Differential Gene Expression in Vascular Smooth Muscle Cells in Primary Atherosclerosis and In Stent Stenosis in Humans
Objective— We sought to identify differentially expressed genes in human in stent stenosis (ISS) to provide insights into the mechanism of disease.
Methods and Results— Using representation difference analysis, we examined differential gene expression between cultured normal human medial vascular smooth muscle cells (VSMCs) and cells from primary atherosclerotic plaques or ISS sites. Specific groups of genes were overexpressed in ISS and plaque VSMCs, including cell cycle regulatory proteins and cell matrix and contractile proteins. Differential expression was validated by virtual Northern analysis, reverse transcriptase-polymerase chain reaction, in situ hybridization, and immunohistochemistry. All ISS genes were expressed by normal intima and had even higher expression in primary plaque VSMCs.
Conclusions— ISS VSMCs have a stable gene expression profile reflecting an intimal pattern, intermediate between normal medial and primary plaque VSMCs. Differential expression profiling may identify markers of disease that are overexpressed in ISS and also help elucidate the origin of the ISS lesion.
The major limitation of intracoronary stenting is in stent stenosis (ISS), manifesting as late arterial renarrowing at the intervention site. The success of new interventional devices has (until recently) contrasted with the failure of drugs to prevent ISS.1 This failure may reflect lack of differential gene expression in ISS compared with the normal vessel that can be manipulated therapeutically.
Previous studies have identified differences in biological properties in human angioplasty or ISS vascular smooth muscle cells (VSMCs) versus VSMCs from normal vessels. Histological studies describe hyperplasia of VSMCs in restenosis or a fibroproliferative phenotype,2,3⇓ and cultured VSMCs from restenosis lesions proliferate faster than cells from primary plaques.4,5⇓ Restenosis after stenting is attributable mostly to neointimal formation,6 and there is increased collagen and proteoglycan synthesis in restenotic versus primary plaque lesions7,8⇓ and also in culture in some studies.9 Although these studies show that cells in or derived from restenotic lesions show different biological properties, the different gene expression underlying these properties have been little studied in human ISS.
Tremendous advances in genetic techniques have permitted study of differential gene expression in diseased versus normal tissues or cells, identifying gene expression profiles and specific targets for drug development. These methods include subtractive hybridization, differential display, sequential analysis of gene expression, representational difference analysis (RDA), and DNA microarrays. The choice of technique is based in part on the amount of study material; polymerase chain reaction (PCR)-based methods are required for small amounts of tissue or cells. In addition, some techniques can better screen out large numbers of false-positive results and identify true positives, even when expressed at low abundance.
We used RDA to identify differentially expressed genes between human medial VSMCs and both primary plaque or ISS VSMCs. Cultured VSMCs were used so that stable gene expression could be assessed, rather than transient changes associated with the local plaque microenvironment. The use of cultured cells also removes the large numbers of inflammatory cell products that would otherwise appear in primary plaque DNA.
Cell Isolation and Culture
Primary atherosclerotic plaque cells were derived from carotid artery endarterectomies obtained from patients with symptomatic disease. Normal medial VSMCs were obtained from both aorta and coronary artery at renal or cardiac transplantation, respectively. ISS specimens were obtained by directional coronary atherectomy from patients with stable angina. Ethical committee approval and informed consent were obtained. Patients were aged 58 to 83 years (primary plaque, n=14), 48 to 76 years (ISS, n=3), and 33 to 55 years (normal coronary arteries, n=6). VSMCs were explant cultured as before,10 and cells passaged 1:3 by trypsinization at confluence to passages 3 to 8. Screening of cultures with antibodies to CD3 (lymphocytes), CD68 (macrophages), CD31 (endothelial cells), and α-smooth muscle-actin showed pure VSMC cultures.
Representational Difference Analysis
Total RNA isolation, double-stranded cDNA synthesis cDNA, and representational difference analysis (RDA) were performed as previously described11,12⇓ using cDNA from normal VSMCs as driver and primary plaque or ISS VSMCs as tester. Briefly, dsDNA was synthesized from each sample, and equal amounts were pooled to 12 μg. Samples were digested with DpnII, ligated to R-adapters, and PCR amplified to provide ≈300 μg DNA from each representative amplicon. The R adapters were removed from both amplicons by DpnII digestion and J adapters ligated to the tester amplicon only. The representative amplicons then underwent 3 rounds of subtractive hybridization and PCR amplification, generating a final difference product comprising differentially expressed genes unique to the tester.
Difference products were excised from agarose gels, purified, and cloned into pCRII (TA Cloning Kit, Invitrogen) and sequenced by PCR using M13 primers. Sequence comparisons were made with GenBank using the BLAST computer program.
Virtual Northern Blot Analysis
Equal amounts (1 μg or 0.5 μg) of cDNA amplicons from driver or tester samples were separated on a 1.5% agarose gel. The DNA was denatured (1.5 mol/L NaCl, 0.5 mol/L NaOH for 30 minutes) and neutralized (1.5 mol/L NaCl, 0.5 mol/L Tris-HCl pH7.2, 0.001 mol/L EDTA for 2×15 minutes) within the gel then transferred to a nylon membrane (Hybond N+, Amersham). A random primed, 32P-labeled target cDNA probe (1.0×109 dpm/μg) was hybridized for 16 hours at 68°C in 5×SSC, 5× Denhardt’s solution, 0.5% SDS, and denatured sonicated nonhomologous DNA. Unincorporated probes were removed in 2×SSC, 0.1% SDS for 15 minutes, and 0.1× SSC, 0.1% SDS for 30 minutes, each at 65°C.
In Situ Hybridization
Target gene fragments were cloned into a T7/SP6 promoter-containing vector (Riboprobe, Promega) and 0.5 μg linearized DNA used for synthesis of sense and antisense digoxigenin-labeled cRNA probe. Frozen sections of human atherosclerotic or normal artery were incubated in 100 mmol/L glycine in PBS (10 minutes) and 0.3% Triton X-100 in PBS (15 minutes), permeabilized at 37°C with 1 μg/mL RNase-free proteinase K in PE buffer for 30 minutes. After postfixation with 4% PFA in PBS, slides were quenched in TEA buffer (0.1 mol/L triethanolamine, 0.25% acetic anhydride) for 10 minutes and prehybridized (4×SSC containing 50% [vol/vol] deionized formamide) at 37°C for 15 minutes. Hybridizations were performed at 37°C for 16 hours with 5 to 10 ng of labeled RNA probe in hybridization buffer (4×SSC, 10% dextran sulfate, 1×Denhardt’s solution, 1 mg/mL denatured and sheared salmon sperm DNA, 10 mmol/L DTT). Slides were washed for 2×30 minutes with 2×SSC, 1×SSC, and 0.1×SSC each at 37°C.
Confirmation of differential protein expression was performed using immunohistochemistry of rat carotid artery sections, 2 weeks after standard balloon injury.13 Antibodies used were from Serotec (CD81, filamin, HSP-27, cathepsin D), Biocarta (laminin β, collagen α2, tropomyosin), or Santa Cruz (IGFBP2).
Representational Difference Analysis
We used RDA to isolate differentially expressed genes between intimal VSMCs cultured from primary atherosclerotic plaques or ISS versus medial VSMCs from normal vessels. All VSMCs expressed α-smooth muscle actin, calponin, and smooth muscle cell myosin, indicating their VSMC origin, but differed in gross phenotype. In particular, rates of cell proliferation and passage number before senescence were normal>ISS>plaque, and rates of apoptosis were ISS>plaque>normal, similar to other studies.14,15⇓ The first screen compared normal VSMCs from coronary arteries and aorta with VSMCs from primary carotid plaques. Sequencing of difference products identified several genes whose expression was increased in primary plaque VSMCs, including insulin-like growth factor 1 binding protein 2 (IGF-BP2), cartilage link protein, enteric muscle γ-actin, α-tropomyosin, and laminin β1 and also some novel gene sequences (Table 1).
Confirmation of Differential Expression by Virtual Northern Blotting
Virtual Northern blots, which use cDNA as the template instead of mRNA, were used to confirm that the identified sequences showed differential expression in mRNA samples from original cell populations before RDA. Expression of cartilage link protein, α tropomyosin, IGFBP2, laminin β1, II-18 (AC018738), V8 (AC021937), and V22 (AL589792) was increased between 1.5- and 3.1-fold in primary plaque versus normal medial VSMCs (Figure 1, Table 1). Differential expression of enteric muscle γ-actin was not confirmed.
In Situ Hybridization of Plaque Differentially Expressed Genes
In situ hybridization of human coronary atherosclerotic plaques at different stages or histologically normal arteries was used to confirm expression of differentially expressed genes in primary plaques in vivo. Although almost all identified genes were expressed in primary plaques, IGF-BP2, cartilage link protein, α-tropomyosin, and laminin β1 were also expressed in the normal vessel intima (Figure 2). Enteric γ-actin expression was seen uniformly throughout VSMCs in the vessel wall.
RDA Analysis of ISS Versus Normal Medial VSMCs
The second RDA screen examined differential gene expression in VSMCs from ISS sites versus normal medial VSMCs cultured only from coronary arteries, effectively excluding any differences attributable to VSMCs being derived from different arterial beds. This comparison yielded a larger panel of genes whose expression was increased in ISS versus normal medial VSMCs (Table 2). This list includes genes associated with cell proliferation, contractile machinery, and cell/matrix signaling.
Analysis of Differentially Expressed Genes in ISS, Primary Plaque, and Normal VSMCs
Virtual Northern analysis was used again to confirm that the genes identified by RDA were truly differentially expressed in ISS versus normal medial VSMCs and also to examine their relative expression (Figure 3). For most genes, increased expression was confirmed in ISS versus normal medial VSMCs, from 1.2- to 4.0-fold in ISS versus normal VSMCs (Table 2). However, in almost all cases, primary plaque VSMCs showed even higher expression. Thus, the gene profile of ISS VSMCs is intermediate between medial and plaque VSMCs. Differences in size of expressed cDNA were also evident, indicating the possibility of different splice variants (eg, prolactin receptor-associated protein) in ISS versus primary plaque VSMCs.
Uniformity of Differential Expression
Our RDA analysis examined differential gene expression between pooled cDNAs rather than cDNAs isolated from individual patients. Thus, our findings could result from a massive increase in expression of any particular gene in a subset of samples that was not representative of all cDNAs from that group. We therefore examined expression of some differentially expressed genes in random individual patient VSMC cultures. Online Figure I (see http://atvb.ahajournals. org) shows a remarkable uniformity of expression within groups of cDNAs, with differential expression evident between groups. This analysis also demonstrated that many genes identified in the primary plaque versus normal VSMC RDA screen were also overexpressed in ISS VSMCs.
In Situ Hybridization and Immunohistochemistry of ISS Differentially Expressed Genes
Because most genes overexpressed in primary plaque VSMCs were also expressed by the normal intima, we examined expression of ISS genes by in situ hybridization in normal vessels. Figure 4 shows that, similar to primary plaque genes, the ISS genes were also components of the normal intima and in intimal thickenings. To examine whether the genes identified were markers of the intima (or neointima), we examined protein expression of a sample of the identified genes in an animal model of neointima formation after injury, the ballooned rat carotid artery. Online Figure II (see http://atvb.ahajournals.org) demonstrates that protein expression of these gene products was increased in intima versus media in this model.
Atherosclerosis is characterized by an accumulation of vascular smooth muscle cells, intracellular and extracellular lipid, inflammatory cells, and extracellular matrix. Primary plaque VSMCs display characteristic features in vivo, including an undifferentiated or synthetic phenotype with few myofilament bundles at the cell periphery and a high organelle content.16 In culture, plaque VSMCs demonstrate poor proliferation, early senescence, and increased apoptosis compared with medial VSMCs.14,16⇓ Some of the genetic differences underlying these properties have been elucidated, such as an inability to phosphorylate RB (blocking proliferation)17 and defective IGF-1 signaling in plaque VSMCs10 (promoting apoptosis). ISS is caused by neointimal accumulation of VSMCs and their products, extracellular matrix and collagen. VSMCs in ISS lesions also demonstrate an undifferentiated or synthetic phenotype,4 manifest by different expression of matrix proteins and collagen and increased growth rates in culture compared with primary plaque VSMCs.4,15⇓ These previously described growth properties of primary plaque and ISS VSMCs were also evident in this study.
The stability of these properties argues that, like primary plaque VSMCs, ISS VSMCs represent a distinct phenotype with different gene expression, which should be identifiable by differential screening. Previous genetic screens of human vascular tissue have used microarrays to identify genes differentially expressed in whole tissue in artery versus vein,18 ISS versus normal artery,19,20⇓ and ruptured versus unruptured plaques.21 These studies have identified a variety of genes overexpressed in one of the two populations studied, although direct correlation of gene expression with the known properties of plaque or ISS VSMCs has not been performed.
We therefore used RDA to examine differential gene expression between normal and primary atherosclerotic plaque or ISS VSMCs. In contrast to some other methods, RDA is a positive selection technique that couples subtractive hybridization to PCR amplification. Thus, only differentially expressed sequences are amplified, reducing the number of false-positives isolated and comparison of large numbers of hybridization signals. The subtractive hybridization steps generate a difference product representing all of the differentially expressed genes amplified in a single reaction, and unwanted difference products are thus eliminated.11 In addition, cDNA RDA allows the isolation of clones representing rarely expressed mRNAs22 and is not limited to those genes on a microarray.
RDA identified several genes that were overexpressed in plaque compared with normal VSMCs. Differential expression was confirmed by virtual Northern analysis and in situ hybridization. In particular, plaque VSMCs showed 1.5- to 3.0-fold increased expression of some contractile proteins such as skeletal muscle α-tropomyosin, a marker of dedifferentiation,23 but not actin isoforms, which are more associated with differentiated medial VSMCs.24 We also identified increased IGF-BP2 expression in plaque VSMCs. This confirms our earlier observations that human plaque VSMCs secrete higher levels of IGFBP2, 3, and 4 than normal medial VSMCs and IGF-BP2 directly promotes the increased sensitivity to apoptosis of plaque VSMCs.10 These studies strongly argue that the differences we find in expressed protein result in differences in biological properties of VSMCs from different sources. Of note, laminin β1 was also increased in plaque VSMCs. Laminin-1, a major component of the basement membrane, consists of α1, β1, and γ1 chains. Laminin β1 regulates cell/matrix adhesion, permitting VSMC spreading and migration over basement membrane.25 Thus, the identified genes largely reflected either the change in VSMC phenotype (contractile protein expression) or known properties of plaque VSMCs (IGF-bp2, laminin β1). Importantly, the identified genes are markers of intimal cells, not necessarily plaque VSMCs, because intimal expression was also seen in normal vessels.
RDA of ISS versus normal medial VSMCs identified a larger group of genes, some common to primary plaque VSMCs. These genes could be grouped according to function, reflecting contractile proteins, extracellular matrix (ECM) proteins or synthetic proteins of ECM, basal transcription machinery proteins, proteins associated with receptor-associated signaling, proteins induced during proliferation, membrane tetraspan multifunctional proteins, or mitochondrial or ribosomal proteins. Importantly, none of the genes were unique to ISS VSMCs but were expressed at even higher levels in primary plaque VSMCs and were also expressed in the intima of normal arteries or rat vessels after injury.
We identified many genes overexpressed in ISS VSMCs that encode filament proteins as part of the contractile machinery (cofilin, filamin, and actinin). Actinins are actin-binding proteins involved in bundle formation, anchorage to focal adhesion complexes, and integrin signaling.26 Downregulation of α-actinin is associated with inability to migrate and proliferate on polymerized collagen.26 The increased expression of α-actinin we observe may therefore reflect an altered ability of plaque/ISS VSMCs to respond to monomeric collagen seen after matrix breakdown. In addition, differential expression of extracellular matrix genes (thrombospondins I and II, collagen α2) was seen, possibly reflecting the different composition of ECM synthesized by plaque/ISS VSMCs.
Importantly, several gene products that regulate cell proliferation or are induced when cells proliferate were increased in primary plaque and ISS cells. These included the SWI/SNF complex 60-kDa subunit (BRG1) and the growth factor-induced gene 2A9. BRG1 is a component of the chromatin remodelling complex required for pRB-induced growth arrest, whose function is disrupted by cyclin E binding.27 We have previously demonstrated that primary plaque VSMCs show enhanced RB-mediated growth suppression, which needs to be overcome for proliferation to occur.17 Clearly, BRG1 may mediate the poor proliferation of plaque VSMCs, and sequestration of BRG1 may be a mechanism that ISS cells use to overcome RB-induced cell cycle block. The differential expression of receptor-associated proteins may also reflect differences in signaling after cellular stress (Tyk2) or inflammatory cell cytokines (IRAK) present in the atherosclerotic plaque.
In contrast to other studies examining differential gene expression in stent tissue,19,20⇓ we did not find upregulation of interferon-regulated genes or the rapamycin target protein FKBP12, although our studies did confirm the finding of upregulation of HSPs and thrombospondin 1. This may be attributable in part to different methods used to identify differentially expressed gene products (microarrays versus RDA) and also our focus on differences maintained in cultured VSMCs, which may not identify genes whose expression is a response to local inflammatory cells.
Our results demonstrate that ISS VSMCs exhibit a stable phenotype, intermediate between medial and plaque VSMCs. Such a phenotype may be explained by several mechanisms that are not mutually exclusive. First, ISS VSMCs may arise from migration of medial cells, which may partially dedifferentiate during migration. In contrast, repair of atherosclerotic plaques after angioplasty may be attributable to proliferation of cells from the plaque itself; ISS VSMCs may therefore closely resemble primary plaque VSMCs. However, our observation that ISS genes are expressed in the normal vessel intima also suggests that these VSMCs may arise from the vessel intima, representing a repair phenotype, selected from the normally heterogeneous intimal VSMCs.
In conclusion, we have identified differential gene expression in ISS versus normal medial VSMCs. ISS genes are also overexpressed by primary plaque VSMCs and are markers of the normal arterial intima. Our studies suggest that ISS forms from the intima, either from resident cells representing the normal intima or from the plaque itself.
This study was supported by BHF Grant PG/98045.
Received July 29, 2002; revision accepted September 30, 2002.
- ↵Garratt KN, Edwards WD, Kaufmann UP, Vlietstra RE, Holmes DRJ. Differential histopathology of primary atherosclerotic and restenotic lesions in coronary arteries and saphenous vein bypass grafts: analysis of tissue obtained from 73 patients by directional atherectomy. J Am Coll Cardiol. 1991; 17: 442–448.
- ↵MacLeod DC, Strauss BH, de Jong M, Escaned J, Umans VA, van Suylen RJ, Verkerk A, de Feyter PJ, Serruys PW. Proliferation and extracellular matrix synthesis of smooth muscle cells cultured from human coronary atherosclerotic and restenotic lesions. J Am Coll Cardiol. 1994; 23: 59–65.
- ↵Patel V, Zhang Q-J, Soos M, Siddle K, Weissberg PL, Bennett MR. Defect in insulin-like growth factor 1 signaling underlies increased apoptosis of human atherosclerotic plaque-derived vascular smooth muscle cells. Circ Res. 2001; 88: 895–902.
- ↵Hubank M, Schatz DG. Identifying differences in mRNA expression by representational difference analysis of cDNA. Nucleic Acids Res. 1994; 22: 5640–5648.
- ↵Tyson K, Shanahan C. Use of cDNA representational difference analysis to identify disease-specific genes in human atherosclerotic plaques. In: Baker A, ed. Methods in Molecular Medicine. Vascular Disease: Molecular Biology and Gene Therapy Protocols. Totowa, NJ: Humana Press Inc; 1999: 83–98.
- ↵Bennett M, Lindner V, DeBlois D, Reidy M, Schwartz S. Effect of phosphorothioated oligonucleotides on neointima formation in the rat carotid artery. Arterioscler Thromb. 1996; 17: 2326–2332.
- ↵O’Sullivan M, Scott S, Shapiro L, Bennett M. Aberrant cell cycle proliferation in human coronary in stent stenosis provides a target for specific anti-restenotic therapy. Heart. 2002; 87: 22.Abstract.
- ↵Bennett MR, Macdonald K, Chan SW, Boyle JJ, Weissberg PL. Cooperative interactions between RB and p53 regulate cell proliferation, cell senescence, and apoptosis in human vascular smooth muscle cells from atherosclerotic plaques. Circ Res. 1998; 82: 704–712.
- ↵Adams LD, Geary RL, McManus B, Schwartz SM. A comparison of aorta and vena cava medial message expression by cDNA array analysis identifies a set of 68 consistently differentially expressed genes, all in aortic media. Circ Res. 2000; 87: 623–631.
- ↵Zohlnhofer D, Klein CA, Richter T, Brandl R, Murr A, Nuhrenberg T, Schomig A, Baeuerle PA, Neumann FJ. Gene expression profiling of human stent-induced neointima by cDNA array analysis of microscopic specimens retrieved by helix cutter atherectomy: detection of FK506-binding protein 12 upregulation. Circulation. 2001; 103: 1396–1402.
- ↵Faber BC, Cleutjens KB, Niessen RL, Aarts PL, Boon W, Greenberg AS, Kitslaar PJ, Tordoir JH, Daemen MJ. Identification of genes potentially involved in rupture of human atherosclerotic plaques. Circ Res. 2001; 89: 547–554.
- ↵O’Neill MJ, Sinclair AH. Isolation of rare transcripts by representational difference analysis. Nucleic Acids Res. 1997; 25: 2681–2682.
- ↵Kashiwada K, Nishida W, Hayashi K, Ozawa K, Yamanaka Y, Saga H, Yamashita T, Tohyama M, Shimada S, Sato K, Sobue K. Coordinate expression of alpha-tropomyosin and caldesmon isoforms in association with phenotypic modulation of smooth muscle cells. J Biol Chem. 1997; 272: 15396–15404.
- ↵Shanahan CM, Weissberg PL, Metcalfe JC. Isolation of gene markers of differentiated and proliferating vascular smooth muscle cells. Circ Res. 1993; 73: 193–204.
- ↵Ichii T, Koyama H, Tanaka S, Kim S, Shioi A, Okuno Y, Raines EW, Iwao H, Otani S, Nishizawa Y. Fibrillar collagen specifically regulates human vascular smooth muscle cell genes involved in cellular responses and the pericellular matrix environment. Circ Res. 2001; 88: 460–467.
- ↵Shanahan F, Seghezzi W, Parry D, Mahony D, Lees E. Cyclin E associates with BAF155 and BRG1, components of the mammalian SWI-SNF complex, and alters the ability of BRG1 to induce growth arrest. Mol Cell Biol. 1999; 19: 1460–1469.