Angiotensin receptor type 1 mRNA in human right ventricular endomyocardial biopsies: downregulation in heart failure

Abstract

Objective: Atrial angiotensin II receptors type 1 (AT₁) are downregulated in end-stage human heart failure at mRNA and protein level. The present study investigated whether AT₁ ventricular mRNA content was reduced in myocardial biopsies from heart failure patients. Methods: AT₁ mRNA was quantitated in right ventricular endomyocardial biopsies from 16 patients with decreased left ventricular function (LVEF 36±3%) due to dilated cardiomyopathy (DCM) and in biopsies from 12 patients with suspected myocardial disease but normal cardiac function (LVEF 62±2%). Two biopsies per patient were pooled, RNA was extracted and reverse-transcribed after addition of an AT₁ cRNA standard. AT₁ standard and wild-type RNA were amplified with the same primers in the same PCR tube. The PCR products were hybridized to a microtiter plate and detected and quantitated by an ELISA system. Glyceraldehyde phosphate dehydrogenase (GAPDH) mRNA was determined in the same samples as AT₁ mRNA. Results: In the biopsies from 16 patients with heart failure, a 68% decrease in AT₁ mRNA content was found in comparison with 12 controls (heart failure 94±15 AT₁ mRNA copies/ng RNA; controls 297±45; P<0.001). Relating AT₁ mRNA content to GAPDH mRNA confirmed the specific decrease in AT₁ mRNA (AT₁/GAPDH: heart failure 1.3±0.15; controls 3.4±0.5; P<0.002). The best correlation between AT₁ mRNA content and clinical parameters was found for right ventricular ejection fraction (r = 0.59, P<0.01). Conclusions: The quantitative RT-PCR procedure indicated a loss of ventricular AT₁ mRNA in human heart failure which corresponds to the loss of AT₁ protein described previously. It may underlie the decrease in AT₁ protein expression in human heart failure.

Human, Angiotensin receptor, mRNA, RT-PCR

Time for primary review 25 days.

1 Introduction

Cardiac angiotensin receptors (ATR) mediate the direct effects of angiotensin II (Ang II) on the heart. Cardiac Ang II effects probably play a role in remodelling after myocardial infarction, in the development of ventricular hypertrophy and in the morphological changes associated with heart failure. In the human heart, both ATR subtypes (AT₁ and AT₂) are expressed [1–3]. Ang II, via AT₁, exerts hypertrophic and positive inotropic effects in the myocardium from experimental animals and in human atrial tissue [4–8]. In contrast, functions of AT₂ are still largely unknown. In transfected rat vascular smooth muscle cells and rat coronary endothelial cells, antiproliferative effects of AT₂ have recently been suggested [9, 10]and the induction of apoptosis by Ang II has been found in different cell lines which express abundant AT₂ receptor [11]. This apoptotic function of AT₂ may play a role in developmental biology and pathophysiology.

Regulation of cardiac AT₁ has already been reported in a number of physiological conditions. Upregulation was found in hypertrophic and ischaemic states [12–14], whereas downregulation of cardiac AT₁ was found in rat and human heart failure [1, 15, 16]. Human atrial and ventricular AT₁ differ in expression level, which is higher in the atria [1, 17], and their effects on contractile function [8]. Whereas AT₁ downregulation in human atria at end-stage heart failure has been well described at the mRNA and protein level, fewer data are available on ventricular AT₁ regulation [1, 18, 19]. So far, ventricular AT₁ regulation has only been investigated in explanted hearts from transplant recipients, which limits the conclusions to end-stage heart failure. To assess the regulation of human ventricular AT₁ in heart failure of different stages, particularly in non-terminally-diseased patients, AT₁ mRNA content must be assessed in endomyocardial biopsies. This approach has been used in the present study. A novel quantitative PCR procedure [20]enabled the determination of absolute mRNA copy numbers with a low variance. The study is the first to compare ventricular AT₁ mRNA content in endomyocardial biopsies from patients with heart failure due to dilated cardiomyopathy and normal systolic cardiac function.

2 Methods

AT₁ mRNA was measured in endomyocardial biopsies from 16 patients with heart failure due to dilated cardiomyopathy undergoing diagnostic cardiac catheterization and right ventricular endomyocardial biopsy. All patients had decreased left ventricular function (mean LVEF 36±3%) and increased ventricular volumes in the presence of a normal left ventricular wall thickness (Table 1). Most patients belonged to the NYHA class II (n = 9), 5 to class III, 1 to class IV and 1 to class I. Patients were treated with digitalis (n = 9), ACE inhibitors (n = 14), and diuretics (n = 9). Twelve patients undergoing cardiac catheterization and biopsy for suspected myocarditis based on atypical chest pain, arrhythmia and/or borderline echocardiographic ventricular volumes and function who turned out to have normal right and left ventricular function at cardiac catheterization, without signs of inflammatory disease in the biopsy, were used as controls. These patients were treated with digitalis for supraventricular arrhythmia (n = 5) and with ACE inhibitors (n = 6) and/or diuretics (n = 3) for hypertension. Significant coronary artery disease was excluded by coronary angiography in all cases. Informed consent was obtained from all patients before tissue sampling. The study was approved by the Ethical Committee at the Virchow Clinic. The investigation conforms with the declaration of Helsinki.

Table 1

Patients studied—clinical and hemodynamic data

Group	n	Age (years)	Sex (%male)	Medical therapy (%)					LVEF (%)	RVEF (%)	LVEDVI (ml/m2)	LVESVI (ml/m2)	LVEDP (mmHg)	MPAP (mmHg)
				ACEI	Diu	Digi	β-Bl	CA-A
DCM	16	44±4	87.5	87.5	56.3	56.3	12.5	6.3	36±3	49±2	167±10	108±10	13±2	18±2
Controls	12	44±5	75	50	25	41.7	41.7	16.7	62±2	63±2	107±6	41±4	9±2	16±2

ACEI = ACE inhibitors; Diu = diuretics; Digi = digitalis; β-Bl = β-blockers; CA-A = calcium antagonists; LVEF = left ventricular ejection fraction; RVEF = right ventricular ejection fraction; LVEDVI = left ventricular end-diastolic volume index; LVESVI = left ventricular end-systolic volume index; LVEDP = left ventricular end-diastolic pressure; MPAP = mean pulmonary artery pressure.

Cardiac catheterization in all patients included the measurement of left and right ventricular pressures, cardiac output, coronary angiography and the determination of right and left ventricular volumes and ejection fraction by biplane angiography after injection of 30 (right) or 40 (left) ml of non-ionic contrast medium. Volumes were determined according to the Simpson and Ferlinz method which were previously validated in our setting [21–23]. The left ventricular mass was calculated from the end-diastolic wall thickness in LAO projection and the left ventricular end-diastolic volume.

Right ventricular endomyocardial biopsies were taken at cardiac catheterization by a transfemoral approach using a 7-French Signus Bioptome (2.2 mm×104 cm, Signus GmbH, D-63755 Alzenau, Germany). The bioptome was directed versus the right ventricular septum by standard techniques. Two pieces were taken for histological (HE and van Giesson staining) and immunohistological examination and two were used for mRNA quantitation. Routine immunohistology included staining with antibodies for vimentin, alpha smooth muscle actin, panT-cells (UCL1), B-lymphocytes (CD21) and macrophages (KP1) (Dako, 22047 Hamburg, Germany). Patients with histological signs of myocarditis were excluded from the study.

Biopsies were frozen immediately on dry ice and stored at −70°C until analysis. Biopsies from controls and heart failure patients did not differ in wet weight (Table 2). Two biopsies per patient were pooled, RNA was extracted with RNAzol B (Lorei+Paesel, Germany) and subjected to a DNAse digest [20]. Yields of RNA before and after DNase digest were comparable in the patients with heart failure and controls. From each sample, 250 ng RNA were mixed with internal AT₁ cRNA standard and reverse-transcribed with 300 ng random hexamers and 100 U Superscript™ (Gibco BRL, Germany). To control the efficient digestion of DNA, all samples were run after digestion in a PCR with the specific primers for AT₁ (see below) before reverse transcription and with primers for pyruvate dehydrogenase (PDH) which yield bands of different length at the DNA and RNA level after reverse transcription [1]. All samples were negative for the AT₁ and PDH DNA fragments, indicating complete digestion of DNA.

Table 2

Yield of RNA from endomyocardial biopsies before and after digestion with DNAseI

	Patients (n)	No. of biopsies per patient	wwt (mg)	RNA extract (μg)	RNA after DNAse (μg)	RNA/wwt (μg/mg wwt)
Controls	12	2	11.4	5.8	2.2+	0.21
			±1.0	±3.6	±1.1	±0.02
Heart	16	2	10.6	5.8	2.2	0.22
failure			±1.0	±5	±1.3	±0.02

wwt = wet weight.

The construction of a 836 AT₁ cRNA standard which differed from the corresponding AT₁ wild-type mRNA only in a 6-base-pair deletion has been published previously [20]. The primer pair AT₁-1/AT₁-2 (AT₁-1, 5′-CCTTCGACGCACAATGCTTG-3′; AT₁-2, 5′-AGCCCTATCGGAAGGGTTGA-3′) amplifies a 168 bp fragment containing the 6 bp deletion from the AT₁ standard and a 174 bp fragment from the AT₁ wild type. Amplification of standard and wild type in one tube with the primers AT₁-1/AT₁-2 produced the 168/174 fragments from standard and wild type with equal efficiency (Fig. 1a). For quantitation, 50 ng wild-type cDNA together with a given amount of standard was amplified in a 25 μl PCR assay with primers AT₁-1 and AT₁-2 [20]. Aliquots of PCR products were hybridized to different wells of a microtiter plate which bind specifically standard or wild type. This process was highly specific. Binding of standard amplicons to the wild-type capture probe reached only 0.3% of binding to the standard capture probe and binding of wild-type amplicons to the standard capture probe only 3% of the binding to the wild-type probe. For detection of bound amplicons, a digoxigenin molecule was integrated into the 5′ end of primer AT₁-2. The bound digoxigenin was quantitated with an anti-digoxigenin alkaline phosphatase conjugate in a 96-channel ELISA reader [20].

Fig. 1

Left panel: Amplification efficiency for the 168 bp fragment from AT₁ standard and 174 bp fragment from wild type in a single PCR tube; 500 copies standard and wild type template were co-amplified in one tube. Aliquots were removed after 31, 32, 33, 34, 35, 36 and 37 cycles and the products were quantitated in the microtiter plates. Plotting specific ELISA-OD (optical density, ×1000) against number of cycles in semilogarithmic manner shows identical efficiencies for standard and wild-type amplification. Right panel: Calculated ratios of wild-type/standard for the reverse transcription and co-amplification of increasing amounts of wild type RNA together with 40 000 copies standard in each tube. A linear increase in the ratios of wild-type/standard is observed over the whole measurement range.

About 40 000 copies of the standard cRNA were reverse-transcribed with 250 ng sample RNA. A regression line was established by adding the same amount of standard to increasing wild-type concentrations (50–1000 ng sample RNA). Over a relatively small measuring range, a linear increase in the wild-type/standard ratio resulted (Fig. 1b). Linearity is probably due to the high structural identity and thereby identical amplification efficiency of wild type and standard as well as to the fact that only a small measurement must be covered, since the changes in AT₁ expression are relatively small. Based on this regression line and the known amount of standard in each sample, the amount of wild type was calculated. To ensure that potential changes in the AT₁ mRNA were not due to mRNA degradation, the AT₁ mRNA content in the atria and ventricles was related to the stably expressed GAPDH mRNA. The mRNA for GAPDH was amplified by PCR (22 cycles) and the amplicons were quantitated by HPLC.

To analyze the variability of the different steps of the procedure, 3 biopsies from about 10 mg were taken from an explanted heart. RNA was extracted from each sample, subjected to 3 different reverse transcriptions (RT, resulting in 9 different RTs) and each RT product was subjected to 3 different PCRs (27 PCRs) (Table 3). If one sample was reverse-transcribed in one RT, followed by 3 different PCRs, a mean PCR variability of about 10% resulted. If the 3 tissue samples were extracted and then reverse-transcribed in the same RT reaction (same master mixes and chemicals, identical conditions) and each of them was analysed by a single PCR, the variability between the different tissue samples (including the PCR variability) amounted to about 15%. If we reverse-transcribed each sample in 3 different RT's on 3 different days (with 3 different master mixes) and quantitated each RT product by a single PCR, variability amounted in the mean to 25%. This clearly indicates that the largest variance is introduced by the variability of the RT procedure. We reduced this variance in our study by reverse-transcribing all samples in a single RT using the same chemicals and master mixes.

Table 3

Variance for the quantitation of AT₁ mRNA in ventricular biopsies

Step	n	Coefficient of variance
ELISA detection	108	0.26%
Well-to-well variance	54	3.7%
PCR variance	27	10.1%
RT variance	9	25%
Sample variance	3	15%

Three biopsies of 10 mg were taken from an explanted heart. To determine the variance between the samples (including PCR variance), the 3 tissue samples were extracted and then reverse-transcribed in the same RT reaction. Each of them was analysed by a single PCR. To assess the RT variance (including PCR variance), each sample was reverse-transcribed in 3 different RT's on 3 different days (with 3 different master mixes) and each RT product was quantitated by a single PCR. To obtain PCR variance, one sample was reverse-transcribed in 3 RTs, each followed by 3 different PCRs. To assess well-to-well variance and the variance of ELISA detection, the same PCR amplicons were determined in 2 different wells on the ELISA plate and measurements of the same wells in the ELISA reader were repeated twice.

2.1 Statistics

Data are given as mean and s.e.m. Wilcoxon tests have been used to compare groups. In addition, confidence intervals for the mean values have been calculated using the SPSS programme. Pearson's correlation coefficients were calculated between AT₁ expression and haemodynamic parameters.

3 Results

Patients and controls differed significantly in ventricular AT₁ mRNA expression. The number of AT₁ copies amounted to 94±15 copies/ng RNA in the biopsies from failing hearts versus 297±45 (P<0.001) in the control group (Fig. 2), corresponding to a 68% loss of AT₁ mRNA. 95% confidence intervals ranged from 208–385 copies/μg RNA in the control group and 64–123 copies/μg RNA in the heart failure group.

Fig. 2

Number of AT₁ mRNA copies in right ventricular endomyocardial biopsies from heart failure (HF) and from control patients (Con). AT₁ mRNA is significantly decreased, by 68%, in heart failure. No significant difference in the GAPDH expression was found between both groups, whereas the AT₁/GAPDH ratio was significantly decreased in the heart failure group, indicating the specificity of AT₁ mRNA loss.

The mRNA for GAPDH was unchanged in heart failure and in controls (Fig. 2). If AT₁ mRNA was related to GAPDH mRNA, the difference between controls and heart failure was maintained. The AT₁/GAPDH ratio amounted to 1.3±0.15 in heart failure in comparison with 3.4±0.5 in controls (P<0.002, Fig. 2). Thus, relating AT₁ expression to GAPDH expression confirmed the selective downregulation of AT₁ mRNA.

Weak but significant correlations were detected between AT₁ mRNA expression and right and left ventricular ejection fractions and volumes. The best correlation was found with right ventricular ejection fraction (r = 0.59, P<0.01), followed by left ventricular end-diastolic volume index (r = 0.48, P<0.02) and left ventricular ejection fraction (r = 0.43, P<0.05, Fig. 3). No correlation with left ventricular mass was detected (r = −0.042, P = 0.86, n.s.).

Fig. 3

Correlation between AT₁ mRNA expression and right and left ventricular ejection fraction (RVEV, LVEF). Patients with DCM are depicted as circles and controls are depicted as squares. The number of AT₁ copies is plotted against LVEF and RVEF and the equation of the regression line is indicated in the figure. Correlation coefficients are r = 0.43, P<0.05 for AT₁/LVEF and r = 0.59, P<0.01 for AT₁/RVEF. There is no significant correlation between LV- or RVEF and AT1 in the DCM group (LVEF/AT1, r = −0,30, P = 0.3, n.s.; RVEF, r = +0,40, P = 0.19, n.s.).

To determine whether AT₁ downregulation was associated with medical therapy, we compared AT₁ mRNA expression in the control group, where 6 patients were treated with ACE inhibitors and 6 were not. No significant difference was found (287±55 versus 306±52 copies/ng RNA).

4 Discussion

Using a novel quantitative PCR technique, we determined the AT₁ mRNA content in human right ventricular endomyocardial biopsies and found a significant (68%) decrease in AT₁ mRNA in human heart failure. Relating AT₁ mRNA to the mRNA of the stably expressed GAPDH mRNA confirmed the specific downregulation of ventricular AT₁ in heart failure. AT₁ mRNA content was significantly correlated with functional parameters of the right and left ventricle.

To determine the AT₁ mRNA content, we used a quantitative RT-PCR procedure based on an internal AT₁ cRNA standard, liquid-phase hybridization of PCR products to a 96-well microtiter plate and quantitation by an ELISA system [20]. The PCR-ELISA system has a low variance of about 10%, whereas a variance of about 25% is due to the combination of RT and PCR. The method allows for the determination of absolute copy numbers although limited by the assumptions that have to be made for calculations. Precise quantitation of the cRNA standard remains a problem that is inherent to most quantitative PCR approaches as well as the difficulty of ensuring that standard and wild type are reverse-transcribed with the same efficiency. The high structural similarity between standard and wild type including their surrounding DNA sequences favours similar transcription rates. For use in biopsies, we have optimized the RT procedure and the amount of standard to be added and we transcribed all samples in a single RT reaction.

To exclude non-specific degradation of mRNA, we related AT₁ to GAPDH mRNA, confirming the selective downregulation of AT₁ mRNA. GAPDH has already been used as a reference gene for mRNA quantitation in failing hearts and was found unchanged [24]. An increase in extracellular matrix proteins would not affect the results since the AT₁ mRNA decrease is related to total RNA content and thus to the cellular components of the heart.

In different animal models of cardiac dysfunction, upregulation [13, 25]as well as downregulation [15]of ventricular AT₁ has been described. Disease-dependent regulation of human myocardial AT₁ has been studied by binding assays in atrial and ventricular membranes. Most investigators agree that the density of AT₁ in the normal and failing human heart is low in comparison with β-adrenergic receptors and that AT₁ is downregulated in heart failure [1, 2, 19]. Ventricular AT₁ protein content has so far only been determined in explanted end-stage failing hearts, donor hearts or autopsy tissue, restricting conclusions to a narrow spectrum of pathophysiological conditions.

We have previously described a 60% loss of AT₁ binding sites [1]and a 58% decrease of AT₁ mRNA in the atria of end-stage failing hearts [20]. AT₁ mRNA content has now been determined for the first time in right ventricular endomyocardial biopsies from heart failure patients with different degrees of functional impairment. The ventricular AT₁ mRNa loss is comparable to the reduction in the atria. The higher expression of AT₁ protein in the atria in comparison with the ventricles [1, 17]has now found its correspondence at the mRNA level since we determined 94 copies of AT₁ mRNA/ng RNA in the failing ventricle whereas about double the number of copies was previously found in atria from heart failure patients [20]. These data suggest that changes in AT₁ mRNA content correspond to changes in AT₁ protein expression and that the reduced AT₁ mRNA content may underlie the decreased AT₁ protein expression.

The downregulation of cardiac AT₁ mRNA may be caused by an increase in tissue Ang II content. Downregulation of AT₁ by its agonist has already been described in glomerular mesangial cells [26]. Relevant cardiac Ang II levels may persist in heart failure in spite of ACE inhibitor therapy since cardiac ACE is upregulated [27]and not all ACE inhibitor dosages inhibit tissue ACE effectively over 24 h. Furthermore, human heart chymase may generate intracardiac Ang II, particularly in the presence of elevated Ang I concentrations, and thus may maintain relevant cardiac Ang II levels in spite of ACE inhibitor therapy [28]. Continuing action of chymase in the presence of ACE inhibitors may explain the lack of a correlation between AT₁ expression and ACE inhibitor therapy.

The weak correlation between AT₁ expression and clinical parameters is not surprising given the relatively small number of patients studied and the variance in medical therapy. The relatively small number of patients studied within each group together with the clustering of ejection fractions around the mean value in each group is probably the reason for the lack of a correlation between AT₁ expression and ejection fraction within the control or DCM group. There was also no correlation with left ventricular mass or wall thickness, probably because all DCM patients had enlarged ventricles, in the presence of a normal wall thickness and the increase in mass reflected more the process of dilatation than hypertrophy. However, a significant correlation with right and also left ventricular ejection fraction was found, indicating a functional relevance of AT₁-mediated effects in the human heart. Since we have used right ventricular endomyocardial biopsies and a decrease in right ventricular ejection fraction is a hallmark of the progression of dilated cardiomyopathy, the predominant correlation of AT₁ expression with right ventricular ejection fraction seems reasonable.

The fact that we have used right ventricular endomyocardial biopsies to study a disease that is mainly characterized by left ventricular dysfunction represents somewhere a limitation of the study. However, biochemical changes in the right ventricular septum have already been shown to correlate with functional parameters of the left ventricle [29, 30].

A second limitation arises from the fact that AT₂ mRNA was not quantitated. AT₂ seems to play a role in developmental biology and may counteract AT₁ in rat and mouse cell lines, in transfected vascular smooth muscle cells [9], neonatal rat myocytes [31], or rat coronary endothelial cells [10]. However, all cardiac actions of Ang II that have been revealed so far by subtype selective blockers in a number of intact animal models and in humans or have been found in the adult rodent or human heart are mediated by AT₁. Therefore, we focused in the present clinical investigation first on AT₁.

The consequences of the reduced AT₁ expression in failing ventricles are still unknown. Ang II is coupled to calcium influx into cardiac myocytes. Downregulation of AT₁ may thus prevent calcium loading of myocytes and act in a protective manner. Ang II, via AT₁, exhibits positive inotropic actions in the normal human atrium but not in the failing ventricle [8]. Thus, downregulation of AT₁ would not decrease contractility. Based on recent data from rat heart, downregulation of AT₁ may limit specific growth responses [32, 33]and thereby prevent a progression of hypertrophy. In vascular smooth muscle cells, Ang II, via AT₁, elicits contraction and hyperplasia. The reduction of such an effect could be beneficial in a number of pathophysiological states. However, further investigations must be awaited to clarify this issue.

Acknowledgements

We thank Dr. E. Wellnhofer for his advice in the quantitative evaluation of right and left ventricular angiograms. We appreciate the excellent technical assistance of Heike Kallisch and Britta Hannack. The study was supported by DFG Re662/2-2.