500 µM glutamine (Invitrogen), 25 µM mercaptoethanol (Invitrogen), penicillin–streptomycin (Invitrogen) or in Dulbecco’s modified Eagle’s medium/F12 (Invitrogen) containing 5% foetal bovine serum, 1 × N2 supplement A (Stem Cell Technologies), glucose (0.36%, Sigma), bovine serum albumin (0.25%, Invitrogen) and penicillin–streptomycin (Invitrogen). Either culture media could maintain spinal cord cultures containing motor neurons. This 6-day culture period was developed to allow astrocytes time to proliferate in vitro and motor neurons to form an interconnected network prior to exposure to glutamate and the glutamate uptake blocker, l-trans-2,4-pyrrolidine-2,4-dicarboxylic acid (PDC). After the 6 days of culture, glutamate toxicity was induced by the addition of 20 µM glutamate and 100 µM PDC for 4–7 days. For analysis of neuroprotection, the glutamate challenge was preceded by a 2–4 h pretreatment with 1–100 ng/ml recombinant IGF-II (R&D Systems) or 100 ng/ml guanine deaminase (MP Biomedicals, LLC, Solon, OH). Cultures were subsequently maintained for an additional 4–7 days. Identification of motor neurons was done by staining fixed cultures using antibodies against islet-1/2 [1:500 when antibodies were used in combination and 1:100 when used separately, 39.4D5, 40.2D6, Developmental Studies Hybridoma Bank (DSHB), University of Iowa], Neurofilament (150 kDa) (1:500), choline acetyltransferase (1:750, Millipore) and homeodomain protein (MNR2)/homeobox 9 (1:100, 81.5C10, DSBH).

Immunofluorescent staining and stereological procedures

For immunofluorescent staining, coverslips/sections were rinsed with phosphate buffered saline and incubated with blocking buffer (see above) for 1 h. Coverslips/sections were then incubated overnight at 4°C with primary antibodies in blocking buffer. The following antibodies were used: mouse anti-islet-1/2 and rabbit anti-neurofilament (see above), rabbit anti-peripherin (1:100), mouse anti-tyrosine hydroxylase (1:1000), mouse anti-NeuN (1:1000, Millipore), rabbit anti-G protein-coupled inwardly rectifying potassium channel 2 (1:80, Alomone Laboratories), rabbit anti-guanylate cyclase soluble subunit alpha-3 (Gucy1a3) (1:60, Abgent), rabbit anti-placental growth factor (1:30, Proteintech group), rabbit anti-IGF-II (1:100, R&D systems), rabbit anti-early growth response protein 1 (1:100), goat anti-cypin (Guanine Deaminase) (A-20, 1:100, Santa Cruz Biotechnology) and mouse anti-glial fibrillary acidic protein (1:1000, Sigma). Localization of the proteins was done on multiple sections along the cervical spinal cord in multiple animals. The coverslips/sections were then incubated with Alexafluor secondary antibodies for 1 h and rinsed. Hoechst 33342 (4 µg/ml) was used for counterstaining. Confocal analysis was performed using a Zeiss LSM510/Meta Station (Thornwood, NY, http://www.zeiss.com), with optical thickness kept to a minimum and orthogonal reconstructions obtained. The effect of glutamate, IGF-II and/or guanine deaminase on motor neuron survival was carefully evaluated, as was the co-localization of motor neuron markers (see Supplementary material for details).

Quantitative polymerase chain reaction

Quantitative polymerase chain reaction on mRNA extracted from LCMed motor neurons, utilizing SYBR green I (Molecular Probes, Eugene, OR) or specific Taqman probes was used to confirm differential gene expression. For details, see Supplementary material.

Statistical analysis

All experimental data was analysed using Student’s t-test and ANOVA. InStat3 software (GraphPad software inc.) was used for the statistical analyses.

Results

Microarray analysis of motor neurons showed high reproducibility and specificity to the anatomical nuclei

Based on the differential motor neurons loss in motor neuron diseases, we isolated individual motor neurons from CN3/4 (do not degenerate in amyotrophic lateral sclerosis, spinobulbar muscular atrophy or spinal muscular atrophy), CN12 (show vulnerability in amyotrophic lateral sclerosis and spinobulbar muscular atrophy) and from the lateral motor column of the cervical enlargement of the spinal cord (degenerate in all three diseases) using LCM in wild-type rats (Fig. 1A–N). The RNA isolated from motor neurons was hybridized to whole genome rat arrays. The gene expression data showed that individual replicates within a motor neuron nucleus were highly reproducible with an average Pearson’s correlation of 0.93 for CN3/4, 0.94 for CN12 and 0.93 for the cervical spinal cord. When motor neurons from different groups were compared, the average Pearson’s correlation was 0.91 for CN12 versus cervical spinal cord, 0.89 for CN3/4 versus CN12 and 0.86 for CN3/4 versus cervical spinal cord cross comparisons (Supplementary Fig. 2). Additionally, correspondence analysis, hierarchial clustering and gene distance matrix confirmed that individual samples within each nucleus clustered together and showed that motor neurons located in CN12 and in the lateral motor column of the cervical spinal cord had the most similar transcriptomes, whereas motor neurons of CN3/4 showed a more different gene regulation pattern (Fig. 2A and B; Supplementary Fig. 3A). Within the heat map, the locations of peripherin, which was predominantly expressed in motor neurons of the cervical spinal cord, and IGF-II and guanine deaminase, which were restricted to CN3/4 motor neurons, have been indicated (Fig. 2B).

Figure 1

LCM of motor neurons from subpopulations showing differential vulnerability to degeneration in amyotrophic lateral sclerosis. (A) Approximately 40% of cervical spinal motor neurons had degenerated in the SOD1G93A rats at the time of disease onset, as defined by grip strength analysis, while the number of motor neurons in the different brain stem nuclei remained unchanged. (B) Schematic figure depicting the rat brain, brainstem and spinal cord, displaying three nuclei of motor neurons along the rostrocaudal axis of the CNS which show differential vulnerability to degeneration in amyotrophic lateral sclerosis: CN3/4, CN12 and the lateral motor column of the cervical spinal cord. Motor neurons in the (C–F) CN3/4, (G–J) CN12 and (K–N) the cervical spinal cord, visualized by (C, G and K) choline acetyltransferase (ChAT) staining or (D, H and L) HistoGene staining were isolated by LCM (E, F, I, J, M and N).

Figure 2

Global gene expression analysis of motor neurons isolated from CN3/4, CN12 and the cervical spinal cord of normal rats. (A) Correspondence analysis of the gene expression data (22 911 genes) showed that individual samples within each nucleus clustered together and that motor neurons of the CN12 and the cervical spinal cord were closely associated, whereas motor neurons of CN3/4 were less associated with the other motor neuron nuclei. (B) Hierarchial clustering, using Euclidean distance, of the 1968 genes (extracted by ANOVA using a P-value set to 0.01) that were differentially expressed between the three nuclei, showed that individual samples within each somatic motor neuron nucleus clustered closely together, and that motor neurons of CN12 shared more genes with those of the cervical spinal cord compared to CN3/4. Within the heat map, the location of peripherin, which was predominantly expressed in spinal motor neurons, and the CN3/4-restricted genes IGF-II and guanine deaminase (GDA) have been indicated. (C) Heat map displaying the differential expression pattern of Hox genes in motor neuron subpopulations (CN3/4, CN12 and cervical spinal cord) according to their anterior–posterior position within the brain and spinal cord (extracted using two-sided t-test with 1000 permutations, and a P-value set to ≤ 0.05). (D–F) Examples of gene groups that showed a high differential expression level between the three different motor neuron subpopulations. ALS = amyotrophic lateral sclerosis.

Several known motor neuron markers were expressed at similar levels within motor neurons of all three anatomical nuclei (see Supplementary material).

Genes of the Hox cluster provide positional information needed for spatial and temporal pattering of the vertebrate body axis. The known differential expression of the Hox genes, along the anterior posterior axis of the developing hindbrain and spinal cord, was used to validate the microarray data further. Hox A genes in positions 1 and 2, with anterior limits within the midbrain/hindbrain (Carpenter, 2002) were expressed in all three motor neuron nuclei, as would be expected (Fig. 2C). Hox genes in positions 3–5 were identified only in CN12 and the cervical spinal cord as expected based on their known expression in brain stem and spinal cord, but lack thereof in midbrain (Fig. 2C). Finally, Hox genes in position 6–9, with anterior limits at cervical spinal cord levels (Carpenter, 2002), were consequently only identified in motor neurons of the cervical spinal cord (Fig. 2C). The 90 most differentially expressed genes for each cross comparison between the three groups are displayed in heat maps (Supplementary Fig. 2B–G; Supplementary Table 1). There were large differences in the number of regulated genes involved in endoplasmatic reticulum function, mitochondria, ubiquitination, apoptosis regulation, nitrogen metabolism, calcium regulation, transport, cell adhesion and growth (Fig. 2E–F; Supplementary Tables 1 and 2). There were also large differences in the number of genes involved in transcription, RNA metabolic and biosynthetic processing, RNA binding, RNA splicing and regulation of translation (Fig. 2D; Supplementary Tables 2 and 3), functions which have recently been implicated in amyotrophic lateral sclerosis. Furthermore, we identified IGF-I to be preferentially expressed in motor neurons of the CN3/4 (Table 1; Supplementary Table 1). IGF-I can protect spinal motor neurons from degeneration in a mouse model of amyotrophic lateral sclerosis (Kaspar et al., 2003). Our novel finding that IGF-I is preferentially expressed within CN3/4 motor neurons could perhaps further explain the resistance of these cells to degeneration. Functional annotation and pathway analysis showed that several genes, including catalase, neurofilaments, protein phosphatase 3 and tumour protein p53, shown to be involved in amyotrophic lateral sclerosis pathogenesis, were more highly expressed in motor neurons of the cervical spinal cord (Fig. 2F; Supplementary Fig. 4). Furthermore, investigation of ubiquitin mediated proteolysis, a process thought to be involved in the pathogenesis of motor neuron diseases showed that multiple genes were expressed at higher levels in spinal cord motor neurons (Fig. 2F; Supplementary Fig. 5).

Confirmation of differential expression in specific anatomical motor neuron nuclei

Immunofluorescence of the resulting proteins of genes identified as differentially expressed among motor neuron subpopulations displaying differential vulnerability to degeneration confirmed their specific localization and differential expression. The intermediate neurofilament peripherin protein showed a preferential expression within spinal motor neurons (Fig. 3A–C), consistent with the mRNA expression (Fig. 2; Supplementary Fig. 6A). Placental growth factor protein was predominantly localized to spinal motor neurons (Fig. 3D–F), consistent with their microarray data. IGF-II mRNA (Fig. 2; Supplementary Fig. 6B) and protein (Fig. 3G–I) were restricted to motor neurons of CN3/4. Guanine deaminase mRNA and protein were restricted to CN3/4 motor neurons (Fig. 5B–E, Table 1; Supplementary Table 1). Guanine deaminase was also expressed within the striatum, as previously shown (Firestein et al., 1999). The soluble protein Gucy1a3’s mRNA and protein were mainly expressed in motor neurons of CN3/4, but were also detectable in spinal motor neurons. Gucy1a3 protein was also expressed in non-motor neurons within and surrounding CN3/4 (Fig. 3J–L, Supplementary Fig. 7). Early growth response 1 protein was mainly expressed in CN3/4 motor neurons (Fig. 3M–O), consistent with the mRNA expression. The G protein-coupled inwardly rectifying potassium channel 2 mRNA and protein were predominantly expressed in CN3/4 motor neurons. In the cervical spinal cord, the expression appeared more variable, with some neurons displaying a high and others a somewhat lower level of the protein (Supplementary Fig. 8A–D).

Figure 3

Confirmation of protein expression of genes differentially expressed in motor neurons of the cervical spinal cord, CN12 and the CN3/4. Immunofluorescent analysis using confocal microscopy was used to analyse the protein expression of genes found to be differentially expressed between different groups of motor neurons using microarray. Consistent with the microarray RNA data, the protein expression of peripherin in motor neurons of CN3/4 (A) and CN12 appeared low (B), whereas in cervical spinal cord it was high (C and inset). Placental growth factor (PGF) was localized to motor neurons of the spinal cord (F and inset) and to a lesser extent to motor neuron of CN3/4 (D) and CN12 (E). IGF-II protein was expressed in motor neurons of CN3/4 (G and inset), but was below the detection level in motor neurons of CN12 (H) and the cervical spinal cord (I). Gucy1a3 protein was present in motor neurons of the CN3/4. Gucy1a3 was also localized to additional cell types within and surrounding the CN3/4 (J and inset). The level of Gucy1a3 in CN12 was very low (K) whereas it was more easily detected in motor neurons of the cervical spinal cord (L). Early growth response 1 (Egr-1) protein was highly expressed in motor neurons of CN3/4 (M), while the expression level in motor neurons of CN12 and the cervical spinal cord was much lower (N and O). Scale bar: 100 µm (O, applies to A–N).

The CN3/4-restricted genes IGF-II and guanine deaminase protected spinal motor neurons from glutamate-induced toxicity

For analysis of possible neuroprotective properties of differentially expressed candidate genes on somatic motor neurons we developed a primary embryonic spinal cord culture system. The cultures initially contained a majority of neurons, but also a smaller population of astrocytes, which continuously profilerated and thereby constituted the majority of cells at the later parts of the culture time (Hoechst staining in Supplementary Fig. 10 and data not shown). The presence of cell types other than motor neurons provided trophic support, enabling culturing without the addition of growth factors that are necessary if motor neurons are to be cultured alone (Henderson et al., 1993). Motor neurons were present at all times in the culture and displayed large neuritic networks as the culture time progressed. The motor neurons had a healthy appearance and expressed neurofilament and islet-1 (Supplementary Fig. 10). Islet-1 positive cells also expressed homeobox 9 (98.4 ± 1.5% of islet-1 positive cells were homeobox 9 positive) and choline acetyltransferase (Supplementary Fig. 11), confirming their motor neuron identity. At Day 13 of the culture, 9.7 ± 5.3% of all the cells in the culture were motor neurons (homeobox 9 positive, islet-1 positive). Glutamate toxicity could be a general downstream event of degeneration in motor neuron disease. Addition of glutamate (20 µM) and a general glutamate uptake blocker (PDC, 100 µM) induced motor neuron toxicity (Figs. 4A and 5F; Supplementary Fig. 12; *P < 0.001, ANOVA). We selected the CN3/4-restricted genes IGF-II and guanine deaminase for analysis of neuroprotective properties, based on their high differential expression and predominant expression in protected motor neurons and specific cellular functions (Table 1; Supplementary Tables 1 and 2). We hypothesized that the endogenous expression of IGF-II and/or guanine deaminase within CN3/4 motor neurons might protect these cells from glutamate toxicity. IGF-II is a survival factor for motor neurons in some instances and guanine deaminase is a protein important for dendritic branching and synaptic function, but it was not known if either of these proteins could help motor neurons resist high levels of glutamate. Because IGF-II and guanine deaminase are both present extracellularly, which may be of significance and benefit for therapeutic development, we added either of these proteins exogenously to primary spinal cord cultures prior to glutamate insult. Pretreatment with IGF-II at 10–100 ng/ml concentrations protected motor neurons from glutamate-induced toxicity (Fig. 4A–D; Supplementary Fig. 12; *P < 0.001, ANOVA). Confocal analysis of IGF-II pretreated cultures exposed to glutamate for 7 days show healthy motor neurons expressing neurofilament and islet-1 (Fig. 4B–D). Preincubation of the spinal cultures with guanine deaminase (Table 1, Fig. 5A–E) at 100 ng/ml concentrations also protected motor neurons from glutamate-induced toxicity (Fig. 5F–I, P < 0.001, ANOVA).

Figure 4

IGF-II protected spinal motor neurons in primary culture from glutamate-induced toxicity. (A) The number of spinal motor neurons in primary culture was significantly decreased after the addition of 20 µM glutamate (Glu) and 100 µM of the glutamate uptake blocker PDC (P < 0.001, ANOVA). Pretreatment of the cultures with IGF-II (10–100 ng/ml) for 2–4 h prior to glutamate insult protected motor neurons against the toxicity (P < 0.001, ANOVA). Confocal analysis of (B and D) 150 kD neurofilament (NF) and (C and D) islet-1 expression in primary spinal cord cultures pretreated with IGF-II show the presence of large numbers of motor neurons in the cultures after 7 days of combined IGF-II and glutamate treatment. Scale bar: 50 µm (D, applies to B and C). Asterisk signifies a statistically significant difference with a P < 0.01, by ANOVA.

Figure 5

Guanine deaminase was expressed within CN3/4 motor neurons and exogenous delivery protected primary spinal motor neurons from glutamate toxicity. Immunofluorescent analysis using confocal microscopy was used to analyse the protein expression of guanine deaminase (GDA). (A) Striatal tissue was used as a positive control for the antibody staining. (B and E) Consistent with the microarray RNA data, guanine deaminase protein was present in motor neurons of the CN3/4 as well as within other surrounding cell types within this nucleus. The level of guanine deaminase in CN12 and cervical spinal cord was very low (C and D). (F–I) The addition of glutamate (Glu, 20 µM) and the glutamate uptake blocker PDC (100 µM) to primary spinal cord culture decreased the number of motor neurons (P < 0.001, ANOVA). Pretreatment of the cultures with guanine deaminase (100 ng/ml) for 3–4 h prior to glutamate insult protected motor neurons against the toxicity (P < 0.001, ANOVA). Scale bars: 100 µm (D, applies to A–C), 50 µm (E) and 50 µm (I, applies to G and H). ChAT = choline acetyltransferase. Asterisk signifies a statistically significant difference with a P < 0.01, by ANOVA.

Discussion

Difference in vulnerability of specific cellular populations is a prominent feature of neurodegenerative diseases. Understanding such differences by analysing normal function and gene expression within particular neuronal populations could lead to a better understanding of the molecular events underlying degeneration as well as protection.

Relative vulnerability and comparative analysis of differential gene expression among motor neuron subpopulations

Factors intrinsic to motor neurons appear crucial for initiation of motor neuron degeneration in amyotrophic lateral sclerosis (Boillee et al., 2006; Jaarsma et al., 2008) and perhaps also in spinobulbar muscular atrophy and spinal muscular atrophy. To understand why disease is initiated in some, but not all motor neurons we analysed the gene expression profiles of motor neurons isolated from CN3/4 (unaffected in amyotrophic lateral sclerosis, spinobulbar muscular atrophy and spinal muscular atrophy), CN12 (affected in amyotrophic lateral sclerosis and spinobulbar muscular atrophy) and the lateral motor column of the cervical spinal cord (highly affected in amyotrophic lateral sclerosis, spinobulbar muscular atrophy and spinal muscular atrophy). We studied motor neurons using LCM, which allows for isolation of neurons from adult animals and does not require any genetic manipulation.

Our analyses demonstrated that motor neurons in CN12 and the lateral motor column of the cervical spinal cord had more commonality in gene expression levels than those in CN3/4. This higher degree of clustering of CN12 and cervical spinal cord motor neurons could be related to parameters such as the size of the neurons, the lengths of their projections and their respective muscle targets. We believe that our analysis of ‘gene expression dosage’ can give important information about the normal function of a cell and its disease responses. Differences in actual gene dosage are highly linked to selective vulnerability and neurodegenerative diseases. Familial Parkinson’s disease can be caused by point mutations in the α-synuclein gene (Polymeropoulos et al., 1997) or increased gene dosage (Singleton et al., 2003). Data from animal models indicate that Down’s syndrome occurs due to an increased dosage of genes located in a critical region on the triplicated chromosome 21 (Belichenko et al., 2009). Interestingly, in respect to the lack of degeneration of motor neurons in CN12 and CN3/4 in spinal muscular atrophy, genes identified as common expressors between these groups could provide clues to protective mechanisms in this disease where degeneration is caused by a deficiency of SMN1 protein, with altered stoichiometry of small nuclear RNAs and pre-mRNA splicing defects (Zhang et al., 2008). On the other hand, genes that are differentially expressed between these groups, and genes that are shared between the spinal cord and CN12 motor neurons, could reveal why motor neurons in CN12 and spinal cord degenerate in amyotrophic lateral sclerosis and spinobulbar muscular atrophy, while those in CN3/4 do not.

Critically, our analyses revealed that all three motor neuron subpopulations displayed distinct profiles and exhibited genes with unique expression. Our novel identification of differences in Hox gene expression levels in motor neurons isolated from the three anatomical nuclei along the anterior–posterior (A–P) axis of the adult rat matched that of the developing embryo (Carpenter, 2002) and validated the microarray data. This differential Hox gene expression pattern in the adult nervous system indicates that these genes might be important for maintenance of phenotype in addition to providing positional information during development. Consistent with such a role, adult expression of the Hox-like homeoprotein pancreatic and duodenal homeobox 1 (Pdx1) is necessary for the maintenance of pancreatic cells (Holland et al., 2002) and prospero homeobox protein 1 for lymphatic endothelial cells (Johnson et al., 2008). Comparison of groups of genes revealed differences in regulation of genes involved in endoplasmatic reticulum and mitochondrial functions, ubiquitination, apoptosis regulation, nitrogen metabolism, calcium regulation, transport and growth; cellular pathways that have been implicated in the pathogenesis of amyotrophic lateral sclerosis (Collard et al., 1995; Raoul et al., 2002, 2006; Inoue et al., 2003; Pasinelli et al., 2004; Jiang et al., 2005; Kirkinezos et al., 2005; Obal et al., 2006; Nishitoh et al., 2008). The importance of RNA processing for motor neuron survival is evident since loss of SMN1 causes spinal muscular atrophy (Bussaglia et al., 1995; Lefebvre et al., 1995) and mutations in TDP-43 and FUS cause amyotrophic lateral sclerosis (Kabashi et al., 2008; Sreedharan et al., 2008; Kwiatkowski et al., 2009; Vance et al., 2009). Our expression analysis showed large differences in the number of genes involved in transcription, RNA metabolic and biosynthetic processing, RNA binding, RNA splicing and regulation of translation in the different motor neuron subpopulations, further indicating that these processes could in part be responsible for the differential vulnerability observed.

While it is beyond the scope of this article to go into detail for all these pathways, it illustrates the usefulness of a cell phenotype analysis with subsequent verification and exploration of a few differentially expressed genes with implications for motor neuron degeneration.

Differential expression of genes with implications for motor neuron vulnerability

We identified peripherin to be predominantly expressed in spinal motor neurons. Over-expression of peripherin, results in defective axonal transport of neurofilament proteins (Millecamps et al., 2006) and late-onset motor neuron degeneration (Beaulieu et al., 1999). Elevated levels of peripherin splice forms have been detected in spinal cords of patients with familial (Robertson et al., 2003) and sporadic amyotrophic lateral sclerosis (He and Hays, 2004; Xiao et al., 2008). Mutations in the peripherin gene are associated with a small percentage of amyotrophic lateral sclerosis cases (Gros-Louis et al., 2004; Leung et al., 2004). Consequently, a higher level of peripherin within specific motor neurons might predispose these cells to degenerative events.

We identified several genes as selectively expressed within motor neurons of CN3/4, which could play protective roles, e.g. Gucy1a3, early growth response protein 1, IGF-II and guanine deaminase. Gucy1a3 functions as the main receptor for nitric oxide (Zabel et al., 1998). Motor neurons from mutant SOD1 mice show increased susceptibility to exogenous nitric oxide, through upregulation of Fas ligand and subsequent Fas receptor activation. The activation of Fas receptor leads to further nitric oxide synthesis and it has been proposed that chronic low-level activation of the Fas/nitric oxide feedback loop may underlie the progressive motor neuron loss that characterizes familial amyotrophic lateral sclerosis (Raoul et al., 2006). The presence of Gucy1a3 within motor neurons of CN3/4 suggests that these cells will contain less unbound nitric oxide and, as a consequence, might show a lower level of Fas activation. Furthermore, the CN3/4-restricted gene early growth response protein 1 can confer resistance to apoptotic signals by inhibiting Fas expression, and thereby leading to insensitivity to Fas ligand (Dinkel et al., 1997). The higher expression of early growth response protein 1 within motor neurons of CN3/4 could help to explain further why these cells are not affected by degeneration in amyotrophic lateral sclerosis. In addition, the restricted expression of IGF-II to CN3/4 motor neurons could prove beneficial to these cells. IGF-II can act as a survival factor for motor neurons and can support regeneration of motor axons after nerve injury and during normal development (Caroni and Grandes, 1990; Near et al., 1992; Pu et al., 1999). Guanine deaminase catalyses the conversion of guanine to xanthine. Analysis in hippocampal neurons has shown that guanine deaminase can regulate post-synaptic sorting (Firestein et al., 1999) and promote dendritic branching (Akum et al., 2004). Interestingly, the gene TDP-43 can promote dendritic branching, but amyotrophic lateral sclerosis-associated mutations in TDP-43 attenuates the dendritic function (Lu et al., 2009). We therefore reasoned that a high expression of guanine deaminase, a protein important for dendritic branching and synaptic function, could be protective to motor neurons.

None of these gene products have, to our knowledge, previously been identified to have a differential expression within subpopulations of motor neurons.

Functional analysis revealed neuroprotective properties of the CN3/4-restricted genes IGF-II and guanine deaminase

Motor neuron toxicity and protection in response to glutamate was assayed in a system containing neurons and astrocytes. Glutamate toxicity was utilized since it is considered a downstream event in motor neuron degeneration. Motor neurons are usually protected from high levels of glutamate in vivo by surrounding astrocytes. However, astrocytes in the spinal cords of patients with amyotrophic lateral sclerosis and lower motor neuron disease (Rothstein et al., 1995; Sasaki et al., 2000) and in mutant SOD1 mice (Bruijn et al., 1997) and rats (Howland et al., 2002), have been shown to lose the expression of the focal glutamate transporter excitatory amino acid transporter 2, which could decrease their ability to sequester glutamate. In our assay we blocked excitatory amino acid transporters to mimic glutamate over-load in motor neuron disease and to hopefully create a reliable tool in predicting substances that can protect motor neurons in vivo. We believe that the predicted value for finding neuroprotective genes by analysing neurons displaying differential vulnerability to disease for their expression profiles in the normal animal is high, based on our previous approach in Parkinson’s disease models (Chung et al., 2005). Subsequently, we tested the effects of the CN3/4-specific genes IGF-II and guanine deaminase in our assay and found that IGF-II blocked glutamate-induced motor neuron loss completely, while guanine deaminase considerably decreased the loss of motor neurons. Future studies will analyse the utility of additional CN3/4-specific genes for neuroprotective properties in vitro and evaluate the protective properties of IGF-II and guanine deaminase in animal models of motor neuron disorders. These data are promising for selection of future gene therapy and/or drug targets for motor neuron diseases based on a specific gene expression within protected motor neurons. These findings also suggest that future drug screening profiles could analyse the up- or down-regulation of a handful of genes (identified as differentially expressed among motor neurons showing differential vulnerability to degeneration) to enable an evaluation of possible neuroprotective properties of drug candidates and hopefully of potential clinical success.

We believe that our report provides insight into the intrinsic properties of different motor neuron subpopulations and gives important clues to the mechanisms of relative vulnerability. Therefore, this extensive expression analysis could provide a basis for understanding why degeneration in amyotrophic lateral sclerosis, spinal muscular atrophy and spinobulbar muscular atrophy involve some, but not all, motor neuron populations and hopefully be used to develop treatments for these diseases.

Funding

This work was supported by a young investigator award from the ALS Association (E.H.), ALS Research Program Therapeutic Development Award/DOD USAMRAA W81XWH-08-1-0496 (O.I.) the Consolidated Anti-Aging Foundation (O.I.) and the training grant award number T32AG000222-17 from the National Institute On Aging (T.O.).

Supplementary material

Supplementary material is available at Brain online.

Footnotes

  • *Present address: Eva Hedlund, Ludwig Institute for Cancer Research Ltd., Karolinska Institutet, 171 77 Stockholm, Sweden.

  • Abbreviations:
    Abbreviations
    CN
    cranial nerve
    Gucy1a3
    guanylate cyclase soluble subunit alpha-3
    IGF
    insulin-like growth factor
    LCM
    laser capture microdissection
    PDC
    l-trans-2,4-pyrrolidine-2,4-dicarboxylic acid
    SMN1
    survival of motor neuron protein
    SOD1
    superoxide dismutase 1

References

View Abstract