Gene Expression in Macrophage-Rich Inflammatory Cell Infiltrates in Human Atherosclerotic Lesions as Studied by Laser Microdissection and DNA Array
Overexpression of HMG-CoA Reductase, Colony Stimulating Factor Receptors, CD11A/CD18 Integrins, and Interleukin Receptors
Objective— Inflammatory cells play an important role in atherogenesis. However, more information is needed about their gene expression profiles in human lesions.
Methods and Results— We used laser microdissection (LMD) to isolate macrophage-rich shoulder areas from human lesions. Gene expression profiles in isolated cells were analyzed by cDNA array and compared with expression patterns in normal intima and THP-1 macrophages. Upregulation of 72 genes was detected with LMD and included 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, interferon regulatory factor-5 (IRF-5), colony stimulating factor (CSF) receptors, CD11a/CD18 integrins, interleukin receptors, CD43, calmodulin, nitric oxide synthase (NOS), and extracellular superoxide dismutase (SOD). Several of these changes were also present in PMA-stimulated THP-1 macrophages in vitro. On the other hand, expression of several genes, such as VEGF, tissue factor pathway inhibitor 2, and apolipoproteins C-I and C-II, decreased.
Conclusions— Overexpression of HMG-CoA reductase in macrophage-rich lesion areas may explain some beneficial effects of statins, which can also modulate increased expression of CD11a/CD18 and CD43 found in microdissected cells. We also found increased expression of CSF receptors, IRF-5, and interleukin receptors, which could become useful therapeutic targets for the treatment of atherosclerotic diseases.
Macrophages are important for the development of atherosclerotic lesions. cDNA array analysis has become an important tool for studying gene expression patterns because it enables simultaneous analysis of hundreds of genes and can therefore be used to study molecular events during the disease process.1 Gene expression patterns during foam cell formation have been studied in cell culture2 and atherosclerotic events in animal models3 as well as in human samples.4–6 However, cell culture and animal models may not fully reflect molecular events in human lesions. On the other hand, when analyzing entire atherosclerotic vessels, the heterogeneity of arterial tissue, which includes endothelia, smooth muscle cells (SMCs), macrophages, T cells, mast cells, and fibroblasts, makes gene expression patterns difficult to interpret. Recently, laser microdissection (LMD) techniques have enabled isolation of relatively pure cell populations from histological sections.7 Trogan et al8 used LMD to dissect foam cells from apoE-deficient mice to study macrophage-specific genes, but no studies of human lesions using LMD combined with DNA array have been reported.
We have used LMD to isolate macrophage-rich shoulder areas from human atherosclerotic lesions and compared their gene expression patterns to macroscopically normal diffuse intimal thickening (DIT) by a cDNA array. Expression of several genes, such as 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, was increased in lesion macrophages. The results may help in the identification and validation of treatment targets for cardiovascular diseases.
Materials and Methods
Fresh arterial samples (n=13) were obtained from organ donors (n=10) and fast autopsies (n=3) (Table 1). Some patients had received antihypertensive treatment, but none were on statin therapy. Macroscopically normal and atherosclerotic areas were dissected. Samples were divided into two parts: one part was immersion-fixed in 4% paraformaldehyde/15% sucrose (pH 7.4) and embedded in paraffin;9 the other part was frozen in liquid nitrogen and stored for RNA analysis or was frozen in isopenthane and embedded in OCT compound. According to histology, the samples were classified as normal (n=5) and atherosclerotic lesions (n=8).10 All studies were approved by the Ethics Committee of the Kuopio University Hospital.
Histology and Laser Microdissection
Paraffin sections were used for immunostainings for macrophages (CD68), T cells (CD3), endothelia (CD31), and SMCs (HHF35; all monoclonal antibodies from Dako Co.).11 Controls included incubations where primary antibodies were omitted and incubations with class- and species-matched irrelevant immunoglobulins.11
For LMD, 7-μm-thick OCT sections were used. Sections were fixed in 70% ethanol and counterstained with hematoxylin/eosin and dehydrated in graded ethanol and xylene. Macrophage-rich shoulder areas, adjacent lesion cap areas containing no inflammatory cells, and areas of macroscopically normal DIT were dissected (Leica DM LMD system, Leica) from the same individuals.12
Human monocytic THP-1 cells (ATCC TIB-202) were cultured in RPMI1640 medium according to ATCC instructions. Cells were stimulated with 0.1 μmol/L PMA (Sigma) to induce differentiation into macrophages and collected at indicated time points. THP-1 cells were also similarly activated with PMA and incubated with 100 μg/mL oxidized LDL (Ox-LDL) for 72 hours.13
Total RNA was isolated from LMD samples using the Strataprep total RNA microprep kit (Stratagene) and was extracted from cultured cells using Trizol reagent (Gibco BRL). RNA was pooled from two independent cell culture experiments at each time point.
DNA Array Analysis
RNA from LMD cells was quantified with the RiboGreen RNA quantitation kit (Molecular Probes). 100 ng RNA was used for amplification. Amplification was done with T7 polymerase using three cycles as described elsewhere.14 Probes were prepared from 2.0 μg of amplified RNA, random hexamers, and Cy3-dCTP or Cy5-dCTP. The probes were hybridized as described on the Wellcome Trust Sanger Institute web site (http://www.sanger.ac.uk) to microarrays (Sanger center Hver1.2.1) containing approximately 10000 clones from the Integrated Molecular Analysis of Genomes and their Expression (I.M.A.G.E.) Consortium collection, representing approximately 6000 different genes. 50 μg of total RNA from cultured cells was labeled with Cy3-dCTP or Cy5-dCTP. All analyses were repeated three times, including the use of reversed fluorophores. Quantification of signals was performed with QuantArray software (GSI Lumonics), and the background was subtracted from the signal intensities. The linearity of the signal intensities was checked using GeneSpring software (Silicon Genetics) with M-A plot, where M represents the log ratio of the test and control samples and the A is the total log intensity of each spot. Because the data were linear (ie, blot displayed 45° rotation), normalization to the median of the signals was allowed, according to the protocol of the Finnish Center for Scientific Calculation.15 Signals were normalized using GeneSpring software by dividing each gene by the median of its measurements in all samples, and intensity ratios were calculated. Gene expression was considered significantly up- or downregulated when the intensity ratio test/control was ≥2.0 or ≤0.5, respectively. Statistical significance of the differences was calculated according to Claverie.16 To reach the 5% significance level ‖Inttest-Intcontrol‖ must be ≥2.8ςcontrol when analyzed in triplicates, where Inttest and Intcontrol are the averages of normalized signal intensities in test and control samples in repeated experiments and ς is the distribution of signal intensities in the control sample.
Reverse Transcription-Polymerase Chain Reaction (RT-PCR)
DNAase treatment and cDNA synthesis were performed as described.17 Primers for HMG-CoA reductase were 5′-ACAGGATGCAGCACA GAATG-3′ and 5′-CTTTGCATGCTCCTTGAACA-3′. The PCR reaction consisted of 25 cycles of +94°C for 1 minute, +60°C for 1 minute, and +72°C for 90 seconds. Similarly, RT-PCRs were performed for β-actin (5′-CTCTTCCAGCCTTCCTTCCT-3′ and 5-TCTGCTTGCTGATCCACATC-3′), interferon regulatory factor-5 (IRF-5) (5′-TCTTCTGCTTTGGGGAAGAA-3′ and 5′-GATGGAG CTCCTTGAATTGC-3′), VEGF (5′-CCCACTGAGGAGTCCAAC AT-3′ and 5′-GCGAGTCTGTGTTTTTGCAG-3′), nitric oxide synthase (NOS) (5′-GCTCTCCGTGTAGCCTGAAG-3′ and 5′-ACCGCAAGAATGATGTCTCC-3′) and extracellular superoxide dismutase (GC-SOD) (5′-GAGCTTCTCCTCTGCTCCAA-3′ and 5′-GGGAAGATCGTCAGGTCAAA-3′).
Expression of Genes Typical for Macrophages in LMD Lesion Cells
All arterial samples were immunocytochemically characterized for lesion type and the presence of SMCs, macrophages, endothelia, and T cells.11 Typical examples of arterial samples are shown in Figure 1A through 1F. LMD was used to dissect macrophage-rich shoulder areas of the advanced lesions and macroscopically normal DIT from the same patients (Table 1). Examples of the LMD are shown in Figure 1G through 1J. The number of T cells in the dissected shoulder areas was less than 1 to 2%. SMCs represented <10% of the total cell population. The prevalence of foam cells was small (<5%) in the LMD shoulder fractions. LMD DIT samples were almost exclusively composed of SMCs with only sporadic macrophages. Gene expression patterns of the LMD samples were analyzed by cDNA array after the RNA was amplified. To validate the analysis we first looked at genes typical for macrophages. Array analysis detected upregulation of 33 genes that were previously connected to macrophages (Table 2). These genes included cathepsin G; several colony stimulating factor (CSF) and interleukin receptors; iNOS; a receptor for macrophage stimulating 1 receptor (c-met-related tyrosine kinase); RAB33A, which is a GTP binding protein restricted to monocytic lineage; a toll-like receptor that plays a role in interleukin signaling; integrins CD11a/CD18; and sialophorin (CD43). The expression of endothelial cell specific genes such as cadherin 5, endothelin 1, endoglin, and endothelial TEK tyrosine kinase, or T-cell specific genes, such as CD3, CD8, CD6, and CD27 antigens, were not significantly changed between the analyzed samples; the ratios varied from 0.78 to 1.2.
Comparison of Gene Expression in LMD Lesion Cells and Cultured Macrophages
Table 3⇓ presents 62 genes that had altered expression levels in LMD samples and were not previously connected to atherosclerosis as primary products of macrophages. Genes were classified into functional classes as described previously18 and compared with PMA-treated THP-1 macrophages. Table 3⇓ lists genes where changes (either ≥2.0-fold or ≤0.5-fold change) in gene expression profiles were seen in both lesion LMD cells and in PMA-stimulated THP-1 macrophages. Among these were several genes that have a role in signal transduction, for example, serine/threonine protein kinase, calmodulin and adenylate cyclase 7, cytochromes, calponins, and extracellular SOD. Interestingly, HMG-CoA reductase, which is involved in cholesterol biosynthesis, was highly upregulated in lesion macrophage-rich areas. In addition, several KIAA proteins that had unknown functions were up- or downregulated. The expression of HMG-CoA reductase was confirmed by RT-PCR in LMD fractions and PMA-stimulated macrophages (Figure 2). However, HMG-CoA reductase was not induced in Ox-LDL-treated, lipid-loaded THP-1 cells. In addition, the induction of IRF-5, NOS, and extracellular SOD and the downregulation of VEGF were confirmed by RT-PCR (Figure 2).
We used LMD and gene expression profiling for the analysis of macrophage-rich areas in human atherosclerotic lesions. Previously, analysis of gene expression in tissue samples has been mostly based on in situ hybridization and immunocytochemistry19 because it has been difficult to isolate pure cell populations from histological sections. LMD enables the isolation of relatively pure cell populations from well characterized tissue samples.7,12 It is important to note that these macrophage-rich areas also contained some T cells (<1% to 2%), and it is possible that some genes are derived from T cells. However, it can still be concluded that macrophage-rich areas were successfully isolated since (1) several genes typical for macrophages were identified, (2) T-cell content was low, (3) expression changes in endothelial cell and T-cell-specific genes were not significant, and (4) similar changes were found in vitro in PMA-stimulated THP-1 macrophages. When interpreting the results, it should be remembered that DNA arrays do not necessarily reflect regulation of transcription but tissue levels of mRNA (ie, transcriptosome or expression phenotype) at the time of tissue collection, whereas in cell cultures it is possible to analyze changes in expression profiles with time.
Upregulation of a total of 72 genes was identified; 33 genes were previously connected to atherosclerosis and 39 were not. Twenty-three genes were downregulated. LMD results are based on the amplification of a small amount of RNA from individual samples, followed by DNA array analysis and standardization of the gene expression ratios using the GeneSpring program. Every effort was made to avoid any systematic bias in the results. However, we cannot totally exclude the possibility of some bias or that some important genes were not picked up by the analysis.
Upregulation of several important genes was detected. One of these was HMG-CoA reductase, which is a key enzyme in cholesterol biosynthesis.20 It was highly upregulated in LMD samples. The reason for the high expression level may be due to the proliferation and/or differentiation of macrophages because similar findings were obtained from PMA-stimulated THP-1 macrophages. Interestingly, it has been reported that HMG-CoA reductase in macrophages is not very sensitive to downregulation by cholesterol derived from LDL,21 which together with the low prevalence of lipid-loaded foam cells (<5%) in the analyzed LMD fractions may contribute to these findings. On the other hand, when THP-1 macrophages were incubated in the presence of Ox-LDL, downregulation of the HMG-CoA reductase was observed, which indicates that lipid-loaded THP-1 macrophages may not be an accurate model of macrophages present in shoulder areas. However, further studies are needed to clarify these issues.
It has been suggested that HMG-CoA reductase inhibitors, statins, have several beneficial effects that are not directly related to their cholesterol-lowering effects.22 Even though this hypothesis of the pleiotrophic effects of statins has not yet been fully tested, it is possible that macrophage functions in lesion shoulder areas may be attenuated by statins because of the high induction level of HMG-CoA reductase in these cells. HMG-CoA reductase also plays a role in the regulation of cell surface integrins CD11a and CD18, which bind to intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) on endothelial cells. These integrins upregulated in LMD cells, and statins have been shown to downregulate CD11a and CD18 in cell culture.23 In addition, HMG-CoA reductase also functions in cell movements mediated by CD43,24 which was upregulated in LMD cells. Thus, DNA array analysis suggests that statin therapy may have favorable effects through these mechanisms, which should reduce monocyte-macrophage activity and trafficking in atherosclerotic lesions.
Other upregulated genes included CSF receptors, interleukin receptors, calmodulin, iNOS, extracellular SOD, and novel p53 target gene IRF-5. We have also previously demonstrated with DNA array analysis that some well known macrophage genes, such as apoE, TIMPs and VEGF receptor-2, were upregulated in lesion macrophages.6 Activation of these genes indicates that growth factor activities, inflammatory mediators, and oxidative stress play important roles in the microenvironment of lesion macrophages. On the other hand, expression of genes, such as VEGF, tissue factor pathway inhibitor 2, and apolipoproteins C1 and C2, were decreased, indicating less efficient vascular maintenance and increased susceptibility to thrombosis. It is important to note that genes associated with increased lipid uptake, such as scavenger receptors (1.7-fold induction), were not significantly upregulated in the LMD samples.
It is concluded that LMD can be used to dissect small cell populations from atherosclerotic lesions and that these cells can be used for DNA array analysis. Results showed that, among other genes, HMG-CoA reductase was highly upregulated in lesion macrophage-rich areas, which may explain some beneficial effects of statins not related to their lipid-lowering activity. Also, additional potential targets for therapy of cardiovascular diseases were identified.
We thank Seija Sahrio, Anneli Miettinen, and Marja Poikolainen for technical assistance and the staff at the Kuopio University Hospital for the collection of human samples. This study was supported by grants from Finnish Academy, Sigrid Juselius Foundation, and Finnish Foundation for Cardiovascular Research.
- Received August 4, 2003.
- Accepted October 6, 2003.
Shiffman D, Mikita T, Tai JT, Wade DP, Porter JG, Seilhamer JJ, Somogyi R, Liang S, Lawn RM. Large scale gene expression analysis of cholesterol-loaded macrophages. J Biol Chem. 2000; 275: 37324–37332.
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.
Trogan E, Choudhury RP, Dansky HM, Rong JX, Breslow JL, Fisher EA. Laser capture microdissection analysis of gene expression in macrophages from atherosclerotic lesions of apolipoprotein E-deficient mice. Proc Natl Acad Sci U S A. 2002; 99: 2234–2239.
Ylä-Herttuala S, Rosenfeld ME, Parthasarathy S, Glass CK, Sigal E, Witztum JL, Steinberg D. Colocalization of 15-lipoxygenase mRNA and protein with epitopes of oxidized low density lipoprotein in macrophage-rich areas of atherosclerotic lesions. Proc Natl Acad Sci U S A. 1990; 87: 6959–6963.
Stary HC, Chandler AB, Dinsmore RE, Fuster V, Glagov S, Insull W, Jr., Rosenfeld ME, Schwartz CJ, Wagner WD, Wissler RW. A definition of advanced types of atherosclerotic lesions and a histological classification of atherosclerosis. A report from the Committee on Vascular Lesions of the Council on Arteriosclerosis, Am Heart Association. Arterioscler Thromb Vasc Biol. 1995; 15: 1512–1531.
Hakkinen T, Karkola K, Yla-Herttuala S. Macrophages, smooth muscle cells, endothelial cells, and T-cells express CD40 and CD40L in fatty streaks and more advanced human atherosclerotic lesions. Colocalization with epitopes of oxidized low-density lipoprotein, scavenger receptor, and CD16 (Fc gammaRIII). Virchows Arch. 2000; 437: 396–405.
Laurema A, Heikkila A, Keski-Nisula L, Heikura T, Lehtolainen P, Manninen H, Tuomisto TT, Heinonen S, Yla-Herttuala S. Transfection of oocytes and other types of ovarian cells in rabbits after direct injection into uterine arteries of adenoviruses and plasmid/liposomes. Gene Ther. 2003; 10 (7): 580–584.
Laukkanen J, Lehtolainen P, Gough PJ, Greaves DR, Gordon S, Ylä-Herttuala S. Adenovirus-mediated gene transfer of a secreted form of human macrophage scavenger receptor inhibits modified low-density lipoprotein degradation and foam-cell formation in macrophages. Circulation. 2000; 101: 1091–1096.
Tuimala J, Laine MM, eds. DNA Microarray Data Analysis. Espoo, Finland: Finnish IT Center for Science; 2003.
Claverie JM. Computational methods for the identification of differential and coordinated gene expression. Hum Mol Genet. 1999; 8: 1821–1832.
Vajanto I, Rissanen TT, Rutanen J, Hiltunen MO, Tuomisto TT, Arve K, Narvanen O, Manninen H, Rasanen H, Hippelainen M, Alhava E, Yla-Herttuala S. Evaluation of angiogenesis and side effects in ischemic rabbit hindlimbs after intramuscular injection of adenoviral vectors encoding VEGF and LacZ. J Gene Med. 2002; 4: 371–380.
Hwang DM, Dempsey AA, Wang RX, Rezvani M, Barrans JD, Dai KS, Wang HY, Ma H, Cukerman E, Liu YQ, Gu JR, Zhang JH, Tsui SK, Waye MM, Fung KP, Lee CY, Liew CC. A genome-based resource for molecular cardiovascular medicine: toward a compendium of cardiovascular genes. Circulation. 1997; 96: 4146–4203.
Luoma JS, Strålin P, Marklund SL, Hiltunen TP, Särkioja T, Ylä-Herttuala S. Expression of extracellular SOD and iNOS in macrophages and smooth muscle cells in human and rabbit atherosclerotic lesions: Colocalization with epitopes characteristic of oxidized LDL and peroxynitrite-modified proteins. Arterioscler Thromb Vasc Biol. 1998; 18: 157–167.
Palinski W, Napoli C. Unraveling pleiotropic effects of statins on plaque rupture. Arterioscler Thromb Vasc Biol. 2002; 22 (11): 1745–1750.
Yoshida M, Sawada T, Ishii H, Gerszten RE, Rosenzweig A, Gimbrone MAJ, Yasukochi Y, Numano F. Hmg-CoA reductase inhibitor modulates monocyte-endothelial cell interaction under physiological flow conditions in vitro: involvement of Rho GTPase-dependent mechanism. Arterioscler Thromb Vasc Biol. 2001; 21 (7): 1165–1171.
Gubina E, Chen T, Zhang L, Lizzio EF, Kozlowski S. CD43 polarization in unprimed T cells can be dissociated from raft coalescence by inhibition of HMG CoA reductase. Blood. 2002; 99: 2518–2525.