Adrenergic Receptor Polymorphism and Maximal Exercise Capacity after Orthotopic Heart Transplantation

Background Maximal exercise capacity after heart transplantion (HTx) is reduced to the 50–70% level of healthy controls when assessed by cardiopulmonary exercise testing (CPET) despite of normal left ventricular function of the donor heart. This study investigates the role of donor heart β1 and β2- adrenergic receptor (AR) polymorphisms for maximal exercise capacity after orthotopic HTx. Methods CPET measured peak VO2 as outcome parameter for maximal exercise in HTx recipients ≥9 months and ≤4 years post-transplant (n = 41; mean peak VO2: 57±15% of predicted value). Donor hearts were genotyped for polymorphisms of the β1-AR (Ser49Gly, Arg389Gly) and the β2-AR (Arg16Gly, Gln27Glu). Circumferential shortening of the left ventricle was measured using magnetic resonance based CSPAMM tagging. Results Peak VO2 was higher in donor hearts expressing the β1-Ser49Ser alleles when compared with β1-Gly49 carriers (60±15% vs. 47±10% of the predicted value; p = 0.015), and by trend in cardiac allografts with the β1-AR Gly389Gly vs. β1-Arg389 (61±15% vs. 54±14%, p = 0.093). Peak VO2 was highest for the haplotype Ser49Ser-Gly389, and decreased progressively for Ser49Ser-Arg389Arg > 49Gly-389Gly > 49Gly-Arg389Arg (adjusted R2 = 0.56, p = 0.003). Peak VO2 was not different for the tested β2-AR polymorphisms. Independent predictors of peak VO2 (adjusted R2 = 0.55) were β1-AR Ser49Gly SNP (p = 0.005), heart rate increase (p = 0.016), and peak systolic blood pressure (p = 0.031). Left ventricular (LV) motion kinetics as measured by cardiac MRI CSPAMM tagging at rest was not different between carriers and non-carriers of the β1-AR Gly49allele. Conclusion Similar LV cardiac motion kinetics at rest in donor hearts carrying either β1-AR Gly49 or β1-Ser49Ser variant suggests exercise-induced desensitization and down-regulation of the β1-AR Gly49 variant as relevant pathomechanism for reduced peak VO2 in β1-AR Gly49 carriers.


Introduction
Cardiopulmonary exercise testing after orthotopic heart transplantation (HTx) shows that maximal exercise capacity is reduced to the 50-70% level of age-matched healthy controls despite of normal left ventricular (LV) ejection fraction [1]. Recipient age, peak heart rate and blood pressure, pulmonary artery resistance, diastolic LV function, BMI, and transplant vasculopathy explain altogether 51-66% of peak VO 2 variance after transplant [2,3] suggesting the relevance of other parameters.
Human cardiomyocytes predominantly express the β 1 -and β 2 -adrenergic receptor (AR) subtypes [4] which play a pivotal role for exercise-induced increase in cardiac function. In the healthy heart, the polymorphisms β 1 -AR Ser49Gly, β 1 -AR Arg389Gly, β 2 -AR Arg16Gly, β 2 -AR Gln27Glu, and β 2 -AR Thr164Ile show distinct cardiovascular responses to sympathetic activation [5][6][7][8]. However, aortic anastomosis disrupts postganglionic sympathetic innervation and fine-tuned modulation of β-AR activation at the level of the postganglionic nerve-cardiomyocyte synapsis. Consequently, adrenergic stimulation of donor heart function depends on the circulating catecholamine levels [9]. Postganglionic sympathetic fibers of the healthy heart extract large amounts of catecholamines from the circulation [9]. After HTx, postganglionic fibers degenerate, which results in reduced catecholamine retention of the donor heart [10,11] and 3-5 fold increased circulatory catecholamine levels [12,13]. We hypothesized that this unique clinical setting in the HTx recipient may change the characteristics of β-AR variant response to adrenergic stimulation because of the different sensitivity of the individual β-AR variant to downregulation when exposed to increased catecholamine concentration.
The outcome parameter of this correlational study was peak VO 2 measured by cardiopulmonary exercise testing. Peak heart rate and systolic blood pressure are variables that determine maximal exercise capacity after HTx [2,3]. Both variables depend on circulating catecholamines, therefore, all study patients were screened for recipient expression of α 2c -AR Del 322-325 variant, which strengthens activation of the adrenal gland chromaffin cell [14] with subsequent increased spillover of norepinephrine and epinephrine into the circulation [15].

Study design and population
This study complies with the Declaration of Helsinki and was approved by the ethical committee of the canton de Vaud. Written informed consent was obtained from all patients. This is a substudy of the prospective epidemiological cohort study following solid organ transplant recipients in Switzerland (Swiss Transplant Cohort Study, STCS No.0038). Inclusion criteria were stable adult HT recipients !9 months and <4 years post-transplant with a maximal cardiopulmonary exercise test (CPET) as defined by the achievement of a respiratory exchange ratio (RER) >1. 1 [16]. Exclusion criteria were age: 1. <18 years; 2. missing consent; 3. !moderate severity of comorbidity limiting execution of maximal CEPT; 4. presence of !moderate acute allograft rejection (International Society for Heart and Lung Transplantation grade !2R) [17] in the endomyocardial biopsy; 5. !moderate cardiac allograft vasculopathy in the coronary angiogram at the time of exercise testing; or 6. severe valvular dysfunction (insufficiency !III/IV or more than !moderate stenosis) as assessed by standard echocardiography performed by a board-certified cardiologist <1 month before CPET.

Demographic and clinical data
Demographic and clinical data at CPET were collected from electronic charts. Standard laboratory tests included sodium, potassium, creatinine, hemoglobin, and brain natriuretic peptide (BNP) level.

Maximal cardiopulmonary exercise testing
All patients underwent maximal CPET. Patients were tested using an electrically braked cycloergometer (Ergoline 900/911 digital, Ergoline GmbH, Germany) with individualized ramp protocol. Respiratory gas exchange was measured breath-by-breath with determination of peak VO 2 , VO 2 AT (VO 2 at anaerobic threshold), VE/VCO 2 (ventilatory efficiency), ventilatory reserve and RER. Maximal exercise capacity was determined by the highest VO 2 achieved during the last 30 seconds of maximal exercise. Heart rate, blood pressure, and 12-lead ECG were continuously recorded throughout exercise and recovery.

Circumferential strain measurement of the donor heart
Patients underwent cardiovascular magnetic resonance imaging in order to study the influence of the β 1 −49 genotype on LV function. All patients were scanned with a prototype slice followed balanced Steady State Free Precession (bSSFP) CSPAMM (Complementary Spatial Modulation of Magnetization) tagging technique [18,19]. All scans were performed on a clinical MAGNETOM Verio 3T scanner (Siemens Healthcare, Erlangen, Germany). Scout scans were acquired to find the short axis (SA) of the LV and place the tagged slices: for each exam, three SA slices were acquired at basal, mid-ventricular and apical levels of the LV. The basal slice was acquired 1 cm below the mitral valve, the apical slice 1 cm above the endomyocardial border of the apex, and the mid-ventricular slice was centered between the two other slices. The imaging parameters for the SF bSSFP CSPAMM sequence were as follows: TE/TR = 1.38/ 3.2ms, 12°radio frequency excitation angle, a bandwidth of 849 Hz/pixel, a temporal resolution of 45 ms, (82-85) Ã 256 matrix size, with 32 to 33% of phase resolution. The tagged slice thickness was 6 mm, while the imaged slice thickness was 20 mm, 25 and 30 mm, respectively, at apex, mid-ventricle and base in order to keep the tagged slice in the imaged slab at all times. Each slice was acquired in a 16 heartbeats breath hold. All tagged images were analyzed using Harmonic Phase Imaging (HARP) [20], in the Virtue software from Diagnosoft (HARP, v4.1, Diagnosoft Inc., Palo Alto, CA, USA). Circumferential strain measurements could be obtained from the analysis as an average over each slice. All obtained values were adapted to the duration of the systole [21] using a custom Matlab script (The Mathworks, Inc, Natick, MA, USA), in order to create a meaningful comparison. For comparison of circumferential strain measurements between the two β 1 −49 genotypes groups (see above) unpaired Student's t-tests corrected for samples of different sizes were used.

Polymorphism genotyping
Donor genomic DNA was extracted from paraffin-embedded endomyocardial biopsy specimen using Purelink genomic DNA kit (Invitrogen1). Extracted leucocyte DNA provided by the STCS biobank was used for recipient genomic DNA analysis. Detection of AR polymorphism used Taq-Man single-nucleotide polymorphism genotyping assays (Life Technologies1). SNP ID were: rs1801252 for β 1 -AR Ser49Gly, rs1801253 for β 1 -AR Arg389Gly, rs1042713 for β 2 -AR Arg16Gly, rs1042714 for β 2 -AR Gln27Glu and rs1800888 for β 2 -AR Thr164Ile. Genotype assignments were obtained by fluorescence measurement using an ABI Prism 7500 Sequence Detection System with its allelic discrimination software (Life Technologies1).
α 2C -AR polymorphism was examined after amplification of recipient genomic DNA by PCR. Primers for PCR were: 5'-ACGTGGAGCCGGACGAGA-3' (sense) and 5'-GTTCTTCC TGTCGCGCCG-3' (antisense). The PCR consisted of 5 ng of genomic DNA, 1 pmol of each primer, 0.2 mM dNTPs, 1 unit of Gotaq DNA polymerase (Promega1), 4 μl of 5X GoTaq buffer and 5% DMSO in a 20 μl reaction volume. Reactions were started by an initial incubation at 94°C for 4 min, followed by 35 cycles of 94°C for 30 s, 65°C for 30 s, and 72°C for 30 s, followed by a final extension at 72°C for 10 min. PCR products were digested with HaeIII (Life Technologies1) at 37°C for one hour since α 2C Del322-325 results in the loss of one of four HaeIII restriction sites in α 2C .

Statistical analysis
Allele frequencies were computed by standard gene-counting methods. Association of peak VO 2 with demographic and clinical parameters, or AR SNPs was tested using univariate regression analysis with peak VO 2 as the dependent variable. Peak VO 2 was expressed as the percentage of the predicted value (% predicted), which is already adjusted for age, BMI and gender [22]. Thus, these three co-variables were not included in the univariate and multivariate analysis. β 1 −49 genotype was considered as Gly carriers when homozygous or heterozygous, in accordance with the literature [5,23]. The mean peak VO 2 of each AR genotype suggested a dominant and a recessive model for the β 1 −389 and the β 2 genes, respectively. Thus, β 1 −389 genotype was considered as homozygous Arg389Arg or Gly389 carriers; β 2 genotype was considered homozygous Arg16Arg or Gly16 carriers and homozygous Glu27Glu or Gln27 carriers; and α 2C genotype was considered as homozygous WT or carriers of the deletion 322-325. Because of the low number of patients carrying the β 2 Ile164 variant (n = 2/41), this SNP did not enter into the final analysis.
Secondary clinical variables included continuous and categorical clinical variables. Categorical variables were defined as non-treated or treated by a certain drug. Cut-off for BNP and N-terminal propeptide BNP (NT-proBNP) levels were 100 ng/l and 300 ng/l, respectively, in accordance with the heart failure guidelines of the European Society of Cardiology. Continuous clinical data were expressed as mean±S.D.; a p-value <0.05 was considered as statistically significant.
Multivariable logistic regression using backward analysis of parameters correlating with p<0.10 was performed to estimate the association between explanatory variables and peak VO 2. Association of a β-AR genotype with peak VO 2 was considered statistically significant when p was <0.0125 since 4 β-AR SNPs entered the final analysis.
Patient characteristics S1 Table shows baseline characteristics of the 41 HTx recipients (6 females). Gender distribution in this present study corresponds the one reported in the registry of the International Society of Heart and Lung transplantation [24]. Mean age at HTx was 51±12 years with a mean time interval of 560±309 days between HTx and CPET. BMI was 25.6±4.6.
No study patient had histological signs of acute rejection in the endomyocardial biopsy obtained before CPET.

Maximal cardiopulmonary exercise testing
CPET parameters are shown in S2 Table. Mean peak VO 2 was 17.1±6.2 ml/kg/min corresponding to 57.0±14.8% of the predicted value adjusted for age, BMI and gender; the mean peak power output was 107±51 Watts. VO 2 AT was 12.0±4.1 ml/kg/min corresponding to 41±12% of the predicted value of AT. Ventilatory efficiency (measured by the VE/VCO 2 slope) was 33.2 ±5.7; peak ventilatory reserve was 43±16%. The RER was 1.27±0.14 at maximal exercise level.

Predictors of peak VO 2
Peak HR as well as Δsystolic BP were considered as variables depending on chronotropic reserve or peak systolic BP respectively, and were thus not included in the final model. Likewise, creatinine was considered as dependent of the eGFR. Multivariable analysis showed that the β 1 -AR Gly49 variant of the donor heart (p = 0.005), along with ΔHR (p = 0.016) and peak systolic BP (p = 0.031) were independently associated with peak VO 2 (adjusted R 2 = 0.55) ( Table 2). The β 1 -AR Gly49 polymorphism remained correlated to peak VO 2 when β-blocker treatment was forced into the model.

Circumferential strain measurement by tagging
Nine patients (3 women; age 46±14 y) were scanned. Circumferential strain measurements were obtained as described above; one slice with artifact was not included in the analysis (n = 1/27 slices). Patients were grouped into carriers of the β 1 -AR Gly49 variant (n = 4) and patients with the

Discussion
By showing that β-adrenergic receptor polymorphism in the donor heart is a determinant of peak VO 2 , the present study adds to the pathophysiological understanding of limited maximal exercise capacity in HTx recipients with normal left ventricular function at rest.

Study population
The hypothesis that polymorphism of the donor heart β-AR or the recipient α 2c -AR affects maximal exercise capacity after orthotopic HTx was tested in a prospective cohort study design including consecutive patients. HTx recipients were not included when transplanted less than 9 months ago because maximal exercise capacity increases after transplantation reaching a plateau only after the first 9 postoperative months [2,[30][31][32] [2,3,30,32,35]. In theory, the low-dose metoprolol treatment (in the present study: mean dose 24±10 mg/d) administered in almost one third of the study participants at the moment of CPET might decrease chronotropic reserve. However, multivariable analysis did not show interaction between β-blocker treatment and peak  The strain values were obtained using HARP analysis on tagged images acquired with a SF bSSFP CSPAMM tagging technique. All measurements were adapted to the systole duration of each exam. No statistically significant difference could be found between the two groups.

Adrenergic receptor polymorphism and orthotopic heart transplantation
In this study, the deletion variation of the α 2c -adrenoceptor did not interact with peak VO 2 , chronotropic reserve, systolic or diastolic BP suggesting that the minor allele is not associated with a distinct exercise-associated phenotype in HTx recipients. This interpretation is in accordance with the absence of a distinct phenotype of the minor allele in populations with other cardiovascular disease [5] but has to consider the small number of study participants carrying the deletion variant. However, the proportion of study patients with the minor allele compares to the allele frequency reported in other populations [26,27]. Frequencies of the donor heart β 1 -or β 2 -AR alleles correspond to respective reports from the California Donor Transplant Network [37] and the European population [27-29] suggesting the absence of a selection bias in the present cohort. Previous studies have shown that βadrenergic signal transduction in the cardiac allograft is altered as a consequence of a decreased G Sα expression [38] but this change does not explain the individual variation of maximal exercise capacity after HTx. Sustained agonist exposure may differently affect β-AR variant desensitization and down-regulation [39]. Therefore, we hypothesized that increased circulatory catecholamine levels in combination with the specific biochemical characteristics of the tested β-AR variants should explain interindividual variation of maximal exercise capacity after HTx. Especially, the β 1 -49Gly and the β 1 -Arg389 variants exhibit greater agonist-promoted desensitization when exposed to saturating catecholamine concentrations while the β 1 -Ser49 variant is resistant to agonist-promoted downregulation [25,40,41]. In concordance, the haplotype β 1 -Gly49/Arg389Arg shows more rapid agonist-promoted receptor down-regulation and desensitization in vitro when compared with other haplotypes [42]. In the present study, the β 1 -Ser49-Ser variant was associated with significantly higher peak VO 2 values when compared with the β 1 -AR Gly49 variant while there was no significant correlation with the β 1 -AR Arg389 or the various β 2 -AR variants tested. Moreover, the donor heart Ser49Ser+Gly389Gly haplotype was related with the highest peak VO 2 levels while peak VO 2 decreased progressively for the haplotypes Ser49Ser+Arg389Arg > 49Gly+389Gly > 49Gly+Arg389Arg. We thus identified β 1 -AR polymorphism at position 49 and the haplotype combination of β 1 -AR49 + β 1 -AR389 polymorphisms as independent predictors of exercise performance after orthotopic HTx.
Studies in patients with coronary artery disease or heart failure, however, have shown higher peak VO 2 values in carriers of either β 1 -AR 49Gly or β 1 -AR 389Arg variant, or the haplotype 49Gly/389Arg [43,44]. The appraisal of these contrasting results has to consider the 3-5 fold increased level of circulatory catecholamines in HTx recipients at rest, which increases even further with exercise. In the present study, baseline VO 2 , circumferential cardiac fiber shortening kinetics at rest, and chronotropic reserve were not different between carriers of the β 1 -AR Gly49 and β 1 -AR Ser49 variant suggesting that the lower peak VO 2 in β 1 -AR 49Gly carriers should relate to a relative decrease of myocardial contractility at peak exercise. This conclusion is compatible with the substantial down-regulation of β 1 -AR 49Gly membrane expression shown for cells exposed to high catecholamine levels (27). The β 1 -AR Ser49 variant, however, seems to maintain agonist promoted stimulation of myocardial contractility during peak in the HTx recipient because of its resistance to agonist-promoted downregulation (27). Finally, physiological studies in HTx recipients have shown the important role of the β 2 -AR for heart rate (36, 37), which can explain why β 1 -AR Gly49Ser variants impact on myocardial contractility at peak exercise without affecting peak heart rate.

Limitations of the study
This study collective includes only a small number of patients. However, demographic and clinical characteristics of the study population as well as distribution of allele frequencies are in accordance with reports from other populations. Despite of the fact that this study provides solid indirect evidence suggesting that exercise induces desensitization and down-regulation of the β 1 -AR Gly49 variant, direct proof is missing. However, proof of β-AR down-regulation with peak exercise in vivo or measurement of circumferential shortening at maximal exercise in the cardiac MRI is not feasible. Quantification of circumferential cardiac kinetics by echocardiography may be an alternative but may miss differences due to larger standard deviation.

Conclusion
Reduced exercise capacity remains a concern after HTx because more recent advance in immunosuppression permits long survival, which is why quality of life aspects such as daily physical activity gain importance. This study demonstrates that maximal exercise capacity as measured by peak VO 2 is reduced in HTx recipients carrying a donor heart expressing the β 1 -AR 49Gly variant. The results of the present study suggest that HTx recipients carrying this polymorphism in their donor heart should benefit from high-dose β 1 -AR blockade.