Heterogeneity of gene expression in human atheroma unmasked using cDNA representational difference analysis

Kerry L. Tyson, Peter L. Weissberg and Catherine M. Shanahan

Department of Medicine, Division of Cardiovascular Medicine, Addenbrooke’s Centre for Clinical Investigation, Addenbrooke’s Hospital, Cambridge CB2 2QQ, United Kingdom


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The rupture of an atherosclerotic plaque can have profound consequences, such as myocardial or cerebrovascular infarction. The complex interactions of vascular smooth muscle cells (VSMCs) with inflammatory and immune cells are thought to contribute to both plaque genesis and stability. Key to our understanding of these processes is the identification of genes expressed in human atheromatous lesions. We have employed cDNA representational difference analysis (RDA) to investigate the differences in gene expression between normal and atherosclerotic human vessels. Thirty-one cDNA clones representing sequences expressed in atheroma were isolated, many of which encoded components of inflammatory and immune pathways. The reciprocal experiment, to identify genes expressed in the healthy vasculature, identified two genes associated with the contractile functions of VSMCs. Semiquantitative RT-PCR analysis of expression of these genes in forty samples, derived from healthy and atheromatous vessels, demonstrated marked heterogeneity of gene expression between lesions, although several of the genes were preferentially expressed in atherosclerotic lesions. In situ hybridization identified subsets of macrophages at sites of neovascularization within the lesion and intimal VSMCs as expressing the disease-associated genes. In conclusion, cDNA RDA is a useful, fast, and efficient technique for studying differential gene expression particularly when clinical material is limiting.

atherosclerosis; inflammation; smooth muscle; macrophages


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
THE TEMPORAL AND SPATIAL REGULATION of gene expression underlies developmental, physiological, and pathological processes. The identification of genes that are differentially expressed in disease can provide information on the pathological process and also identify possible therapeutic targets. Such studies are, however, often hampered by the availability of human clinical samples and the quantity or quality of RNA that can be obtained from them. To overcome these problems, investigators have used cultured human cells or animal models of disease as an alternative source of RNA for analyzing differential gene expression (12, 13, 26, 29). However, both of these approaches are associated with their own limitations. Studies on cultured human cells are unlikely to reflect the entire disease process, as changes in gene expression resulting from the interaction of different cell types and physiological processes in vivo are unlikely to be reiterated in vitro. Similarly, animal models of disease may not accurately represent human pathologies. Both of these problems are pertinent to the study of atherosclerosis.

The formation of an atherosclerotic plaque is a complex process, occurring over several decades and involving the interaction of cells in the blood vessel wall, namely vascular smooth muscle cells (VSMCs) and endothelial cells, with inflammatory cells, such as macrophages and T cells. In addition, environmental and genetic factors are also known to influence the disease process. Although some plaques are asymptomatic throughout their evolution, others provoke clinical events, leading to morbidity and mortality. The erosion and rupture of atheromatous plaques, and the subsequent formation of thrombus, have been recognized as the major mechanisms underlying most acute coronary and cerebrovascular events (6, 8, 9), and thus plaque integrity and stability have become important concepts in the treatment of vascular disease. Histological analysis of ruptured plaques indicates that typically these lesions have a high inflammatory cell content and are covered with a thin fibrous cap comprising relatively few VSMCs (6). In contrast, stable lesions have much thicker fibrous caps with a higher VSMC content and typically contain fewer inflammatory cells. It is likely that the genes expressed by plaque cells contribute both to the formation of the lesion and also to the pathological process of plaque rupture. Clearly, the complex interaction between inflammatory cells and the blood vessel wall can only be studied in whole organisms and not by in vitro cell culture. To date, however, animal models of atherosclerosis that mimic all aspects of human disease have not been developed, and so the most informative studies must be performed on human clinical material (23).

In this study we have investigated the utility of cDNA representational difference analysis (cDNA RDA) as a technique for examining gene expression in human atheroma (14). cDNA RDA couples PCR amplification to subtractive hybridization and avoids the high incidence of false-positive clones associated with other PCR-based screening techniques, such as differential display (17), by amplifying only those sequences that are differentially expressed. Using minimal amounts of RNA obtained from healthy and diseased blood vessels, we isolated several clones that represented genes expressed in atheroma. Subsequent analyses confirmed that some of these genes were preferentially expressed in atherosclerotic lesions and in particular in symptomatic, " active" lesions. Furthermore, the genes identified were found to be expressed by different cell types within the plaque, demonstrating the utility of cDNA RDA in identifying differentially expressed genes from tissue samples comprising a mixture of cell types.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Tissues.
Human aortic and carotid tissue was obtained, with required consent and ethical approval, from organ transplant donors and patients undergoing carotid endarterectomy, respectively. Samples from each tissue piece were snap-frozen in liquid nitrogen for sectioning and also used in explant culture. The outer adventitial layer was dissected from the remainder, and the smooth muscle layer was dispersed by digestion at 37°C with 3 mg/ml collagenase and 1 mg/ml elastase (Sigma). The resulting cell suspension was washed with PBS and subsequently used for RNA isolation.

Cell culture.
Human aortic VSMCs were explanted in medium 199 (GIBCO) supplemented with 20% fetal calf serum, antibiotics, and 4 mM L-glutamine. Cells were confirmed as smooth muscle by positive staining with monoclonal antibodies against smooth muscle (SM) {alpha}-actin (A2547, Sigma) and calponin (C2687, Sigma).

Human peripheral blood monocytes from buffy coat preparations (Blood Transfusion Service, Cambridge, UK) were sedimented with Histopaque (Sigma) and separated from other leukocytes by adherence to plastic (2). The monocytes were differentiated into macrophages by culturing in medium 199 supplemented with 20% fetal calf serum for 21 days.

Isolation of RNA.
RNA was isolated from enzymatically dispersed tissue, cultured cells, and frozen tissue sections by lysis in 150 mM NaCl, 10 mM Tris (pH 7.4), 1 mM MgCl2, and 0.5% Nonidet P-40. Following phenol extraction, the RNA was precipitated with ethanol and treated with RNase-free DNase (Promega) to remove contaminating genomic DNA.

cDNA RDA.
cDNA RDA comprises three steps: the generation of " representative amplicons" from the sample containing the target genes (tester) and control sample containing genes common to both cDNA populations (driver); the hybridization and selective amplification of differentially expressed genes; and the cloning, sequencing, and identification of "difference products" (Fig. 1). RDA was performed as described previously (14, 24), using double-stranded cDNA prepared from 5 µg total RNA isolated from a normal aorta (driver, 49-yr-old female) and from an atherosclerotic aorta (tester, 57-yr-old male). cDNAs from both samples were digested with DpnII and ligated to double-stranded adaptors containing binding sites for PCR primers. The DpnII fragments were then amplified by PCR to generate an amplicon of DpnII fragments that "represent" the genes from which they are derived.



View larger version (22K):
[in this window]
[in a new window]
 
Fig. 1. Schematic representation of cDNA representational difference analysis (RDA). Driver and tester cDNAs are represented by dashed and solid lines, respectively. Following digestion with DpnII, double-stranded (ds) adaptors (black boxes) are ligated onto both samples and the fragments are amplified by PCR to generate the representative amplicons, using PCR primers that anneal to the adaptors. The adaptors are then removed by digestion with DpnII, and new adaptors (hatched boxes) are ligated onto the tester amplicon, allowing selective amplification of tester sequences following hybridization with an excess of driver (tester:driver = 1:100), which results in the first difference product (DP1). Second and third difference products (DP2 and DP3) are generated by reintroducing DP1 (or DP2) at the hybridization stage, and increasing the tester to driver ratio to 1:800 and 1:400,000, respectively. DP3 was cloned and sequenced.

 
The representative amplicons were then digested with DpnII to remove the adaptors, and new adaptors containing binding sites for different PCR primers were ligated to the tester amplicon fragments only, thus allowing the selective amplification of tester-derived sequences in following PCR steps. The tester amplicon was then hybridized to an excess of driver (tester:driver ratio 1:100). Sequences common to both cDNA populations formed tester:driver hybrids and were amplified linearly in the subsequent PCR step, whereas sequences unique to the tester formed tester:tester duplexes and were amplified exponentially. Driver:driver duplexes were not amplified. Thus PCR amplification following hybridization enriched the pool of fragments for tester unique sequences and generated the first difference product (DP1). Further enrichment was achieved by performing a further two rounds of hybridization and amplification to generate difference products DP2 and DP3.

The DP3 fragments were subsequently digested with DpnII and shotgun cloned into BamHI-digested pBluescript KSII+ (Stratagene), prior to sequencing. Clones were identified using the NETBLAST program of the Wisconsin Genetics Computer Group (GCG) package (version 9.1) and the National Center for Biotechnology Information (NCBI) human genome Blast server to search the public databases for sequence homologies.

A reciprocal experiment, to identify genes expressed in healthy vasculature, used cDNA derived from the healthy aorta as tester and from the diseased sample as driver.

Southern analysis.
One-microgram quantities of tester and driver representative amplicons were electrophoresed on a 1% agarose gel. The DNA was denatured by soaking the gel in 1.5 M NaCl/0.5 M NaOH, then neutralized in 3.0 M NaCl/0.5 M HCl. Following overnight transfer to Hybond NX membrane (Amersham), the blots were hybridized to [{alpha}-32P] dCTP-labeled probes derived from the cloned DP3 fragments and washed under high-stringency conditions.

Virtual Northern analysis.
Double-stranded cDNAs derived from tester and driver RNAs were amplified using the Clontech SMART PCR system. First-strand synthesis of cDNA was performed using a modified oligo-dT primer and MMLV reverse transcriptase which adds deoxycytidine residues to the 3' end of the cDNA strand. Second-strand synthesis was then primed by annealing an oligonucleotide containing an oligo(G) stretch at the 3' end (SMART oligo) to the deoxycytidine residues added to the first strand. Both the SMART oligo and oligo-dT primers contained binding sites for PCR primers. The double-stranded cDNAs were then PCR amplified to generate a pool of full-length cDNAs representative of the starting RNA sample and to provide sufficient material for differential expression analysis of multiple clones by "virtual" Northern.

To confirm differential expression, 0.5-µg quantities of each amplified cDNA pool were fractionated on a 1% agarose gel, transferred to membranes, probed, and washed as described for Southern analysis.

RT-PCR analysis.
Semiquantitative RT-PCR was used to analyze gene expression in a panel of 40 cDNA samples derived from normal and atherosclerotic vessels. Five micrograms of total RNA was reverse transcribed using AMV reverse transcriptase and an oligo-dT primer. cDNA, 2.5 µl, was used as a template in 20-µl reactions containing gene-specific primers, reagents for PCR, and Taq DNA polymerase (Promega). All gene-specific primer pairs were designed to span an intron to detect amplification from contaminating DNA (Table 1). The identity of all PCR products was confirmed by sequencing. Preliminary experiments using a range of cycle numbers were performed to ensure that PCR amplification was within the linear range for each of the genes tested (not shown). Aliquots of 10 µl of the PCRs were fractionated on a 1% agarose gel, Southern blotted, and hybridized to labeled probes as described above. The amount of amplified product was measured by electronic autoradiography (Instant Imager, Packard) and normalized to a ß2-microglobulin control performed for each sample.


View this table:
[in this window]
[in a new window]
 
Table 1. Primers used for RT-PCR analysis of gene expression

 
Immunohistochemistry.
Serial sections from the two tissue samples used in the RDA, from representative lesions in the panel samples, and from nine carotid arteries used for in situ hybridization were stained for SM {alpha} -actin (VSMC), CD68 (macrophages), Q-Bend 10 and CD31 (endothelial cells), or CD3 (T cells, Dako), then stained by von Kossa’s method to detect calcified deposits and by oil red O for lipid.

In situ hybridization.
RNA isolated from 20 serial sections cut from nine carotid endarterectomy specimens was used to confirm gene expression in the lesions by RT-PCR. Adjacent sections were used for in situ hybridization, performed as described previously (21). Linearized plasmids containing cloned DP3 fragments were used to generate 35S-UTP-labeled sense and antisense cRNA probes by in vitro transcription. After washing, the slides were coated in Ilford K5 emulsion and exposed in the dark at 4°C for 3 wk before development and counterstaining with hematoxylin and eosin.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Figure 2 shows the aortic tissue samples used as the source of RNA for RDA. Driver cDNA was derived from RNA extracted from a normal aorta (Fig. 2A) that had no visible fatty streaks or atherosclerotic lesions. Immunohistochemistry confirmed that the vessel wall contained few macrophages and T cells with no lipid or calcified deposits (Fig. 3). In contrast, the aorta used as the source of tester cDNA had several atherosclerotic lesions (Fig. 2B). Immunohistochemistry of sections through the lesions confirmed that they comprised a fibrous cap of VSMCs overlying a necrotic core containing macrophages, T cells, lipid, and calcified material (Fig. 3).



View larger version (79K):
[in this window]
[in a new window]
 
Fig. 2. En face view of aortic tissue samples used as source RNA for the screen. A: healthy aorta (49-yr-old female), with no visible signs of fatty streak or lesion formation. B: atherosclerotic aorta (57-yr-old male) exhibiting fatty streaks (arrowheads) and several large fibrolipid plaques (arrows).

 


View larger version (69K):
[in this window]
[in a new window]
 
Fig. 3. Immunohistochemical analysis of aortic tissue used in RDA. Sections taken through the healthy aorta [top, media (m) is delineated by arrows] and one of the fibrous plaques in the diseased aorta [bottom, intima (i) is delineated by arrows] were stained with hematoxylin and eosin (A and B), antibodies against smooth muscle (SM) {alpha} -actin (C and D), CD68 (E and F; arrows indicate macrophages), CD3 (G and H; arrows indicate T cells), or with von Kossa’s stain for calcium (I and J; arrows indicate calcium).

 
Thirty-one clones representing lesion-associated genes were isolated by cDNA RDA (Table 2). Sequence analysis revealed that 18 clones were novel sequences with no matches in the GenBank, EMBL, and NCBI human genome nucleotide and expressed sequence tag (EST) databases. It is not known how many different genes are represented by these clones, but preliminary analyses indicate that the genes from which these clones are derived are expressed at very low levels in aortic tissue. The remaining 12 clones were known cDNAs and included the genes for intercellular adhesion molecule 1 (ICAM-1); tumor necrosis factor-{alpha} (TNF{alpha}) receptor type 2 (TNFR-2); the B- and C-chains of the complement protein C1q; plasminogen activator inhibitor type 2 (PAI-2); calcyclin, an S100 protein; Marco, a novel macrophage scavenger receptor; the {gamma}-subunit of Fc receptors (FcR{gamma}), an intracellular signaling molecule that mediates cellular activation in response to binding immune complexes; and Dap12, a natural killer cell receptor associated signaling molecule.


View this table:
[in this window]
[in a new window]
 
Table 2. Differentially expressed genes identified by cDNA RDA

 
In the reciprocal experiment, to identify genes expressed in the normal vasculature, cDNA from the diseased aorta was used as the driver, and cDNA from the normal aorta as tester. Two clones representing the genes for SM {alpha}-actin and {alpha}B-crystallin, a small heat shock-like protein, were isolated. Several clones representing ribosomal protein S4 and 28S rRNA were isolated from each RDA experiment. As expected, these clones were subsequently shown not to be differentially expressed between the tester and driver amplicons (see below): their presence in the final difference product (DP3) is most likely due to the use of total RNA rather than poly(A)+ RNA as the template for cDNA synthesis.

Confirmation of differential gene expression.
Differential expression of the clones was confirmed in several ways (Fig. 4). First, the abundance of the cloned sequences in the representative amplicons derived from the healthy and diseased tissues was investigated by Southern analysis. Second, expression of the full-length cDNAs in the two samples was determined by virtual Northern analysis. However, the expression of two genes, Dap12 and Marco, could not be detected by this technique. Presumably this was due to low levels of expression of Marco and Dap12 in the tissue samples, as differential expression was subsequently confirmed by RT-PCR analysis.



View larger version (31K):
[in this window]
[in a new window]
 
Fig. 4. Differential expression of the cDNAs isolated from the RDA screen was confirmed by Southern analysis of the representative amplicons (Dap12, {alpha}B-crystallin, calcyclin), virtual Northern analysis (ICAM-1, FcR{gamma}, TNFR-2, ß-actin control), and RT-PCR analysis of gene expression (Marco). Left lane, normal aorta. Right lane, atherosclerotic aorta.

 
All of the disease-associated genes, with the exception of calcyclin, exhibited an absolute difference in expression between the two cDNA samples. In contrast, SM {alpha}-actin, {alpha}B-crystallin and calcyclin were expressed in both the normal and diseased cDNA samples: calcyclin was more highly expressed in the diseased sample, whereas higher levels of expression of SM {alpha}-actin and {alpha}B-crystallin were observed in the normal cDNA sample (Fig. 4).

Expression patterns in atherogenesis.
To establish whether the association of these genes with atherosclerotic disease was consistent over a larger number of lesions, semiquantitative RT-PCR was used to investigate the levels of mRNA in 40 cDNA samples derived from human aortic and carotid endarterectomy tissue (Table 3). Group 1 consisted of 20 samples derived from normal and fatty streak aortic tissue, whereas group 2 consisted of 10 samples from advanced aortic atherosclerotic lesions and from 10 carotid endarterectomy specimens. Immunohistochemistry was used to assess the numbers of macrophages and T cells and the amount of calcification in each sample (Table 3). Although some macrophages were detected in the group 1 samples, particularly in the fatty streak tissues, these samples contained few T cells and had no calcified deposits. In contrast, all of the samples in group 2 had high numbers of T cells and macrophages and contained deposits of calcified material. Furthermore, the carotid samples in group 2 also displayed evidence of plaque rupture including the presence of organized thrombus, indicating that these lesions represent unstable atheroma. The aortic samples in group 2, however, had fibrous caps and showed no evidence of thrombus formation.


View this table:
[in this window]
[in a new window]
 
Table 3. Aortic and carotid endarterectomy panel samples used in RT-PCR analysis

 
Expression of all genes isolated by RDA was heterogeneous within each group of samples analyzed (Fig. 5A). Indeed, there was marked heterogeneity of gene expression both within groups and between groups for the genes isolated. When between group comparisons were made, expression of calcyclin, C1qB, and Dap12 was not significantly different between normal and atherosclerotic vessels. However, expression levels of Marco, PAI-2, TNFR-2, and FcR{gamma} were all significantly increased in atherosclerotic lesions compared with normal/fatty streak vessels (Fig. 5B). In contrast, levels of {alpha}B-crystallin mRNA were significantly lower in the atherosclerotic samples (Fig. 5). A similar pattern of expression was observed for SM {alpha}-actin, an SM-specific protein (not shown).



View larger version (30K):
[in this window]
[in a new window]
 
Fig. 5. Gene expression patterns in human panel samples. A: RT-PCR showing expression of genes in 5 normal, 5 aortic atherosclerotic, and 5 carotid atherosclerotic samples demonstrating differential expression of genes identified by RDA. B: mRNA expression of the genes indicated was examined in normal/fatty streak aortas (group 1, n = 20; open bars) atherosclerotic aortas and atherosclerotic carotid arteries (group 2, n = 20; gray bars) by semiquantitative RT-PCR. The y-axis shows the mean level of expression of each gene normalized to a ß-microglobulin control. Bars represent the means ± SE for each group. ANOVA was used to determine significance. **P < 0.05. Small inset graphs show levels of gene expression in each individual lesion (group 1, open bars; group 2, gray bars); ns, not significant.

 
Tissue expression in vitro and in vivo.
RT-PCR on cultured human VSMCs and macrophages was used to investigate which cell types within vascular lesions may potentially express the cDNAs isolated by the screen (Fig. 6). C1qB, Dap12, and FcR{gamma} were expressed in cultured macrophages but not VSMCs, whereas PAI-2 and {alpha}B-crystallin were detected only in VSMCs. Calcyclin, Marco, and TNFR-2 mRNA were present in both cell types.



View larger version (35K):
[in this window]
[in a new window]
 
Fig. 6. RT-PCR analysis of gene expression in cultured human vascular smooth muscle cells (VSMCs) (passage 3) (right lane) and macrophages (day 21) (left lane).

 
To identify the types of lesions and the cell types within atheromatous plaques that were expressing the lesion-associated genes, we used RT-PCR on RNA derived from tissue sections, and in situ hybridization, to examine the distribution of Marco, TNFR-2, PAI-2, and FcR{gamma} mRNA in nine carotid lesions removed by endarterectomy because of symptomatic disease. Histological analysis revealed that two of these lesions did not have the morphology typical of an unstable atheromatous plaque. In both the " stable" lesions, one SMC-rich and fibrocellular and one heavily calcified, the expression of all four genes was barely detectable by both RT-PCR (not shown) and in situ hybridization. These two lesions were also devoid of macrophage-derived foam cells. In the seven " unstable" lesions, expression of all four genes was high and found to be highest in subsets of macrophages and foam cells particularly as sites of neovascularization within the plaque (Fig. 7). Marco, PAI-2, and TNFR-2 were also expressed in association with organized thrombus, whereas low levels of PAI-2 and TNFR-2 expression were detected in intimal VSMCs.



View larger version (103K):
[in this window]
[in a new window]
 
Fig. 7. Immunohistochemistry and in situ hybridization of carotid endarterectomy specimens. A: in situ hybridization light-field (LF) and dark-field (DF) views of Marco (b and b‘, x20) and PAI-2 (c and c‘, x20). Note that immunohistochemistry of adjacent sections shows that expression of both these genes is predominantly by macrophages (d, x40) localized at sites of neovascularization (arrows) (a, x40). B: in situ hybridization showing FcR{gamma} expression (arrow) (c and c‘, LF and DF microscopy, respectively). Expression is predominantly by macrophages (b) present at sites of neovascularization (a). The boxed area in c is enlarged in a and b to show details of microvessels and macrophages in this region of the plaque; c‘, negative sense control for in situ showing some autofluorescence of calcium deposits; d, negative control for immunohistochemistry.

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The aim of this study was to investigate the utility of cDNA RDA as a technique for studying differential gene expression in human atheromatous lesions. As atherosclerosis is a disease involving several cell types, where the expression of inflammatory mediators by macrophages and foam cells within the plaque may influence gene expression in cells in the vessel wall, i.e., VSMCs and endothelial cells, we reasoned that studies on human pathological specimens would be more informative than studying these cell types in isolation by using cultured cells. However, the quantity and particularly the quality of RNA obtained from human vascular material is generally unsuitable for cDNA library construction or probe generation for cDNA array analysis. To overcome this problem, we employed PCR-based methods of differential screening, as such methods require much less material.

Initial experiments using differential display RT-PCR (DD RT-PCR) (17) proved unsuccessful, due to the high incidence of false-positive clones isolated (Tyson KL and Shanahan SM, unpublished results). However, cDNA RDA avoids this problem by only amplifying those sequences that are differentially expressed (14). Indeed, we isolated only two false-positive clones from our experiments, which represented ribosomal protein S4 and 28S rRNA. cDNA RDA is usually performed using poly(A)+ RNA as the template for cDNA synthesis (14), but to conserve material we used total RNA. The two false-positive clones are major transcript species in total RNA, which may account for their presence in the final difference products. All of the other genes isolated were confirmed as being differentially expressed, with most of the genes showing an absolute difference in expression between the two amplicons. The high driver-to-tester ratio used in the final round of subtractive hybridization (400,000:1) will bias the subsequent PCR step in favor of amplification of absolute differences rather than amplification of cDNAs that show a change in the level of expression. By altering the hybridization conditions or eliminating the final subtraction, it should be possible to isolate sequences present at different levels in the two samples.

Some of the lesion-associated genes isolated in the screen did not show a consistent association with atherosclerotic disease when examined over a larger sample size, reflecting an inherent variability in gene expression between different lesions and individuals. Nevertheless, {alpha}B-crystallin and SM {alpha}-actin, isolated as preferentially expressed in normal vessels, both showed significantly decreased expression in atherosclerotic disease. Similarly, four genes, Marco, PAI-2, TNFR-2, and FcR{gamma}, demonstrated elevated expression in diseased blood vessels compared with healthy vascular tissues. Moreover, in situ hybridization demonstrated that this elevated expression was preferentially observed in " unstable" carotid plaques. It is likely that these changes in expression reflect in part the increased number of inflammatory cells in lesions. However, the expression of these four genes in association with subsets of macrophages at sites of neovascularization may point to a role for this process in contributing to instability. Whether the products of these genes contribute to the pathological process of plaque rupture will require further studies.

Genes identified by the screen: properties and expression in healthy and diseased vessels.
Several of the genes isolated by the screen, including Marco, FcR{gamma}, PAI-2, and TNFR-2, are potential mediators of inflammatory and immune reactions, emphasizing important roles for these processes in atherogenesis.

Marco is a type I scavenger receptor, which has been implicated in host defense against bacterial infection (4, 5, 25). The role of infectious organisms in the progression of atherosclerosis has received attention recently (19); thus the detection of Marco expression in human lesions is enticing. Marco has also been detected in murine atherosclerotic lesions (20). Expression of Marco in carotid lesions was highest in macrophage-derived foam cells, and we have also detected Marco mRNA in cultured VSMCs (data not shown). It is tempting to speculate that Marco may play a role in both foam cell formation and VSMC function.

FcR{gamma} chain is a membrane-bound signaling molecule that is common to several Fc receptors, which bind to the Fc domains of antibodies (3). Upon binding immune complexes, the intracellular domain of FcR{gamma} triggers a phosphorylation cascade that leads to cellular activation. FcR{gamma} was expressed by macrophages in the carotid lesions, and Fc receptors on this cell type have been implicated in foam cell formation and the activation of cytokine release (11, 18, 27). More recently, FcR{gamma} has been shown to be part of the platelet collagen receptor, and so this protein may have multiple roles in cellular activation in plaque cells (22).

TNF{alpha} is a potent pro-inflammatory cytokine produced by macrophages, T cells, and SMCs in atherosclerotic plaques. The type 2 TNF{alpha} receptor, TNFR-2, is less well characterized that TNFR-1. However, signaling by both receptors can lead to activation of NF-{kappa}B, a transcription factor that regulates expression of inflammatory mediators such as adhesion molecules, growth factors, and cytokines. Thus the effects of TNF{alpha} on the artery wall are largely considered to be pro-atherogenic. The contribution of TNFR-2-mediated signaling to atherogenesis and plaque cell apoptosis must await further studies (16).

PAI-2 is a secreted protein that regulates fibrinolysis by inhibiting urokinase-type plasminogen activator (u-PA) activity. In this study, in situ hybridization showed that PAI-2 was expressed by intimal VSMCs and macrophages, particularly in association with organized thrombus. In a rat model of periodontal wound healing, PAI-2 expression was also found to be closely associated with infiltrating inflammatory cells and the deposition of a fibrin clot (28). Thus PAI-2 expression may be associated with the wound-healing process following plaque rupture.

In conclusion, cDNA RDA has enabled us to identify lesion-associated genes from small amounts of material. The technique is quick, extremely sensitive, able to isolate genes expressed at low levels and by only a small subset of cells within the tissue sample. Increasingly, microarrays are being used to analyze gene expression and these systems offer the ability to rapidly analyze the expression of a large number of genes in a single experiment. However, microarrays represent a "closed" system, as only the expression of those genes represented on the array can be interrogated. In contrast, cDNA RDA, is an "open" system, in which no assumptions are made about the genes expressed by the tissue under investigation (10). Indeed, RDA has the potential to interrogate the expression of all genes in a cDNA population that contain an amplifiable DpnII fragment, and can also lead to the discovery of novel sequences. Interestingly, several groups have more recently combined the techniques of RDA and gene expression arrays, using RDA to enrich for differentially expressed clones that are subsequently spotted onto an array and probed with labeled cDNA (1, 7, 15). This approach combines the ability of RDA to select for differentially expressed genes, with the capacity of arrays for screening large numbers of sequences simultaneously. It may be that such combinatorial methods might prove useful in the future, particularly where the quality of the starting RNA is unsuitable for array analysis, and such approaches help to reduce the number of clones that require secondary screening and characterization.

In terms of the analysis of differential gene expression in atherosclerosis, RDA alone led to the identification of absolute differences in gene expression between two samples. However, subsequent expression analysis in a large panel of lesions unmasked heterogeneity between lesions, clearly demonstrating the importance of the careful selection of tester and driver RNA. In future experiments the pooling of cDNA samples for the generation of representative amplicons should enable us to isolate more robust markers of atherosclerotic disease. In particular, the isolation of markers of plaque instability would be useful for identifying patients at risk of a clinical event. This could be best achieved by careful prior screening of lesion histology for signs of plaque stability vs. rupture before the incorporation of their RNA into a screen. The small quantity of RNA required for RDA would make such a process feasible.


    ACKNOWLEDGMENTS
 
We acknowledge the important contributions of Mr. P. Kirkpatrick of the Department of Neurosurgery, Addenbrooke’s Hospital, and the Addenbrooke’s Transplant Team.

This research is supported by project grants from SmithKline Beecham Pharmaceuticals and the British Heart Foundation to C. M. Shanahan and P. L. Weissberg.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: C. M. Shanahan, Dept. of Medicine, Division of Cardiovascular Medicine, Addenbrooke’s Centre for Clinical Investigation, Box 110, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, UK (E-mail: cs131{at}mole.bio.cam.ac.uk).

10.1152/physiolgenomics.00116.2001.


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Boeuf S, Klingenspor M, van Hal NL, Schneider T, Keijer J, and Klaus S. Differential gene expression in white and brown preadipocytes. Physiol Genomics 7: 15–25, 2001. First published August 8, 2001;10.1152/physiolgenomics.00048.2001.[Abstract/Free Full Text]
  2. Bøyum A. Isolation of mononuclear cells and granulocytes from human blood: isolation of mononuclear cells by one centrifugation by combining centrifugation and sedimentation at 1g. Scand J Clin Lab Invest 21: 77–89, 1986.
  3. Däeron M. Fc receptor biology. Annu Rev Immunol 15: 203–234, 1997.[ISI][Medline]
  4. Elomaa O, Kangas M, and Sahlberg C. Cloning of a novel bacteria-binding receptor structurally related to scavenger receptors and expressed in a subset of macrophages. Cell 80: 603–609, 1995.[ISI][Medline]
  5. Elomaa O, Sankala M, Pikkarainen T, Bergmann U, Tuuttila A, Raatikainen-Ahokas A, Sariola H, and Tryggvason K. Structure of the human macrophage Marco receptor and characterization of its bacteria-binding region. J Biol Chem 273: 4530–4538, 1998.[Abstract/Free Full Text]
  6. Falk E, Shah PK, and Fuster V. Coronary plaque disruption. Circulation 92: 657–671, 1995.[Free Full Text]
  7. Frohme M, Scharm B, Delius H, Knecht R, and Hoheisel JD. Use of representational difference analysis and cDNA arrays for transcription profiling of tumor tissue. Ann NY Acad Sci 910: 85–104, 2000.[Abstract/Free Full Text]
  8. Fuster V, Badimon L, Badimon JJ, and Chesebro JH. The pathogenesis of coronary artery disease and the acute coronary syndromes. N Engl J Med 326: 242–250, 1992.[ISI][Medline]
  9. Fuster V and Lewis A. Conner Memorial Lecture. Mechanisms leading to myocardial infarction: insights into studies of vascular biology. Circulation 90: 2126–2146, 1994.[Abstract]
  10. Green CD, Simons JF, Taillon BE, and Lewin DA. Open systems: panoramic views of gene expression. J Immunol Methods 250: 67–79, 2001.[ISI][Medline]
  11. Griffith RL, Virella GT, Stevenson HC, and Lopes-Virella MF. Low density lipoprotein metabolism by human macrophages activated with low density lipoprotein immune complexes. J Exp Med 168: 1041–1059, 1988.[Abstract]
  12. Haley KJ, Lilly CM, Yang JH, Feng Y, Kennedy SP, Turi TG, Thompson JF, Sukhova GH, Libby P, and Lee RT. Overexpression of eotaxin and the CCR3 receptor in human atherosclerosis: using genome technology to identify a potential novel pathway of vascular inflammation. Circulation 102: 2185–2189, 2000.[Abstract/Free Full Text]
  13. Horrevoets AJ, Fontyn RD, van Zonneveld AJ, de Vries CJ, ten Cate JW, and Pannekoek H. Vascular endothelial genes that are responsive to tumor necrosis factor-alpha in vitro are expressed in atherosclerotic lesions, including inhibitor of apoptosis protein-1, stannin and two novel genes. Blood 93: 3418–3431, 1999.[Abstract/Free Full Text]
  14. Hubank M and Schatz DG. Identifying differences in mRNA expression by representational difference analysis of cDNA. Nucleic Acids Res 22: 5640–5648, 1994.[Abstract]
  15. Kim S, Zeller K, Dang CV, Sandgren EP, and Lee LA. A strategy to identify differentially expressed genes using representational difference analysis and cDNA arrays. Anal Biochem 288: 141–148, 2001.[ISI][Medline]
  16. LeBoeuf RC and Schreyer SA. The role of tumour necrosis factor-{alpha} receptors in atherosclerosis. Trends Cardiovasc Med 8: 131–138, 1998.[ISI]
  17. Liang P and Pardee AB. Differential display of eukaryotic messenger RNA by means of the polymerase chain reaction. Science 257: 967–971, 1992.[ISI][Medline]
  18. Lopes-Virella MF, Binzafar N, Rackley S, Takei A, La Via M, and Virella G. The uptake of LDL-IC by human macrophages: predominant involvement of the Fc{gamma}RI receptor. Atherosclerosis 135: 161–170, 1997.[ISI][Medline]
  19. Noll G. Pathogenesis of atherosclerosis: a possible relation to infection. Atherosclerosis 140: S3–S9, 1998.[ISI][Medline]
  20. Sakaguchi H, Takeya M, Suzuki H, Hakamata H, Kodama T, Horiuchi S, Gordon S, van der Laan LJ, Kraal G, and Ishibashi S. Role of macrophage scavenger receptors in diet-induced atherosclerosis in mice. Lab Invest 78: 423–434, 1998.[ISI][Medline]
  21. Shanahan CM, Cary NRB, Metcalfe JC, and Weissberg PL. High expression of genes for calcification-regulating proteins in human atherosclerotic plaques. J Clin Invest 93: 2393–2402, 1994.[ISI][Medline]
  22. Tsuji M, Ezumi Y, Arai M, and Takayama H. A novel association of the Fc receptor gamma chain with glycoprotein VI and their co-expression as a collagen receptor in human platelets. J Biol Chem 272: 23528–23531, 1997.[Abstract/Free Full Text]
  23. Tung WS, Lee JK, and Thompson RW. Simultaneous analysis of 1176 gene products in normal human aorta and abdominal aortic aneurysms using a membrane-based complementary cDNA expression array. J Vasc Surg 34: 143–150, 2001.[ISI][Medline]
  24. Tyson K and Shanahan CM. Use of cDNA representational difference analysis to identify disease-specific genes in human atherosclerotic plaques. In: Vascular Disease: Molecular Biology and Gene Therapy Protocols edited by Baker AH. Totowa, NJ: Humana, 1999, p. 83–98.
  25. van der Laan LJW, Dopp EA, Haworth R, Pikkarainen T, Kangas M, Elomaa O, Dijkstra CD, Gordon S, Tryggvason K, and Kraal G. Regulation and functional involvement of macrophage scavenger receptor Marco in clearance of bacteria in vivo. J Immunol 162: 939–947, 1999.[Abstract/Free Full Text]
  26. Vasile E, Tomita Y, Brown LF, Kocher O, and Dvorak HF. Differential gene expression of thymosin beta-10 by early passage and senescent vascular endothelium is modulated by VPF/VEGF: evidence for senescent endothelial cells in vivo at sites of atherosclerosis. FASEB J 15: 458–466, 2000.[Abstract/Free Full Text]
  27. Virella G, Muñoz JF, Galbraith GP, Gissinger C, Chassereau C, and Lopes-Virella MF. Activation of human monocyte-derived macrophages by immune complexes containing low-density lipoprotein. Clin Immunol Immunopathol 75: 179–189, 1995.[ISI][Medline]
  28. Xiao Y, Li H, Bunn C, and Bartold PM. The expression of plasminogen activator system in a rat model of periodontal wound healing. J Periodontol 72: 849–857, 2001.[ISI][Medline]
  29. Yasuda O, Zhang SH, Miyamoto Y, and Maeda N. Differential expression of {alpha}1 type VIII collagen gene by smooth muscle cells from atherosclerotic plaques of apolipoprotein-E-deficient mice. J Vasc Res 37: 158–169, 2000.[ISI][Medline]