Stimulation of myocardial β1-adrenoceptors (AR) is a major mechanism that increases cardiac function. We investigated the functional consequences of genetic β1-AR knockdown in three-dimensional engineered heart tissue (EHT). For β1-AR knockdown, short interfering RNA (siRNA) sequences targeting specifically the β1-AR (shB1) and a scrambled control (shCTR) were subcloned into a recombinant adeno-associated virus (AAV)–short hairpin RNA (shRNA) expression system. Transduction efficiency was ∼100%, and radioligand binding revealed 70% lower β1-AR density in AAV6-shB1–transduced EHTs. Force measurements, performed over the culture period of 14 days, showed paradoxically higher force generation in AAV6-shB1 compared with shCTR under basal (0.19 ± 0.01 versus 0.13 ± 0.01 mN) and after β-AR-stimulated conditions with isoprenaline (Δfractional shortening: 72 ± 5% versus 34 ± 4%). Large scale gene expression analysis revealed that AAV6-shCTR compared with nontransduced EHTs showed only few differentially regulated genes (<20), whereas AAV6-shB1 induced marked changes in gene expression (>250 genes), indicating that β1-AR knockdown itself determines the outcome. None of the regulated genes pointed to obvious off-target effects to explain higher force generation. Moreover, compensational regulation of β2-AR signaling or changes in prominent β1-AR downstream targets could be ruled out. In summary, we show paradoxically higher force generation and isoprenaline responses after efficient β1-AR knockdown in EHTs. Our findings 1) reveal an unexpected layer of complexity in gene regulation after specific β1-AR knockdown rather than unspecific dysregulations through transcriptional interference, 2) challenge classic assumptions on the role of cardiac β1-AR, and 3) may open up new avenues for β-AR loss-of-function research in vivo.
In the heart β-adrenoceptor (AR) signaling is essential for the “fight-or-flight” response mediated by the sympathetic nervous system through norepinephrine and epinephrine, acting on cardiac myocytes mainly via β-AR. In rodents, activation of β1-AR largely accounts for the increase in cardiac contractility (inotropy), frequency (chronotropy), and relaxation velocity (lusitropy) (Brodde et al., 2001; Rockman et al., 2002; Xiang and Kobilka, 2003; El-Armouche and Eschenhagen, 2009). β-AR stimulation leads to the activation of cAMP-dependent protein kinase A (PKA) via stimulation of G-proteins (Gs), activation of adenylyl cyclases (AC), and production of cyclic adenosine-3′,5′-monophosphate (cAMP). This results in the phosphorylation of a set of key regulatory proteins that control the excitation-contraction coupling cycle, such as sarcolemmal L-type Ca2+ channels (LTCC), sarcoplasmic ryanodine receptors (RyR2), and phospholamban (PLB), which together coordinate significant increases in cardiac inotropy, chronotropy, and lusitropy (Rapundalo, 1998; Bers, 2002).
Transgenic overexpression of β1-AR in mice resulted in a higher sensitivity to catecholamines, elevated basal heart rate and contractility in young mice, but heart failure and sudden death during aging (Engelhardt et al., 1999). In contrast, genetic deficiency of the β1-AR led to prenatal death of the majority of the mice, indicating the important role of the receptor during development (Rohrer et al., 1998). Hearts of mice that survived were morphologically and functionally not distinguishable from wild-type, including normal heart rate and blood pressure. But as expected, β1-AR-knockout (KO) mice lacked positive inotropic and chronotropic effects of the mixed β1-/β2-AR agonist isoprenaline, substantiating the dominant role of this receptor subtype. β2-AR-mediated hypotensive reaction was conserved. Heart rate in β1-AR-deficient mice was regulated by the parasympathetic nervous system, and interestingly cardiovascular function during exercise (e.g., treadmill experiments) did not differ from wild-type mice (Rohrer et al., 1998). These findings underscore the necessity and utility of studying β1-AR functions through acute loss-of-function approaches in vitro and in vivo.
In this study, we assessed functional consequences of acute genetic β1-AR silencing in an experimental setup of three-dimensional engineered heart tissue (EHT). EHTs represent three-dimensional heart-tissue-like cardiac myocyte cultures that allow the determination of standard parameters of contractile function under isometric conditions. EHTs provide cells with a relatively physiologic three-dimensional environment are stable for weeks and are easy to infect with adeno-associated virus (AAV) with high efficiency (Hansen et al., 2010; Cattin et al., 2013). Contrary to our expectations, β1-AR knockdown was associated with higher basal and β-AR isoprenaline–mediated force generation. Large-scale gene expression analysis by microarray technology identified almost 280 differentially regulated genes attributed to specific β1-AR knockdown rather than unspecific AAV or RNA interference effects. No obvious off-target effects explaining our results or compensatory changes in, e.g., β2-AR could be identified.
Materials and Methods
The cell culture media, horse serum, fetal calf serum, and penicillin/streptomycin, were obtained from Gibco (Invitrogen, Carlsbad, CA). The following cell culture media were used: minimal essential medium (MEM) from Biowest (Nuaillé, France) and Dulbecco’s modified Eagle’s medium (DMEM) from Biochrom (Berlin, Germany). Cell culture plates were obtained from Nunc (Thermo Fisher Scientific, Roskilde, Denmark). The following chemicals, substances, reagents, enzymes, and antibodies were obtained from Sigma-Aldrich (St. Louis, MO): 5′-bromo-2′-deoxyuridine, aprotinin, insulin, thrombin, Triton X-100, isoprenaline, propranolol, and anti-α-actinin monoclonal antibody (Ab). ICI 118,551 hydrochloride [(±)-1-[2,3-(dihydro-7-methyl-1H-inden-4-yl)oxy]-3-[(1-methylethyl)amino]-2-butanol hydrochloride] was obtained from Tocris Bioscience (Bristol, UK). Antibodies Alexa Fluor 546 goat anti-mouse IgG secondary Ab and DAPI were obtained from Invitrogen. Restriction enzyme MluI was obtained from New England Biolabs (Ipswich, MA). Oligonucleotide sequences were obtained from Eurofins MWG Operon (Ebersberg, Germany); short interfering RNA (siRNA) sequences and RNA extraction kit TRIzol were obtained from Ambion (Invitrogen). Different TaqMan Gene Expression Assays, TaqMan Universal PCR-Master Mixes, High Capacity cDNA Reverse Transcription Kit, and Illumina TotalPrep RNA Amplification Kit were obtained from Applied Biosystems Germany (Darmstadt, Germany). Maxima SYBR Green/ROX qPCR Master Mix was obtained from Thermo Fisher Scientific Baltics (Vilnius, Lithuania). RNeasy Mini Kit was obtained from Qiagen (Hilden, Germany). RatRef-12 Expression BeadChip Array (Illumina BD-27-302), BeadArray Reader, BeadScan software, and GenomeStudio software 2010 were from Illumina, Inc. (San Diego, CA). Bovine serum albumin and Histofix were obtained from Carl Roth (Karlsruhe, Germany). Radioligand [3H]CGP 12177 was obtained from Amersham (GE Healthcare Europe, Freiburg, Germany), Multiscreen polyvinylidene difluoride filter plate was from Merck Millipore (Billerica, MA), and scintillation cocktail Ultima Gold was from PerkinElmer Life and Analytical Sciences (Waltham, MA). Fluoromount G was obtained from Southern Biotech (Birmingham, AL).
Cardiomyocytes and Analysis.
Neonatal rat cardiac myocytes (NRCM) were isolated from 1- to 3-day-old neonates (Wistar and Lewis rats of mixed sex) by a fractionated DNase/trypsin digestion protocol as described earlier (El-Armouche et al., 2003). Experimental procedures were reviewed and approved by Ethics Committee, University of Hamburg, and were in accordance with the Declaration of Helsinki. Two-dimensional cultures of NRCM were maintained in culture medium (MEM, 10% fetal calf serum, 1% penicillin/streptomycin, and 1 mM 5′-bromo-2′-deoxyuridine) and in 37°C, 7% CO2 atmosphere. Cells were transduced with AAV-6 encoding bicistronically short hairpin RNA (shRNA) and enhanced green fluorescent protein (eGFP) (3 × 103, 1 × 104, 3 × 104, 1 × 105 virus genomes/cell). After 72 hours, transduction efficiencies were determined by GFP-epifluorescence using Carl Zeiss confocal microscope (Zeiss LSM 510 Meta Axiovert 200; Jena, Germany), and cells were harvested for further mRNA analysis.
Evaluation of siRNA Sequences Targeting β1-AR.
Based on three sequences targeting the β1-AR, siRNA (siB1)-, shRNA (shB1)-, and AAV6-shRNA (AAV6-shB1), expression vectors were generated. Knockdown efficiency of each sequence compared with the corresponding scrambled sequence (siCTR, shCTR, AAV6-shCTR) was assessed by quantitative real-time polymerase chain reaction (qRT-PCR) and [3H]CGP 12177 binding assay in a stable β1-AR cell line [human embryonic kidney (HEK)-B1; Supplemental Fig. 1]. For detailed information see Supplemental Methods.
Generation of Adeno-Associated Viral Vectors.
An shRNA expression cassette containing the H1-promoter sequence and the specific shB1/shCTR sequence was amplified by PCR using specific primers to insert MluI restriction sites at both ends. After restriction with MluI, shRNA expression cassette was subcloned into pdsAAV-cytomegalovirus (CMV)–enhanced green fluorescent protein vector that was linearized with MluI before. Recombinant AAV genome is depicted in Supplemental Fig. 2A. Production, purification, and titration of high titer adeno-associated viral vectors of serotype 6 (AAV-6) has been described previously (Raake et al., 2013). Virus titers were approximately 1.5 × 1012 virus genomes per milliliter.
Engineered Heart Tissue and Analysis.
Fibrin-based engineered heart tissues (EHT) from neonatal rat heart cells were generated and cultured as previously described (Hansen et al., 2010). In brief, for each EHT, a 100 μl-reconstitution mix containing 4 × 105 cells/EHT, bovine fibrinogen, aprotinin, and DMEM was mixed with 3 μl of thrombin and pipetted around two elastic silicone posts (produced by Siltec GmbH & Co. KG, Weiler-Simmerberg, Germany). EHTs were transduced with AAV-6 encoding shB1- or shCTR (105 virus genomes/cell), which was directly applied to the reconstitution mix. EHTs were cultured up to 21 days in culture medium (DMEM, 10% horse serum, 2% chick embryo extract, 1% penicillin/streptomycin, 10 μg/ml insulin, and 33 μg/ml aprotinin). Contraction measurements were performed by video optical recording on days 6, 8, 10, 12, 14, 16, and 18, as previously described (Hansen et al., 2010). Average force, frequency, contraction, and relaxation times were calculated from the recorded contractions by an algorithm that takes into account the elastic properties of the silicone posts. EHTs represent three-dimensional heart tissue–like structures that exhibit spontaneous, regular, and synchronous beating and allow measurement of contractile force under isometric conditions. After 21 days in culture, fractional shortening (FS in %) was measured in an IonOptix-based system custom-adapted for the measurement of loaded auxotonic contractions of EHTs. During measurement, constructs were continuously perfused with gassed Tyrode’s solution and electrically stimulated at a frequency of 2 Hz. In this system inotropic responses to low and high external concentrations of calcium (0.1 and 1.8 mM) and isoprenaline bolus (100 nM) were recorded. To exclude functional upregulation of β2-AR in AAV6-shB1 constructs, we analyzed concentration-frequency response to unselective stimulation with isoprenaline (3 × 10−9 to 3 × 10−6 M) in presence and absence of the selective β2-AR antagonist ICI 118,551 (100 nM).
Quantitative Real-Time RT-PCR.
Expression levels of mRNAs were determined by qRT-PCR (El-Armouche et al., 2008). Total RNA was isolated from NRCM or EHT with TRIzol, and ~200 ng of total RNA was reverse-transcribed with High-Capacity cDNA Reverse Transcription Kit according to manufacturer protocols. qRT-PCR was performed by using commercially available TaqMan Gene Expression Assays for β1-AR (rat, human), calsequestrin 2 (Casq2) (rat), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (human) and TaqMan Universal PCR-Master with the TaqMan system (ABI Prism 7900 Sequence Detection System; Applied Biosystems, Foster City, CA), or by using specific primers against the following rat target genes: β2-AR (forward: 5′-GAG CAC AAA GCC CTC AAG AC-3′, reverse: 5′-CTG GAA GGC AAT CCT GAA ATC-3′, transcript size: 209 bp); β1-AR kinase 1 (Bark1) (forward: 5′-AGA CGG AGG AGG AAC GTG T-3′, reverse: 5′-ATG GCT GGA AGA GAT CTG GA-3′, transcript size: 159 bp); G-protein α inhibiting activity polypeptide 2 (Gnai2) (forward: 5′-AGT ACA CAG GGG CCA ACA AG-3′, reverse: 5′- TCA GAA GAG GCC ACA GTC CT-3′, transcript size: 200 bp); phospholamban (forward: 5′-CGA TCA CAG AAG CCA AGG CCT C-3′, reverse: 5′-CGC GCT TGC TGG GGC ATT TC-3′, transcript size: 199 bp); sarcoendoplasmic reticulum calcium ATPase 2a (Serca2a) (forward: 5′-TGC TGG AAC TTG TGA TCG AG-3′, reverse: 5′-AGC GTT TCT CTC CTG CCA TA-3′, transcript size: 191 bp); and Casq2 (forward: 5′-TCA AAG ACC CAC CCT ACG TC-3′, reverse: 5′- AGT CGT CTG GGT CAA TCC AC-3′, transcript size: 200 bp) and Maxima SYBR Green/ROX qPCRMaster. Specificity of primers was determined by dissociation (melt) curve analysis.
Gene Expression Analysis–Microarray Data.
For microarray analysis, the total RNA from EHT was purified by DNase digestion (RNeasy Mini Kit) and further converted to biotinylated cRNA via in vitro transcription according to the manufacturer’s protocol (Illumina TotalPrep RNA Amplification Kits). Subsequently, cRNA samples (n = 6) from each experimental group were hybridized on Illumina RatRef-12 Expression BeadChip Array containing 21,910 probes (Illumina BD-27-302) and were analyzed with a BeadArray Reader and BeadScan software (Illumina). Gene expression levels were calculated using GenomeStudio software 2010 (Illumina), data were fit to a quantile normalization model, and approximately 21,791 expressed genes were identified. Differentially regulated genes were detected by applying Illumina custom error model; multiple testing correction was performed using Benjamini and Hochberg false discovery rate (Luchtefeld et al., 2011). Genes with a fold change (FC) ≥ 1.5 or FC ≤ 0.66 and a corrected P value < 0.05 were considered significant. Fold changes of selected genes were visualized in scatter plots.
β1-AR density was analyzed by radioligand binding assay (Joseph et al., 2004) in membrane proteins of cell homogenates or directly in intact EHT by using the hydrophilic, nonselective β-AR agonist [3H]CGP 12177. Either 30 ng of membrane protein or whole-mount EHT (4 × 105 cells) were incubated with [3H]CGP 12177 (3 nM) in assay buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM MgCl2, 0.5 mM EDTA) for 90–180 minutes at 4°C. Binding assay of membrane proteins was performed in a MultisScreen polyvinylidine fluoride filter plate (Merck Millipore). Unspecific binding was determined by additional incubation with the competitive, nonradioactive β-AR antagonist propranolol (3 μM). After stringent washing with ice-cold assay, buffer radioligand binding was determined by liquid scintillation counting (Counter Wallac 1409; PerkinElmer Life and Analytical Sciences) with the scintillation cocktail Ultima Gold. Specific binding was calculated from the difference between total and unspecific binding (counts per minute) and according to the scintillation efficiency, specific ligand activity, and protein amount, the decays per minute and β1-AR density [femtomoles per milligram protein] were computed. Each measurement was performed in triplicate.
Histology and Immunochemical Analysis.
EHTs were analyzed by immunofluorescence using confocal imaging. Therefore, EHTs were rinsed with phosphate-buffered saline (PBS) and fixed with Histofix overnight at 4°C. EHTs were carefully detached from the silicon posts and treated for 24 hours with blocking solution [0.05 M Tris buffered saline (TBS), pH 7.4, 10% fetal calf serum (FCS), 1% bovine serum albumin (BSA), 0.5% Triton X-100] at 4°C. Immunofluorescence staining was performed with mouse anti-α-actinin monoclonal Ab (1:800) and Alexa Fluor 546 goat anti-mouse IgG secondary Ab (1:800) in antibody solution (0.05 M TBS, pH 7.4, 1% BSA, 0.5% Triton X-100). Nuclei were stained with DAPI (1:1000). Incubation time for first and secondary antibody was 24 hours, respectively. Finally, EHTs were fixed between coverslip and glass slide with Fluoromount G, and fluorescence signal was analyzed by using Carl Zeiss confocal microscope (Zeiss LSM 510 Meta Axiovert 200).
Differences between groups were assessed using Student’s t test, one-way analysis of variance (ANOVA) or two-way ANOVA. Average data are presented as mean ± S.E.M. Differences were considered significant when *P < 0.05, **P < 0.01, and ***P < 0.001.
AAV6-Mediated shRNA Expression Against β1-AR in NRCM and EHTs.
We identified one of three sequences mediating a robust and strong β1-AR knockdown in HEK-B1 cells (Supplemental Fig. 1), and therefore, this sequence was used for AAV6-shB1 generation (Supplemental Fig. 2A) and all further experiments. To establish efficient gene silencing of β1-AR, NRCM were transduced with increasing amounts of viral genomes per cells, 3 × 103 to 1 × 105 of either AAV6-shB1 or AVV6-shCTR. After 72 hours, transduction efficiency was visualized by GFP-epifluorescence (Supplemental Fig. 2B) and knockdown efficiency was tested on the mRNA level (Supplemental Fig. 2C). Increasing GFP signals correlated with an increasing reduction in β1-AR mRNA by up to 62% compared with AAV6-shCTR.
To study functional consequences of β1-AR knockdown, we transduced EHTs with 105 viral genomes per cell of AAV6-shB1/shCTR. The highest transduction efficiency was achieved when applying virus directly to the EHT reconstitution mixture (data not shown). GFP epifluorescence showed a high and constant transduction rate of both viral vectors in EHTs (Fig. 1A, upper panel). A more detailed immunofluorescence analysis showed predominant colocalization of the GFP signal with the cardiomyocyte-specific marker α-actinin (Fig. 1A, lower panel), indicating preferential transduction of cardiomyocytes compared with nonmyocytes that are also present in EHTs. Accordingly, β1-AR mRNA and receptor density were reduced by 62% and 72% (Fig. 1B, C), respectively, whereas β2-AR mRNA (Fig. 1D) was not affected, proving selective and robust effectiveness of our knockdown strategy.
Functional Consequences of β1-AR Downregulation in EHTs.
Virally transduced EHTs developed properly and were indistinguishable from the nontransduced controls (data not shown). Contractility was analyzed regularly over the entire culture period by video-optical recording. The most important parameters measured at day 14 are depicted in Fig. 2. Beating rate as well as contraction and relaxation times (T1, T2) did not differ from control (Fig. 2, A and B), but unexpectedly force development (Fig. 2C) was higher in β1-AR knockdown than in control EHTs (force at 1.8 mM calcium: 0.19 ± 0.01 mN versus 0.13 ± 0.01 mN in AAV6-shCTR, P < 0.001). Likewise, the inotropic response to β-adrenergic stimulation with 100 nM isoprenaline (Fig. 2D; measured at low external calcium and electrical stimulation at 2 Hz) was higher (FS increase 72 ± 5% versus 34 ± 4%, P < 0.001). Results were obtained and reproduced in four different time series with total numbers of n = 32 (AAV6-shB1) and n = 26 (AAV6-shCTR). Concentration response curves for calcium and isoprenaline showed no difference in the sensitivity to external calcium (Supplemental Fig. 3) and in the inotropic potency of isoprenaline (Supplemental Fig. 4). Thus, diametrically opposed to our expectations, genetic knockdown of β1-AR was associated with higher force generation and larger inotropic responses to isoprenaline.
To test whether a functional upregulation of β2-AR may have compensated for β1-AR downregulation we analyzed the chronotropic response to the β1/β2-AR agonist isoprenaline in the presence and absence of the selective β2-adrenergic antagonist ICI 118,551 (100 nM). In the presence of ICI 118,551 the maximal chronotropic effect of isoprenaline was preserved, but the curve was shifted to the right by approximately a half-log unit in both groups (Fig. 3, A and B). The shift was significantly less than what is to be expected of a β2-AR-mediated effect [two log units (Hoffmann et al., 2004)], suggesting that the small shift was due to the incomplete selectivity of ICI 118,551 and partial occupancy of β1-AR at this concentration. These results argue against functional upregulation of β2-AR, a conclusion which is also supported by the lack of β2-AR mRNA regulation (Fig. 1D).
Genetic Consequences of β1-AR Downregulation in EHTs.
To address potential involvement of AAV and/or shRNA-related adverse or off-target effects caused either by viral transduction, shRNA, or by β1-AR knockdown itself, we performed a microarray gene expression analysis. Results show that virus treatment (AAV6-shCTR compared with nontransduced CTR) did not induce strong changes in the gene expression profile (Fig. 4A) demonstrating that AAV and/or shRNA-related adverse or off-target effects have little impact. On the other hand β1-AR knockdown resulted in marked differential regulation of genes compared with AAV6-shCTR (Fig. 4B). In particular, analysis of AAV6-shB1 versus AAV6-shCTR displayed a total number of 277 differentially regulated genes (Fig. 4B), of which 109 were upregulated (FC ≥ 1.5) and 168 were downregulated (FC ≤ 0.66). The strongest regulated genes are depicted in Table 1 (for all regulated genes see Supplemental Table 1). There was no regulation of prominent β-AR signaling targets, e.g., β2-adrenoceptor (FC = 0.92), β-adrenergic receptor kinase 1 (FC = 0.91), SR-calcium ATPase (FC = 0.99), phospholamban (FC = 1.23), and stimulatory/inhibitory G-protein α subunit (FC = 1.09, FC = 0.93), which eventually may have resulted in higher contractility. qRT-PCR analysis confirmed array data for these genes (Supplemental Fig. 2D). Prominent examples of upregulated genes are angiopoetin 2 (Angpt2, FC = 2.76), Iroquois homeobox 5 (Irx5, FC = 2.34), sodium channel β4 subunit (Scn4b, FC = 2.05) α1B-adrenergic receptor (Adra1b, FC = 1.92), as well as tissue factor pathway inhibitor 2 (Tfpi2, FC = 0.24) and regulator of G-protein signaling 2 (Rgs2, FC = 0.41), while α1 skeletal muscle actin (Acta1, FC = 0.47) and natriuretic peptide precursor A (Nppa, FC = 0.49) were strongly downregulated.
Cardiac β1-AR plays a key role in the regulation of heart function. In this study, we assessed functional consequences of subacute genetic β1-AR silencing in spontaneously beating and force-generating three-dimensional EHT. The feasibility of the EHT system as a test bed for pharmacological (Hansen et al., 2010; Schaaf et al., 2011) and genetic interventions (El-Armouche et al., 2003, 2007; Cattin et al., 2013; Stöhr et al., 2013) has been previously proven. Here, we developed for the first time an RNA interference–mediated approach to provide posttranscriptional gene silencing of β1-AR. We achieved highly efficient transduction of cardiac myocytes in EHTs (∼100%) and reproducible β1-AR downregulation by ∼70%. Contrary to our expectations, this was associated with higher force generation and augmented responses to acute isoprenaline stimulation, which clearly challenges the classic concept of β1-AR signaling (see below). Sufficiently high numbers of EHTs (n = 32) and consistent results in four different series indicate a robust and real effect. Thus, β1-AR knockdown in cardiac myocytes induced a (beneficial) phenotype in this model, which is opposite to the cardiomyopathic phenotype seen in hearts of β1-AR–overexpressing transgenic mice (Engelhardt et al., 1999) and similar to hearts of patients chronically treated with β-blockers. It seems likely that the phenotype in β1-AR knockdown EHTs is the consequence of the altered gene expression program, but direct cause-effect relationships cannot be established with this method.
Because our loss-of-function approach did not achieve complete β1-AR depletion, the remaining receptors were likely sufficient for maintaining β1-AR inotropic (Fig. 2D) and chronotropic (Fig. 3B) responses. Indeed, a high receptor reserve has been reported in rat heart compared with human myocardium (Brown et al., 1992). However, receptor reserve does not explain the significant increase in basal contractility and increased positive inotropic response. A compensatory upregulation of β2-AR as a mechanism superimposing β1-AR downregulation was ruled out by demonstrating unaffected β2-AR mRNA levels and also functionally with the selective β2-AR antagonist ICI 118,551 (Figs. 1D and 3B). This is consistent with findings in β1-AR-KO mice (Rohrer et al., 1998). In addition, gene expression analysis did not indicate differential regulation of other elements of the β-AR signaling cascade, including β-AR receptor kinase 1, SR-calcium ATPase, phospholamban, or stimulatory/inhibitory G-protein α subunit. Thus, no obvious candidate was identified that explains the increased force and response to isoprenaline.
In principle, the phenotype could be just an artifact, e.g., by off-target effects of the si/shRNA sequence or the expression vector itself. This cannot be completely excluded but seems unlikely for two reasons. First, nonspecific effects have almost always resulted in decreases of contractile force and isoprenaline responses in EHTs up to now. Second, care was taken to minimize the chance for off-target effects. A single mismatch can result in the complete loss of function of the relevant siRNA sequence (Brummelkamp et al., 2002; Schubert et al., 2005) or in the loss of specificity toward the target mRNA (Agrawal et al., 2003), and thereby siRNAs may cross-react with targets of limited sequence similarity (Saxena et al., 2003). In this context, direct silencing of nontargeted genes containing as few as eleven contiguous nucleotides of identity to the siRNA can occur (Jackson et al., 2003). Besides sequence identity, off-targeting can be associated with the matches between parts of the 3′-untranslated region (UTR) and with the seed region of the antisense strand of the siRNA (Birmingham et al., 2006). Therefore, minimization of those unwanted effects is essential for siRNA design (Qiu et al., 2005; Jackson et al., 2006). In this study, we used “scrambled” control sequences in each experiment for any vector or shRNA-mediated effects. In addition, we compared the results of AAV-shRNA transduced EHTs to untransduced EHTs. Comparison of central housekeeping genes (β-actin and glyceraldehyde-3-phosphate dehydrogenase) and the cardiac-specific gene calsequestrin showed no differences between the groups (data not shown), which indicates no obvious counter-regulation or toxicity. This is further supported by almost identical gene expression profiles in the AAV-shCTR compared with untransduced control group (Fig. 4A).
Thus, more likely than an artifact, the knockdown of β1-AR induced changes in gene expression that, in their totality, may explain the beneficial phenotype in EHTs. This may resemble the paradoxical effect of long-term β-blockade in patients with heart failure. Whereas β-blockers decrease cardiac output short term, they reverse myocardial remodeling and improve contractile function when given long term. This process involves changes in gene expression (Lowes et al., 2002), e.g., downregulation of natriuretic peptide precursor A (ANP), which was also observed in this study. Another gene with proposed relevance in cardiovascular disease, Rgs2 (Tsang et al., 2010), a negative regulator of Gαq/11 and Gαi signaling (Hao et al., 2006), was downregulated. A decrease in Rgs2 may cause sensitization of Gαq/11 and thereby of α1-AR signaling. β1-AR knockdown was associated with almost twofold increased α1B-AR expression. Whereas numerous studies have documented the detrimental consequences of overactivation of the Gαq/11 pathway (D’Angelo et al., 1997; Wettschureck et al., 2001), increasing evidence suggests that α1-AR signaling in fact mediates adaptive and protective effects in the heart. It is noteworthy that β-AR blockade has been suggested to provide an example of α1-AR “gain-of-function” (Jensen et al., 2011). In adult mouse ventricular cardiomyocytes, adrenaline abolished extracellular signal–regulated kinase (ERK) signaling via β-ARs, whereas it augmented ERK signaling via α1-AR—which was cardioprotective—under conditions of β-AR blockade (Huang et al., 2007; Jensen et al., 2011). This indicates that β-AR blockade may mediate favorable effects by augmenting beneficial α1-AR signaling. Thus, highly efficient β1-AR downregulation knockdown may switch signal transduction pathways toward beneficial signaling profiles and thereby to an improved phenotype with greater contractility in our model.
Our in vitro β1-AR loss of function model clearly differs from β1-AR KO mice, in which isoprenaline had simply no effect (Rohrer et al., 1996). However, it shares the observation that loss of β1-AR is compatible with apparently normal cardiac function. β1-KO mice displayed normal basal heart rate and maximal exercise capacity (treadmill experiment), indicating compensatory mechanisms to maintain normal cardiac function and regulation. Even the total loss of β1-AR and β2-AR in β1/β2-AR double KO mice was accompanied by normal basal heart rate, normal blood pressure, and normal exercise capacities (Rohrer et al., 1999). Moreover, β1/β2-AR double KO showed protection from chronic pressure overload–induced cardiac hypertrophy and fibrosis. It is noteworthy that this was associated with preserved cardiac function (Kiriazis et al., 2008). This indicates impressively that chronic β1/β2-AR deficiency rather than chronic overstimulation of the β-AR system is associated with beneficial effects (Lee et al., 2008).
Our surprising finding also opens future avenues for in vivo cardiac β-AR research. Techniques to monitor and image the murine cardiovascular system have developed enormously in the last decade (Ram et al., 2011; Moran et al., 2013). Thus re-evaluation of β1-AR and β1/β2-AR double KO with, e.g., high-resolution Doppler techniques, strain rate-, and four-dimensional echocardiography, as well as micro-CT and micro-MRI, should allow new insights into the structural and functional characteristics. This may help to decipher the differential roles of β1-AR and β2-AR. Also, conditional cardiac-specific β-AR KO models (Cre-Lox-, Tet-, or MerCreMer-System) may help to circumvent the high intrauterine mortality and compensational mechanism. Moreover, crossing of β1-AR KO with dopamine β-hydroxylase KO mice, which are devoid of endogenous norepinephrine and epinephrine (Rapacciuolo et al., 2001), would be very useful to delineate the role of the sympathetic nervous system for the outcome.
In conclusion, our study emphasizes the overall complexity regarding β-AR regulation and underscores the necessity of studying β-AR functions through acute loss-of-function approaches in vivo to understand the mechanisms more properly. In future this may aid development of new and better generations of β-blockers with fewer side effects.
Participated in research design: Neuber, Müller, Hansen, Stoll, Katus, Eschenhagen, El-Armouche.
Conducted experiments: Neuber, Müller, Hansen, Eder, Witten, Rühle.
Performed data analysis: Neuber, Hansen, Eder, Witten, Rühle.
Wrote or contributed to the writing of the manuscript: Neuber, Eschenhagen, El-Armouche.
- Received November 4, 2013.
- Accepted January 13, 2014.
T.E. and A.E.-A. contributed equally to this work.
This study was supported by the Deutsche Forschungsgemeinschaft [DFG FOR 604 (to T.E. and A.E.A.), DFG EL 270/5-1 and SFB 1002 TP-A02 (to A.E.A.)], Deutsche Herzstiftung (to A.E.A.), and DZHK [German Center for Cardiovascular Research (to T.E. and A.E.A.)].
- adeno-associated virus
- adeno-associated virus serotype 6
- Dulbecco’s modified Eagle’s medium
- engineered heart tissue
- fold change
- fetal calf serum
- fractional shortening
- green fluorescent protein
- G-protein α inhibiting activity polypeptide 2
- ICI 118,551 hydrochloride
- (±)-1-[2,3-(dihydro-7-methyl-1H-inden-4-yl)oxy]-3-[(1-methylethyl)amino]-2-butanol hydrochloride
- neonatal rat cardiac myocyte
- polymerase chain reaction
- quantitative reverse transcription-PCR
- regulator of G-protein signaling 2
- short hairpin β1-adrenoceptor
- short hairpin scrambled control
- short hairpin RNA
- short interfering RNA
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics