doi:10.1016/j.neuroscience.2004.08.058 
Copyright © 2005 IBRO Published by Elsevier Ltd.
Carotid atherosclerotic plaques from symptomatic stroke patients share the molecular fingerprints to develop in a neoplastic fashion: A microarray analysis study
R. Vemugantia, 
, 
and R.J. Dempseya
aDepartment of Neurological Surgery, University of Wisconsin-Madison, K4/8 (Mail Stop Code CSC-8660), 600 Highland Avenue, Madison, WI 53792, USA
Accepted 15 August 2004.
Available online 22 January 2005.
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Abstract
Identification of genetic mechanisms that promote the onset of stroke and transient cerebral ischemic attack symptoms in carotid atherosclerotic patients would further our understanding of the pathophysiology of this disease and could lead to new pharmacological and molecular therapies. Using Affymetrix Human Genome 230 GeneChip set, the present study evaluated the gene expression differences in geometrically similar carotid artery plaque samples extricated from six symptomatic stroke patients and four asymptomatic patients. There was no significant difference in the degree of stenosis between the two groups. Of the 44,860 transcripts analyzed, 289 (approximately 0.6% of the total transcripts) were differentially expressed between the plaques from the symptomatic and asymptomatic groups (236 were expressed more abundantly and 53 were expressed less abundantly in the symptomatic group). Of the 236 transcripts expressed more abundantly in the symptomatic plaques, 71% (167 transcripts) indicate an active cell proliferation and neoplastic process. These include oncogenes, growth factors, tumor promoters, tumor markers, angiogenesis promoters, transcription factors, RNA splicing factors, RNA processing proteins, signal transduction mediators and those that control the metabolism. Real-time polymerase chain reaction confirmed the increased expression of 63 transcripts in the symptomatic plaques. The other groups of transcripts expressed more abundantly in the symptomatic plaques are those that control ionic homeostasis, those that participate in the progression of degenerative neurological diseases (Alzheimer’s disease, amyotrophic lateral sclerosis and Huntington’s disease) and epilepsy. This indicates that symptomatic plaques are molecularly and biochemically more active than the asymptomatic plaques, or active plaque growth precipitates stroke symptoms.
Key words: atherosclerosis; gene expression; neoplasia; neurodegenerative disease; plaque; stroke
Abbreviations: ACAS, Asymptomatic Carotid Atherosclerosis Study; D, decrease; DAVID, Database for Annotation, Visualization and Integrated Discovery; EASE, Expression Analysis Systematic Explorer; EST, expressed sequence tag; I, increase; MD, marginal decrease; MI, marginal increase; UPS, ubiquitin-proteasome system
Fig. 1. Histogram showing the number of differentially expressed transcripts in each functional class between the symptomatic and the asymptomatic groups. Of the 44,860 human transcripts represented on the Human Genome Chip 230 set, 289 were observed to be differentially expressed between the two groups; 236 transcripts were more abundantly expressed in the symptomatic group over the asymptomatic group and 53 transcripts were more abundantly expressed in the asymptomatic group over the symptomatic group.
 |
Fig. 2. Hierarchical clustering analysis of the transcripts expressed more abundantly in the symptomatic over the asymptomatic group. The individual expression signal of each transcript in each chip was clustered using CLUSFAVOR 6.0 software (developed by Leif E. Peterson, Baylor College of Medicine; free software available at http://mbcr.bcm.tmc.edu/genepi/) using the Euclidean distance function. The transcripts analyzed were arranged as ordered by clustering algorithm so that transcripts with the most similar expression pattern were placed adjacent to each other. The color codes in the heat map are continuous with blue as lowest and red as highest. The transcripts could be divided in to six clusters based on the expression level. The dendrograms (tree diagrams) shows the grouping of arrays and genes according to the order in which they were joined during clustering. The array dendrograms shows that the six symptomatic and the four asymptomatic plaque samples were clustered together as two groups.
Fig. 3. Magnified view of clusters 1, 2 and 3 of Fig. 1. Increased mRNA abundance of approximately 50% transcripts in each cluster (names shown in blue) was confirmed by real-time quantitative PCR analysis. Factor analysis was performed assuming the genes as the variables, with expression values across the arrays serving as the records. The n×n correlation matrix R was calculated. Eigenanalysis is performed to extract eigenvalues and eigenvectors from R; only eigenvalues ≥1 were extracted and sorted. Factor loadings were calculated for each gene using eigenvalues and eigenvectors. Next, varimax orthogonal rotation was performed on the factor loadings so that each gene expression profile “mostly” loads on a single factor.
Fig. 4. Magnified view of clusters 4, 5 and 6 of Fig. 1. Remaining legend is same as in Fig. 3.
Table 1.
Patient characteristics
a Stroke is the only characteristic that is different between the two groups. Neuropathological examination indicated no significant differences in the plaque characteristics (calcification, gross lipid, hemorrhage and ulceration) between the two groups.
Table 2.
Oncogenes and growth-inducing transcripts expressed more abundantly in the symptomatic plaques compared to the asymptomatic plaquesa
a BDNF, brain derived neurotrophic factor; EGF, epithelial growth factor; FGF, fibroblast growth factor; Maf, musculoaponeurotic fibrosarcoma; TGF, transforming growth factor. Δ Fold is the expression level in symptomatic plaques over the asymptomatic plaques. Each value is a mean of 24 pair-wise cross-comparisons between the six symptomatic and the four asymptomatic plaque samples. In all cases, the SD was <20%. All the fold changes are significant compared to the asymptomatic group (
P<0.05 by Wilcoxon’s signed-rank test). RAS, Rous avian sarcoma.
Table 3.
Putative tumor promoters and markers expressed more abundantly in the symptomatic plaques compared to the asymptomatic plaquesa
a DDR1 expressed in NPC, developmentally down-regulated 1 expressed in neural precursor cells; ECM, extracellular matrix; etif, eukaryotic translation initiation factor 4A; GADD, Growth arrest and DNA damage-inducible β; GAT9, galactosaminyltransferase 9; GRB, growth factor receptor bound; IGF-BP, insulin-like growth factor binding protein; IRQ, iroquis; IRQ-3, IRQ-class homeodomain protein 3; LAMP, lysosomal-associated membrane protein 1; MAD2, mitotic arrest deficient, yeast, homolog-like2; MADH2, mothers against decapentaplegic,
Drosophila; MALAT-1, metastasis associated in lung adenocarcinoma transcript 1; MEN1, multiple endocrine neoplasia type 1; MIDA, mucinous intraductal adenoma; MMP, matrix metalloproteinase; RNP, ribonucleoprotein; SUMO, small ubiquitin-like modifier; UDP, uridine diphosphate.
b Another transcript of MALAT-1 (AW005982) also showed a 4.2-fold higher abundance (rank: 2) in the symptomatic group over the asymptomatic group.
c Two other β2 Human MEN1 region clone εβ transcripts (BE675516 and BG170478) also showed 2.5- and 2.9-fold higher abundance (both rank 2) in symptomatic over asymptomatic group. Rest of the legend is same as in
Table 2.
Table 4.
Transcription factors and other nuclear function-related transcripts expressed more abundantly in the symptomatic plaques compared to the asymptomatic plaquesa
a Δ Fold is the expression level in symptomatic over the asymptomatic group. Rest of the legend is same as in
Table 2. CTCF, CCCTC-binding factor; PAI-1 plasminogen activator inhibitor-1
Table 5.
RNA function-related transcripts more abundantly in the symptomatic plaques compared to the asymptomatic plaquesa
a EF, elongation factor; HnRNP, heterogeneous nuclear ribonucleoprotein.
b Another transcript of transformer-2α (AI268231) also showed a 3.1-fold higher abundance (rank 2) in the symptomatic over asymptomatic plaque samples. Rest of the legend is same as in
Table 2. Arg/ser, arginine/serine; EF, elongation factor; Hn, human.
Table 6.
Comparison of the fold changes observed in GeneChip and real-time PCR analysisa
a Real-time PCR was conducted in triplicate for each transcript. The primer sequences designed using the Primer Express Software (Applied Biosystems) are given as additional data. The real-time PCR values were normalized using 18S rRNA as a housekeeping control. Δ Fold is the mean expression level in symptomatic plaques over the asymptomatic plaques and all the fold changes are statistically significant (
P<0.05, Student’s
t-test).
Table 7.
Disease-related transcripts expressed more abundantly in the symptomatic plaques compared to the asymptomatic plaquesa
a ALS, amyotrophic lateral sclerosis. All the fold changes reported are statistically significant compared with the asymptomatic plaque samples (
P<0.05 by Wilcoxon’s signed-rank test).
b BLAST search showed 98% sequence homologies between NM_001642 and BC000373 and between NM_000113 and BC000674. Rest of the legend is same as in
Table 2.
Table 8.
Gene ontology and functional annotations based on biological process of the transcripts expressed more abundantly in the symptomatic groupa
a The functional annotations are generated by subjecting the data to DAVID software with the GoCharts (gene ontology charts) function using biological process defining broad biological goals such as mitosis or purine metabolism, that are accomplished by ordered assemblies of molecular functions as classification types at level 3 annotation (covering physiological processes, response to external stimulus and response to biotic stimulus).
Table 9.
Gene ontology and functional annotation based on molecular function of the transcripts expressed more abundantly in the symptomatic groupa
a The functional annotations presented in this table are generated by subjecting the data to DAVID software with the GoCharts function using molecular function defining the tasks performed by individual gene products (examples are transcription factor and DNA helicase) as classification types at level 3 annotation (covering physiological processes, response to external stimulus and response to biotic stimulus).
Abstract
Purchase PDF (1547 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (10)
Purchase PDF (332 K)
