Oligonucleotide Microarrays Reveal Regulated Genes Related to Inward Arterial Remodeling Induced by Urokinase Plasminogen ActivatorPlekhanova O.a · Berk B.C.b · Bashtrykov P.a · Brooks A.I.c · Tkachuk V.a, b · Parfyonova Y.a
aCardiology Research Center, Moscow, and bLomonosov Moscow State University, Moscow, Russia; cAab Cardiovascular Research Institute, University Rochester Medical Center, Rochester, N.Y., USA Corresponding Author
Accumulating evidence suggests that urokinase plasminogen activator (uPA) is involved in vascular remodeling and lumen stenosis after angioplasty and stenting. We have shown previously that increased uPA expression greatly promotes neointima formation and inward arterial remodeling after balloon injury. To evaluate the role of inflammation in early mechanisms responsible for inward arterial remodeling induced by uPA and elucidate the mechanisms of remodeling, we characterized changes in the expression profiles of 8,799 genes in injured rat carotid arteries 1 and 4 days after recombinant uPA treatment compared to vehicle. We used a standard model of the balloon catheter injury of the rat carotid followed by periadventitial application to the injured vessel of either uPA dissolved in Pluronic gel, or plain gel. Vessels were harvested and analyzed by immunohistochemistry, morphometry, microarray gene expression profiling and quantitative RT-PCR. Periadventitial application of uPA significantly reduced lumen size and vessel area encompassed by the external elastic lamina at both 1 and 4 days after treatment. Inflammatory cells accumulated in the arterial adventitia at both 1 and 4 days after uPA treatment. On the 4th day, increases in the areas and arterial cell numbers of all arterial layers were found. Among 79 differentially expressed known genes 1 day after uPA application, 12 proinflammatory genes, including TNF-α and TACE, and 15 genes related to mitochondrial metabolism and oxidative stress regulation were identified. Four days after injury in uPA-treated arteries, 3 proinflammatory and 2 oxidation-related genes were differentially expressed. We conclude that uPA likely promotes inward arterial remodeling by regulating oxidative stress and inflammation after arterial injury.
Copyright © 2008 S. Karger AG, Basel
Inward vessel remodeling and intima-media thickening are the fundamental mechanisms responsible for lumen narrowing  in a number of pathological conditions including restenosis after transluminal coronary angioplasty and intraluminal stenting [2, 3]. We and others have demonstrated previously that urokinase plasminogen activator (uPA) is a potent stimulator of neointima growth and unfavorable arterial remodeling [4,5,6,7]. Despite multiple studies focused on the problem of arterial remodeling, the exact mechanisms of uPA effects in the vessel wall are not fully understood.
Vascular inflammation induced by proinflammatory cytokines is one of the key mechanisms in the development of neointimal hyperplasia, restenosis and atherosclerosis [8, 9]. In particular, leukocytes contribute to the generation of reactive oxygen species (ROS) and augment oxidative stress and restenosis . Moreover, monocytes/macrophages represent the main source of cytokines playing the key roles in the progression of vascular pathologies. In our previous studies in the rat model of balloon catheter injury of the carotid artery we found that uPA augmented inflammatory cell accumulation in the arterial adventitia . These data are consistent with findings in uPA knockout mice that uPA mediated infiltration of leukocytes during arterial remodeling and neointima formation after injury .
We hypothesized that the inflammatory mediators could be involved in the effects of urokinase in the injured arterial wall. To evaluate the role of inflammation in early mechanisms responsible for inward arterial remodeling induced by uPA, we characterized changes in gene expression profiles in injured rat carotid arteries after the local application of recombinant uPA (r-uPA).
Materials and Methods
Male Wistar-Kyoto rats (4–5 months old) were obtained from a colony maintained under standard colony conditions including 12-hour light/12-hour dark cycle, 22 ± 2°C temperature and free access to food and water. Their left common carotid artery was subjected to balloon catheter injury using surgical procedures approved by the Cardiology Research Center’s Animal Experimentation Committee.
Following balloon catheter injury, 20 nmol/kg of r-uPA dissolved in Pluronic gel or 500 μl of plain Pluronic gel F-127 (BASF) were applied to the adventitial side of left common carotid arteries . One and four days later, the animals were killed and total RNA was isolated from left and right carotids of both ballooned and sham-operated as well as nonoperated animals (9 rats per group). Changes in gene expression profiles were analyzed using U34A microarray hybridization (Affymetrix, Santa Clara, Calif., USA) and Quantitative RT- PCR (Applied Biosystems).
Briefly, after anesthetizing the rats with ketamine (100 mg/kg body weight), the common carotid arteries were injured with an inflated Fogarty 2 F balloon catheter, the external carotid arteries were ligated, then uPA or vehicle in Pluronic gel were applied around the common carotids as previously described . One or four days after surgery, the animals were perfusion-fixed, the removal of endothelium was confirmed with the solution of Evans blue dye, carotids were harvested, embedded in paraffin , and cross-sections for morphometry (10 μm) and immunohistochemistry (5 μm) were made as described . For microarray technique left and right common carotids were rapidly removed, rinsed with cold saline solution to remove blood, cleaned of extraneous material and frozen in liquid nitrogen.
For morphometry, cross-sections were stained with toluidine blue and analyzed using a Zeiss microscope coupled to a ProgRess-3008 camera (Kontron Elektronik) and a computer with a UTHSCSA ImageTool, version 3.0, image processing and analysis program (San Antonio, Tex., USA) as previously described .
Immunohistochemical staining was performed using either the anti-PCNA monoclonal antibody (4.8 μg/ml), the anti-ED-1 monoclonal antibody (3.3 μg/ml), the anti-α smooth muscle actin monoclonal antibody (3.6 μg/ml) or the mouse control nonimmune IgGs in concentrations coinciding with those of each immune IgG as previously reported .
Total RNA was extracted from homogenized samples using the Qiagen RNAeasy Micro kit. RNA quality was assessed by gel electrophoresis and spectrophotometric analysis prior to cDNA synthesis. For each hybridization experiment, the probe was generated using material pooled from 3 animals; 3 hybridizations were performed per treatment group. The amplification and biotinylation of probes were done according to Affymetrix recommendations for microarray analysis . Affymetrix rat genome U34A oligonucleotide microarrays, representing 8,799 known genes and expressed sequence tags (ESTs), were used (www.affymetrix.com) at the University of Rochester Microarray Core Facility. Hybridization, staining, washing, and scanning were performed as per the manufacturer’s protocol. Streptavidin-phycoerythrin stain (SAPE, Molecular Probes) was used as the fluorescent conjugate to detect hybridized target sequences. The detection and quantitation of target hybridization were performed with a Gene Array Scanner (Hewlett Packard/Affymetrix). All arrays were assessed for ‘array performance’ prior to data analysis. Control transcripts were spiked into the hybridization mixture to control for hybridization efficiency and sensitivity. To evaluate inter-array variability, the Affymetrix software (Microarray Analysis Suite; MAS 5.0) was employed and the array data were normalized by a global scaling approach to a target intensity of 500 across all probe sets . To reduce the errors due to cross-hybridizing probes and image contamination and lower the estimated false detection rate and the number of targets included in the further statistical analyses, we calculated model-based expression indexes (MBEI) for perfect match (PM)-only arrays using DNA-Chip Analyzer (dChip) software package . The invariant set normalization method was applied prior to analysis. The percentage of probe sets called ‘Present’ in all arrays was > 40%. The ‘Array-outlier’ percentage in all arrays did not exceed 5%, indicating the absence of dramatic image contamination or sample hybridization problems of our arrays. We also performed gene filtering to exclude genes that are ‘Absent’ in the majority of the samples, restricting the data analyses to targets called ‘Present’ in more than 20% of arrays. We confirmed that there was negligible inter-array variability (SD less than 2%) of the mean signal (with 5% of signals trimmed from both the high and low ends) across the 4,458 targets that passed the presence/absence filter. After dChip-filtering, we used significance analysis of microarrays (SAM) software as we previously described  to identify genes with statistically significant changes in expression. We also performed statistical comparison analysis using parametric ANOVA or Welch’s approximate t test for non-equal variances (GeneSpring Software; Silicon Genetics, Redwood City, Calif., USA). To obtain highly confident and robust data, we considered significant only genes confirmed by both analyses (SAM and Welsh’s t test) .
The statistical algorithms of Microarray Suite 5.0 were used with the default parameters to generate signals. DNA-Chip Analyzer software package was kindly provided by Drs. Cheng Li and Wing Hung Wong, Computational Biology Lab, Department of Biostatistics, Harvard School of Public Health (www.dchip.org). Data generated by Microarray Suite and DNA-Chip Analyzer were exported to Microsoft Excel for further analysis. SAM runs within Excel. SAM was kindly provided by the Department of Statistics, Stanford University (http://www-stat.stanford.edu/∼tibs/SAM/). GeneSpring 5.0 software was obtained from Silicon Genetics (http://www.silicongenetics.com/).
Quantitative RT-PCR (qPCR) analyses were performed using the ABI Prism 7900HT sequence detection system (Applied Biosystems) . Double-stranded cDNA template preparation and purification were performed with the Ambion MessageAmp aRNA kit. The qPCR primers and Master Mix from RT2 Real-Time Gene Expression Assay kits (SuperArray) were obtained for 6 rat genes: β-actin (V01217), manganese-containing SOD-2 (AA926129), TNF-α-converting enzyme (TACE) (AJ012603), TNF-α (L00981), MMP-2 (U65656), cAMP phosphodiesterase 4B (PDE4B) (M25350). Each reaction contained 2 μl of SYBR Green PCR Master Mix (Applied Biosystems) and 2 μl diluted cDNA (1:100). The PCR consisted of an initial enzyme activation step at 95°C for 15 min, followed by 40 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 30 s.
All results are reported as mean ± SEM. Statistical tests were performed with Jandel SigmaStat. Comparisons were made by the t test or ANOVA for repeated measurements, as appropriate. A value of p < 0.05 was considered statistically significant.
To evaluate the effects of uPA on arterial remodeling after injury, we evaluated the areas, encompassed by the external elastic lamina (EEL), lumen areas, media/lumen ratio (M/L), and intima-media thickness (IMT), important parameters reflecting vessel remodeling after balloon injury to the rat carotid artery. Periadventitial application of uPA progressively decreased both lumen and EEL areas 1 and 4 days after injury (p < 0.05; fig. 1a, b). uPA significantly increased the M/L ratio 1 and 4 days after injury (p < 0.05; fig. 1c) and IMT 4 days after injury (p < 0.05; fig. 1d). These parameters show that the vessel remodeling response induced by uPA leads to lumen narrowing (fig. 1e), a decrease in EEL and an increase in IMT.
|Fig. 1. Bar graphs summarizing the effects of perivascularly applied r-uPA 1 and 4 days after balloon injury on lumen area (a), EEL cross-sectional area (b), M/L (c) and IMT (d) 1 and 4 days after injury to carotid arteries. ‘Gel’ represents control vessels treated only with Pluronic gel/saline. For lumen and EEL areas, all injured vessels were compared with their corresponding right uninjured carotid artery in the same animals and the differences calculated. Results are the means ± SEM of 6 or 7 animals in each group. * p < 0.05 from control. e Histological figures depicting the effects of perivascular administration of r-uPA on the vessel structure 1 and 4 days after balloon catheter injury of the carotid artery. ‘Gel’ represents control vessels which only received Pluronic gel/saline. Sections are stained with toluidine blue; magnification ×63. Scale bar 500 μm.|
To identify the downstream mechanisms underlying uPA-stimulated neointima formation and inward remodeling, in vivo gene expression profiling was performed. The expression of 8,799 genes was analyzed 1 and 4 days after arterial injury followed by treatment with either exogenous recombinant uPA or vehicle (Plain gel/saline). To confirm successful hybridization, we analyzed the expression of 7 housekeeping genes including ribosomal proteins L18, L18a and S9, α-tubulin, thymosin β4, ubiquitin, β-actin, which consistently gave comparably positive signals in all samples including intact artery, sham-operated, operated controls and experimental arteries (fig. 2). Analysis of signal intensities revealed that 1 day after the operation in arteries treated with uPA 150 transcripts (62 known genes, table 1A) differed significantly compared to control (p ≤ 0.05) confirmed by two statistical analyses. Among these genes, 34 (54.8%) were found to be downregulated, and 28 (45.2%) upregulated after uPA treatment compared to control (plain gel/saline). At day 4, 65 mRNAs (31 known genes, table 1B) were significantly different in the uPA group (complete information is available from the authors upon request). Among these genes, 8 (25.8%) were downregulated and 23 (74.2%) were upregulated compared to control.
|Table 1.A Transcription profiles of genes differentially expressed in rat carotid artery 1 day after balloon injury treated with uPA compared to control vessels, which only received Pluronic gel/saline. B Transcription profiles of genes differentially expressed in rat carotid artery 4 days after balloon injury treated with uPA compared to control vessels, which only received|
|Fig. 2. Transcription profiles of highly expressed housekeeping genes in rat carotid arteries. Each column represents a gene expression pattern for one replicate. For each gene, the signal intensity level normalized to the average signals of 7 housekeeping genes is represented by a gray value according to the signal intensity scale at the bottom.|
To evaluate the role of inflammatory mediators in inward arterial remodeling induced by uPA early after arterial injury, we analyzed changes in gene expression profiles in injured arteries after uPA application. Clusters of genes participating in inflammatory responses were found to be activated in injured artery both 1 and 4 days after uPA treatment. Proinflammatory gene patterns may come from infiltrating inflammatory cells such as macrophages and neutrophils (for example CD45 and Nramp 2, table 1A) or lymphocytes (for example CCR10, table 1B) or may be induced in vascular smooth muscle cells (VSMC) (for example interleukin-1β-converting enzyme (ICE, caspase-1) and TACE, table 1A). Caspase-1 enzymatically activating the IL-1β and IL-18 precursors is the founding member of the caspase family of enzymes, which are involved in maturation of cytokines and in initiation and execution of apoptotic processes . Since IL-1β and TNF-α are central mediators in the cytokine network, they may act as the potent activators of cardiovascular cells involved in the pathogenesis of cardiovascular diseases including atherosclerosis and restenosis. The selective upregulation of TACE in uPA-treated vessels compared with vehicle-treated arteries deserves special attention since TNF-α is one of the most potent proinflammatory cytokines secreted by monocytes/macrophages. Moreover, a number of studies demonstrated the role of TNF-α signaling for proliferative arterial diseases and restenosis [18, 19].
Of greatest interest, among genes differentially expressed 1 day after injury and uPA treatment were 13 genes related to mitochondrial metabolism and oxidative stress (table 2). Among these genes, 8 encoding mitochondrial proteins were downregulated, and 5 were upregulated (table 2). At day 4 after injury, the expression of 2 genes with putative relevance to oxidative stress was elevated in uPA-treated arteries (table 2). Thus, microarray gene analysis showed that uPA may cause unfavorable arterial remodeling at least in part through the upregulation of oxidative stress after arterial injury. These data are consistent with our previous findings that uPA can increase ROS production in VSMC and stimulate VSMC growth in a ROS-dependent manner .
|Table 2. Transcription profiles of genes related to mitochondrial metabolism and oxidative stress regulation differentially expressed in rat carotid artery 1 day after balloon injury treated with uPA compared to control vessels, which only received Pluronic gel/saline|
Among genes differentially expressed after uPA treatment we also found several clusters of genes involved in mediating the key processes of vascular remodeling such as cell migration, proliferation and apoptosis, as well as adhesion and extracellular matrix (ECM) remodeling (table 1A, B). At the same time, the significant up-regulation of uPA mRNA (GenBank No. X63434) was observed (p ≤ 0.0005) that may reflect the mechanism of uPA self-regulation as positive feedback, enhancing and elongating the uPA effects in the injured artery. The expression of matrix metalloproteinase-2 (MMP-2) was also up-regulated (table 1B), consistent with our previous findings that perivascular urokinase can increase the content and activity of MMP-2 in the injured artery . Importantly, MMP-2 participates in vascular remodeling not only providing cell migration, but is also associated with the inflammatory response after injury .
In a separate experiment we validated our microarray data using the same samples of aRNA from the carotid arteries. For validation of hybridization signals by qPCR, we selected several genes of interest: TACE, TNF-α, MMP-2, manganese-containing SOD-2, cAMP phosphodiesterase 4B (PDE4B) and β-actin (table 3). Microarray analysis and qPCR showed similar results, indicating the confidence of microarray data.
|Table 3. Gene expression in rat carotid arteries by Microarray and qPCR|
Since the expression of proinflammatory genes may originate from infiltrating macrophages, we evaluated the accumulation of monocytes/macrophages in arterial layers 1 and 4 days after balloon injury and uPA treatment. On the 1st day after uPA application, the frequency of ED1-positive cells in the media did not differ significantly between groups (p > 0.05; table 4), but was increased in the adventitia (p < 0.05; table 4). Four days after the operation, the proportion of ED1-positive cells was significantly increased only in the adventitia of uPA-treated arteries . There were similar trends in the neointima and media, but changes were not statistically significant.
|Table 4. Total cell numbers and percentages of cells expressing PCNA or ED-1 in injured arteries 1 day after treatments|
To further elucidate the processes induced by uPA-regulated genes, we compared the proportion of PCNA-positive VSMC in arterial layers of uPA- and vehicle-treated arteries. On the 1st day after uPA treatment, there was a small but not significant increase in the percentage of PCNA-positive cells in the adventitia without changes in the proportion of PCNA-positive cells in the media (p > 0.05; table 4). Four days after injury, uPA significantly increased the frequency of PCNA-expressing cells in the neointima .
We also evaluated the effects of uPA on cell accumulation in arterial layers. One day after the operation, uPA increased adventitial cell numbers by nearly 20% compared to control arteries (p < 0.05; table 4) without affecting total cell numbers in the media (p > 0.05; table 4). On the 4th day, perivascular uPA significantly increased the numbers of cells in the neointima (155.9 ± 33.4 vs. 84.7 ± 8.6 in control, p < 0.05), media (234.3 ± 57.6 vs. 152.4 ± 12.9 in control, p < 0.05) and adventitia (303.7 ± 69.7 vs. 171.8 ± 13.8 in control arteries, p < 0.05).
In the present study, changes in gene expression profiles after balloon catheter carotid injury and local administration of recombinant uPA were analyzed. Four major results were obtained. First, exogenous uPA significantly reduced the size of the arterial lumen area both 1 and 4 days after the operation. Second, we observed increases in the percentage of monocytes/macrophages and total cell numbers in the adventitia, suggesting a proinflammatory effect of uPA. The increased inflammation in uPA-treated arteries may be at least in part mediated by the group of proinflammatory genes upregulated by urokinase based on microarray analysis and qPCR including such potent cytokine as TNF-α. Third, changes in oxidation-related gene expression suggest that uPA promotes oxidative stress, which is consistent with our previous study showing increased Nox1 and Nox4 expression in response to uPA . Taking into account that oxidative stress may regulate vascular tone and induce vasoconstriction (uPA itself possesses vasoactive properties ), we may hypothesize that the day 1 decrease in lumen and EEL areas after uPA could reflect changes in vascular tone. Fourth, the changes in gene expression were much greater at day 1 than day 4, possibly reflecting the disappearance or inactivation of uPA .
As we previously reported [7, 11], on the 4th day after injury exogenous uPA treatment induced inward arterial remodeling by stimulating several processes such as neointima formation, VSMC accumulation, intima-media thickness and decrease in EEL area. We also observed the accumulation of adventitial monocytes/macrophages. In uPA-treated arteries, microarray study revealed changes in the expression of 3 proinflammatory and 2 oxidation-related genes, indicating that these systems still play a role in uPA-induced changes 4 days after injury. We may suppose a trigger mechanism of action of uPA in the injured artery as soon as its effects on gene expression develop longer than exogenous uPA is detected there .
Our data on inflammatory cell accumulation after uPA treatment both 1 and 4 days after injury are consistent with previous studies in uPA knockout mice which showed that uPA mediates infiltration of leukocytes during arterial remodeling and neointima formation after injury . Inflammation is the primary response to vessel wall injury during angioplasty and stenting. The important role for leukocyte accumulation is postulated for neointima formation and restenosis . It was shown that inhibition of theinflammatory response in the adventitia limited the developmentof arterial constrictive remodeling by reducing adventitialconstrictive fibrosis . Moreover, macrophages express multiple proteases, including metalloproteinases and urokinase, that degrade the extracellular matrix and secrete numerous other effectors including ROS, TNF-α and interleukin-1 . The present findings indicate that monocyte/macrophage accumulation in the adventitia is one of the earliest effects of uPA in the injured vessel wall, which may represent an important step in the effects of urokinase effects on inward arterial remodeling.
The microarray results suggest mechanisms of action of uPA leading to dramatic arterial remodeling. Of great interest are the significant increases in TNF-α and its converting enzyme (TACE) compared with control arteries. TNF-α has pleiotropic functions in cardiovascular diseases . The periprocedural release of plaque-derived TNF-α was supposed to represent the amount and activity of the atherosclerotic process and to be a predictor for restenosis . The upregulation of TNF-α is a potent mechanism for leukocyte recruitment to the injured region . Moreover, TNF-α stimulates VSMC migration, one of the key processes of vascular stenotic lesion formation . It was also shown that TNF-α may stimulate proliferation of VSMC  and enhance VSMC proliferation induced by monocytes , which is important for neointima growth and vessel remodeling. Finally, TNF-α was shown to be a key transcriptional inducer of several MMPs [32, 33], active participants of arterial response to intravascular injury promoting VSMC migration and replication [34, 35].
Our microarray study revealed changes in oxidation-related gene expression after r-uPA application compared with control arteries. The potential ability of uPA to regulate oxidative stress after arterial injury may represent one of the important mechanisms of inward remodeling of uPA-treated arteries. The oxidative stress induced by injury contributes to unfavorable vessel remodeling and plays an important role in a number of cardiovascular pathologies [36, 37]. ROS are important signaling molecules regulating VSMC proliferation and tone . We have previously shown increased superoxide production after balloon injury and demonstrated the beneficial effects of inhibiting ROS on vascular remodeling . Previous studies showed that uPA contributes to superoxide production  and release  by neutrophils. Our microarray data are consistent with our recent findings that uPA increases ROS generation by VSMC . Taken together, these data suggest that uPA has the potential ability to regulate oxidative stress in the injured artery that may promote inward arterial remodeling.
The cross-interaction between inflammation-related and oxidation-related gene patterns may occur in the injured artery, for example, TNF-α is able to cause oxidative stress in a variety of cell types  and to increase mitochondrial ROS . In a recent study, a model for an interaction between ROS signaling and pathways activated by upregulation of the TACE/TNF-α system leading to vascular remodeling was proposed .
In conclusion, the major findings of the present study are that the early mechanisms leading to inward arterial remodeling induced by uPA in the balloon injury model involve the increased recruitment of leukocytes, early upregulation of the potent proinflammatory TACE/TNF-α system and changes in oxidation-related gene expression. Our results provide mechanistic insights into the role of uPA in remodeling, and provide a basis for combination therapy targeting both uPA and inflammatory mediators to prevent unfavorable arterial remodeling.
This study was supported by National Institutes of Health grant HL-62836 to B.C.B., by NHLBI funds for exchanges of scientists under the auspices of the US-Russia Joint Agreement in Cardiopulmonary Research, by NATO LST. CLG.979209, by Russian Basic Research Foundation grants No. 06-04-49303-а, No. 05-04-49108-а, No. 06-04-08138-ofi and state contracts with BSU No. 73/2006 and No. 60/2008. The authors are grateful for assistance with research of Drs. Oleg Viborov, Mikhail Menshikov, Marina Solomatina, Victoria Stepanova and Sergey Prokhorov.
Prof. Yelena Parfyonova
Molecular Endocrinology Laboratory
Institute of Experimental Cardiology
Cardiology Research Center, Moscow 121552 (Russia)
Tel. +7 495 414 67 13, Fax +7 495 414 67 12 (19), E-Mail firstname.lastname@example.org
Received: January 13, 2008
Accepted after revision: March 18, 2008
Published online: September 24, 2008
Number of Print Pages : 11
Number of Figures : 2, Number of Tables : 4, Number of References : 44
Journal of Vascular Research (Incorporating 'International Journal of Microcirculation')
Vol. 46, No. 3, Year 2009 (Cover Date: April 2009)
Journal Editor: Pohl U. (Munich), Meininger G.A. (Columbia, Mo.)
ISSN: 1018–1172 (Print), eISSN: 1423–0135 (Online)
For additional information: http://www.karger.com/JVR