After an approval by the Human Ethical Committee of the Aachen University of Technology and informed consent of the caregivers or the patients, if applicable, 20 consecutive patients (seven females and 13 male) with a median age of 13.8 years (minimum: 3 months, maximum 26.5 years; interquartile range, IQR: 12.3 years) were enrolled in this prospective study. Eight patients were younger than 12 years and pre-pubertal, 12 were older than 12 years and had reached puberty. No patient was on estrogen/progestogen combination. Table 1 summarizes the characteristics of the subjects.

All surgical procedures were performed by the same pediatric cardiac surgeon. In all cases, conventional general anesthesia consisted of isoflurane and sufentanyl. Dexamethasone (1 mg/m² body surface area) was given before the sternotomy. Perioperative antibiotic prophylaxis was carried out with cefuroxime. Before the institution of a cardiopulmonary bypass (CPB), a right atrial biopsy was taken. After the institution of a moderate hypothermic low-flow CPB, the aorta was cross-clamped, and cardiac arrest was instituted by intra-aortal injection of 4 °C cold cardioplegic solution (Bretschneider, 30 ml/kg body weight) that was re-aspirated in the right atrium. After the intra-cardiac repair, the patient was weaned from CPB under progressive re-warming. Epicardiac pacemaker leads and pericardial and mediastinal drains were placed before chest closure.

The arterial blood pressure and central venous pressure were continuously monitored via an arterial and a central venous line, respectively.

Inotropic support, which consisted of dobutamine, was given to maintain a normal mean arterial blood pressure for age and volume therapy by injections of crystalloid solutions, if requested. The patient was transported to the intensive care unit where weaning from artificial ventilation was begun as early as possible. The ratio between the arterial partial pressure of oxygen (PaO₂) and the fraction of inspired oxygen (FiO₂) was used to assess oxygenation. Diuresis was continuously monitored via a bladder catheter, and water balance was calculated hourly.

The routinely performed laboratory investigations included the determination of blood gases, blood concentration of lactate, glycemia, complete blood count, serum creatinine, aspartate aminotransferase (AST), troponin T, and coagulation parameters and were measured at least 4 and 24 h post-operatively.

Biopsies taken for the detection of messenger ribonucleic acid (mRNA) were immediately snap-frozen in liquid nitrogen and stored at −80 °C until analysis.

Total ribonucleic acid (RNA) was extracted from the atrial myocardium by using the RNeasy kit (QIAGEN Inc., Hilden, Germany). The RNA (100 ng) was reverse-transcribed to complementary deoxyribonucleic acid (DNA) with random hexamers. A 2-μl cDNA sample was incubated with 20 μl of QuantiTect Mix containing fluorescence dye SYBR® Advantage® qPCR premix from Clontech (Takara Bio Inc. Otsu, Shiga, Japan).

The expression of target genes was normalized to the levels of 18S-mRNA and calculated with 2^−ΔCT. Besides the expression of mRNA coding for ERα and ERβ, the expression of mRNA coding for the pro-inflammatory cytokine tumor necrosis factor-a (TNFα), interleukin (IL)-1β, for the regulator of the acute phase reaction that shares pro- and anti-inflammatory properties, IL-6, for chemokine IL-8, for the anti-inflammatory cytokine IL-10, for the growth factor and major regulator of fibrosis tissue growth factor (TGF)-β, for the main growth factor of cardiomyocytes cardiotrophin (CT)-1, and for the early marker of myocardial stress brain natriuretic peptide (BNP) was quantified. The primers used are listed in Table 2.

Results are given as median and interquartile range (median, IQR). Data analysis was done using the Statistical Product and Service Solutions program, version 25 (SPSS®, IBM, USA). The non-normal distribution of data was verified, and non-parametric tests were applied. Correlations between two independent parameters were analyzed by using Spearman's rank correlation test, and the results were reported by Spearman's rank coefficient (r_s). The Mann–Whitney U-test was used to compare the clinical and biological parameters in two different groups. P ≤ 0.05 were considered significant, while p < 0.1 indicated a tendency toward significance.

The myocardial expression of mRNA coding for ERα and ERβ was detected in all patients, independently of gender, age, and achieved puberty, respectively. In the whole cohort, the expression of ERβ-mRNA was significantly higher than that of ERα-mRNA [4.89 (0.49) vs. 4.22 (0.40), p < 0.0001]. The myocardial expression of mRNA coding for BNP and for all tested cytokines was also detected in all patients and was not influenced by gender or age except that of TGF-β that was significantly higher in males than in females (p = 0.024) and correlated with age (r_s: 0.490, p = 0.028). TGF-β-mRNA was also higher in patients who had achieved puberty than in the others (p = 0.047).

The ERα-mRNA and ERβ-mRNA levels correlated with each other (r_s: 0.922, p < 0.0001) and with the levels of mRNA coding for inflammatory cytokines (TNFα, IL-1β, IL-6, IL-8, and IL-10), growth factors (CT-1 and TGF-β), and the marker of myocardial stress (BNP), respectively (Table 3). Figure 1 shows the correlations between the myocardial levels of ERα-mRNA and IL-6-mRNA and between ERα-mRNA and IL-10-mRNA, respectively. Figure 2 shows the correlation between the myocardial levels of ERα-mRNA and BNP-mRNA.

The outcome variables were not different in males than in females except the creatinine concentration at 4 and 24 h post-operatively that was lower in females (p = 0.007 and p = 0.01, respectively). The outcome variables were not correlated with age except the creatinine concentration at 4 and 24 h post-operatively (r_s: 0.812 and 0.804, respectively; p < 0. 0001, respectively).

The expression of ERα-mRNA correlated negatively with the troponin T concentration at 24 h after the operation (r_s:−0.505, p = 0.032) (Figure 3), positively with the ratio PaO₂/FiO₂ calculated at 4 h post-operatively (r_s: 0.453, p = 0.059), negatively with the AST levels measured 24 h post-operatively (r_s: −0.489, p = 0.047), and negatively with the water balance at 4 h post-operatively (r_s: −0.487, p = 0.048).

The patients were divided in two groups (n = 10 each) depending on whether their levels of ERα- and ERβ-mRNA expression was greater or less than the median value (percentile, P, 50) and were compared to each other with respect to the post-operative outcome variables. The patients with ERα- and ERβ-mRNA expression >P50 showed lower troponin T levels at 4 and 24 h po than the others [0.53 ng/ml (0.69) and 0.28 ng/ml (0.09) vs. 0.88 ng/ml (2.69) and 0.79 ng/mL (0.78); p = 0.003, respectively] and lower AST concentrations at 4 and 24 h po [51.5 IU/L (19.7) and 58 IU/L (14); p = 0.043, respectively]. Figure 4 shows the troponin T levels in patients with ERα-mRNA expression less than or greater than P50.

We show for the first time that mRNA coding for ERα- and ERβ is expressed in the right atrium of children and young adults with congenital heart defects independently of age, gender, and achieved puberty, respectively.

We also confirm our previous results demonstrating the myocardial expression of mRNA coding for pro- and anti- inflammatory cytokines, factors regulating cell growth and fibrosis, and proteins involved in the early cellular stress responses as a consequence of the activation of inflammatory pathways by hemodynamic overload (14). In this series, the expression of TGF-β increased with age, pointing to the importance of the duration of hemodynamic load on maladaptive myocardial remodeling and myocardial fibrosis (15).

Furthermore, the concentrations of mRNA coding for the different proteins tested correlated well with each other, suggesting an interplay between ERs and the mediators of inflammation, growth, fibrosis, and stress response of myocardial cells and therefore the participation of ERs in the complex mechanisms of myocardial remodeling in patients with congenital cardiac disease (13). In particular, the relationship between the expression of ER-mRNA and BNP-mRNA might indicate a role of hemodynamic overload in the upregulation of ERs as it has been shown in adult patients with aortic stenosis (12). Moreover, the fact that healthy women have higher BNP plasma levels than healthy men of the same age group suggests that estrogens may induce BNP expression throughout ER signaling (16). This is supported by the observation of increasing BNP blood levels after estrogen replacement therapy (17).

Our results show further that the mRNA expression of ERs correlated positively with the mRNA expression of pro- and also of anti-inflammatory cytokines. This apparent contradictory result might be explained by the complexity of the induction of inflammatory cytokines by the mechanical stimulation of cardiomyocytes. Indeed the early stress response leads to a sustained induction of pro-inflammatory cytokines that, in turn, initiates the gene expression of anti-inflammatory cytokines (14). Besides this, ERs are involved in the modulation of the inflammatory response by estrogens (18) and either activate or repress gene expression depending on local estrogen concentrations (19). Thus, higher estradiol (E2) levels downregulate pro-inflammatory cytokines such as TNF-α and upregulate anti-inflammatory cytokines such as IL-10 in different cell types, whereas low E2 levels stimulate TNF-α and IL-1β expression (20). The anti-inflammatory effect of E2 might be related to ERα that blocks TNF-α-induced IL-6 synthesis by interfering with nuclear factor kappa B (NFκB) (21). ERs inhibit multiple NFκB pathways and are therefore considered anti-inflammatory proteins (22–26).

The mechanisms by which low and high physiological concentrations of estrogens differentially affect ER activity to influence the expression of inflammatory genes are unclear. One possibility is that the low and high levels of E2 induce distinct transcriptional complexes and that this activates different pathways to promote or dampen inflammation (27). Our observation that the expression of ER-mRNA was not related to gender, age, or achieved puberty and that it was present in very young infants suggests an E2-independent mechanism for the activation of ERs in the myocardium of children. Indeed a large number of substances are likely to bind ERs such as xenoestrogens, in particular, phytoestrogens present in a wide spectrum of food constituents (28).

In our series, the myocardial concentrations of ERβ-mRNA were higher than those of ERα-mRNA. While a differential expression of both ERα and ERβ in the human right atrial myocardium has not been described that far, animal studies performed in neonatal and adult female rats have shown higher ERα-mRNA concentrations in the oldest animals (29). This suggests an influence of age on ERα expression in animal cardiomyocytes. In our patients, however, ERα expression was not age dependent. The role of inflammatory cytokines induced in the myocardium by hemodynamic overload on the upregulation of ERβ and down-regulation of ERα as has been documented previously (30, 31) remains speculative in our patient cohort who was heterogeneous in terms of quality and severity of cardiac defects and hemodynamic overload (14).

Besides this, hypoxemia could also have impacted ER expression as demonstrated in human breast cancer cell lines where hypoxia represses ERα (32).

Both the repression of ERα and the upregulation of ERβ involve the activation of hypoxia-inducible factor (HIF)-1α (33, 34) that is increased in the myocardium of children with cyanotic congenital cardiac disease, as we have shown previously (35). The putative influence of pre-operative hypoxemia on ER expression in the myocardium of patients with congenital cardiac disease was not investigated in our study, owing to the fact that only three patients of this series were cyanotic.

The secondary objective of this study was to address the question of whether myocardial ER expression may provide cardio-protection to patients undergoing cardiac surgery for congenital cardiac disease and be related to better post-operative outcome.

Our results showing that a higher expression of ERα-mRNA was associated to lower myocardial damage, improved lung function, lower water retention, and lower cytolysis in the early post-operative period might support the assumed protective role of ERs in this particular patient population (22–26). The reason why the expression of ERβ-mRNA did not correlate significantly with the outcome variables but with reduced fluid retention might be related to the statistical rank correlation analysis performed on the small patient group.

Open cardiac surgery in adults and children is associated with a systemic inflammatory reaction that relates to post-operative myocardial cell damage and multiple organ dysfunction syndrome being a severe issue (36). In this context, troponin release correlates with the importance of systemic inflammation, in particular, with the levels of circulating IL-6 (37). Inflammatory proteins such as complement proteins are present in the circulation immediately after connection to the extracorporeal circuit and initiate the synthesis of pro- and anti-inflammatory cytokines by circulating and tissue cells (38). An adequate anti-inflammatory balance is thought to be necessary to limit and/or terminate inflammation and protect from organ injury (39).

The anti-inflammatory potential of myocardial ERs may provide peri-operative organ protection against operative and inflammatory stress. In an experimental sepsis model classically associated with a systemic inflammatory reaction, ERβ agonists provide increased survival and reduced tissue damage and modify the genomic sepsis signature with a decreased expression of pro-inflammatory genes (40). Besides these nuclear-mediated effects of ERβ, ERα initiates the activation of acute protective pathways via non-nuclear mechanisms involving the activation of kinases that enhance the phosphorylation of eNOS, PI3K/Akt, and ERK1/2 (41, 42).

While our results, taken together, might indicate that gender- and age-independent expression of ERα and ERβ in the myocardium of young patients undergoing cardiac surgery works as protective, more experimental data are needed to answer the question of whether the modulation of ER expression in the myocardium would participate to improve peri-operative organ protection in this patient group. For this purpose, an animal model of cardiac surgery for congenital cardiac disease with hemodynamic overload (15) involving pre-operative induction of ERs by pharmacological or genetic engineering procedures should be established.

Our study has several limitations. The small number of patients investigated and their heterogeneity in terms of cardiac diagnosis did not allow us to analyze the role of the quality of hemodynamic overload and of the degree of hypoxemia on ER-mRNA expression.

Furthermore, the limited size of the myocardial samples was insufficient to quantify protein synthesis of our target genes and prejudge the biological activity of ER induction.

Finally, we describe an association between higher myocardial ER-mRNA expression and lower clinical and biological markers of post-operative organ damage but are not able to give evidence of the organ-protective role of ERs during cardiac surgery for congenital cardiac defect at this stage. This would require experimental studies involving the modulation of ER expression in a model of cardiac surgery for congenital cardiac disease.