Value of Serum-Soluble Intercellular Adhesion Molecule-1 for the Noninvasive Risk Assessment of Transplant Coronary Artery Disease, Posttransplant Ischemic Events, and Cardiac Graft Failure
Background—Adhesion molecules on arterial endothelium have been implicated in spontaneous atherosclerosis and transplant coronary artery disease (CAD). We studied whether elevated serum-soluble intercellular adhesion molecule-1 (sICAM-1) during the immediate posttransplant period was a risk factor for CAD, posttransplant ischemic events, or cardiac graft failure.
Methods and Results—We initially studied serum sICAM-1 in a subset of 16 cardiac allograft recipients (5.5±0.7 samples per patient) to determine a cutoff point that best correlated with presence of arterial and arteriolar endothelial ICAM-1 in matching endomyocardial biopsies. The cutoff value was 308 ng/mL. Subsequently, we prospectively evaluated serum sICAM-1 in serial samples (5.3±0.1 per patient) obtained during the first 3 months after transplantation in a validation subset of 130 recipients and correlated early sICAM-1 levels with long-term outcome. Serum sICAM-1 >308 ng/mL correlated significantly with ICAM-1 on arterial and arteriolar endothelium (P=0.02). Cardiac allograft recipients with serum sICAM-1 >308 ng/mL had 2.67 (95% CI, 1.28 to 5.59, P=0.009) times greater risk of CAD and 3.63 (95% CI, 1.05 to 12.5, P=0.04) times greater risk of graft failure. Recipients with sICAM-1 >308 ng/mL also developed more severe CAD (P=0.009) and more ischemic events (P=0.03) after transplantation.
Conclusions—Serum sICAM-1 levels can be used to noninvasively assess risk of transplant CAD, posttransplant ischemic events, and cardiac graft failure.
Transplant coronary artery disease (CAD) is a leading cause of allograft failure in cardiac transplant recipients after the first year after transplantation.1 Several risk factors have been implicated in the development of spontaneous disease, and some of the traditional risk factors have been associated with development of transplant CAD.1 2 3 Adhesion molecule expression on the endothelium may be a risk factor for developing both spontaneous and transplant CAD.4 5 6
Adhesion molecules are expressed on arterial endothelium from lesions of transplant CAD.5 7 8 The development of transplant CAD is preceded by expression of intercellular adhesion molecule-1 (ICAM-1) on arterial and arteriolar endothelium,5 and adhesion molecules on arterial endothelium have been directly implicated in neointimal formation.7 Soluble adhesion molecules seem to be relevant, because the extent of human atherosclerosis is correlated with circulating levels.7 Interestingly, plasma-soluble ICAM-1 (sICAM-1) is a risk factor for coronary occlusion and myocardial infarction, supporting the idea that endothelial activation occurs early in atherothrombosis.6 We evaluated whether increased serum sICAM-1 during the first 3 months after transplantation was a risk factor for subsequent transplant CAD, ischemic events, or graft failure.
Initially, we studied serum samples and corresponding endomyocardial biopsies (5.5±0.7 per patient) from 16 consecutive adult cardiac allograft recipients who received transplants between January and October 1989 to determine a value of sICAM-1 that best correlated with immunohistochemical findings for ICAM-1 within the biopsies. We then prospectively studied an additional 130 consecutive patients who received transplants between October 1989 and July 1997 to evaluate the relationship between sICAM-1 during the first 3 months after transplantation and subsequent outcome. Patients were enrolled if they survived at least 6 months after transplantation, had pretransplantation and serial endomyocardial biopsy specimens taken during the first 3 months after transplantation for light microscopy and immunohistochemical studies, and had angiographic or histopathological evaluations of coronary arteries.5 9 10 Coronary angiography was performed yearly and averaged 3.6±0.2 angiograms per patient. This study was approved by the Methodist Research Institute’s Institutional Review Board, and all subjects signed a consent form.
Triple-drug immunosuppressive therapy consisted of prednisone, azathioprine or mycophenolate mofetil, and cyclosporine. Prednisone was administered at an initial dose of 1 mg · kg−1 · d−1, tapered to 0.5 mg · kg−1 · d−1 during the first month, 0.2 mg · kg−1 · d−1 during months 1 through 2, and 0.1 mg · kg−1 · d−1 during months 3 through 12 after transplantation. Patients were maintained on this dose unless they developed steroid complications. Azathioprine was administered at a dose of 1.5 to 2.0 mg · kg−1 · d−1 and mycophenolate mofetil at a dose of 1 g twice a day. Cyclosporine was administered at an initial dose of 7 to 10 mg · kg−1 · d−1, tapered to 3 to 5 mg · kg−1 · d−1 to maintain a blood level of 300 to 480 ng/mL during the first 3 months, 180 to 360 ng/mL during months 3 through 6, 90 to 180 ng/mL during months 6 through 12, and 75 to 120 ng/mL at >12 months after transplantation, depending on renal function. Grades 3 and 4 rejection episodes were treated with steroids and rabbit antithymocyte globulin or OKT3. Rejection with hemodynamic compromise was defined as a decreased ejection fraction and clinical signs of low cardiac output treated with methylprednisolone and/or inotropic agents. Higher-dose immunosuppressants, calcium channel blockers, or lipid-lowering agents (statins) were used at the physician’s discretion without knowledge of serum or immunohistochemical data for ICAM-1. Ejection fractions were measured by radionuclide ventriculography. Ischemic events were defined as either (1) presence of nontraumatic sudden cardiac death, (2) need of a revascularization procedure (eg, coronary artery bypass surgery, percutaneous transcoronary atherectomy, or stent), (3) presence of clinically evident ischemia (eg, clinical myocardial infarction, anginal symptoms, segmental wall motion abnormalities at cardiac catheterization supplied by a stenotic or occluded coronary artery, or ischemia by dobutamine stress test), or (4) any combination thereof. Graft failure was defined as death associated with unexplained cardiac allograft dysfunction or biopsy-proven cellular rejection or need for a second transplant.
A control biopsy from the right ventricle was obtained before transplantation from all donor hearts. Endomyocardial biopsies were obtained by right cardiac catheterization at 7 to 10 days, every 2 weeks during the first 2 months, and at 3 months after transplantation. Cellular infiltrates were graded according to the International Society for Heart Transplantation.5 9 10 Cytomegalovirus disease was defined as clinical disease with evidence of tissue invasion by cytopathological and/or tissue culture criteria. Cytomegalovirus prophylaxis with ganciclovir was used in seronegative recipients with seropositive donors.
Determinations of sICAM-1
Serial serum samples for each recipient were obtained at the time endomyocardial biopsies were performed and stored at −75°C. Samples were thawed and assayed in duplicate for sICAM-1 with an ELISA (R&D Systems). The minimum detectable concentration (sensitivity) of sICAM-1 was 0.35 ng/mL. Laboratory personnel were unaware of immunohistochemical data from biopsies or patient outcome.
Criteria for Diagnosis of CAD
CAD was diagnosed as any decrease in luminal diameter, whether in the left main coronary artery or primary or branch vessels, and was classified as mild, moderate, or severe on the basis of the most severe CAD reported in each coronary angiogram.11 Annual arteriograms were compared with identical projections in serial studies and evaluated by side-by-side comparisons. The presence and severity of disease were determined by a consensus of 2 experienced angiographers blinded to the results of the serological or immunohistochemical studies for ICAM-1. Progression of disease was evaluated by comparing serial angiograms side-by-side with the baseline angiogram obtained the first year after transplantation. To reduce the possibility of donor-transmitted CAD, recipients having a normal angiogram the first year after transplantation were studied during subsequent follow-up, and the first annual angiogram was considered baseline. Coronary arteries were examined histopathologically in recipients who died before their first annual angiogram. The degree of luminal narrowing was estimated visually and classified as mild, moderate, or severe.12
Antibodies and Control Experiments
Endomyocardial biopsies were studied immunohistochemically for ICAM-1 with monoclonal antibody LB-2 (Becton Dickinson). Arteries and arterioles were identified with monoclonal antibody to smooth-muscle–specific α-actin (1A4, Biomakor). Secondary antibodies consisted of affinity-purified fluorochrome-labeled F(ab′)2 antibody fragments to mouse immunoglobulins (Protos ImmunoResearch). Control experiments were performed as described.5 9 10
Biopsies were embedded in OCT compound (Miles), snap-frozen in liquid nitrogen, and stored at −20°C. Cryostat sections (4 μm) were air-dried overnight without chemical fixation. Antibody experiments were performed with primary antibodies and fluorochrome-labeled secondary antibodies. Immunohistochemical data were evaluated by 2 investigators unaware of the clinical outcome. The precise vascular localization of ICAM-1 was performed as described.5 The proportion of biopsies with arterial and arteriolar endothelial ICAM-1 during the first 3 months after transplantation was then calculated.5
Receiver operating characteristic analysis determined a cutoff point for sICAM-1 that corresponded to the immunohistochemical detection of arterial and arteriolar endothelial ICAM-1 in the matching biopsy among the initial subset of 16 allograft recipients. Recipients in the validation subset (n=130) with sICAM-1 at or under the cutoff point and those with sICAM-1 over the cutoff point were compared for incidence and severity of CAD and rates of graft failure. Wilcoxon rank-sum tests (for continuous and ordered variables) and Fisher’s exact tests (for discrete variables) were used to compare demographic and clinical-laboratory characteristics of the 2 groups. Any significant variable was included in subsequent Cox regression (for the interval to CAD and graft failure) and logistic regression (for severity of disease and ischemic events) to determine the significance of sICAM-1. Kaplan-Meier curves illustrated the rates of CAD and graft failure of the 2 groups over time. Values of P<0.05 were considered statistically significant. Summary statistics were reported as mean±SEM. Statistical analyses were performed with SAS version 6.12.
Among the initial subset of 16 allograft recipients, a cutoff point for sICAM-1 was determined that best correlated with the presence of arterial and arteriolar endothelial ICAM-1 in the biopsies (5.5±0.7 per patient) because such reactivity was never found in donor hearts before transplantation. Sensitivity and specificity of this association were maximized when a cutoff of 308 ng/mL was used (Figure 1⇓). With this cutoff point, serial determinations (5.3±0.1 per patient) of sICAM-1 and matching endomyocardial biopsies obtained during the first 3 months after transplantation were prospectively analyzed in the subsequent 130 recipients. Thirty-three recipients had sICAM-1 remaining ≤308 ng/mL (average sICAM-1, 185.2±10.5 ng/mL). However, 97 recipients had ≥1 samples exceeding the cutoff point (average sICAM-1, 355.1±13.7 ng/mL). Most samples from these recipients (52.8%; 2.8±0.2 per patient) showed serum sICAM-1 >308 ng/mL, and 61 of 97 (62.9%) already had sICAM-1 >308 ng/mL 7 to 10 days after transplantation. The average within-patient standard deviation was 92.3±7.8 ng/mL, and no relationship between time after transplantation and sICAM-1 was found during the first 3 months after transplantation (P=0.68). A significant correlation (P=0.02) was found between increased sICAM-1 and ICAM-1 on arterial and arteriolar endothelium of matching serial endomyocardial biopsies (Table 1⇓). Recipients with sICAM-1 >308 ng/mL had a significantly higher percentage of endomyocardial biopsies with arterial and arteriolar endothelial ICAM-1 (Figure 2⇓).
Demographic and clinical-laboratory data of the studied population are shown in Table 2⇓. Significant differences were found for sex of recipient, age of donor, HLA-DR mismatches, and presence of cytomegalovirus disease.
We studied the relationship between sICAM-1 determinations performed during the first 3 months after transplantation and subsequent development of transplant CAD in 130 recipients followed up for 44.9±2.5 months after transplantation (Table 3⇓). Recipients with sICAM-1 >308 ng/mL developed significantly more CAD (P=0.02), developed CAD earlier (P=0.009), developed more severe disease (P=0.009), and had more disease progression (P=0.04) than recipients with sICAM-1 remaining at or under the cutoff point.
To reduce the possibility of donor-transmitted disease, we evaluated allograft recipients who had a normal angiogram the first year after transplantation (n=97) and followed up this subgroup of patients over a period of 37.2±3.0 months after the first angiogram. Recipients who initially had sICAM-1 >308 ng/mL (n=67) developed significantly more CAD (P=0.004) and more severe disease (P=0.002) than recipients with sICAM-1 levels ≤308 ng/mL (n=30) (Figure 3⇓).
Kaplan-Meier estimates indicated that 1-year CAD rates in recipients with sICAM-1 remaining ≤308 ng/mL were 3.3±3.3%, compared with 23.0±4.3% among recipients with sICAM-1 >308 ng/mL (Figure 4⇓).
The average sICAM-1 level during the first 3 months after transplantation for allografts that remained free of CAD was 274.5±7.8 ng/mL, and for allografts that developed angiographically detectable CAD, the average sICAM-1 level was 344.0±18.4 ng/mL (P<0.001). Interestingly, these values were, respectively, under and over the cutoff of 308 ng/mL calculated with an initial subset of 16 allograft recipients.
We subsequently analyzed the relationship between sICAM-1 and development of ischemic events during follow-up. Recipients with sICAM-1 >308 ng/mL during the first 3 months after transplantation experienced significantly more events (P=0.03) than recipients with sICAM-1 ≤308 ng/mL (Table⇑ 3).
We then studied the relationship between sICAM-1 and subsequent allograft failure (Table 3⇑). Recipients with sICAM-1 >308 ng/mL during the first 3 months after transplantation experienced more graft failure (P=0.03) and their grafts failed earlier (P=0.04) than recipients with sICAM-1 ≤308 ng/mL. Interestingly, 87.1% of grafts that failed developed CAD, detected either angiographically or histopathologically.
Kaplan-Meier estimates indicated that graft survival at 5 years was 86.3±7.5% in recipients with sICAM-1 ≤308 ng/mL and 69.8±5.5% in recipients with sICAM-1 >308 ng/mL (Figure 5⇓).
Multivariate analysis indicated that sICAM-1 was a significant risk factor for CAD and graft failure after adjustment for demographic and clinical parameters found to be different between the 2 groups (Table 4⇓); therefore, our results were unlikely to be due to confounding effects. Recipients with sICAM-1 >308 ng/mL during the first 3 months after transplantation were at significantly greater risk of CAD (relative risk, 2.67; 95% CI, 1.28 to 5.59) and graft failure (relative risk, 3.63; 95% CI, 1.05 to 12.5) than recipients with sICAM-1 remaining ≤308 ng/mL.
We found that elevated serum sICAM-1 levels (>308 ng/mL) during the first 3 months after transplantation were associated with ICAM-1 in graft arterial endothelium and with subsequent development of transplant CAD, posttransplant ischemic events, and allograft failure (primarily associated with development of CAD). Immunopathological studies show increased expression of adhesion molecules in atherosclerotic plaques,7 and adhesion molecules are implicated in transplant CAD.7 8 Prevention of disease with monoclonal antibodies to ICAM-18 and reduction in neointimal formation in a carotid artery injury model of P-selectin–deficient mice7 support the involvement of adhesion molecules in developing atherosclerosis-like lesions. The association between atherosclerosis and increased soluble adhesion molecules also suggests that they are involved in arterial disease.4 6 Recent studies demonstrated that ICAM-1 has significant prognostic value for patients at risk of developing transplant CAD,5 cardiac allograft failure,5 or myocardial infarction.6
Adhesion molecule–mediated leukocyte–endothelial cell interactions were implicated in the development of spontaneous or transplant CAD.5 Cytokines could further facilitate expression of endothelial adhesion molecules, because increased expression of cell adhesion molecules or increased concentrations of soluble adhesion molecules have been related to graft rejection.13 However, this relationship is controversial.14 Our findings suggest that upregulation of ICAM-1 in cardiac transplant recipients is not caused by cellular rejection, because recipients with increased sICAM-1 did not have more rejection episodes. Adhesion molecules could be upregulated by preformed anti-endothelial antibodies15 associated with development of transplant CAD.16
Increased sICAM-1 could be associated with a viral infection. Virus-infected endothelial cells enhance ICAM-1 expression,17 and cytomegalovirus disease is associated with development of transplant CAD.1 However, other investigators showed lack of association between cytomegalovirus disease and transplant CAD.18 Although our data showed a relationship between sICAM-1 and cytomegalovirus disease, multivariate analysis failed to show an association between cytomegalovirus disease and CAD or graft failure. sICAM-1 increases with age19 and decreases with the presence of estradiol.20 However, donor age and recipient sex are probably not major factors in our population because (1) no difference in vascular ICAM-1 was found in donor hearts before transplantation, (2) a decreased male/female ratio was found in the group with higher sICAM-1 concentrations, and (3) multivariate analyses showed that sICAM-1 was still an independent significant risk factor for CAD and graft failure after adjustment for these variables.
Endothelial adhesion molecules could be upregulated by ischemia and reperfusion. This is supported by the fact that 62.9% of the recipients who had sICAM-1 >308 ng/mL already had elevated sICAM-1 at 7 to 10 days after transplantation. Tissue hypoxia enhances induction of ICAM-1 in human endothelial cells, and these changes are inhibited with anti–ICAM-1 antibodies or antisense oligodeoxynucleo-tides.21 Furthermore, increased endothelial cell–surface and soluble adhesion molecules are found after ischemia and reperfusion,22 23 and blocking expression of these molecules significantly improves allograft outcome.8
Basal determinations for donor sICAM-1 were not available to determine whether donors differed in sICAM-1 before transplantation, because hypotensive brain death induces endothelial ICAM-1 expression.24 However, endomyocardial biopsies obtained from all donor hearts before transplantation showed similar immunohistochemical levels of ICAM-1 irrespective of type of donor brain death. Because samples were stored at −75°C for up to 10 years, we cannot exclude the possibility of protein degradation. However, no negative correlations were found between specimen age and either sICAM-1 values or the likelihood of having sICAM >308 ng/mL. The unavailability of a continuous evaluation of the status of the coronary arteries since transplantation precludes the precise determination of the time course for association between sICAM-1 and CAD. However, the finding of increased sICAM-1 at 7 to 10 days after transplantation in a significant proportion of recipients who subsequently developed CAD suggests that ICAM-1 upregulation precedes CAD. The lack of ultrasonographic studies in all recipients could be another limiting factor. However, although intravascular ultrasonography is more sensitive for detecting CAD in epicardial arteries,1 this technique lacks accessibility to peripheral arteries, which can be assessed by coronary angiography.10 Whatever the technique used, the measurement error would be similar in all recipients.
Our findings suggest that sICAM-1 could derive from graft endothelium. It could be released by proteolytic cleavage by neutrophil elastase.25 This is particularly relevant because cardiac allografts with myocardial cell damage during the immediate posttransplant period show neutrophil infiltration.10 The demonstration that antibodies or antisense oligodeoxynucleotides to adhesion molecules attenuate ischemia-reperfusion injury and subsequent development of CAD26 and the beneficial effect of vitamin E27 or salicylates28 on reperfusion injury lesions by downregulating expression of adhesion molecules suggests that antiadhesion therapies may provide a novel approach to prevent transplant CAD. Finally, our data raise the possibility that soluble adhesion molecules can serve as molecular markers to assess risk for transplant CAD and graft failure. Although the clinical implications for sICAM-1 are uncertain, closer surveillance for transplant CAD in heart transplant recipients with elevated sICAM-1 may be warranted. This is particularly important when we consider that these determinations can be performed by use of a simple test that is low-risk, inexpensive, and convenient for the patients.
This research was supported by Methodist Research Institute, Clarian Health Partners; the Showalter Foundation, Indianapolis, Ind; and the American Heart Association (Midwest affiliate). The authors thank Sandra Mitchell, Alicia Currin, and Sandra Lemmon for their assistance; Roula Antonopoulos for excellent technical assistance; Beatriz Labarrere for her invaluable help; and Karen Spear, Judie Aitken, Jennifer Shaver, and Heather Richardson for editing and typing the manuscript.
Reprint requests to Dr Carlos A. Labarrere, Methodist Research Institute, Clarian Health Partners (Methodist, Indiana University, Riley Hospitals), 1812 N Capitol Ave, Indianapolis, IN 46202.
- Received March 8, 2000.
- Revision received May 3, 2000.
- Accepted May 4, 2000.
- Copyright © 2000 by American Heart Association
Hwang SJ, Ballantyne CM, Sharrett AR, et al. Circulating adhesion molecules VCAM-1, ICAM-1, and E-selectin in carotid atherosclerosis and incident coronary heart disease cases: the Atherosclerosis Risk In Communities (ARIC) study. Circulation. 1997;96:4219–4225.
Fuster V, Poon M, Willerson JT. Learning from the transgenic mouse: endothelium, adhesive molecules, and neointimal formation. Circulation. 1998;97:16–18.
Suzuki J, Isobe M, Yamazaki S, et al. Inhibition of accelerated coronary atherosclerosis with short-term blockade of intercellular adhesion molecule-1 and lymphocyte function-associated antigen-1 in a heterotopic murine model of heart transplantation. J Heart Lung Transplant. 1997;16:1141–1148.
Tanio JW, Basu CB, Albelda SM, et al. Differential expression of the cell adhesion molecules ICAM-1, VCAM-1, and E-selectin in normal and post-transplantation myocardium: cell adhesion molecule expression in human cardiac allografts. Circulation. 1994;89:1760–1768.
Waldman WJ, Knight DA, Huang EH. An in vitro model of T cell activation by autologous cytomegalovirus (CMV)-infected human adult endothelial cells: contribution of CMV-enhanced endothelial ICAM-1. J Immunol. 1998;160:3143–3151.
Rohde LE, Hennekens CH, Ridker PM. Cross-sectional study of soluble intercellular adhesion molecule-1 and cardiovascular risk factors in apparently healthy men. Arterioscler Thromb Vasc Biol. 1999;19:1595–1599.
Zund G, Uezono S, Stahl GL, et al. Hypoxia enhances induction of endothelial ICAM-1: role for metabolic acidosis and proteasomes. Am J Physiol. 1997;273:C1571–C1580.
Kalawski R, Bugajski P, Smielecki J, et al. Soluble adhesion molecules in reperfusion during coronary bypass grafting. Eur J Cardiothorac Surg. 1998;14:290–295.
Champagne B, Tremblay P, Cantin A, et al. Proteolytic cleavage of ICAM-1 by human neutrophil elastase. J Immunol. 1998;161:6398–6405.
Zund G, Dzus AL, Pretre R, et al. Endothelial cell injury in cardiac surgery: salicylate may be protective by reducing expression of endothelial adhesion molecules. Eur J Cardiothorac Surg. 1998;13:293–297.