Author + information
- Received November 17, 2008
- Revision received February 13, 2009
- Accepted February 19, 2009
- Published online May 19, 2009.
- Scott R. Auerbach, MD*,
- Cedric Manlhiot, BS§,
- Sushma Reddy, MD*,
- Caroline Kinnear, MS*,
- Marc E. Richmond, MD*,
- Dorota Gruber, MS*,
- Brian W. McCrindle, MD, MPH§,
- Liyong Deng, MS‡,
- Jonathan M. Chen, MD†,
- Linda J. Addonizio, MD*,
- Wendy K. Chung, MD‡ and
- Seema Mital, MD§,* ()
- ↵*Reprint requests and correspondence:
Dr. Seema Mital, Division of Cardiology, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada
Objectives This study sought to investigate the influence of recipient renin-angiotensin-aldosterone system (RAAS) genotype on cardiac function, rejection, and outcomes after heart transplantation.
Background The RAAS influences cardiac function and up-regulates inflammatory/immune pathways. Little is known about the effect of recipient RAAS polymorphisms in pediatric cardiac transplantation.
Methods Patients <25 years of age, after cardiac transplantation, were enrolled (2003 to 2008) and genotyped for polymorphisms in genes associated with RAAS upregulation: AGT-G, ACE-D, AGTR1-C, CYP11B2-G, and CMA-A. Presence of at least 1 high-risk allele was defined as a high-risk genotype. Univariable and multivariable associations between genotypes and outcomes were assessed in time-dependent models using survival, logistic, or linear regression models. Biopsy samples were immunostained for interleukin (IL)-6, transforming growth factor (TGF)-β, and tumor necrosis factor (TNF)-α during rejection and quiescence.
Results A total of 145 patients were studied, 103 primary cohort and 42 replication cohort; 81% had rejection, 51% had graft dysfunction, and 13% had vasculopathy, 7% died and 8% underwent re-transplantation. A higher number of homozygous high-risk RAAS genotypes was associated with a higher risk of graft dysfunction (hazard ratio [HR]: 1.5, p = 0.02) and a higher probability of death (HR: 2.5, p = 0.04). The number of heterozygous high-risk RAAS genotypes was associated with frequency of rejection (+0.096 events/year, p < 0.001) and rejection-associated graft dysfunction (+0.37 events/year, p = 0.002). IL-6 and TGF-β were markedly upregulated during rejection in patients with ≥2 high-risk RAAS genotypes.
Conclusions Recipient RAAS polymorphisms are associated with a higher risk of rejection, graft cytokine expression, graft dysfunction, and a higher mortality after cardiac transplantation. This may have implications for use of RAAS inhibitors in high-risk patients after transplantation.
In recent years, survival after pediatric heart transplantation has improved to >90% at 1 year and >70% at 5 years (1). Despite improved survival and quality of life, post-transplantation cardiac function is not normal. Almost 30% of patients show evidence of cardiac dysfunction by 10 years after transplantation (2), accounting for up to 20% of deaths in children after transplantation and causing significant morbidity in survivors (1).
The causes of graft dysfunction are many and include systemic hypertension, allograft rejection, and graft vasculopathy. The latter two are the leading causes of mortality (1,3). Neurohormonal systems may play an important role in the pathogenesis of graft dysfunction. A major such system regulating cardiac function and hemodynamics is the renin-angiotensin-aldosterone system (RAAS). Activation of the RAAS causes peripheral vasoconstriction and also acts at the cellular level to cause cardiomyocyte hypertrophy, apoptosis, and fibrosis (4–6). In addition, RAAS activation is associated with up-regulation of cytokines and inflammatory pathways that may increase the risk of rejection and exacerbate immune-mediated myocardial injury (7–10). The different steps in the RAAS signaling pathways are regulated by individual genes, which are polymorphic. Polymorphisms that cause up-regulation of the RAAS have been associated with cardiac dysfunction and heart failure in previous studies in both children and adults (6,11,12). In addition, some polymorphisms in the RAAS pathway genes have been associated with a higher incidence of graft vasculopathy in the adult cardiac transplantation population (13,14). Whether this is related to increased cytokine expression in the graft is not known. The effect of RAAS polymorphisms on cardiac allograft function, cytokine expression, and outcomes in children has not been previously studied. Also, although most studies have evaluated the effect of polymorphisms in individual RAAS genes, few have evaluated the effect of multiple RAAS polymorphisms in the same patient (6,15). The purpose of this study was to determine the influence of recipient RAAS genotype on rejection, graft cytokine expression, graft dysfunction, graft vasculopathy, and survival in children after cardiac transplantation.
This was a cohort study of patients <25 years of age with a history of cardiac transplantation who were followed up in the pediatric cardiac transplantation clinic at 2 institutions: Morgan Stanley Children's Hospital of New York Presbyterian between 2003 and 2006 (primary cohort, n = 103) and Hospital for Sick Children, Toronto, between 2007 and 2008 (replication cohort, n = 42). The study protocol was approved by the Institutional Review Board of Columbia University Medical Center and the Hospital for Sick Children. Written informed consent was obtained from all subjects before enrollment.
Clinical data were obtained from a review of the medical records. All patients underwent serial echocardiography and cardiac catheterization with endomyocardial biopsy every 3 to 6 months as part of surveillance monitoring for rejection, as well as annual coronary angiograms to assess graft vasculopathy. Data on the age at listing; sex; race/ethnicity; pre-transplantation diagnosis; date of transplantation; need for pre- or post-transplantation mechanical circulatory support, including extracorporeal membrane oxygenation or ventricular assist device; date of death; the number of episodes of rejection; use of angiotensin-converting enzyme inhibitors or β-blockers; and echocardiographic variables were obtained from the medical records. Hemodynamic data obtained at the time of cardiac catheterization included central venous pressure, pulmonary artery pressure, pulmonary capillary wedge pressure, cardiac index, and indexed systemic vascular resistance. Rejection was defined as International Society of Heart Lung Transplantation grade 1B or greater (advanced 1R or higher per new classification) (16) or clinical rejection. Graft dysfunction was defined as a shortening fraction <28%, pulmonary capillary wedge pressure of >12 mm Hg, or a cardiac index of <2.5 l/min/m2. Studies have shown that a pulmonary capillary wedge pressure of >12 mm Hg can be prognostic for both rejection and graft vasculopathy (17,18). Graft vasculopathy was defined as pruning or stenosis on coronary angiography and was based on the Stanford classification system (19).
A blood or saliva sample was obtained for deoxyribonucleic acid extraction. Patients were genotyped for polymorphisms in genes associated with RAAS up-regulation. The following polymorphisms were evaluated: 1) a M235T missense variant in angiotensinogen (AGT); 2) a deletion variant of angiotensin-converting enzyme (ACE), in intron 16; 3) an A/G polymorphism at nucleotide 1,903 of the cardiac chymase a gene (CMA); 4) an A/C polymorphism at nucleotide 1,666 of angiotensin II type 1 receptor (AGTR1); and 5) a C/T polymorphism at nucleotide 344 in aldosterone synthase (CYP11B2)(20). The clinicians caring for the patients were unaware of the patients' genotypes. Candidate gene polymorphisms in the RAAS neurohormonal axis were selected based on previous association studies, known functional effects of the polymorphisms, and allele frequency in the population (4,11,15). The RAAS genotypes were determined in all subjects by pyrosequencing assays for AGTR1, CYP11B2, AGT, and CMAand electrophoresis of polymerase chain reaction products for the ACEassay as previously described (6). Primer sequences are shown in Table 1.
Immunostaining of endomyocardial biopsy samples for cytokine expression
Paired biopsy samples obtained from a subset of 31 patients during episodes of rejection and quiescence were immunostained for cytokine expression. Paraffin-embedded sections of myocardium were immunostained using a rabbit polyclonal to interleukin (IL)-6 (1:400), a mouse monoclonal to transforming growth factor (TGF)-β (1:30), and a rabbit polyclonal to tumor necrosis factor (TNF)-α (1:150) (all from Abcam, Cambridge, Massachusetts) as primary antibodies. A horseradish peroxidase conjugated streptavidin and diamino benzidine method (Histostain kit, Zymed Laboratories, Invitrogen, Carlsbad, California) was used according to the manufacturer's protocol. IL-6, TGF-β, and TNF-α staining were visualized in tissue sections at 200× magnification. For the purpose of this comparison, high-risk was defined as the presence of ≥2 homozygous high-risk RAAS genotypes (1 of which was AGT), and low-risk as <2 homozygous high-risk RAAS genotypes. The distribution and intensity of staining was compared qualitatively between the high- and low-risk groups.
Data are presented as means with standard deviation, medians with minimum and maximum, and frequencies as appropriate. Time-dependent clinical outcomes (death, graft vasculopathy, and graft dysfunction) were modeled with Kaplan-Meier, nonparametric survival models, whereas rejection episodes with or without graft dysfunction were expressed as number of events per patient-year. Effect of the high-risk allele was evaluated in both the heterozygous and homozygous forms (additive model). Univariable associations between individual gene polymorphisms and frequency of rejection and frequency of rejection with graft dysfunction were assessed in linear regression models, whereas associations with freedom from cardiac dysfunction, graft vasculopathy, retransplantation, and death were assessed in survival models. Individual gene polymorphisms found to be significantly associated with outcomes in univariable models were then included in multivariable analyses to obtain a final model for each outcome. Parameter estimates of associations between outcomes and individual RAAS genes (from the univariable analysis) were converted from the original regression models to odds ratio or hazard ratio (HR) as appropriate to allow for meaningful comparison across outcomes. Exponential mathematical transformations were performed on frequency of rejection and frequency of rejection associated with graft dysfunction to account for the skewed distribution of those outcomes (Shapiro-Wilk test for transformed data p > 0.05 in both cases). To facilitate interpretation of exponentially transformed data, reverse-transformation was used to obtain mean differences and 95% confidence intervals (CIs). A replication cohort of 42 patients from the Hospital for Sick Children enrolled from 2007 to 2008 underwent RAAS genotyping to assess the effect of the RAAS genotype on graft function. All statistical analyses were performed using SAS Statistical Software version.9.1 (SAS Institute Inc., Cary, North Carolina).
Of 222 patients who underwent transplantation, 103 survivors of transplantation who were younger than 25 years of age and followed up at our institution were enrolled in the study as the primary cohort. Year of transplantation ranged from 1989 to 2006, with a mean follow-up of 5.7 ± 3.8 years. Enrollment occurred at variable times after transplantation. Maintenance immunosuppression consisted of a combination 2 to 6 mg/kg/day cyclosporine that was titrated to maintain appropriate serum cyclosporine levels. Patients were switched from cyclosporine to tacrolimus (0.05 to 0.1 mg/kg/day) if rejection was noted in the setting of therapeutic cyclosporine levels. Antiproliferative agents included azathioprine (1 to 2 mg/kg/day) or mycophenolate mofetil (25 to 50 mg/kg/day). Prednisone was initiated at 1 to 2 mg/kg/day and then tapered. Prednisone was discontinued by 1 year post-transplantation if rejection free. Induction therapy was not routinely used at our institution during the study period. Demographics and patient characteristics are listed in Table 2,and echocardiographic and hemodynamic data at last follow-up are listed in Table 3.During the follow-up period, 81% of patients had rejection (34 had grade 1B, 2, or 2R rejection, 37 had grade ≥3A or 3R rejection, and 14 had clinical rejection). Graft dysfunction developed in 51 patients (51%), graft vasculopathy developed in 11 (12%), death occurred in 7 (7%), and retransplantation was performed in 8 (8%). Frequency of rejection episodes per year was 1.05 ± 0.93. Frequency of rejection episodes with graft dysfunction was 0.44 ± 0.61 per year.
At 5 years post-transplantation, 24% patients were receiving angiotensin-converting enzyme inhibitors or angiotensin receptor blockers and 5% were receiving β-blockers. There was no difference in angiotensin-converting enzyme inhibitor or β-blocker use across genotypes. Kaplan-Meier survival curves for graft dysfunction, graft vasculopathy, and survival for the entire cohort are shown in Figures 1A,1B, and 1C, respectively.
Frequency of RAAS polymorphisms
The high-risk allele frequency was: ACE51%, AGTR123%, AGT59%, CYP11B235%, and CMA42%. All genotypes were in Hardy-Weinberg equilibrium; the distribution of genotypes is shown in Table 4.There was no difference in genotype frequencies based on race, sex, or diagnosis. The frequency of at least 1 homozygous high-risk RAAS allele was 73%, with a median of 1 homozygous high-risk gene present.
Cumulative effect of multiple high-risk RAAS genotypes on outcomes
Initial analysis assessed the cumulative effect of multiple high-risk RAAS genes on outcomes. Freedom from graft dysfunction was 75% at 1 year and 57% at 5 years. The presence of left ventricular hypertrophy and severity of left ventricular hypertrophy at 1 month after transplantation was not associated with graft dysfunction (parameter estimate: −0.137, SE: 0.450, p = 0.76, HR: 0.87, 95% CI: 0.36 to 2.11, and parameter estimate: −0.057, SE: 0.128, p = 0.66, HR: 0.94, 95% CI: 0.73 to 1.21, respectively). Freedom from graft dysfunction for patients with no high-risk RAAS genotype versus patients with at least 1 high-risk RAAS genotype is shown in Figure 2A.A higher number of high-risk RAAS genotypes was significantly associated with a higher risk of graft dysfunction (HR: 1.50, 95% CI: 1.08 to 2.07, p = 0.02) (Fig. 2B), higher frequency of rejection (+0.10 episodes/year, 95% CI: 0.05 to 0.15, p < 0.001) (Fig. 3A),higher frequency of rejection episodes with graft dysfunction (+0.37 episodes/year, 95% CI: 0.12 to 0.70. p = 0.002) (Fig. 3B), and higher risk of death (HR: 2.53, 95% CI: 1.08 to 5.94. p = 0.04). The RAAS genotype did not influence the severity of rejection. However, subjects with the high-risk genotypes were more likely to have a graft dysfunction develop during an episode of rejection, independent of the severity of rejection. There was no association between the number of high-risk RAAS genotypes and grade of rejection, graft vasculopathy, or retransplantation. Although the association with frequency of rejection episodes was observed even in the presence of a single high-risk allele, that is, heterozygous high-risk genotypes, the association with graft dysfunction and survival was seen only with the homozygous high-risk genotypes, suggesting a gene-dosing effect.
Effect of individual high-risk RAAS genotypes on outcomes
Associations between individual RAAS genotypes and clinical outcomes are shown in Figure 4.On univariable analysis, polymorphisms in individual RAAS genes were associated with several adverse outcomes, including rejection, graft dysfunction, and death. On multivariable analysis, AGTwas most consistently associated with adverse outcomes, including a higher risk of graft dysfunction, rejection, and rejection with graft dysfunction. In addition, AGTwas associated with a markedly higher risk of graft vasculopathy, with an HR of 4.7 (95% CI: 1.12 to 9.61, p = 0.04). Of the other genotypes, there was an association between CMAand rejection with graft dysfunction in both univariable and multivariable analyses.
Graft cytokine expression
Patients with ≥2 high-risk RAAS genotypes, all of whom were also homozygous for the AGTpolymorphism (n = 16), showed significant up-regulation of IL-6 and TGF-β during rejection episodes compared with patients with <2 high-risk genotypes (n = 15) (Figs. 5Aand 5B). Expression of TNF-α did not change significantly during rejection in any group (Fig. 5C).
The characteristics of the replication cohort are shown in Table 5.The allele frequencies were in Hardy-Weinberg equilibrium. The high-risk gene frequencies were as follows: ACE62%, AGT41%, AGTR23%, CYP11B267%, and CMA43%. Like the primary cohort, a higher number of high-risk RAAS genes was associated with a greater likelihood of graft dysfunction (HR: 1.3, 95% CI: 1.0 to 1.8, p = 0.05). This association was strongest with the AGTRgenotype (HR: 2.3, 95% CI: 1.1 to 4.8, p = 0.03). The AGThigh-risk genotype was associated with a higher risk of graft dysfunction, but this did not reach statistical significance (HR: 1.6, 95% CI: 0.9 to 2.6, p = 0.09).
Although outcomes after cardiac transplantation continue to improve, there remains a high rate of graft failure, dysfunction, and retransplantation. With over 3,000 children receiving transplants in the past 10 years (1), there is a significant health burden associated with graft dysfunction and need for retransplantation in the context of limited donor availability. As we begin to understand how genetics plays a role in heart disease, it is becoming clear that risk stratification based on genotyping is possible and potentially beneficial. This is the first study to identify recipient genetic predictors of graft function and survival after pediatric cardiac transplantation as well as to provide a potential mechanism for genetic risk. A larger number of RAAS high-risk polymorphisms was associated with a higher incidence of graft dysfunction, rejection, and rejection associated with graft dysfunction, as well as lower survival. Polymorphisms in the AGTgroup were most highly associated with adverse outcomes after cardiac transplantation. The association of both heterozygous and homozygous carriers of the high-risk genotypes with adverse outcomes suggests a multiplicative effect of high-risk alleles. The observation of a cumulative effect of multiple RAAS genes highlights the importance of studying genes that do not just regulate a single step, but that act additively or synergistically by influencing several steps in a biologic pathway. This is also the first study to show up-regulation of cytokine expression in the grafts of patients with a high-risk RAAS genotype, providing a mechanistic link between the RAAS and the immune system.
Although the adverse effects of chronic RAAS activation on myocyte hypertrophy, apoptosis, and fibrosis are well known, this is the first study to report an association between the RAAS genotype and an increased susceptibility to rejection and graft vasculopathy in the donor heart after cardiac transplantation. This association between high-risk RAAS genotypes and a higher incidence of graft dysfunction occurred both during rejection and in the absence of rejection, and was likely related to cytokine up-regulation. Previous studies have reported an association between RAAS activation and induction of inflammatory cytokines, most notably TGF-β, IL-6, and TNF-α in other conditions. Upregulation of TGF-β in response to increased angiotensin II increases cardiac fibrosis, extracellular matrix deposition, and myocyte hypertrophy (8–10). Other studies have reported that increased IL-6 and TNF-α expression is associated with an increased incidence of allograft rejection and graft vasculopathy (8,21,22). The up-regulation of IL-6 and TGF-β during rejection in patients with a higher number of high-risk RAAS genotypes, in particular AGT, provides evidence of a link between the immune system and the RAAS; AGTwas most consistently associated with adverse outcomes, and the increased levels of IL-6 and TGF-β activity in these patients may be responsible for both rejection and immune-mediated myocardial injury, resulting in graft dysfunction and graft vasculopathy. This association supports the idea that increased susceptibility of the transplanted heart to graft dysfunction during rejection in patients with the high-risk RAAS genotypes may be related to increased expression of these cytokines and adverse cardiac morphometric changes. Increased expression of cytokines may lead to a greater degree of myocyte damage independent of the severity of cellular infiltration of the myocardium. Cytokine up-regulation in biopsy samples may provide a new tissue biomarker for identifying patients at risk for graft dysfunction as well as patients likely to benefit from cytokine inhibition. The possibility of a haplotype effect related to an interaction between RAAS genes and other immune-response-encoding genes requires further investigation.
Although AGTpolymorphisms have been shown to be associated with elevated levels of circulating angiotensinogen, cardiac hypertrophy, and a higher prevalence of heart failure, hypertension, and atherosclerosis (23–26), there are limited data on the association of RAAS genotypes with post-transplantation outcomes. The ACEand AGTR1high-risk genotypes have previously been associated with graft vasculopathy, but we know of no other reports of association of AGTwith graft vasculopathy (13,14,27). Although overall results of studies of AGTin other solid organ transplants have been conflicting, 2 studies have shown associations between the AGTand both chronic allograft nephropathy and a shorter time to renal graft loss (28,29). Chronic allograft nephropathy is histologically characterized by vasculopathy, glomerulopathy, tubular atrophy, and interstitial fibrosis, with vascular lesions showing smooth muscle cell proliferation and intimal thickening (30). Therefore, the pathogenesis of chronic allograft nephropathy may be similar to that of cardiac allograft vasculopathy, with both conditions being secondary to immune-mediated inflammatory processes. Further study of the association between AGTand graft vasculopathy is warranted.
The results from the replication cohort support the results from the original cohort by showing an increased likelihood of graft dysfunction with a greater number of higher-risk RAAS genotypes. The smaller sample size and the shorter follow-up time resulted in bias toward survivors; therefore, we only evaluated the association of RAAS genotypes with graft dysfunction and not with mortality or other outcomes.
Our findings suggest that knowledge of the RAAS genotype can be used to identify patients at risk for rejection, graft dysfunction, graft vasculopathy, and early mortality after transplantation. This knowledge may be used in the future to tailor monitoring as well as treatment based on the genotype. Studies in adults with heart failure reported that survival can be improved in the high-risk ACE DD(deletion/deletion) subset of patients by using high-dose angiotensin-converting enzyme inhibitors compared with standard doses (12). The response to angiotensin receptor blockers in hypertension is also influenced by the RAAS genotype (31). Whether use of high-dose RAAS inhibitors in patients with a high-risk RAAS genotype can improve outcomes after transplantation warrants further investigation. Consideration should be given to the use of RAAS inhibitors as a form of long-term immunosuppression.
There were several limitations in the study. The transplanted heart may likely have a different genotype than the recipient, which may independently affect graft function. Donor genotype was not known in our study. Therefore, the effect of graft genotype and the potential for interaction between the recipient and graft genotypes was not assessed. The sample size did not permit us to investigate the role of variations in genes that encode the immune pathway that may act synergistically with the RAAS genes or be inherited with the RAAS genes in a haplotype model. Furthermore, the study was not powered to detect an interaction between RAAS genotype and effect of angiotensin-converting enzyme inhibition. The cohort study design began enrolment in the last 5 years; therefore, the study population was biased toward long-term survivors of transplantation.
Patients with a higher number of high-risk RAAS gene alleles are more likely to experience adverse outcomes after cardiac transplantation. Early identification of RAAS genotypes in transplantation recipients may be beneficial for risk stratification, surveillance, and potentially, tailoring treatment regimens.
Dr. McCrindle is a consultant (minor) for Roche and Abbott.
This work was presented in part at the American Academy of Pediatrics National Conference and Exhibition, October 26, 2007, and American Heart Association Annual Meeting, November 5, 2007.
- Abbreviations and Acronyms
- angiotensin-converting enzyme
- angiotensinogen M235T
- angiotensin II type 1 receptor
- confidence interval
- cardiac chymase A
- aldosterone synthase
- hazard ratio
- renin-angiotensin-aldosterone system
- transforming growth factor
- tumor necrosis factor
- Received November 17, 2008.
- Revision received February 13, 2009.
- Accepted February 19, 2009.
- American College of Cardiology Foundation
- Hao J.,
- Wang B.,
- Jones S.C.,
- Jassal D.S.,
- Dixon I.M.C.
- Abdallah A.N.,
- Billes M.A.,
- Attia Y.,
- Doutremepuich C.,
- Cassaigne A.,
- Iron A.
- Kurland L.M.H.,
- Karlsson J.,
- Kahan T.,
- et al.
- McNamara D.M.,
- Holubkov R.,
- Postava L.,
- et al.
- Yousufuddin M.H.S.,
- Starling R.C.,
- Tuzcu E.M.,
- et al.
- Ortlepp J.R.,
- Vosberg H.P.,
- Reith S.,
- et al.
- Eun L.Y.,
- Gajarski R.J.,
- Graziano J.N.,
- Ensing G.J.
- Gao S.Z.,
- Alderman E.L.,
- Schroeder J.S.,
- Silverman J.F.,
- Hunt S.A.
- Perkins M.J.,
- Van Driest S.L.,
- Ellsworth E.G.,
- et al.
- van Rijn M.J.E.,
- Bos M.J.,
- Isaacs A.,
- et al.
- Cunningham D.A.,
- Crisp S.J.,
- Barbir M.,
- Lazem F.,
- Dunn M.J.,
- Yacoub M.H.
- Weis M.,
- von Scheidt W.