Transcriptional and Posttranscriptional Regulation of Cyclooxygenase-2 Expression by Fluid Shear Stress in Vascular Endothelial Cells
Objective— Fluid shear stress induces cyclooxygenase (COX)-2 gene expression in vascular endothelial cells. We investigated the underlying mechanism of this induction.
Methods and Results— Exposure of human umbilical vein endothelial cells to laminar shear stress in the physiological range (1 to 30 dyne/cm2) upregulated the expression of COX-2 but not COX-1, a constitutive isozyme of COX. The expression of COX-2 mRNA began to increase within 0.5 hour after the loading of shear stress and reached a maximal level at 4 hours. Roles of the promoter region and the 3′-untranslated region in the human COX-2 gene were evaluated by the transient transfection of luciferase reporter vectors into bovine arterial endothelial cells. Shear stress elevated luciferase activity via the region between −327 and 59 bp. Mutation analysis indicated that cAMP-responsive element (−59/−53 bp) was mainly involved in this response. On the other hand, shear stress selectively stabilized COX-2 mRNA. Moreover, shear stress elevated luciferase activity when a 3′-untranslated region of COX-2 gene containing 17 copies of the AUUUA mRNA instability motif was inserted into the vector.
Conclusions— Transcriptional activation and posttranscriptional mRNA stabilization contribute to the rapid and sustained expression of COX-2 in response to shear stress.
Vascular endothelial cells are always exposed to a wide variety of biochemical and biomechanical stimuli, including fluid shear stress caused by blood flow. Shear stress modulates several endothelial functions, such as control of vascular tone, maintenance of antithrombotic surfaces, regulation of inflammation, protection against oxidative stresses, and regulation of endothelial cell proliferation and apoptosis.1
Cyclooxygenase (COX), a rate-limiting enzyme for prostaglandin (PG) biosynthesis, comprises 2 isozymes, COX-1 and COX-2.2,3⇓ COX-1 is constitutively expressed in most cell species, whereas COX-2 is an inducible enzyme whose expression is regulated differently among cell types. Growing evidence indicates that COX-2 plays a key role in several biological processes, such as inflammation, tumorigenesis, development, and atherogenesis.4–9⇓⇓⇓⇓⇓ Laminar shear stress upregulates COX-2 gene expression.10 COX-2 is involved in lipopolysaccharide-stimulated production of prostacyclin (PGI2) in endothelial cells11 and is also involved in PGI2 biosynthesis in healthy humans.12 Previously, we have reported that shear stress promotes the production of PGD2 in endothelial cells by stimulating the expression of lipocalin-type PGD2 synthase (L-PGDS), whereas PGI2 synthase was constitutively expressed but did not respond to shear stress.13 Therefore, the induction of COX-2 expression by shear stress may be involved in the production of PGI2 and PGD2 in endothelial cells.
Three cis-acting elements, namely, the nuclear factor (NF)-κB binding site, the NF–interleukin-6 (NF-IL6) binding site, and the cAMP-responsive element (CRE), reside in the region between base pairs −327 and +59 (−327/+59) in the human COX-2 gene promoter. Their involvement in COX-2 gene transcription varies among cell species.14–24⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓ Recently, the COX-2 gene has been reported to be posttranscriptionally regulated through its 3′-untranslated region (3′-UTR) containing 17 copies of the AUUUA motif, which is assumed to promote mRNA degradation.25–30⇓⇓⇓⇓⇓ However, the mechanism underlying shear stress–induced COX-2 gene expression remains to be elucidated, although it has been recently reported that shear stress stimulates the transcription of the COX-2 gene in murine osteoblastic MC3T3-E1 cells that produce PGE2 but not PGI2 or PGD2.24
In the present study, we investigated the molecular mechanism for the shear stress–induced expression of COX-2 in vascular endothelial cells. The gene expression of COX-2 was more sensitive to shear strength than that of L-PGDS. We found that shear stress induces COX-2 expression not only at the transcriptional level but also at the posttranscriptional level through the 3′-UTR, which would make it possible to rapidly and persistently induce COX-2 expression in response to shear stress.
Human umbilical vein endothelial cells (HUVECs) were isolated and cultured as described.31 Bovine arterial endothelial cells (BAECs) were grown in DMEM supplemented with 10% FCS (Flow), 100 U/mL penicillin, and 100 μg/mL streptomycin.16
Shear Stress Apparatus
HUVECs and BAECs were plated on a gelatin-coated polyester sheet (Plastic Suppliers), and flow experiments were performed in a parallel-plate flow chamber as described.32
Western Blot Analysis and Electrophoretic Mobility Shift Assay
Total RNA was isolated by using the acid guanidinium thiocyanate procedure. RNAs were then subjected to electrophoresis. The cDNA probes for COX-1, COX-2, and GAPDH have been described previously.16 The levels of mRNA and 28S rRNA were calculated on the basis of hybridization signals and ethidium bromide–staining intensities as measured with the imaging analyzers Fujix BAS 2500 and FLA 2000, respectively (Fuji Photo Film Co).
The following control vectors were used: for luciferase (Luc), pGV-C (Toyo); for β-galactosidase (β-gal), pCMV-βgal; and for green fluorescent protein (GFP), pEGFP-N1 (Clontech).11 The human COX-2 genomic clone λhPESII95 was digested by EcoRI, and then these fragments were subcloned into pBluescript II SK(+). Digestion of the subclone 7k1-2 with MroI and EcoRI yielded a fragment containing part of the coding region (57 bp) and the full-length 3′-UTR, which contains 17 copies of the ATTTA motif followed by 3 copies of the polyadenylation signal (AATAAA).33 This fragment was blunted and ligated into the blunted–PflM I site of pGV-C, which is located downstream from the Luc coding region. In this clone, designated pG-3UCOX2, Luc mRNA is expressed under the control of the SV40 enhancer/promoter and the 3′-UTR of the human COX-2 gene. Construction of other COX-2 reporters has been described previously.11,16⇓
Transfection of BAECs with plasmids has been described previously.11 In this experiment, 200 μL DMEM containing 2 μg COX-2 Luc reporter vector, 0.2 μg pCMV-βgal, and 0.1 μg pEGFP-N1 was mixed with 200 μL DMEM containing 10 μL Trans IT-LT-1 (Pan Vera) and incubated at room temperature for 15 minutes. This DNA/reagent complex was added to semiconfluent BAECs growing in a 90-mm dish containing 8 mL complete growth medium that was changed 1 day before the transfection. After 5 hours of transfection, cells were incubated with new complete growth medium for 19 hours and placed in the flow shear stress chamber. After incubation in the chamber for 24 hours, cells were subjected to fluid shear stress for 5 or 17 hours by using the apparatus described above. The cells were lysed in Reporter lysis buffer (Promega) to release the Luc and β-gal for their activity assays. The numerical readings from the Luc assay were normalized to those of the β-gal assay.16
Results are expressed as mean±SD. Statistical significance was assessed by the Student t test.
Laminar Shear Stress Induces COX-2 Expression in HUVECs
To determine the effect of shear stress on the expression of COX genes, we performed Western and Northern blot analyses with the use of protein and RNA extracted from HUVECs. COX-2 protein was expressed after the loading of shear stress (15 dyne/cm2) for 24 hours but not in cells cultured under static conditions (please see online Figure IA, available at http://atvb.ahajournals.org). As shown in Figure 1A, the expression levels of COX-2 mRNA were very low under static conditions. Exposure of the cells to shear stress (15 dyne/cm2) for the periods indicated led to an increase in the amount of COX-2 mRNA. On the other hand, COX-1 mRNA was constitutively expressed independently of shear stress. Because the expression level of COX-2 mRNA reached an almost maximal value at 6 hours, we further examined the time course of COX-2 induction immediately after loading shear stress (Figure 1B and online Figure IB). The level of COX-2 mRNA began to increase within 0.5 hours and reached a plateau at 4 hours.
Figure 2 shows the dependence of the expression of COX-2 mRNA on the strength of shear stress. The expression of COX-2 mRNA was maximally upregulated even at 5 dyne/cm2, which corresponds to the level of shear stress in veins. Therefore, we performed an additional experiment to determine whether shear stress <5 dyne/cm2 is able to stimulate COX-2 expression. As shown in Figure 2A, shear stress as low as 1 dyne/cm2 significantly upregulated the expression.
Laminar Shear Stress Activates COX-2 Promoter
To examine whether shear stress stimulates the human COX-2 promoter, we used phPES2 (−327/+59), a plasmid that expresses firefly Luc under the control of the human COX-2 gene promoter (−327/+59). Because DNA transfection efficiency was much higher in BAECs than in HUVECs, phPES2 (−327/+59) was transfected into BAECs. Online Figure IIA (available at http://atvb.ahajournals.org) demonstrates BAECs transfected with pEGFP-N1, a GFP expression plasmid. The transfection efficiency seemed to be sufficient, and shear stress induced cell elongation along with the flow direction, which indicated that shear stress was successfully loaded on the cells. Moreover, that bovine COX-2 mRNA expression is also inducible in response to shear stress was confirmed by Northern blotting, as shown in online Figure IIB. Bovine COX-2 mRNA was induced by shear stress after 5 and 17 hours in duplicate experiments with independent materials. However, bovine COX-1 mRNA was not induced, again similar to human COX-1. As shown in online Figure IIC, the promoter activity of phPES2 (−327/+59) was markedly elevated by shear stress loaded for 5 and 17 hours.
The human COX-2 promoter region (−327/+59) contains 3 cis-acting elements, namely, an NF-κB binding site, an NF-IL6 binding site, and a CRE, all of which have been shown to be involved in the regulation of COX-2 gene transcription. When we destroyed all of the 3 consensus sequences as shown in Figure 3A, the shear stress–induced elevation of the promoter activity completely disappeared, suggesting that the elements responsive to shear stress are in these 3 consensus sites. To identify the elements, we then destroyed each of the 3 motifs as shown in Figure 3B. A mutation in CRE (−59/−53) markedly reduced the basal and shear stress–induced promoter activities. Mutations at the NF-κB site (−223/−214) or NF-IL6 site (−132/−124) did not affect the response to shear stress compared with effects of the CRE, although the basal promoter activity was significantly reduced by the NF-IL6 mutation and tended to be reduced by the NF-κB mutation. Electrophoretic mobility shift assay showed that proteins specifically binding to the CRE (−59/−53) were enhanced by the shear stress (see online Figure III, available at http://atvb.ahajournals.org). These data suggest that the CRE (−59/−53) is mainly involved in the shear stress–increased COX-2 promoter activity.
Laminar Shear Stress Stabilizes COX-2 mRNA Through 3′-UTR
To examine whether posttranscriptional mechanisms are involved in shear stress–induced COX-2 expression, we chased the decay of COX-2 mRNA after the addition of actinomycin D, an inhibitor of transcription. As shown in Figure 4, COX-2 mRNA increased by shear stress loaded for 2 hours was almost completely degraded 2 hours after the addition of actinomycin D in the absence of stress. However, when shear stress was loaded simultaneously with actinomycin D, the decay of COX-2 mRNA was significantly delayed, and the enhanced expression was still retained at 4 hours after the addition of actinomycin D.
The entire 3′-UTR of the human COX-2 gene contains 17 copies of the ATTTA motif, which is found in many immediate-early genes and has been shown to promote mRNA degradation.34,35⇓ To determine whether the 3′-UTR of COX-2 mediates shear stress–induced mRNA stabilization, we examined the effect of the insertion of the COX-2 3′-UTR into the Luc expression vector. As shown in Figure 5, transfection of pG-3UCOX2 resulted in an ≈50% reduction in the promoter activity compared with that of pGV-C in cells cultured under static conditions, suggesting that the 3′-UTR of COX-2 destabilized the Luc mRNA. Shear stress did not influence the Luc activity in cells transfected with pGV-C, whereas it markedly elevated the activity by 2.8-fold in cells transfected with pG-3UCOX2. This suggested that exposure to laminar shear stress prevented Luc mRNA breakdown through the 3′-UTR of the COX-2 gene.
The present study demonstrated for the first time that laminar shear stress within the physiological range induces COX-2 gene expression through transcriptional and posttranscriptional mechanisms in vascular endothelial cells. Shear stress as low as 1 dyne/cm2 was able to induce COX-2 mRNA expression in HUVECs, as in murine osteoblastic MC3T3-E1 cells.24 However, it was not able to stimulate COX-1 mRNA expression in either HUVECs or BAECs. A similar result, ie, that laminar shear stress induced COX-2 but not COX-1 expression, has been reported in a study using HUVECs,10 although higher shear stress (24 dyne/cm2) has been reported to induce COX-1 mRNA in HUVECs,36 and mechanical cyclic strain also has been shown to induce COX-1 in human vascular smooth muscle cells.37 COX-2 has been reported to have a higher affinity for arachidonate than COX-1,38 and moreover, COX-2 preferentially cooperates more with PGI2 synthase than with COX-1.39 When these findings are taken together, COX-2 but not COX-1 appears to be mainly involved in PGI2 formation in blood vessels loaded with laminar shear stress in the physiological range. This assumption is consistent with recent studies reporting that selective COX-2 inhibitors suppress the systemic biosynthesis of PGI2 in healthy humans12 and in patients with atherosclerosis.40 There has been a report that glucocorticoids do not depress the excretion of urinary PGI2 metabolite.41 We found that COX-2 expression was not suppressed by dexamethasone, especially in vascular endothelial cells, because of the lower expression of the glucocorticoid receptor.11 There are variable reports of COX-2 expression ex vivo in endothelial cells. It will be difficult to evaluate the precise expression levels of COX-2 ex vivo after postmortem examination because COX-2 mRNA and protein are very unstable.
We reported that laminar shear stress also induces the expression of L-PGDS.13 However, the induction of gene expression by shear stress was more rapid and sensitive for COX-2 than for L-PGDS, whose mRNA expression increases depending on shear strength between 5 and 30 dyne/cm2. PGI2 synthase is constitutively expressed in endothelial cells, and its expression level is not influenced by shear stress.13 Therefore, we hypothesize that there are 2 steps in COX-2–mediated arachidonate metabolism in endothelial cells. Step 1, which operates under low shear stress conditions, predominantly produces PGI2, and step 2, which operates under high shear stress conditions, produces not only PGI2 but also PGD2. In this context, PGD2 may also play a role in preventing the formation of atherosclerotic lesions by being converted to 15-deoxy-Δ12,14-PGJ2, which has been reported to display several antiatherogenic effects on cultured vascular cells.42
In the mean time, macrophages were reported to express augmented levels of COX-2 in atherosclerotic lesions.43,44⇓ This abnormally elevated COX-2 expression in macrophages may be related to inflammation in the lesions. Because the PGs produced in macrophages are different from those produced in endothelial cells, the regulation of COX-2 and downstream enzymes should be very different between endothelial cells and macrophages. In this context, we have recently reported that COX-2 expression is negatively regulated by nuclear receptor peroxisome proliferator–activated receptor-γ and its ligand candidate 15-deoxy-Δ12,14-PGJ2 in macrophages but not in endothelial cells.23 Whether changes in the pattern of blood flow, such as turbulence, influence the PG species produced in endothelial cells should be examined.
In the present study, mutation analysis and gel-retardation assay indicated that the shear stress responsiveness of the COX-2 promoter was mainly dependent on CRE (−59/−53). The promoter activity of the fragment −1432/+59 was similar to that of the fragment −327/+59 induced by shear stress (data not shown), as reported in our previous studies using different stimuli.11,16⇓ Previous reports have indicated that CRE, CCAAT/enhancer binding protein-β (same as NF-IL6), and activator protein-1 motifs play major roles in the shear stress induction of the COX-2 gene in osteoblasts, which is consistent with our results because of multiple binding activity of the CRE for CRE binding protein, CCAAT/enhancer binding proteins, and activator protein-1.16,20⇓ However, a triple mutation assay in Figure 3A suggested that other cis-acting elements also play a role in the COX-2 promoter activity.
Actinomycin D chase experiments have indicated that shear stress stabilizes COX-2 mRNA levels in HUVECs. Similar experiments have also shown granulocyte-macrophage colony–stimulating factor (GM-CSF) mRNA also to be stabilized by shear stress,45 although there was no induction of the promoter activity of the GM-CSF gene. The COX-2 and GM-CSF genes contain 17 and 8 copies, respectively, of the AUUUA motif in their 3′-UTRs.33,34⇓ Conserved AUUUA sequences have been identified within the 3′-UTR of many short-lived mRNAs and regulate their stability. The 3′-UTR of the COX-2 gene contains multiple copies of the AUUUA motif implicated in posttranscriptional regulation. Several investigators have shown that insertion of the 3′-UTR of the COX-2 gene into the 3′-UTR of the reporter gene alters its expression.25–30⇓⇓⇓⇓⇓
The present study demonstrated for the first time that shear stress stabilizes COX-2 mRNA through the 3′-UTR of the COX-2 gene. Recently, it has been reported that the 3′-UTR of the murine COX-2 gene contains multiple regulatory elements that alter mRNA stability and translational efficiency even in the steady state.30 On the other hand, the AUUUA motif found in GM-CSF mRNA has been considered not to function as a shear stress–responsive element, because protein binding to this motif was not influenced by shear stress.45 Further study is necessary to resolve these issues.
This study was partly supported by grants from the Ministry of Education, Science, Sports, and Culture, Japan (grants-in-aid for scientific research Nos. 11838021 and 13670110); from the Ministry of Health and Welfare, Japan (research grants for cardiovascular diseases Nos. 11C-1 and 12C-3); and from the Mochida Memorial Foundation for Medical and Pharmaceutical Research.
Received March 4, 2002; revision accepted June 20, 2002.
- ↵Dubois RN, Abramson SB, Crofford L, Gupta RA, Simon LS, Van De Putte LB, Lipsky PE. Cyclooxygenase in biology and disease. FASEB J. 1998; 12: 1063–1073.
- ↵Topper JN, Cai J, Falb D, Gimbrone MA Jr. Identification of vascular endothelial genes differentially responsive to fluid mechanical stimuli: cyclooxygenase-2, manganese superoxide dismutase, and endothelial cell nitric oxide synthase are selectively up-regulated by steady laminar shear stress. Proc Natl Acad Sci U S A. 1996; 93: 10417–10422.
- ↵McAdam BF, Catella-Lawson F, Mardini IA, Kapoor S, Lawson JA, FitzGerald GA. Systemic biosynthesis of prostacyclin by cyclooxygenase (COX-2): the human pharmacology of a selective inhibitor of COX-2. Proc Natl Acad Sci U S A. 1999; 96: 272–277.
- ↵Taba Y, Sasaguri T, Miyagi M, Abumiya T, Miwa Y, Ikeda T, Mitsumata M. Fluid shear stress induces lipocalin-type prostaglandin D2 synthase expression in vascular endothelial cells. Circ Res. 2000; 86: 967–973.
- ↵Sirois J, Simmons DL, Richards JS. Hormonal regulation of messenger ribonucleic acid encoding a novel isoform of prostaglandin endoperoxide H synthase in rat preovulatory follicles: induction in vivo and in vitro. J Biol Chem. 1992; 267: 11586–11592.
- ↵Inoue H, Yokoyama C, Hara S, Tone Y, Tanabe T. Transcriptional regulation of human prostaglandin-endoperoxide synthase-2 gene by lipopolysaccharide and phorbol ester in vascular endothelial cells. J Biol Chem. 1995; 270: 24965–24971.
- ↵Yamamoto K, Arakawa T, Ueda N, Yamamoto S. Transcriptional roles of NF-κB and NF-IL-6 in the tumor necrosis factor α-dependent induction of cyclooxygenase-2 in MC3T3-E1 cells. J Biol Chem. 1995; 270: 31315–31320.
- ↵Kim Y, Fischer SM. Transcriptional regulation of cyclooxygenase-2 in mouse skin carcinoma cells: regulatory role of CCAAT/enhancer-binding proteins in the differential expression of cyclooxygenase-2 in normal and neoplastic tissues. J Biol Chem. 1998; 273: 27686–27694.
- ↵Xie W, Herschman HR. v-src induces prostaglandin synthase 2 gene expression by activation of the c-Jun N-terminal kinase and the c-Jun transcription factor. J Biol Chem. 1995; 270: 27622–27628.
- ↵Murono S, Inoue H, Tanabe T, Joab I, Yoshizaki T, Furukawa M, Pagano JS. Induction of cyclooxygenase-2 by Epstein-Barr virus latent membrane protein 1 is involved in vascular endothelial growth factor production in nasopharyngeal carcinoma cells. Proc Natl Acad Sci U S A. 2001; 98: 6905–6910.
- ↵Saunders MA, Sansores-Garcia L, Gilroy DW, Wu KK. Selective suppression of CCAAT/enhancer-binding protein β binding and cyclooxygenase-2 promoter activity by salicylate in quiescent human fibroblast. J Biol Chem. 2001; 276: 18897–18904.
- ↵Inoue H, Tanabe T, Umesono K. Feedback control of cyclooxygenase-2 expression through PPARγ. J Biol Chem. 2000; 275: 28028–28032.
- ↵Ogasawara A, Arakawa T, Kaneda T, Takuma T, Sato T, Kaneko H, Kumegawa M, Hakeda Y. Fluid shear stress-induced cyclooxygenase-2 expression is mediated by C/EBPβ, cAMP-response element-binding protein, and AP-1 in osteoblastic MC3T3-E1 cells. J Biol Chem. 2001; 276: 7048–7054.
- ↵Srivastava SK, Tetsuka T, Daphna-Iken D, Morrison AR. IL-1β stabilized COX-2 mRNA in renal mesangial cells: role of 3′-untranslated region. Am J Physiol. 1994; 267: F504–F508.
- ↵Dixon DA, Kaplan CD, McIntyre TM, Zimmerman GA, Prescott SM. Post-transcriptional control of cyclooxygenase-2 gene expression: the role of the 3′-untranslated region. J Biol Chem. 2000; 275: 11750–11757.
- ↵Xu K, Robida AM, Murphy TJ. Immediate-early MEK-1-dependent stabilization of rat smooth muscle cell cyclooxygenase-2 mRNA by Galpha(q)-coupled receptor signaling. J Biol Chem. 2000; 275: 23012–23019.
- ↵Cok SJ, Morrison AR. The 3′-untranslated region of murine cyclooxygenase-2 contains multiple regulatory elements that alter message stability and translational efficiency. J Biol Chem. 2001; 276: 23179–24185.
- ↵Kosaka C, Sasaguri T, Ishida A, Ogata J. Cell cycle arrest in the G2 phase induced by phorbol ester and diacylglycerol in vascular endothelial cells. Am J Physiol. 1996; 270: C170–C178.
- ↵Akimoto S, Mitsumata M, Sasaguri T, Yoshida Y. Laminar shear stress inhibits vascular endothelial cell proliferation by inducing cyclin-dependent kinase inhibitor p21(Sdi1/Cip1/Waf1). Circ Res. 2000; 86: 185–190.
- ↵Okahara K, Sun B, Kambayashi J. Upregulation of prostacyclin synthesis-related gene expression by shear stress in vascular endothelial cells. Arterioscler Thromb Vasc Biol. 1998; 18: 1922–1926.
- ↵Feng Y, Yang JH, Huang H, Kennedy SP, Turi TG, Thompson JF, Libby P, Lee RT. Transcriptional profile of mechanically induced genes in human vascular smooth muscle cells. Circ Res. 1999; 85: 1118–1123.
- ↵Murakami M, Kambe T, Shimbara S, Kudo I. Functional coupling between various phospholipase A2s and cyclooxygenases in immediate and delayed prostanoid biosynthetic pathways. J Biol Chem. 1999; 274: 3103–3115.
- ↵Ueno N, Murakami M, Tanioka T, Fujimori K, Tanabe T, Urade Y, Kudo I. Coupling between cyclooxygenase, terminal prostanoid synthase, and phospholipase A2. J Biol Chem. 2001; 276: 34918–34927.
- ↵Belton O, Byrne D, Kearney D, Leahy A, Fitzgerald DJ. Cyclooxygenase-1 and -2–dependent prostacyclin formation in patients with atherosclerosis. Circulation. 2000; 102: 840–845.
- ↵Sebaldt RJ, Sheller JR, Oates JA, Roberts LJII, FitzGerald GA. Inhibition of eicosanoid biosynthesis by glucocorticoids in humans. Proc Natl Acad Sci U S A. 1990; 87: 6974–6978.
- ↵Miwa Y, Sasaguri T, Inoue H, Taba Y, Ishida A, Abumiya T. 15-Deoxy-Δ12,14-prostaglandin J2 induces G1 arrest and differentiation marker expression in vascular smooth muscle cells. Mol Pharmacol. 2000; 58: 837–844.
- ↵Baker CS, Hall RJ, Evans TJ, Pomerance A, Maclouf J, Creminon C, Yacoub MH, Polak JM. Cyclooxygenase-2 is widely expressed in atherosclerotic lesions affecting native and transplanted human coronary arteries and colocalizes with inducible nitric oxide synthase and nitrotyrosine particularly in macrophages. Arterioscler Thromb Vasc Biol. 1999; 19: 646–655.
- ↵Kosaki K, Ando J, Korenaga R, Kurokawa T, Kamiya A. Fluid shear stress increases the production of granulocyte-macrophage colony-stimulating factor by endothelial cells via mRNA stabilization. Circ Res. 1998; 82: 794–802.