C4 MAKING THE MOST OF YOUR MICROARRAY
Mirnics K1, Sisodia S2
1Vanderbilt Kennedy Center for Research on Human Development and Department of Psychiatry, Nashville, USA, 2The Center for Molecular Neurobiology, Chicago, USA
E‐mail address for correspondence: [email protected]
DNA microarrays can perform whole transcriptome assessment in a single experiment. However, the analysis of microarray data is challenging, and the obtained dataset usually contains both type I and type II errors. As a result, validation of the findings with an independent method is strongly recommended. This is especially important in brain transcriptome profiling experiments, where the magnitude of the mRNA expression change at the tissue level most often does not exceed 50%. Technical replicates of a microarray dataset, starting with new cDNA synthesis from the same RNA, are very helpful for estimation of assay noise and false discovery rate (FDR). For identification of differentially expressed genes, implementing dual statistical criteria, based on both magnitude of change and probability of change, coupled with a permutation analysis of the data can usually uncover the critical expression differences while keeping FDR at a low level. However, regardless of the analysis performed, negative data should be cautiously interpreted as microarrays often do not detect all real, biologically important expression differences.
Once the differentially expressed genes are identified, the data analysis moves to a pattern mining phase. In the first step genes are grouped together based on structure, function or common motifs in the DNA sequence. In the second step, the dataset is assessed for differentially expressed genes that show common structural or functional characteristics that reach beyond what would be expected by chance. These analyses often uncover altered molecular pathways or coregulatory patterns that are not obvious at the single‐gene analysis level.
The true power of these approaches can be best demonstrated using DNA microarray studies with a complex, converging design. For example, to determine presenilin‐1 (PS1) regulated genes, first we compared the neocortical and hippocampal transcriptome of PS1 conditional KO mice to those of wild‐type littermates. Next, we compared the transcriptomes of transgenic mice carrying the wild‐type human PS1 to that of transgenic mice carrying the familial Alzheimer disease‐linked delE9 mutant human PS1. Cross‐correlating findings revealed a number of transcripts showing differential expression across these two datasets. These expression changes, involving many early‐immediate gene (IEG) transcripts, were also regulated in the amyloid‐depositing APPswexdelE9PS1 mutant mice and showed a strong pattern reversal when the same mice were subjected to environmental enrichment. The overall data, obtained across five different animal models, suggest that PS1 is a potent regulator of EIG expression in an amyloid‐dependent fashion and that environmental enrichment prevents amyloid accumulation in the brain tissue through a mechanism closely linked to the IEG transcript network. Finally, based on recently obtained cortical transcriptome data in brain‐derived neurotrophic factor (BDNF) KO mice, we speculate that the PS1‐amyloid system in the brain tissue is strongly influenced by BDNF‐mediated molecular cascades.
C5 GENE EXPRESSION PROFILE OF SPINAL MOTOR NEURONS IN SPORADIC AMYOTROPHIC LATERAL SCLEROSIS
Tanaka F, Jiang YM, Yamamoto M, Huang Z, Katsuno M, Adachi H, Niwa JI, Doyu M, Sobue G
Department of Neurology, Nagoya University Graduate School of Medicine, Nagoya, Japan
E‐mail address for correspondence: [email protected]
Background: The causative pathomechanism of sporadic amyotrophic lateral sclerosis (SALS) is not clearly understood. There have been extensive studies using animal models and culture systems for familial ALS, especially with SOD1 mutations, but no similar approach is available for studying SALS. Microarray analysis is one of the appropriate approaches to understand the pathological pathway related to the neuronal degeneration process in sporadic neurological disorders. However, in the lesions of SALS spinal cords, there are reduced numbers of motor neurons with glial cell proliferation, making it difficult to examine motor neuron‐specific gene expression by conventional microarray analysis.
Objectives: For the purpose of elucidating the pathogenesis of SALS, we reveal motor neuron‐specific gene expression profile in SALS using microarray technology combined with laser‐captured microdissection (LCM).
Methods: Fresh specimens of lumbar spinal cord from 14 SALS patients and 13 neurologically normal patients were obtained at autopsy. The pulsed laser microbeam cut precisely around the targeted motor neurons in the spinal ventral horn. RNA was extracted from LCM‐isolated cells and reverse transcription and T7 RNA polymerase amplification of RNA were performed prior to DNA microarray analysis. RNA was extracted as well from the total homogenates of ventral horn gray matter of spinal cords. The data for each differential gene expression level obtained from microarray analysis (BD Atlas Glass Microarray System: Clontech) was reconfirmed by real time reverse transcription polymerase chain reaction and in situ hybridization.
Results: Spinal motor neurons showed a distinct gene expression profile from the whole spinal ventral horn. Three percent of genes examined were down‐regulated, and 1% were up‐regulated in motor neurons. Down‐regulated genes included those associated with cytoskeletal/axonal transport, transcription, and cell surface antigens/receptors, such as dynactin, microtubule‐associated proteins, and early growth response 3 (EGR3). In contrast, cell death‐associated genes were mostly up‐regulated. Promoters for cell death pathway, death receptor 5, cyclins A1 and C, and caspases‐1, ‐3, and ‐9, were up‐regulated, whereas cell death inhibitors, acetyl‐CoA transporter, and NF‐κB were also up‐regulated. Moreover, neuroprotective neurotrophic factors such as ciliary neurotrophic factor (CNTF), hepatocyte growth factor (HGF), and glial cell line‐derived neurotrophic factor were up‐regulated. Inflammation‐related genes, such as those belonging to the cytokine family, were not, however, significantly up‐regulated.
Discussion and conclusions: Microarray analysis on the laser‐captured motor neurons provided us with significant information about motor neuron degeneration and dysfunction in SALS. Such information cannot be obtained by whole spinal cord tissue microarray assay. In addition, we are now analyzing the sequential motor neuron‐specific gene expression in terms of motor neuron degeneration process, which will provide an avenue for new molecular targeted therapy for SALS through developing the animal or cell models mimicking these molecular events determined in human SALS patients.
C6 GENE EXPRESSION PROFILE OF SPINAL MOTOR NEURONS IN THE G93A SOD1 MOUSE MODEL OF ALS AT DIFFERENT TIME POINTS IN THE DISEASE COURSE
Ferraiuolo L, Heath PR, Holden H, Baptista MJ, Kirby J, Shaw PJ
University of Sheffield, Sheffield, UK
E‐mail address for correspondence: [email protected]
Background: Amyotrophic lateral sclerosis (ALS) is one of the most common adult onset neurodegenerative diseases, and while 90% of the cases are sporadic (SALS), 10% are familial (FALS). The most common cause of FALS is mutation of the SOD1 gene. Transgenic mice carrying mutant forms of SOD1 develop a neuromuscular disease very similar to human ALS in both phenotype and histopathological hallmarks.
Objectives: 1) To investigate changes in gene expression profiles of degenerating spinal motor neurons (MN) isolated from human G93A SOD1 mice, human WT SOD1 mice and non transgenic littermates (LM) at different stages of the disease (60, 90 and 120 days); 2) To identify pathways involved in the development of the neurodegenerative process.
Methods: Approximately 1000 motor neurons have been isolated from the lumbar spinal cord of each animal. RNA was extracted using Picopure kit (Arcturus), amplified using the RiboAmp Amplification kit (Arcturus) and labelled using the BioArray High Yield RNA Transcript Labelling Kit (Enzo). 10 µg cRNA was applied to the Affy MOE430A GeneChip, and data analysis was performed using ArrayAssist (Iobion).
Results: At 60 days, 258 genes were differentially expressed when comparing G93A SOD1 mutant mice with their non‐transgenic LM. The G93A transgenic mice show a significant increase in both transcriptional and translational functions. Significant increases occur in the expression of genes relating to carbohydrate metabolism, the electron transport chain and three subunits of ATP synthase. At 90 days, transcripts involved in carbohydrate metabolism are still up‐regulated, while genes involved in transcription and mRNA processing are down‐regulated. At 120 days, the transcription profile of the G93A SOD1 mutant mice shows a significant change in 167 genes. At this late disease stage, the transgenic mice show a marked degree of transcriptional repression, involving key genes, e.g. NDN, TAF9 and RNA polymerase 1‐1, while many cyclins regulating the first steps of the cell cycle, e.g. cyclin L1, E2 and D2 show increased expression. Interestingly, given the hypothesis that oxidative stress may play a role in ALS, genes involved in antioxidant activity and stress response are significantly decreased, underlining a deficit in this important cellular defence. No significant alterations in gene expression have been found comparing WT SOD1 mice and their littermate controls.
Discussion and conclusions: The up‐regulation of genes involved in the transcription and translation processes as well as in carbohydrate metabolism suggests the activation of a strong cellular adaptive response in the first stage of the disease. At the 120‐day time point, consistent with results obtained by the analysis of NSC34 cell line transfected with vector expressing human G93A SOD1 (1), the transgenic mice present a strong impairment in transcriptional function. Another interesting aspect of this stage is the up‐regulation of the cyclin family. This result suggests that what might be occurring in ALS is the same mechanism found in other neurological pathologies (2,3), with unsuccessful re‐entry into the cell cycle and consequent cell death.
References
- Kirby J, Halligan E, Baptista M, et al. Brain 2005;128:1686–706.
- Yang Y, Mufson EJ, Herrup K J. Neurosci. 2003; 23:2557–63.
- Yang Y, Herrup K J. Neurosci. 2005;25:2522–9.
C7 WHOLE GENOME MICROARRAY ANALYSIS OF MOTOR NEURON VULNERABILITY IN G93A‐SOD1 AND P301L‐TAU TRANSGENIC MOUSE MODELS OF ALS
Kudo LC, Karsten SL, Wiedau‐Pazos M
UCLA, David Geffen School of Medicine, Department of Neurology, Los Angeles, CA, USA
E‐mail address for correspondence: [email protected]
Background: Amyotrophic lateral sclerosis (ALS) is a devastating neurodegenerative disease that selectively affects motor neurons. A small subset of familial ALS patients carries a mutation of the copper/zinc superoxide dismutase 1 gene (SOD1). Another subset of patients with ALS suffers from familial frontotemporal dementia with ALS, caused by mutations of tau. Transgenic animal models expressing SOD1 have been extensively studied and factors that may contribute to motor neuron degeneration have been identified, such as disruption of axonal transport, glutamate metabolism, oxidative stress, copper metabolism, and growth factors. However, how these and other yet unknown factors cause motor neuron degeneration in sporadic ALS remains unknown. Studying motor neuron degeneration in a transgenic animal model expressing mutant protein tau linked to frontotemporal dementia with ALS in addition to SOD1 mice may help to close this crucial gap in knowledge.
Objective: We aimed to characterize the molecular events involved in the initiation of motor neuron degeneration that may be relevant in both familial ALS linked to SOD1 mutations and FTD with ALS linked to TAU‐P301L. We compared early motor neuron‐specific gene expression in two mouse models of ALS before the onset of any known pathological changes in the spinal cord. Mutant SOD1 (G93A‐SOD1) and mutant tau (P301L‐TAU) mice were studied to identify overlapping differentially expressed genes. The expected gene set may elucidate the cause of motor neuron degeneration and may also function as biomarkers for the disease, potentially aiding earlier diagnosis of ALS.
Methods: DNA microarray technology in combination with laser‐capture microdissection (LCM) was used to detect gene expression changes at the early stages of the disease prior to neurodegeneration. Spinal cords were dissected from 3‐month‐old female transgenic mice and their non‐transgenic littermates. Axial cryostat sections from lumbar spinal cords were fixed in ethanol and stained with Cresyl violet. Motor neurons from the ventral horns were microdissected on a PixCell Arcturus LCM instrument. RNA extracted from the collected motor neurons was subjected to microarray experiments using Agilent's Mouse Whole Genome Oligonucleotide Microarray featuring 45,000 gene probes.
Results and discussion: We generated a list of differentially expressed genes for each mouse strain and identified overlapping and genotype specific sets of genes. The overlapping genes are likely to be relevant in the etiology of motor neuron degeneration independent of its genetic cause. Selected gene expression changes will be confirmed using in situ hybridization and RT‐PCR. Their relevance to ALS will be confirmed in functional studies using animal models and post‐mortem ALS patient samples.
C8 MOLECULAR PATHWAY ANALYSIS OF AMYOTROPHIC LATERAL SCLEROSIS BY GENOME‐WIDE EXPRESSION PROFILING OF HUMAN BLOOD
Saris CGJ1, van Vught PWJ1, Yigittop H1, Groot Koerkamp M1, Holstege FCP2, Ophoff RA2, van den Berg LH1
1UMC Utrecht, Utrecht, 2Rudolf Magnus Institute, Utrecht, The Netherlands
E‐mail address for correspondence: [email protected]
Background: Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease of motor neurons, the etiology and pathogenesis of which still remain unknown. Incidence rates for ALS in Europe and North America range between 1.47 and 2.7 per 100,000/year. Life expectancy after the start of clinical symptoms is about two to five years, where respiratory insufficiency is a major problem in the end‐stage of the disease. In approximately 10% of patients with ALS, a mostly autosomal dominant mode of inheritance is observed. The diagnosis of ALS is based on clinical features, electrophysiology and exclusion of ALS mimics, which usually takes from six to 12 months.
With regard to diagnosis or disease progression in ALS, neither a definitive diagnostic test nor surrogate marker is available. Recently, proteomic profiling of biofluids by mass spectrometry identified protein species with altered levels in ALS patients. It is hypothesized that blood genomic fingerprinting may be a way to find candidate markers.
Objectives: By using expression profiling we looked at the possibility of using blood as a surrogate tissue in the diagnostic phase and looked for candidate genes involving pathogenesis and disease progression.
Methods: Specific expression profiles of whole blood from ALS patients and healthy donors were compared using oligo‐array and Illumina Sentrix HumanRef‐8 Expression BeadChip in order to find marker genes and to find novel pathways that influence disease severity and progression.
Results: Results of the genome‐wide expression profiling will be presented. Part of this analysis consists of the combination of profiles of both platforms. Interim analysis of the expression profiles of 19 patients versus 19 healthy controls with oligo‐array method results in significant differential expression of 74 genes. Patients' genes tend to cluster together. Their function varies from regulation of transcription, cellular transport and protein activation. The function of 20 genes is still unknown. Clinically interesting are genes involved in apoptosis, genes with axonogenese activity and a gene containing the copper binding domain of SOD1.