DNA Arrays in The Research of Atherosclerosis

We have analyzed gene expression in normal arteries and in different types of im-munohistologically characterized human atherosclerotic lesions using a nylon-filter-based DNA-array method. Radioactively labelled probes were generated from 1.0 ^g of mRNA from normal arteries, fatty streaks and advanced lesions and hybridized to filters of 18,376 cDNA clones (Incyte Genomics). Intensities were analyzed as duplicates with pairs of filters comparing pooled normal samples vs. fatty streaks and advanced lesions. Scores were calculated as described in formula 1. The sensitivity of the filter arrays currently is at the level of one molecule in 100,000, which together with limitations in phosphoimaging allows detection of differential expression in excess of 1.5-2 folds. Only those genes/ESTs detected in three repeated pairs of arrays for each lesion type and showing > 1.5 fold increase or decrease were processed further [30]. We first validated the array by detecting a group of genes (n=17) that were already known to be connected to atherogenesis.

Tab. 9.2 Some examples of differentially expressed genes in advanced atherosclerotic lesions

Gene

GB access

Function

Score

p-value

a) Genes upregulated in advanced lesions

Melanoma adhesion molecule MCAM

R79246

Signaling

56.5

p<0.05

neuronal PAS domain protein

R78870

Expression

49.8

p<0.05

HS solute carrier family 31 (copper transport)

R68089

Transport

44.8

p<0.05

member 2

EST

R68091

Unknown

40.8

p<0.05

Proteasome (prosome, macropain) subunit,

H12633

Expression

37.6

p<0.05

alpha type 2

Oligopherin 1

R81942

Unclassified

34.4

p<0.05

Platelet/ endothelial cell adhesion molecule-1

R33252

Signaling

29.5

p<0.05

(PECAM-1) CD31 antigen

EST

R36114

Unknown

29.0

p<0.05

Ubiquitin-conjugating enzyme E2D 1

H12682

Metabolism

21.3

p<0.05

Lectin, mannose-binding, 1

R62532

Metabolism

21.3

p<0.05

NADH-ubiquinone oxidoreductase, 51 kDa

R67754

Metabolism

11.6

p<0.05

subunit

Human non-muscle myosin alkali light chain

R70035

Unclass ified

10.1

p<0.05

Zinc finger protein 7

AA005168

Expression

6.0

p<0.05

Proteasome, chain 7

R80719

Expression

4.5

p<0.05

b) Genes downregulated in advanced lesions

transcript asssociated with monocyte to

R34270

Signaling

66.5

p<0.05

macrophage differentiation

Human cleavage and polyadenylation specifity

R82814

Expression

63.6

p<0.05

factor mRNA

LIM binding domain 2

R36692

Unclassified

61.8

p<0.05

EST

H02191

Unknown

57.3

p<0.05

Deoxiribonuclease 1-like 1

R32385

Metabolism

56.7

p<0.05

growth differentiation factor 11, bone

T81804

Unclassified

52.4

p<0.05

morphogenetic protein 11

EST

R80390

Unknown

51.3

p<0.05

EST, FLJ1321

R69260

Unknown

40.5

p<0.05

Diphosphoinositol polyphosphate

H13795

Metabolism

37.0

p<0.05

phosphohydrolase

EST, Weakly similar to reverse transcriptase

H12636

Unclassified

36.7

p<0.05

RecQ protein like-5

R31058

Unknown

34.4

p<0.05

EST

R81144

Unknown

31.1

p<0.05

protein phosphatase 2, regulatory subunit B,

R65749

Metabolism

30.7

p<0.05

alpha isoform PPP2R2A

EST, similar to mus musculus AT3 gene for

R77725

Unknown

26.9

p<0.05

antithrombin antithrombin

These genes included e.g. apoE, CD68, and tissue inhibitor of metalloproteinase (TIMP). Next we detected 150 differentially expressed genes that were previously not connected to atherogenesis. Among these genes we found upregulation of 82 genes, 63 of which were known genes and downregulation of 68 genes, 33 of which were known genes [30]. Tab. 9.2 presents some examples of the genes up or downregulated in advanced lesions. In Fig. 9.2 the expression intensities of all arrayed genes in advanced lesion are blotted against the intensities in normal artery, showing that the expression level of the majority of the genes has not changed.

It is evident that cellular composition of the analyzed lesions has a major impact on the pattern of gene expression. For that reason, arterial samples were im-munostained for the presence of SMC, macrophages and T-cells. Samples analyzed using the DNA array were confirmed to represent typical atherosclerotic lesions with only a moderate infiltration of inflammatory cells. Thus, it is likely that if extensive infiltration of macrophages or T-cells had been present the results from

Fig. 9.2 Expression intensities in advanced lesions blotted against the expression intensities in normal artery.

gene expression profiling could have identified activation of different genes. Since advanced lesions frequently involve microdomains of complex pathology it is likely that laser microdissection of lesions will improve the accuracy of the analysis. Results obtained from atherosclerotic lesions using DNA arrays should always be confirmed by in situ hybridization and/or RT-PCR analyses of the same lesions since current DNA arrays may still produce false results. It is also important to obtain information about the localization of mRNA species in different cell types.

Macrophage-rich shoulder areas of human atherosclerotic lesion were laser mi-crodissected and gene expression profiles were compared to normal intima. Upre-gulation of several macrophage-specific genes e.g. IL-4 and additionally, upregula-tion of many known e.g. pancreatic lipase and unknown genes were detected (Tuomisto et al 2002, unpublished results).

Studies published so far regarding gene expression profiles in atherogenesis have mainly focused on cultured cells or animal models. Foam cell formation has been mimicked by treating macrophages with oxidized low density lipoprotein and changes in gene expression have been studied using a microarray of 9,808 genes in timepoints ranging from 30 min to 4 days. 268 of the genes showed 2fold differential expression for at least 1 time points e.g. adipophilin, heparin-binding growth factor like growth factor and thrombomodulin [5]. Cluster analysis was used to further interpret the data. Vascular inflammation was mimicked by treating human vascular SMC with tumor necrosis factor a (TNFa), and subsequent array analysis revealed the importance of eotaxin and its receptor in inflammatory cell recruitment [31]. Endothelial cell senescence was demonstrated using cultured endothelial cells of early and late passages, and it was found that the expression of thymosin /5-10 was decreased, which might contribute to the senescent phenotype by reducing the endothelial cell plasticity [32]. Wuttge et al. analyzed the gene expression during atherogenesis in apoE deficient mouse and found that atherogenesis is not a linear process with a maximal expression at advanced lesions stage, instead the gene expression has its peaks at the intermediate lesion stage [33]. The gene expression profiles between aorta and vena cava were studied in macaques, and 68 differentially expressed genes out of 4,048 genes had elevated expression in aorta [34]. The expression patterns in fibrous cap versus adjacent media in human atherosclerotic lesions was studied by array of 588 clones, and the induction of Egr-1 expression was detected also in a mouse model [35]. The expression differences between human stable and ruptured plaque was studied using suppression subtractive hybridization and macroarrays, and upregulated expression of periphilin, a phosphoprotein involved in process of lipolysis, in ruptured plaques was detected [36]. Human activated SMC expression was studied by differential display, and 10 known and 30 novel "smooth muscle activation-specific genes" were identified [37].

Analysis of atherogenesis with genomics techniques is still in a very early stage. However, there is no doubt that DNA array techniques will become a valuable tool for studying gene expression patterns in atherosclerosis and identifying novel candidate genes. Laser microdissection will further improve the accuracy of the technique in the analysis of local microdomains in atherosclerotic lesions.

Was this article helpful?

0 0

Post a comment