Vascular Cell Adhesion Molecule–1 Is Induced on Vascular Endothelia and Medial Smooth Muscle Cells in Experimental Cardiac Allograft Vasculopathy
Background Cardiac allograft vasculopathy (CAV) is the major cause of late death among heart transplant recipients. The pathogenesis of CAV is poorly understood.
Methods and Results To better characterize CAV, we performed immunohistochemical analysis of vascular lesions in a previously described murine model of CAV. The B10.A strain hearts were transplanted heterotopically into B10.BR strain recipients. The cardiac allografts were harvested from 1 to 2 months after implantation. The majority of epicardial and intramyocardial coronary arteries in explanted hearts had developed intimal thickening. The cellular infiltrate of the intimal thickening, major histocompatibility (MHC) antigens, intracellular adhesion molecule–1 (ICAM-1), and vascular cell adhesion molecule–1 (VCAM-1) expression were studied with the use of immunohistochemistry. In experimental CAV in mice, the cellular infiltrate of expanded intima consisted of macrophages, T lymphocytes, and smooth muscle cells. A substantial number of macrophages and T lymphocytes within the expanded intima expressed MHC class II antigen, a marker of cellular activation. The vessel wall cells also appeared to be activated due to their expression of endothelium-leukocyte adhesion molecules. The vascular endothelium of cardiac allografts displayed ICAM-1, VCAM-1, and unmatched MHC antigen (MHC class I in this model) upregulation. The medial smooth muscle cells also expressed VCAM-1 and unmatched MHC antigen.
Conclusions These findings suggest that (1) the cellular infiltrate of the expanded intima in experimental CAV is similar to that of human CAV, (2) experimental CAV is a local immune-mediated process requiring active participation of donor vessel wall cells and recipient mononuclear cells, and (3) coexpression of adhesion molecules and unmatched MHC antigen identifies endothelial cells as immune targets for activated host mononuclear cells. Furthermore, the presence of both VCAM-1 and unmatched MHC antigen supports a central role for medial smooth muscle cells as allogeneic immune stimulator.
Cardiac allograft vasculopathy (CAV), an accelerated form of arteriosclerosis in transplanted hearts, has emerged as the leading cause of death among heart transplant recipients after the first year.1 The incidence of clinically significant CAV increases with survival, approaching 50% at 5 years after transplantation, with a resultant mortality rate of 25%.2 3 Pathologically, it is characterized by concentric intimal proliferation of large and medium-sized coronary arteries and veins.4 In human CAV, the cellular infiltrate of the intimal proliferative lesions consists of modified smooth muscle cells, macrophages-monocytes, and T lymphocytes.5 6 7
The pathogenesis of CAV is poorly understood, although an immunological process is believed to play a significant role in this process.6 Hyperlipidemia and/or CMV infection may contribute to the development of CAV as cofactors.8 9 CAV is believed to involve a chronic immune response of the host to the allogeneic donor vasculature, resulting in the production of cytokines that elicit smooth muscle cell migration and proliferation.10 The targets of the allogeneic immune response as well as the details of interaction between host mononuclear cells and donor vasculature are unknown.
A reproducible animal model can provide important insights into the pathogenesis of this form of accelerated arteriosclerosis. We have previously described a murine model of CAV without immunosuppression that uses heterotopic transplantation of B10.A strain hearts into B10.BR strain recipients.11 The histological appearance of the vascular lesions was comparable to CAV seen clinically. We used immunohistochemistry to further characterize the cellular infiltrate of the vascular lesions and to study the expression of major histocompatibility (MHC) antigens, intracellular adhesion molecule–1 (ICAM-1), and vascular cell adhesion molecule–1 (VCAM-1). The findings provide insight into the pathogenesis of experimental CAV.
B10.A and B10.BR strains of mice (7 to 11 weeks old) were purchased from Jackson Laboratories and housed under conventional conditions. They were fed rodent chow (Purina Mills) and water ad libitum. The B10.A and B10.BR strains of mice differ at the D and L loci of MHC class I antigens but are matched at other loci.12 All animals received humane care in compliance with the “Principles of Laboratory Animal Care” and the “Guide for the Care and Use of Laboratory Animals” prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH publication No. 86-23, revised 1985).
The intra-abdominal heterotopic heart transplantation was performed with the use of the microsurgical techniques described by Corry et al.13 The donor ischemia time varied between 30 and 60 minutes. The overall success rate exceeded 90%. Function of the allografts was assessed daily through abdominal palpation and scored on a scale of 0 to 4 (4 indicates normal beating, and 0 indicates absence of contractions).
In experimental groups 1, 2, and 3, B10.A strain hearts were transplanted into B10.BR strain recipients. Cardiac allografts were harvested on day 30 (group 1), 45 (group 2), or 60 (group 3) after implantation. In the control group (group 4), B10.BR strain hearts were transplanted into B10.BR recipients and were harvested at 60 days after implantation. There were six animals in each group. No immunosuppressive therapy was administered. At harvest, the cardiac allografts and the native hearts of the recipient mice were explanted and immediately snap-frozen in OCT compound (Tissue Tek) in liquid nitrogen and stored at −70°C.
The primary antibodies that were used were biotinylated primary murine antibodies, clone 36-7-5 for MHC class I and clone 11-5.2 for MHC class II antigens; hamster anti-mouse clone 3E2 for ICAM-1; rat anti-mouse clone MVCAM.A(429) for VCAM-1; rat anti-mouse MAC-1 clone M1/70 for macrophages; rat anti-mouse clone RN-4-5 for L3T4 and clone 53-6.7 for Ly-2 recognizing CD4+ and CD8+ T lymphocytes, respectively; and rabbit anti-tropomyosin antiserum for smooth muscle cells. All primary antibody reagents were purchased from Pharmingen except for clone M1/70 (Boehringer-Mannheim) and anti-tropomyosin (Sigma Chemical Co).
Immunohistochemistry was performed on cryostat sections with the use of an avidin-biotin-peroxidase technique (Vector Elite detection reagents; Vector Inc) with AEC as the chromogen (Biomeda) by following the manufacturers’ instructions. Serial cryostat sections, 6 μm thick, were air-dried and frozen at −70°C in sealed slide boxes with desiccant. For staining, sections were rehydrated in wash buffer (0.1 mol/L phosphate-buffered saline, pH 7.4) for 15 minutes, fixed for 5 minutes in either chilled acetone or methanol depending on the primary antibody to be used, washed again, and then incubated in a blocking solution (5% nonfat dry milk in wash buffer plus 1% serum of species from which the secondary antibody reagent was derived). Primary antibodies were used at concentrations of 1 or 10 μg/mL and incubated with sections overnight at 4°C, except for the anti-tropomyosin, which was used at a 1:1000 dilution and incubated for 1 hour at 37°C. As controls, adjacent sections were handled similarly but were incubated with nonimmune rat or hamster IgG preparations at the same concentrations. Specificity of antitropomyosin for smooth muscle cells was shown by selective strong reactivity with medial smooth muscle cells of coronary arteries and lack of reactivity with macrophages or lymphocytes. In cardiac allografts, arterial endothelial cells and cardiac myocytes were sometimes weakly reactive, but these were histologically distinguishable from the underlying intimal or medial cells in coronary arteries.
Relative cell distribution in each intimal lesion was determined by counting the number of Ly-2, L3T4, MAC-1, and tropomyosin-reactive cells in adjacent sections and expressing each as the percentage of the sum of all four. The regional intensity of MHC antigen, ICAM-1, and VCAM-1 expression was scored on a blinded basis by one investigator with the use of a semiquantitative method (0 indicates no reactivity; 1, weak; 2, moderate; and 3, strong reactivity). An average of four affected vessels per heart were analyzed for all determinations.
Paired t test was used to compare the mean values of expression of MHC antigens, ICAM-1, and VCAM-1 on cardiac allografts versus native hearts of recipient mice (since both cardiac allografts and native hearts of recipient mice were explanted from the same group of animals). ANOVA (posthoc t test) was used to compare the mean values of expression of these antigens on cardiac allografts versus isografts as well as the relative distribution of the cellular infiltrate in the intimal lesions of groups 1 through 3.
General Characteristics and Histological Analysis of Vascular Lesions
All cardiac allografts were contracting well at the time of harvest, and none were lost to rejection during the study period. All allografts were scored as 3 or 4 as a result of daily abdominal palpation. The histological grading of cellular rejection in cardiac allografts of experimental groups ranged from ISHLT 1A to 3A.14 Cardiac isografts of the control group did not demonstrate cellular rejection. As reported previously, a majority (>55%) of epicardial and intramyocardial coronary arteries in this murine model of CAV develop significant intimal thickening within 30 days.11 Histologically, the vascular lesions are characterized by subendothelial mononuclear cell infiltration, smooth muscle cell proliferation, frequent disruption of internal elastic lamina, and some perivascular inflammation.11 Medial disruption and/or destruction by inflammatory cells was relatively common in vessels with CAV, a process showing progression over time.
Characterization of Cellular Infiltrate
The expanded intima of coronary arteries in experimental CAV contains T lymphocytes, macrophages, and smooth muscle cells (Fig 1a⇓ through 1d). T lymphocyte subtype identification revealed CD4+ and CD8+ T lymphocytes in relatively equal numbers scattered within the expanded intima. The relative contribution of T lymphocytes to the intimal infiltrate was 11% at 30 days (Fig 2a⇓). At 45 days, it was 22% and remained essentially unchanged at 60 days (P<.05). Macrophages were present consistently within the intimal lesions as well as in the perivascular area. The proportion of macrophages in intimal cellular infiltrate remained stable at 30, 45, and 60 days after implantation, comprising approximately 40% of intimal cells (P>.05) (Fig 2b⇓). Within the expanded intima, the smooth muscle cells constituted 46% of cellular infiltrate at 30 days, 32% at 45 days, and 39% at 60 days (P>.05) (Fig 2c⇓). A perivascular adventitial cell population consisting of T lymphocytes and macrophages was consistently noted around vessels with intimal lesions.
The native hearts of recipient mice and the cardiac isografts did not display intimal lesions or perivascular infiltrate.
MHC Antigen Expression
The immunohistological analyses of MHC antigen, ICAM-1, and VCAM-1 expression did not differ significantly when cardiac allografts of different experimental groups were compared. Therefore, we refer to the cardiac allografts of experimental groups 1 through 3 collectively.
As noted previously, the donor and recipient strains differ at the D and L loci of MHC class I antigens but are identical for class II and minor histocompatibility antigens. The MHC class I (H-2K in mouse) antigens were consistently expressed on endothelial cells and medial smooth muscle cells of coronary arteries in cardiac allografts but not in native hearts of recipient mice (Table⇓). The expression of MHC class I antigens on cardiac isografts was weak on endothelial cells and absent on medial smooth muscle cells.
MHC class II antigens could not be detected on the endothelial cells or smooth muscle cells of any of the examined cardiac allografts, isografts, or native hearts of recipient mice. The expression of MHC class II antigens appeared to be limited to the intimal macrophages and T lymphocytes in the cardiac allografts of the experimental groups (Fig 3⇓).
VCAM-1 and ICAM-1 Expression
The endothelial cells of cardiac allografts in all experimental groups uniformly expressed VCAM-1 and ICAM-1. On endothelial cells, the intensity of ICAM-1 expression appeared mild, whereas that of VCAM-1 was strong (Table⇑, Figs 4a⇓ and 5a⇓). Within the intimal lesions, VCAM-1 appeared to be expressed only by smooth muscle cells. Of particular interest was the expression of VCAM-1 on medial smooth muscle cells of cardiac allografts and the lack of ICAM-1 expression. The native hearts of recipient mice also displayed ICAM-1 and VCAM-1 expression on endothelial cells (Figs 4b⇓ and 5b⇓). In contrast to cardiac allografts, the native hearts did not express VCAM-1 on medial smooth muscle cells (Fig 5b⇓). ICAM-1 and VCAM-1 could also be weakly detected on endothelial cells of cardiac isografts but not on medial smooth muscle cells (Table⇑, Figs 4c⇓ and 5c⇓).
In experimental CAV, the intimal cellular infiltrate consists of macrophages, T lymphocytes (CD4+ and CD8+ T lymphocytes in nearly equal numbers), and smooth muscle cells. A majority of macrophages and T lymphocytes of the expanded intima express MHC class II antigens, a sign of cellular activation. The vascular endothelium of both cardiac allografts and native hearts of recipient mice displayed ICAM-1 and VCAM-1 induction. In contrast, the medial smooth muscle cells of coronary vessels in only cardiac allografts expressed VCAM-1. The expression of adhesion molecules on the vessel walls of the cardiac allografts was uniformly accompanied by expression of foreign MHC antigens (MHC class I in this model).
The cellular characteristics of the vascular lesions in this experimental model of CAV appear to be similar to those of human CAV. The immunohistochemical characterization of human CAV has identified smooth muscle cells, macrophages, and T lymphocytes as the cellular constituents of intimal thickening.5 6 7 The number of CD4+ and CD8+ T lymphocytes has varied from nearly equal to a predominance of CD8+ T lymphocytes.5 6 In experimental CAV, the cellular infiltrate of the expanded intima also consists of macrophages, T lymphocytes, and smooth muscle cells. The intimal cell population is composed of a mixture of donor cells (smooth muscle cells) and blood-borne inflammatory recipient cells (macrophages and T lymphocytes). The macrophages and smooth muscle cells were the predominant cellular constituents of the intimal lesions. In contrast to upregulation of MHC class II antigens on endothelial cells of human CAV, the endothelial cells of the intimal lesions in this experimental model (where donor and recipient are matched at MHC class II–determined antigens) did not express MHC class II antigens. This finding is consistent with previous reports demonstrating a lack of MHC class II antigen expression on mouse coronary artery endothelial cells, even in the setting of exogenously administered γ-interferon.15 16 The resemblance of the intimal cellular infiltrate phenotype in this murine model of CAV to that of human CAV underlines the usefulness of this model in future studies.
The etiology of CAV is incompletely elucidated; however, it is likely that the immune system plays a central role. The development of CAV in cardiac allografts of animal models with some histocompatibility mismatch, the lack of development in isografts, and the limitation of vascular lesions to the allograft vascular tree are among the evidence supporting an immune-mediated mechanism. The details of the immune-mediated process are unknown. The findings of the present study provide some insight into the pathogenesis of CAV.
CAV is a chronic, localized, immune-mediated process representing active interaction between donor vessel wall cells and host mononuclear cells.17 A substantial number of host macrophages and T lymphocytes within the expanded intima appeared to be activated due to their expression of MHC class II antigens.18 The endothelial cells and medial smooth muscle cells of the intimal lesions expressed ICAM-1 and VCAM-1, two cytokine-inducible activation markers.19 20 Colocalization of activated recipient mononuclear cells with activated endothelial cells and smooth muscle cells within the vascular lesions was a consistent observation. This finding gives support to the hypothesis that CAV is a local immunological reaction involving active participation of host mononuclear cells and donor vessel wall cells.21 In fact, the resemblance of the vascular lesions to a chronic, immune-mediated, delayed-type hypersensitivity reaction was striking. The hypothesis that CAV represents a form of chronic, delayed-type hypersensitivity reaction had been proposed by Libby et al.10 Interaction between vessel wall cells and the host infiltrating mononuclear cells, mediated by a network of cytokines, yields a local environment favoring intimal proliferation. The current observations provide evidence for the presence of in situ immune-activated donor vessel wall cells and host mononuclear cells in the expanded intima of CAV.
VCAM-1 expression (in association with foreign MHC antigen) on endothelial cells and medial smooth muscle cells of vessels with intimal proliferation suggests that VCAM-1 participates in mononuclear cell recruitment and development of CAV. The adherence of leukocytes to vascular endothelium is a prerequisite for transmigration. The adhesive properties of vascular endothelium are altered by induction of endothelial-leukocyte adhesion molecules. VCAM-1 is a mononuclear leukocyte–selective adhesion molecule that interacts with VLA-4 and is expressed on monocytes, lymphocytes, eosinophils, and basophils but not neutrophils. Due to its functional specificity and selective expression pattern, it has been suggested that VCAM-1 participates in the development of native vessel atherosclerosis.22 In experimental diet-induced atherosclerosis, endothelial cell VCAM-1 induction precedes monocyte attachment and migration.23 In advanced human atherosclerotic plaques, VCAM-1 expression has been documented on plaque smooth muscle cells and neovasculature endothelium, a potential site for inflammatory cell recruitment in advanced atherosclerosis.24 In human CAV as well as this experimental model of CAV, macrophages and T lymphocytes constitute the majority of host leukocytes within the expanded intima. Therefore, due to its selective recruitment potential, VCAM-1 appears to be a likely candidate in recruitment of mononuclear cells into the allograft and development of CAV.
In this model of CAV, VCAM-1 and ICAM-1 were expressed on endothelial cells of cardiac allografts and native hearts of recipient mice. Because the native hearts of recipient mice did not display any vascular lesions, the sole expression of VCAM-1 and ICAM-1 on endothelial cells does not appear to play a pathogenic role. A possible explanation for the upregulation of adhesion molecules on the vascular endothelium of both cardiac allografts and native hearts of recipient mice is the release of systemic inflammatory mediators due to cardiac allograft transplantation.25 In addition to adhesion molecule induction, vascular endothelial cells of cardiac allografts, but not the native hearts of recipient mice, also expressed foreign MHC antigens (MHC class I antigens in this model). T-cell stimulation requires engagement of the T-cell receptor/CD3 complex with MHC antigens as well as a costimulatory signal. The interaction of VCAM-1 and ICAM-1 with their respective ligands has been shown to mediate costimulation of T cells and regulate cytokine release.26 Thus, vascular endothelium of cardiac allografts appears to be capable of mediating adhesion and activation of host T cells through coexpression of foreign MHC antigens and vascular adhesion molecules.
More interestingly, VCAM-1 expression was significantly induced on medial smooth muscle cells of vascular lesions in cardiac allografts. In contradistinction, VCAM-1 expression could not be detected on medial smooth muscle cells of native hearts of recipient mice or cardiac isografts. As noted previously, medial smooth muscle cells of cardiac allografts also displayed a significant upregulation of foreign MHC antigens (MHC class I antigens in this model). It is intriguing to speculate that host mononuclear cell interaction with donor endothelial cells expressing foreign antigens and costimulator adhesion molecules activates and recruits the recipient’s macrophages and T lymphocytes. The recruited mononuclear cells then encounter the medial smooth muscle cells expressing VCAM-1 and foreign MHC antigens, enhancing allogeneic stimulation. Local release of cytokines from activated mononuclear cells may contribute to smooth muscle cell proliferation and the development of CAV.
Two groups have proposed that an alloreactive immune response directed against the allograft endothelial cells is the inciting event in the development of CAV.10 27 In vitro studies have demonstrated that human aortic endothelial cells, when stimulated by allogeneic lymphocytes, are capable of producing a panel of growth factors to modulate smooth muscle cell proliferation.27 The present study demonstrates that there is a significant expression of two adhesion molecules on endothelial cells of cardiac allografts that are known to recruit inflammatory cells and facilitate allogeneic recognition. These in vivo findings further substantiate a role for allograft endothelial cells in the development of CAV.
The role of allograft vessel wall smooth muscle cells in the evolution of CAV is not as well understood. The identification of VCAM-1 on smooth muscle cells is not unprecedented. Cultured vascular smooth muscle cells can be induced to express VCAM-1 by exposure to γ-interferon and tumor necrosis factor–α, products of activated mononuclear cells.28 VCAM-1 has been observed on smooth muscle cells of native human atherosclerotic plaques.24 Smooth muscle cells are also known to function well in antigen presentation.29 The observation of VCAM-1 expression in association with foreign MHC antigens suggests a pivotal role for smooth muscle cells in immune activation and retention of infiltrating host leukocytes in the development of CAV.
In conclusion, the present study provides an immunohistochemical characterization of CAV in an experimental model that resembles human CAV. It demonstrates that CAV is a local immune-mediated process that involves activated vessel wall cells and host mononuclear cells. Furthermore, it substantiates that allograft endothelial cells and medial smooth muscle cells work in concert to provoke an alloreactive immune response by coexpressing foreign MHC antigens and costimulatory adhesion molecules.
Reprint requests to Hillel Laks, MD, Division of Cardiothoracic Surgery, UCLA Medical Center, CHS 62-182A, 10833 LeConte Ave, Los Angeles, CA 90024.
- Received January 9, 1995.
- Revision received January 24, 1995.
- Accepted January 27, 1995.
- Copyright © 1995 by American Heart Association
Uretsky BF, Murali S, Reddy PS, Rabin B, Lee A, Griffith BP, Hardesty RL, Trento A, Bahnson HT. Development of coronary artery disease in cardiac transplant patients receiving immunosuppressive therapy with cyclosporine and prednisone. Circulation. 1987;76:827-834.
Olivari MT, Homans DC, Wilson RF, et al. Coronary artery disease in cardiac transplant patients receiving triple immunosuppressive therapy. Circulation. 1989;80(suppl III):III-111-III-115.
Barnhart GR, Pascoe EA, Mills AS, Szentpetery S, Eich DM, Mohanakumart T, Hastillo A, Thompson JA, Hess HL, Lower RR. Accelerated coronary arteriosclerosis in cardiac transplant recipients. Transplant Rev. 1987;1:31-46.
Winters GL, Kendall TJ, Radio SJ, Wilson JE, Costanzo-Nordin MR, Switzer BL, Remmenga JA, McManus BM. Post-transplant obesity and hyperlipidemia: major predictors of severity of coronary arteriopathy in failed human heart allografts. J Heart Lung Transplant. 1990;9:364-371.
Billingham ME, Cary NR, Hammond ME, Kemnitz J, Marboe C, McCallister HA, Snover DC, Winters GL, Zerbe A. A working formulation for the standardization of nomenclature in the diagnosis of heart and lung rejection: Heart Rejection Study Group. J Heart Lung Transplant. 1990;9:587-593.
Skoskiewicz MJ, Colvin RB, Schneeberger EE, Russell PS. Widespread and selective induction of major histocompatibility complex-determined antigens in vivo by gamma-interferon. J Exp Med. 1985;162:1645-1664.
Ardehali A, Drinkwater DC, Laks H, Drake T. Cardiac allograft vasculopathy. Am Heart J. 1994;127:1498-1502.
Burmester GR, Jahn B, Gramatzki M, Zacher J, Kalden JR. Activated T cells in vivo and in vitro: divergence in expression of Tac and Ia antigens in the nonblastoid small T cells of inflammation and normal T cells activated in vitro. J Immunol. 1984;133:1230-1234.
Bevilacqua MP, Pober JS, Mendrick DL, Cotran RS, Gimbrone MA Jr. Identification of an inducible endothelial-leukocyte adhesion molecule. Proc Natl Acad Sci U S A. 1987;84:9238-9242.
Oppenheimer-Marks N, Davis LS, Bogue BT, Ramberg J, Lipsky PE. Differential utilization of ICAM-1 and VCAM-1 during the adhesion and transendothelial migration of human T-lymphocytes. J Immunol. 1991;147:2913-2921.
Cybulsky MI, Gimbrone MA. Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis. Science. 1991;251:788-791.
Li H, Cybulsky MI, Gimbrone MA Jr, Libby P. An atherogenic diet rapidly induces VCAM-1, a cytokine-regulatable mononuclear leukocyte adhesion molecule, in rabbit aortic endothelium. Arterioscler Thromb. 1993;13:197-204.
O’Brien KD, Allen MD, McDonald TO, Chait A, Harlan JM, Fishbein D, McCarty J, Ferguson M, Hudkins K, Benjamin CD. Vascular cell adhesion molecule-1 is expressed in human coronary atherosclerotic plaques: implications for the mode of progression of advanced coronary atherosclerosis. J Clin Invest. 1993;92:945-951.
Fanslow WB, Sims JE, Sassenfeld H, Morrissey PJ, Gillis S, Dower SK, Widmer MB. Regulation of alloreactivity in vivo by a soluble form of the interleukin-1 receptor. Science. 1990;248:739-742.
Van Seventer GA, Newman W, Shimizu Y, Nutman TB, Tanaka Y, Horgan KJ, Gopal TV, Ennis E, O’Sullivan D, Grey H. Analysis of T cell stimulation by superantigen plus major histocompatibility complex class II molecules or by CD3 monoclonal antibody: costimulation by purified adhesion ligands VCAM-1, ICAM-1, but not ELAM-1. J Exp Med. 1991;174:901-913.
Wagner CR, Morris TE, Shupley GD, Hosenpud JD. Regulation of human aortic endothelial cell-derived mesenchymal growth factors by allogeneic lymphocytes in vitro: a potential mechanism for cardiac allograft vasculopathy. J Clin Invest. 1993;92:1269-1277.
Couffinhal T, Duplaa C, Moreau C, Lamaziere JMD, Bonnet J. Regulation of vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 in human vascular smooth muscle cells. Circ Res. 1994;74:225-234.