Essential role of sympathetic endothelin A receptorsfor adverse cardiac remodelingLorenz H. Lehmanna,b, Julia S. Rostoskya,b, Sebastian J. Bussb,c, Michael M. Kreussera,b, Jutta Krebsa,b, Walter Mierd,Frank Enseleite, Katharina Spigerb,f, Stefan E. Hardtb,c, Thomas Wielandb,f, Markus Haassg, Thomas F. Lüschere,Michael D. Schneiderh, Rosanna Parlatoi,j, Hermann-Josef Grönek, Uwe Haberkornd, Masashi Yanagisawal,m,n,Hugo A. Katusb,c, and Johannes Backsa,b,1

aResearch Unit Cardiac Epigenetics, Department of Cardiology, University of Heidelberg, D-69120 Heidelberg, Germany; bDZHK (German Centre forCardiovascular Research), Partner Site Heidelberg/Mannheim, D-69120 Heidelberg, Germany; cDepartment of Cardiology, University of Heidelberg, 69120Heidelberg, Germany; dDepartment of Nuclear Medicine, University Hospital, D-69120 Heidelberg, Germany; eDivision of Cardiology, University HospitalZürich, CH-8091 Zürich, Switzerland; fInstitute of Experimental and Clinical Pharmacology and Toxicology, Heidelberg University, Medical Faculty Mannheim,D-68169 Mannheim, Germany; gDepartment of Cardiology, Theresienkrankenhaus, 68165 Mannheim, Germany; hBritish Heart Foundation Centre of ResearchExcellence, National Heart and Lung Institute, Faculty of Medicine, Imperial College London, London WC2A 3LJ, United Kingdom; iInstitute of AppliedPhysiology, University of Ulm, D-89081 Ulm, Germany; jInstitute of Anatomy and Cell Biology, University of Heidelberg, D-69120 Heidelberg, Germany;kDepartment of Molecular Pathology, German Cancer Research Center, D-69120 Heidelberg, Germany; lHoward Hughes Medical Institute and mDepartmentof Molecular Genetics, University of Texas Southwestern Medical Center, Dallas, TX 75390-9050; and nInternational Institute for Integrative SleepMedicine (WPI-IIIS), University of Tsukuba, Tsukuba 305-8575, Japan

Edited by David W. Russell, University of Texas Southwestern Medical Center, Dallas, TX, and approved August 1, 2014 (received for review May 15, 2014)

In preclinical studies, endothelin receptor A (ETA) antagonists(ETAi) attenuated the progression of heart failure (HF). However,clinical HF trials failed to demonstrate beneficial effects of ETAi.These conflicting data may be explained by the possibility thatestablished HF drugs such as adrenergic receptor blockers inter-fered with the mechanism of ETAi action in clinical trials. Here wereport that mice lacking ETA only in sympathetic neurons (SN-KO)showed less adverse structural remodeling and cardiac dysfunc-tion in response to pathological pressure overload induced bytransverse aortic constriction (TAC). In contrast, mice lacking ETAonly in cardiomyocytes (CM-KO) were not protected. TAC led toa disturbed sympathetic nerve function as measured by cardiacnorepinephrine (NE) tissue levels and [124I]-metaiodobenzylguani-dine-PET, which was prevented in SN-KO. In a rat model of HF, ETAiimproved cardiac and sympathetic nerve function. In cocultures ofcardiomyocytes (CMs) and sympathetic neurons (SNs), endothelin-1(ET1) led to a massive NE release and exaggerated CM hypertro-phy compared with CM monocultures. ETA-deficient CMs gaineda hypertrophic response through wild-type SNs, but ETA-deficientSNs failed to mediate exaggerated CM hypertrophy. Furthermore,ET1 mediated its effects indirectly via NE in CM-SN coculturesthrough adrenergic receptors and histone deacetylases, resultingin activation of the prohypertrophic transcription factor myocyteenhancer factor 2. In conclusion, sympathetic ETA amplifies ET1effects on CMs through adrenergic signaling pathways. Thus, anti-adrenergic therapies may blunt potentially beneficial effects ofETAi. Taken together, this may indicate that patients with β blockerintolerance or disturbed sympathetic nerve function could be eval-uated for a potential benefit from ETAi.

β adrenergic signaling | sympathetic nervous system | norepinephrinereuptake | epigenetic regulation | HDACs

Heart failure (HF) is common and, despite therapeuticadvancements, is linked to a poor prognosis (1). Various

neurohormones are elevated and correlate with the severity ofHF (2, 3). Among them, the role of endothelin-1 (ET1) has beenextensively studied (4, 5). ET1 was shown to be elevated in bloodsamples of HF patients, to predict outcome and to play a crucialrole for adverse cardiac remodeling (6). In 1996, it was demon-strated that endothelin receptor inhibition is cardioprotective ina preclinical model of HF (7), which was supported by others (8–10). However, despite convincing preclinical data, unspecificET1 receptor and specific ETA inhibitors (ETAi) showed nobeneficial effects in clinical HF trials (11, 12). In contrast toanimal studies, HF patients in clinical trials are usually cotreated

with other HF drugs such as angiotensin-converting enzymeinhibitors or adrenergic receptor antagonists, raising the pos-sibility that these drugs interfere with the effects of ETAi.Conflicting data came also from genetic studies showing thatmice lacking ETA only in cardiomyocytes (CM-KO) were notprotected when exposed to cardiac stress (13). Thus, thequestion has arisen of whether ETA in nonmyocytes affectscritical CM signaling pathways via cell–cell communicationmechanisms.ETA is also expressed in neural crest derived tissue such as

sympathetic neurons (SNs) that innervate the heart and releasenorepinephrine (NE) (14, 15), one of the neurotransmitters thatregulate cardiac function and growth. The success of HF therapywith adrenergic receptor antagonists that inhibit the target re-ceptor of NE in adult CMs underscores the pivotal role of NEin HF (16). Consequently, the mechanisms by which NE is re-leased from sympathetic neurons impact on signaling cascades inCMs (17–19). We now hypothesized that ETAi mediates their

Significance

In failing hearts, norepinephrine (NE) net release and endo-thelin-1 (ET1) levels are increased. ET1 receptor antagonistswere successfully used in preclinical heart failure (HF) studies,but clinical studies did not show beneficial effects in HFpatients. We found that mice lacking endothelin receptor A(ETA) only in sympathetic neurons but not in cardiomyocyteswere protected from the development of HF. Mechanistically,the presynaptic reuptake of NE within the heart was preservedin mice lacking ETA only in sympathetic neurons. These dataprovide an explanation of why patients in clinical studies donot benefit from ET1 receptor antagonists, because thesepatients are usually treated with β blockers, which interferewith the mechanism of action identified here.

Author contributions: L.H.L., T.W., M.H., H.-J.G., U.H., H.A.K., and J.B. designed research;L.H.L., J.S.R., S.J.B., M.M.K., J.K., W.M., K.S., S.E.H., H.-J.G., and J.B. performed research;F.E., T.F.L., M.D.S., R.P., and M.Y. contributed new reagents/analytic tools; L.H.L., J.S.R.,S.J.B., M.M.K., W.M., S.E.H., T.W., M.H., H.-J.G., U.H., and J.B. analyzed data; and L.H.L.,H.A.K., and J.B. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: johannes.backs@med.uni-heidelberg.de.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1409026111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1409026111 PNAS | September 16, 2014 | vol. 111 | no. 37 | 13499–13504

MED

ICALSC

IENCE

S

beneficial effects not through the inhibition of ETA on CMs buton SNs and consequently indirectly through an attenuated acti-vation of α-adrenergic receptors (αARs) and β-adrenergicreceptors (βARs). In CMs, ETA, and also αARs and βARs, sig-nals to kinases such as protein kinase D (PKD) and calciumcalmodulin dependent kinase II (CaMKII). These kinases inducephosphorylation-dependent nucleo-cytoplasmic shuttling of classII histone deacetylases (HDACs) (20, 21), resulting in the acti-vation of the transcription factor myocyte enhancer factor 2(MEF2), a crucial inductor of pathological cardiac remodeling (22).In this study, we investigated two KO mouse lines, one lacking

ETA only in SNs (SN-KO) and another one lacking ETA only inCMs (CM-KO). We show that SN-ETA but not CM-ETA con-tributes to adverse cardiac remodeling.

ResultsDeficiency of ETA Specifically in SNs but Not in CMs Protects fromPathological Cardiac Remodeling. To investigate the role of ETAin specific cell types, we generated conditional mutant micelacking ETA selectively in SNs (SN-KO) vs. CMs (CM-KO) (Fig.S1A). Quantitative RT-PCR revealed that transgenic expressionof Cre under the control of the dopamine-β-hydroxylase promoter(DBH-Cre) (23) led to a substantial loss of ETA in sympatheticganglia of SN-KO mice, whereas transgenic Cre expression underthe control of α-myosin heavy chain promoter (αMHC-Cre) ledto a comparable deletion of ETA in the myocardium of CM-KOmice (Fig. S1B). Both lines developed normally without anyobvious anatomical abnormalities. We then challenged theselines with pathological pressure overload induced by transverseaortic constriction (TAC) and compared them to control litter-mates (Ctrl. LM). The degree of TAC was similar in all experi-mental groups (Fig. S2). CM-KO mice did not show any signs ofcardioprotection with regard to cardiac hypertrophy (Fig. S3 A,D, and E), interstitial fibrosis (Figs. S3 B and F and S4), function(Fig. S3 C and G), and pulmonary congestion (Fig. S3H) or geneexpression (Fig. S3I). Strikingly, SN-KO mice did not developmassive cardiac hypertrophy (Fig. 1 A, D, and E), interstitial fi-brosis (Fig. 1 B and F and Fig. S4), or cardiac dysfunction (Fig. 1

C and G). Pulmonary congestion as an additional sign of severeleft ventricular dysfunction was diminished in SN-KO (Fig. 1H).Moreover, dysregulation of genes that are typically associatedwith adverse cardiac remodeling (including nppa, myh7, Col3a,and Serca) was attenuated in SN-KO after TAC compared withCtrl. LM (Fig. 1I).

ETA Impairs Sympathetic Nerve Function in SN-KO After TAC. To testwhether sympathetic nerve function is preserved in SN-KO afterTAC, sympathetic indices including NE tissue levels and meta-iodobenzylguanidine (MIBG) uptake were determined. Depletedmyocardial NE stores indicated a substantial increase in NE netsecretion as an estimate for overactivation of the local cardiacsympathetic nervous system (24, 25). Accordingly, we found leftventricular NE tissue levels to be decreased after TAC in Ctrl.LM but not in SN-KO (Fig. 2A). To address the question ofwhether depleted myocardial NE stores were mediated—at leastin part—by a reduced NE reuptake via the NE transporter(NET), we used the [124I]-MIBG-PET method to assess cardiacNET activity in vivo (17, 18, 26). This method does not directlymeasure cardiac sympathetic activity but serves as an index forcardiac sympathetic activity. We found that TAC leads to a re-duction of MIBG uptake in Ctrl. LM but not SN-KO, stronglysuggesting that ET1 impairs sympathetic nerve function in vivovia ETA and NET-mediated NE reuptake (Fig. 2 B and C).

ETA Inhibitors Improve Sympathetic Nerve Function in a Rat Model ofHF. To investigate whether the findings in the mouse modelstranslate to larger rodents, we first perfused isolated hearts ofhealthy rats with ET1 and found a significantly reduced NE up-take, which could be diminished by pretreatment with darusentan(ETAi) (Fig. 3A). To study whether ETAi normalizes sympatheticnerve function in a rat model of HF, we used salt-sensitive (DS) vs.salt-resistant (DR) Dahl rats, both fed on a high-salt diet (27). DSrats displayed an impaired NE uptake (Fig. 3B) and depletedcardiac NE stores (Fig. 3C). In accordance with the genetic de-letion of ETA in TAC mice, ETAi significantly improved NE up-take and NE stores (Fig. 3 B and C), confirming an improved

Fig. 1. Sympathetic ETA is essential for the development of adverse cardiac remodeling. (A) Representative four-chamber view of mice lacking ETA insympathetic neurons (SN-KO) vs. control littermates with normal ETA expression (Ctrl. LM) 7 wk after sham or TAC surgery. (Scale bar, 1 mm.) (B) Repre-sentative trichrome-stained sections. (Scale bar, 100 μm.) (C) Representative transthoracic echocardiography of the left ventricle (M-Mode, midventricular). (D)Heart weight/body weight (HW/BW) ratios (n = 5–7 per group). (E) Quantification of cardiomyocyte size (>80 cells per sample from more than five fields ofview; n ≥ 3). (F) Quantification of fibrosis in relation to Ctrl. LM (n ≥ 3 per group). (G) Fractional shortening 3 wk after TAC (n = 5–7 per group). (H) Lungweight/body weight (LW/BW) ratios 7 wk after TAC surgery (n = 5–7 per group). (I) Transcript levels in hearts from Ctrl. LM and SN-KO 7 wk after TAC surgeryfrom genes as indicated in relation to sham-operated Ctrl. LM (n = 5–7 per group). Values are presented as mean ± SEM. *P < 0.05; **P < 0.01.

13500 | www.pnas.org/cgi/doi/10.1073/pnas.1409026111 Lehmann et al.

sympathetic nerve function. The improvement of sympatheticnerve function by ETAi was associated with an attenuation ofcardiac hypertrophy, a prevention of pulmonary congestion, anda reduced mortality (Fig. 3 D–F). Moreover, we observed furtherimprovements of hemodynamic characteristics and the HF bio-marker atrial natriuretic peptide (ANP) (Fig. 3G). Because wefocused in the present study on cardiac function, we cannotrule out that the reduced mortality is—at least in part—due toan improved function of the kidney or other organs that areinnervated by sympathetic neurons.

Sympathetic Neurons Regulate Cardiomyocyte Hypertrophy ThroughSympathetic ETA. Next, we aimed at recapitulating the in vivofindings in isolated cell culture systems. Treatment of rat sym-pathetic neuron-cardiomyocyte cocultures (SN-CM) with 10 nMET1 led to a fourfold increase of NE net release into the culturemedium, which was inhibited by ETAi (Fig. S5A). Although wecannot formally rule out that this effect was mediated by ET1-mediated exocytotic NE release, we propose based on previousexperiments (28) that this effect is rather due to inhibition of NEreuptake. The presence of SNs in the SN-CM coculture wasconfirmed by Western blot analysis using an antibody against theSN-specific marker tyrosine hydroxylase (TH) (Fig. S5B). Re-markably, ET1 led to exaggerated CM hypertrophy in CM-SNcocultures (Fig. S5C). To further support the requirement of thesympathetic ETA for CM hypertrophy, we also performed CM-SN coculture experiments with selective ETA deletions eitherin CMs or SNs. First, we adenovirally expressed Cre in isolatedmurine SNs from ETA

fl/fl mice (SNKO) and cocultured these cells

with WT rat CMs (CMWT). The deletion of SN-ETA abolishedexaggerated CM hypertrophy in SNKO-CMWT cocultures (Fig. S5D and E). Immunocytochemistry with an antibody against ETAillustrates not only a sufficient deletion of ETA through Cretransduction but also that the concentration of ETA on SNs islargely higher than on CMs (Fig. S5D). Second, we adenovirallyexpressed Cre in isolated murine CMs from ETA

fl/fl mice (CMKO)and cocultured these cells with WT rat SNs (SNWT). The deletionof CM-ETA inhibited ET1-induced hypertrophy in monocultures(CMKO) but the addition of SNWT (CMKO-SNWT cocultures) re-stored the hypertrophic response of CMKO, formally proving thatETA on CMs is dispensable for ET1-induced CM hypertrophy inthe presence of SNs (Fig. S5 F and G). Deletion of ETA in CMKO

was confirmed by Western blot (Fig. S6).

Sympathetic ETA Regulates CM Hypertrophy Through AdrenergicReceptors. To test whether the SN-dependent increase in CMsize depends on adrenergic receptors, we used pharmacologicalinhibitors for αARs and βARs (αARi and βARi). Strikingly,ET1-induced CM hypertrophy (Fig. 4 A and B) and proteinsynthesis (Fig. 4C) in CM-SN coculture, but not in CM mono-culture, could be attenuated by βARi, whereas β1ARi seems tobe sufficient to inhibit exaggerated hypertrophy (Fig. S7).However, αARi was ineffective in this regard. As a control, ETAisufficiently blocked CM hypertrophy in both culture systems.

Sympathetic ETA Activates the HDAC–MEF2 Axis in CMs. ETA sig-naling via protein kinases was suggested to result in a phos-phorylation-dependent nucleo-cytoplasmic shuttling of class IIHDACs, which in turn results in activation of the transcriptionfactor MEF2 and gene programs that lead to cardiomyocytehypertrophy (20, 22). In SN-CM, ET1 led to a significant in-creased nuclear export of adenovirally expressed HDAC4 andHDAC5 (Fig. 5 A–C). HDAC4 nuclear export was significantlyreduced by βARi, whereas HDAC5 nuclear export was signifi-cantly reduced by αARi. Moreover, βAR signaling leads specif-ically to 14-3-3 binding to HDAC4 but not HDAC5 (Fig. S8).ETAi prevented nuclear export of both HDAC4 and HDAC5.Using MEF2 reporter assays, we found that ET1 induced ahigher increase in MEF2 activity in SN-CM coculture comparedwith CM monoculture (Fig. 5D). A hallmark of transcrip-tional remodeling is a switch in the expression of myosinheavy chain genes from myh6 to myh7 (29). ET1 caused amoderate myh6/myh7 switch in CM monoculture but, remarkably,a dramatic switch in SN-CM coculture (Fig. 5E).

DiscussionIn this study, we demonstrate that ET1 mediates adverse cardiacremodeling not through CM-ETA but through SN-ETA.

Fig. 2. Sympathetic ETA impairs sympathetic nerve function. (A) ELISA-based quantification of NE within the left ventricle of SN-KO 7 wk after TACin relation to sham-operated mice (n = 4–7 per group). (B) RepresentativePET scan images of [124I]-MIBG uptake in the frame 45–60 min after MIBGinjection into sham- and TAC-operated Ctrl. LM vs. SN-KO. (C) Quantitativeanalysis of the influxes (K1) of [124I]-MIBG via PET scan 7 wk after TAC(controls, n = 4; TAC, n = 6). Values presented as mean ± SEM. *P < 0.05.

Fig. 3. ETA inhibitors improve sympathetic nervefunction in a rat model of HF. (A) [3H]-NE uptake inisolated perfused hearts in the presence/absence ofET1 and ETAi. (B) [

3H]-NE uptake with/without ETAi(darusentan) treatment in salt-sensitive Dahl rats(DS) compared with salt-resistant rats (DR). Almostcomplete inhibition by desipramine (DMI) indicatesthat the majority of NE is taken up by the NEtransporter (NET). (C) Left ventricular NE (tissuelevel) in DS vs. DR. ETAi treatment attenuates thedepletion of left ventricular NE (n = 5–9 per group).(D) ETAi attenuates cardiac hypertrophy (HW/BW)and (E) pulmonary congestion (LW/BW) in DS (n >10 per group). (F) Kaplan–Meier analysis of DStreated with/without ETAi (n = 10 per group). (G)Left ventricular ejection fraction (EF), left ventricu-lar enddiastolic pressure (LVEDP), contractility index (dP/dT+), relaxation index (and dP/dT−), and plasma levels ANP in DS treated with/without ETAi. Valuesare presented as mean ± SEM. *P < 0.05; **P < 0.01; and §P < 0.05 vs. DS.

Lehmann et al. PNAS | September 16, 2014 | vol. 111 | no. 37 | 13501

MED

ICALSC

IENCE

S

Sympathetic ETA Is Required for Abnormal Sympathetic Nerve Functionand Pathological Cardiac Remodeling.We have shown previously inisolated perfused hearts that ET1 impairs NET-dependent NEreuptake but not exocytotic NE release (28). However, isolatedheart perfusion in combination with ETAi did not allow the de-termination of the underlying critical cell types or the inves-tigation of disease-relevant remodeling processes. Now wedemonstrate that SN-ETA determines the cardiac vulnerabilityto pathological pressure overload. Based on our previous data(28) and on the MIBG uptake measurements, we conclude thatinhibition of SN-ETA leads to an attenuation of adrenergicneurotransmission and to cardioprotection. In contrast to SN-KOs, CM-KOs were not protected from pathological cardiacstress. The latter finding is consistent with the prior finding thatCM-KO mice did not show an improvement of cardiac functionafter angiotensin or isoproterenol infusion (13) and highlightsthe critical role of ETA in nonmyocytes. We cannot exclude thatETB in CMs plays a critical role during pathological cardiacremodeling because the role of the ETB has not been in-vestigated in conditional KO models. However, some evidence isprovided from mice lacking ETB globally. These mice developintestinal aganglionosis and megacolon with early death, whichcould be rescued by reexpression of the ETB under the control ofthe dopamine-β-hydroxylase promoter, indicating that the SN-

ETB may also be important for sympathetic nerve function (15,30). Previous pharmacological data indicate that sympatheticETB does not influence NE reuptake but inhibits exocytotic NErelease in the heart (28). Through the use of CM-SN cocultures,we provide additional evidence that sympathetic ETA is requiredfor the transmission of ET1-dependent adrenergic and hyper-trophic signals to CMs. These results explain how ETAi improveslocal NE homeostasis as also shown in DS rats on a high-saltdiet. However, why did selective ETAi not result in reducedplasma NE levels in HF patients (11, 31)? This discrepancy mightbe explained by the idea that plasma NE levels are not idealsurrogates for the local cardiac adrenergic drive but rather serveas a marker for a generalized sympathetic activation. In fact, itwas shown that the cardiac adrenergic drive independently altersfrom the generalized sympathetic tone (32).

Role of Nonmyocytes in the Regulation of ET1-Dependent SignalTransduction in CMs. Until now, it was assumed that ET1 medi-ates its hypertrophic effects through ETA on CMs (33), whichmainly is based on the observation that ET1 exerts hypertrophyof cultured CMs. However, CM primary cultures are heteroge-nous and contain—dependent on the type of isolation and du-ration of cultivation—variable amounts of nonmyocytes such asfibroblasts, endothelial cells, smooth muscle cells, and intrinsic

Fig. 4. Sympathetic ETA regulates CM hypertrophy through adrenergic receptors. (A) Immunocytochemistry of neonatal rat ventricular myocytes (NRVMs,CMs) and cocultures of CMs together with neonatal rat sympathetic neurons (SNs) was performed using antibodies recognizing sarcomeric α-actinin (red) andTH (green). Cultured cells were treated for 24 h with 10 nM ET1 in the absence or presence of the inhibitors darusentan (1 μM; ETAi), prazosin (1 μM; αARi), orpropranolol (1 μM; βARi). (B) CM cell surface area was quantified. #P < 0.05 vs. ET1-treated; *P < 0.05 ET1-treated SN-CM vs. CM (n = 100–150/per group). (C)[3H]-leucin incorporation in CMs vs. SN-CM with treatment as indicated (ET1: 10 nM 18 h; αAR: 1 μM prazosin; βAR: 1 μM propranolol; 1 h before ET1treatment). n = 4. *P < 0.05 vs. nontreated SN-CM; #P < 0.05 vs. ET1-treated.

Fig. 5. Sympathetic ETA activates the HDAC-MEF2axis in CMs. (A) CMs were transduced with adeno-viruses harboring Flag-HDAC4 or -HDAC5, starvedfor 4 h, and then treated for 12 h with 10 nM ET1 inthe absence or presence of the inhibitors dar-usentan (1 μM; ETAi), prazosin (1 μM; αARi), orpropranolol (1 μM; βARi). Immunocytochemistrywas performed using antibodies recognizing Flag.(B and C) Quantitative analysis of cytosolic vs. nu-clear localized HDACs. *P < 0.05 ET1-treated vs.nontreated group and SN-CM vs. CM. #P < 0.05 in-hibitor treated vs. ET1-treated group (n ≥ 40 cellsper group). (D) MEF2 activity was measured byMEF2 luciferase reporter assay (n = 4). *P < 0.05. (E)Transcripts of myh6 and myh7 after 24-h treatmentwith10 nM ET1 (n = 6). *P < 0.05. (B–E) Values arepresented as mean ± SEM. (F) Working model: ET1stimulates ETA on CMs but to a larger extent on SNs.ET1 leads via SN-ETA to a substantial increase of NEnet release by inhibiting NET-dependent NE reup-take, which in turn leads to an activation of αARs and βARs and nuclear export of HDAC5 and HDAC4 through distinct signaling pathways (pathwayhighlighted in red). Thus, the involvement of adrenergic receptors amplifies ET1-driven activation of MEF2.

13502 | www.pnas.org/cgi/doi/10.1073/pnas.1409026111 Lehmann et al.

neurons (34). This heterogeneity is potentially one of the reasonsfor variable results from isolation to isolation as experienced bymany investigators. The results of this study demonstrate thatSN-ETA is required for pathological remodeling of the heart andthat CM-ETA is dispensable in the presence of SNs. In thisregard, previous studies on ET1-induced CM hypertrophy usu-ally used relatively high concentrations of ET1 (10–100 nM).Accordingly, we observed robust CM hypertrophy in CM mon-ocultures only with 100 nM ET1. In cocultures, by contrast,10 nM ET1 was sufficient to induce robust CM hypertrophy,suggesting that the ETA on SNs is more sensitive than on CMs.The reason might lie in the relative expression pattern. Fig. S5Dimplies that the ETA is expressed to a larger extent in SNs ascompared to CMs. Likewise, we reported previously that ET1concentrations around 100 pM affect NE reuptake in isolatedperfused hearts (28). Thus, the possibility exists that physiolog-ical ET1 levels that were reported to lie in the picomolar rangepredominantly act via SN-ETA rather than CM-ETA. Thesefindings are not surprising from a developmental point of view:SNs are derived from neural crest cells and loss-of-functionstudies have unmasked that the deletion of ETA in neural crestresults in severe cardiac phenotypes (35), whereas the CM de-letion of ETA did not result in an obvious phenotype in classicalpathological models (13).

Divergent and Convergent Signaling of ET1.ETA and αAR signal viaPKD toward HDAC5 to derepress MEF2 target genes that driveadverse cardiac remodeling (20, 36). We now show that ET1signals via SN-ETA indirectly to CM-αAR and then convergeswith CM-ETA on HDAC5 signaling. However, our results alsorevealed that SN-ETA transduces its signals via βARs toHDAC4, indicating the involvement of other kinases, becausePKD is not a downstream kinase of βARs (37, 38). These dataare consistent with our previous report that CaMKII selectivelysignals to HDAC4 (21, 39). In contrast to PKD, CaMKII is wellknown to be activated by βARs (40, 41). How βAR signaling actson HDAC5 is under debate (42–44). In our hands, βAR acti-vation leads to more 14-3-3 binding (as a measure of cytosolicaccumulation of HDACs) of HDAC4 but to less 14-3-3 bindingto HDAC5, suggesting that βARs lead to (i) cytosolic accumu-lation of HDAC4 possibly by CaMKII and (ii) nuclear accumu-lation of HDAC5. The latter might be explained by a previouslyreported inhibitory effect of protein kinase A (a major down-stream mediator of βAR signaling) on PKD (44, 45). Based onour data, we propose a model (Fig. 5F), by which CM-ETA sig-nals specifically through PKD to HDAC5, whereas SN-ETAsignals indirectly via NE through αARs to HDAC5 but throughβARs to HDAC4. Why does ET1 signaling diverge to adrenergicreceptors that in turn converge on distinct class II HDAC familymembers? Both, HDAC4 and HDAC5 have been demonstratedto inhibit MEF2 (46). Thus, our findings provide a mechanisticbasis by which SN-ETA amplifies de-repression of MEF2 by re-moving more than one HDAC family member from the nucleus.

Potential Clinical Implications. An improved NE reuptake by ETAileads to restoration of NE stores, which was reported to be as-sociated with an increased exercise tolerance in patients (24).Thus, we hypothesize that exercise tolerance may be an addi-tional clinical readout for future ETAi studies. Based on thefindings of the present study, we hypothesized that individualpatients not tolerating β blockers might benefit from ETAi.Likewise, studies in which only up to 56% of the patients re-ceived β blockers showed a significant benefit from ETAi orunspecific endothelin receptor inhibitors (Table S1). In studiesin which more patients received antiadrenergic therapies, nosignificant benefit was observed, suggesting that β blockers in-deed blunt the potential therapeutic effects of ETAi. We alsoaimed at retrieving supporting information from the largest ETAi

trial on HF, namely the endothelin-a receptor antagonist trial inheart failure trial (11). However, the size of the subgroups (inparticular patients without antiadrenergic therapy) did not allowa statistically reliable analysis. Thus, an ETAi trial with this smallHF patient cohort might be an interesting approach to test thehypothesis that patients without antiadrenergic therapy mightbenefit from ETAi. Moreover, imaging techniques such as [123I]-MIBG uptake (17, 18, 26) that allow an assessment of the sym-pathetic nerve function in patients may serve as a tool for theselection of patients that benefit in particular from ETAi.In conclusion, we unmasked, by the use of cell type-specific in

vivo and ex vivo gene deletion tools, a critical role for sympa-thetic ETA in the development and progression of HF. ET1amplifies its effect on CMs indirectly by the stimulation of adren-ergic neurotransmission and the activation of αARs and βARs andsubsequently distinct HDAC- and MEF2-dependent signalingpathways. Other than hypothesized before, ETA on CMs doesnot seem to critically contribute to the development of HF.

Materials and MethodsExperimental Animals. All experiments were performed in accordance withrelevant guidelines and regulations. This investigation followed the Principleof Laboratory Animal Care (NIH Publication No. 86-23, revised 1985) and wasapproved by the authorities of the Regierungspräsidium Karlsruhe. For de-tailed information, see SI Materials and Methods.

[124I]MIBG-PET. The precursor resin for the production of ultratrace [124I]MIBGwas a kind gift from James F. Kronauge (Molecular Insight Pharmaceuticals,Cambridge, MA). To prevent eventual saturation of the norepinephrinetransporter by MIBG in the animal model, MIBG of high specific activity wasproduced. For this purpose, the solid labeling approach was applied (47). Fordetailed information, see SI Materials and Methods.

Isolated Heart Perfusion. Hearts were perfused according to the Langendorffmethod as previously published (27). For detailed information, see SI Materialsand Methods.

Determination of Neurohormones.Measurement of neurohormones followedthe protocol as previously described (27). For detailed information, see SIMaterials and Methods.

Transthoracic Echocardiography. Echocardiographic evaluation was per-formed using a VisualSonicsVevo 2100 echocardiograph. For detailed in-formation, see SI Materials and Methods.

Isolation of Primary Cells. SNs were prepared and cultured from newbornSprague–Dawley rats or ETAfl/fl mice as described previously (48). Briefly,freshly isolated superior cervical ganglia from 1- to 3-d-old pups were dis-sected, incubated for 45 min in 1.5 mg/mL collagenase, titrated, and thenpreplated on cultural plastic to permit the attachment of nonneuronal cells.After 1 h, SNs were plated with CM medium (SI Materials and Methods) andcocultured for 3–5 d with an additional 10 ng/mL NGF (rat recombinant).After 24 h in coculture, 1 μM cytosine arabinofuranoside was added to stopcell division of potential contaminating fibroblasts. The isolation of CMs isdescribed in detail in SI Materials and Methods.

Plasmids and Reagents. For detailed information, see SI Materials and Methods.

Indirect Immunofluorescence, Western Blot Analysis and Coimmunoprecipitation,Histology, RNA Analysis, Reporter Assays, [3H]-Leucine Incorporation, andAdenovirus Production. For detailed information, see SI Materials and Methods.

Statistics. The results are expressed as the mean ± SEM. Statistical analysiswas performed using SPSS 12.0 software (SPSS). Differences between groupswere tested by one-way ANOVA with post hoc comparisons by Bonferroni’smultiple comparison test or unpaired Student t test where appropriate.Kaplan–Meier survival analysis was performed using the log-rank test. In theimmunofluorescence experiments, the nucleocytoplasmatic distribution ofHDAC4 and -5 was analyzed in >40 cells per experiment, whereas a ratio wasbuilt between fluorescence in nucleus vs. cytosol using ImageJ software. Inall tests, P < 0.05 was considered statistically significant.

Lehmann et al. PNAS | September 16, 2014 | vol. 111 | no. 37 | 13503

MED

ICALSC

IENCE

S

ACKNOWLEDGMENTS. We thank Michaela Oestringer, Claudia Heft, andUlrike Oehl for expert technical assistance and David Stanmore for editingthe manuscript. Darusentan was a gift from Klaus Muenter. The precursorresin for the production of ultratrace MIBG was a gift from James F. Kronauge.Adenovirus for Cre recombinase was a gift from Oliver Müller. Floxed ETAmice were forwarded from Rohini Kuner. J.B. was supported by DeutscheForschungsgemeinschaft Grants BA 2258/2-1 and SFB 1118, European Com-mission Grants FP7-Health-2010 and MEDIA-261409, the Deutsches Zentrum

für Herz-Kreislauf-Forschung [DZHK (German Centre for Cardiovascular Re-search)], and the German Ministry of Education and Research. L.H.L. wassupported by the “Karin-Nolte-Stiftung” and is recipient of the Heidelberg Re-search Center for Molecular Medicine Career Development fellowship. J.S.R.was supported by the Otto Hess Scholarship of the German Cardiac Society.M.Y. was supported by the Japan Society for the Promotion of Sciencethrough the Funding Program for World-Leading Innovative R&D on Scienceand Technology, initiated by the Council for Science and Technology Policy.

1. Roger VL, et al.; American Heart Association Statistics Committee and Stroke StatisticsSubcommittee (2012) Executive summary: Heart disease and stroke statistics—2012update: A report from the American Heart Association. Circulation 125(1):188–197.

2. Benedict CR, et al.; SOLVD Investigators (1994) Relation of neurohumoral activationto clinical variables and degree of ventricular dysfunction: A report from the Registryof Studies of Left Ventricular Dysfunction. J Am Coll Cardiol 23(6):1410–1420.

3. Rockman HA, Koch WJ, Lefkowitz RJ (2002) Seven-transmembrane-spanning re-ceptors and heart function. Nature 415(6868):206–212.

4. Yanagisawa M, et al. (1988) A novel potent vasoconstrictor peptide produced byvascular endothelial cells. Nature 332(6163):411–415.

5. Barton M, Yanagisawa M (2008) Endothelin: 20 years from discovery to therapy. Can JPhysiol Pharmacol 86(8):485–498.

6. McMurray JJ, Ray SG, Abdullah I, Dargie HJ, Morton JJ (1992) Plasma endothelin inchronic heart failure. Circulation 85(4):1374–1379.

7. Sakai S, et al. (1996) Inhibition of myocardial endothelin pathway improves long-termsurvival in heart failure. Nature 384(6607):353–355.

8. Chen XC, et al. (2001) Effects of endothelin receptor A antagonist FR139317 on ratswith congestive heart failure. Acta Pharmacol Sin 22(10):896–900.

9. Mulder P, et al. (2002) Long-term survival and hemodynamics after endothelin-a re-ceptor antagonism and angiotensin-converting enzyme inhibition in rats with chronicheart failure: Monotherapy versus combination therapy. Circulation 106(9):1159–1164.

10. Yamauchi-Kohno R, et al. (1999) Role of endothelin in deterioration of heart failuredue to cardiomyopathy in hamsters: Increase in endothelin-1 production in the heartand beneficial effect of endothelin-A receptor antagonist on survival and cardiacfunction. Circulation 99(16):2171–2176.

11. Anand I, et al.; EARTH investigators (2004) Long-term effects of darusentan on left-ventricular remodelling and clinical outcomes in the EndothelinA Receptor Antago-nist Trial in Heart Failure (EARTH): Randomised, double-blind, placebo-controlledtrial. Lancet 364(9431):347–354.

12. Coletta A, Thackray S, Nikitin N, Cleland JG (2002) Clinical trials update: Highlights ofthe scientific sessions of The American College of Cardiology 2002: LIFE, DANAMI 2,MADIT-2, MIRACLE-ICD, OVERTURE, OCTAVE, ENABLE 1 & 2, CHRISTMAS, AFFIRM,RACE, WIZARD, AZACS, REMATCH, BNP trial and HARDBALL. Eur J Heart Fail 4(3):381–388.

13. Kedzierski RM, et al. (2003) Cardiomyocyte-specific endothelin A receptor knockoutmice have normal cardiac function and an unaltered hypertrophic response to an-giotensin II and isoproterenol. Mol Cell Biol 23(22):8226–8232.

14. Isaka M, Kudo A, Imamura M, Kawakami H, Yasuda K (2007) Endothelin receptors,localized in sympathetic nerve terminals of the heart, modulate norepinephrine re-lease and reperfusion arrhythmias. Basic Res Cardiol 102(2):154–162.

15. Lehmann LH, Stanmore DA, Backs J (2014) The role of endothelin-1 in the sympatheticnervous system in the heart [published online ahead of print March 13, 2014]. Life Sci,10.1016/j.lfs.2014.03.005.

16. Anonymous (1999) Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF). Lancet353(9169):2001–2007.

17. Nakata T, et al. (1998) Cardiac death prediction and impaired cardiac sympatheticinnervation assessed by MIBG in patients with failing and nonfailing hearts. J NuclCardiol 5(6):579–590.

18. Ogita H, et al. (2001) Prognostic significance of cardiac (123)I metaiodobenzylguanidineimaging for mortality and morbidity in patients with chronic heart failure: A prospectivestudy. Heart 86(6):656–660.

19. Kreusser MM, et al. (2006) Injection of nerve growth factor into stellate ganglia im-proves norepinephrine reuptake into failing hearts. Hypertension 47(2):209–215.

20. Vega RB, et al. (2004) Protein kinases C and D mediate agonist-dependent cardiachypertrophy through nuclear export of histone deacetylase 5. Mol Cell Biol 24(19):8374–8385.

21. Backs J, Song K, Bezprozvannaya S, Chang S, Olson EN (2006) CaM kinase II selectivelysignals to histone deacetylase 4 during cardiomyocyte hypertrophy. J Clin Invest116(7):1853–1864.

22. Kim Y, et al. (2008) The MEF2D transcription factor mediates stress-dependent cardiacremodeling in mice. J Clin Invest 118(1):124–132.

23. Parlato R, Otto C, Begus Y, Stotz S, Schütz G (2007) Specific ablation of the tran-scription factor CREB in sympathetic neurons surprisingly protects against de-velopmentally regulated apoptosis. Development 134(9):1663–1670.

24. Rundqvist B, Eisenhofer G, Elam M, Friberg P (1997) Attenuated cardiac sympatheticresponsiveness during dynamic exercise in patients with heart failure. Circulation95(4):940–945.

25. Backs J, et al. (2001) The neuronal norepinephrine transporter in experimental heartfailure: Evidence for a posttranscriptional downregulation. J Mol Cell Cardiol 33(3):461–472.

26. Carrió I, Cowie MR, Yamazaki J, Udelson J, Camici PG (2010) Cardiac sympatheticimaging with mIBG in heart failure. JACC Cardiovasc Imaging 3(1):92–100.

27. Buss SJ, et al. (2006) Spironolactone preserves cardiac norepinephrine reuptake insalt-sensitive Dahl rats. Endocrinology 147(5):2526–2534.

28. Backs J, Bresch E, Lutz M, Kristen AV, Haass M (2005) Endothelin-1 inhibits the neu-ronal norepinephrine transporter in hearts of male rats. Cardiovasc Res 67(2):283–290.

29. Nadal-Ginard B, Mahdavi V (1989) Molecular basis of cardiac performance. Plasticityof the myocardium generated through protein isoform switches. J Clin Invest 84(6):1693–1700.

30. Hosoda K, et al. (1994) Targeted and natural (piebald-lethal) mutations of endothe-lin-B receptor gene produce megacolon associated with spotted coat color in mice.Cell 79(7):1267–1276.

31. Lüscher TF, et al.; Heart Failure ET(A) Receptor Blockade Trial (2002) Hemodynamicand neurohumoral effects of selective endothelin A (ET(A)) receptor blockade inchronic heart failure: The Heart Failure ET(A) Receptor Blockade Trial (HEAT). Circu-lation 106(21):2666–2672.

32. Rundqvist B, Elam M, Bergmann-Sverrisdottir Y, Eisenhofer G, Friberg P (1997) In-creased cardiac adrenergic drive precedes generalized sympathetic activation in hu-man heart failure. Circulation 95(1):169–175.

33. Vega RB, Bassel-Duby R, Olson EN (2003) Control of cardiac growth and function bycalcineurin signaling. J Biol Chem 278(39):36981–36984.

34. Armour JA (2008) Potential clinical relevance of the ‘little brain’ on the mammalianheart. Exp Physiol 93(2):165–176.

35. Clouthier DE, et al. (1998) Cranial and cardiac neural crest defects in endothelin-Areceptor-deficient mice. Development 125(5):813–824.

36. Fielitz J, et al. (2008) Requirement of protein kinase D1 for pathological cardiac re-modeling. Proc Natl Acad Sci USA 105(8):3059–3063.

37. Avkiran M, Rowland AJ, Cuello F, Haworth RS (2008) Protein kinase d in the cardio-vascular system: Emerging roles in health and disease. Circ Res 102(2):157–163.

38. Cuello F, et al. (2007) Protein kinase D selectively targets cardiac troponin I and reg-ulates myofilament Ca2+ sensitivity in ventricular myocytes. Circ Res 100(6):864–873.

39. Backs J, et al. (2009) The delta isoform of CaM kinase II is required for pathologicalcardiac hypertrophy and remodeling after pressure overload. Proc Natl Acad Sci USA106(7):2342–2347.

40. Bers DM (2010) CaMKII inhibition in heart failure makes jump to human. Circ Res107(9):1044–1046.

41. Grimm M, Brown JH (2010) Beta-adrenergic receptor signaling in the heart: Role ofCaMKII. J Mol Cell Cardiol 48(2):322–330.

42. Chang CW, et al. (2013) Acute β-adrenergic activation triggers nuclear import ofhistone deacetylase 5 and delays G(q)-induced transcriptional activation. J Biol Chem288(1):192–204.

43. Haworth RS, Stathopoulou K, Candasamy AJ, Avkiran M (2012) Neurohormonalregulation of cardiac histone deacetylase 5 nuclear localization by phosphorylation-dependent and phosphorylation-independent mechanisms. Circ Res 110(12):1585–1595.

44. Sucharov CC, Dockstader K, Nunley K, McKinsey TA, Bristow M (2011) β-Adrenergicreceptor stimulation and activation of protein kinase A protect against α1-adrenergic-mediated phosphorylation of protein kinase D and histone deacetylase 5. J Card Fail17(7):592–600.

45. Haworth RS, Roberts NA, Cuello F, Avkiran M (2007) Regulation of protein kinase Dactivity in adult myocardium: Novel counter-regulatory roles for protein kinaseCepsilon and protein kinase A. J Mol Cell Cardiol 43(6):686–695.

46. Lu J, McKinsey TA, Nicol RL, Olson EN (2000) Signal-dependent activation of the MEF2transcription factor by dissociation from histone deacetylases. Proc Natl Acad Sci USA97(8):4070–4075.

47. Hunter DH, Zhu X (1999) Polymer-supported radiopharmaceuticals: [131I]MIBG and[123I]MIBG. J Labelled Comp Radiopharm 42(7):653–661.

48. Lockhart ST, Turrigiano GG, Birren SJ (1997) Nerve growth factor modulates synaptictransmission between sympathetic neurons and cardiac myocytes. J Neurosci 17(24):9573–9582.

13504 | www.pnas.org/cgi/doi/10.1073/pnas.1409026111 Lehmann et al.

Supporting InformationLehmann et al. 10.1073/pnas.1409026111SI Materials and MethodsExperimental Animals. Conditional ETA KO mice (ETAfl/fl)were described previously (1). ETAfl/fl were crossed with trans-genic mice overexpressing Cre recombinase under the control ofthe dbh promoter, DBH-Cre (2), to obtain mice lacking ETAonly in SNs (ETAfl/fl;DBH-Cre, abbreviated as SN-KO) or underthe control of themyh6 promoter, αMHC-Cre (3), to obtain micelacking ETA only in CMs (ETAfl/fl;αMHC-Cre, abbreviated asCM-KO). Control littermates (Ctrl. LM) consisted of ETAfl/fl,ETA+/fl, ETA+/+, and ETA+/+;DBH-Cre littermates for SN-KO and ETAfl/fl, ETA+/fl, ETA+/+, and ETA+/+;αMHC-Crelittermates for CM-KO. All these mouse lines and Dahl salt-resistant (DR), Dahl salt-sensitive (DS), Sprague–Dawley, andWistar rats (Charles River Laboratories) were housed understandard conditions and maintained on commercial mouse or ratchow and water ad libitum. The environment was maintained at22 °C with a 12-h light–12-h dark cycle. For preparation ofneonatal CMs or SNs, pups were killed between days 1 and 3postnatal. Male DR and DS (9 wk of age; mean weight, 200 g;M&B) were fed on a high-salt diet [8% (vol/vol) NaCl] for 40 d.Treated DS rats received darusentan (30 mg/kg/d; DS + DA inthe drinking water (was changed twice the day) with onset of thediet. Water and drug intake was controlled on a daily basis.

Transverse Aortic Constriction. Transverse aortic constriction (TAC)was performed in 9- to 10-wk-old male SN-KO and CM-KO andcontrol littermates from the same strain as described previously(4, 5). Briefly, the animals were anesthetized with ketamine(120 mg/kg, i.p.) and xylazine (15 mg/kg, i.p.). The mice wereorally intubated with 20-gauge tube and ventilated (HarvardApparatus) at 120 breaths/min (0.2-mL tidal volume). The aorticconstriction was created via a lateral thoracotomy through thesecond intercostal space. A suture (Prolene 7-0) was placedaround the transverse aorta between the brachiocephalic and leftcarotid artery. The suture was ligated against a 27-gauge needle.The needle was removed, leaving a discrete stenosis. Afterward,the chest was closed. Integrity of aortic banding was confirmedby inspection of the surgical constriction and by visualization ofmarked differences in the caliber of the right and left carotidarteries in TAC (6). A sham procedure in which the thoracicaorta was not banded was also performed. Mice were killed fromabout 7 wk after surgery.

Echocardiography. For measurements of chamber dimensions andcalculation of fractional shortening or ejection fraction, we useda Tracing software from VisualSonics on the parasternal long axisvia the following formula: 100 × [(LVEDV − LVESV)/LVEDV)],where LVEDV is left ventricular end-diastolic volume and LVESDis left ventricular end-systolic volume. To quantify the stenosisafter TAC surgery, echocardiographic data were achieved at heartrates between 450 and 550 beats/min, and the evaluation was per-formed with short-acting isoflurane anesthesia. Arterial bloodpressure of awake animals and left ventricular/intraventricularpressure were recorded as previously described in detail (7). Leftventricular end-diastolic pressure (LVEDP), left ventricular con-tractility [(+)dP/dtmax], and left ventricular relaxation [(–)dP/dtmax]were calculated using chart software package (ADInstrumentsPty. Ltd.). Transthoracic echocardiography was performed aspreviously described in detail (8). The investigator conductingthe echocardiography was blinded to the treatment status.

Metaiodobenzylguanidine-PET. All chemicals were purchased fromSigma–Aldrich unless otherwise noted. 124I was purchased fromEckert & Ziegler. Poly-(3-{dibutyl[2-(3-and-4-vinylphenyl)ethyl]stannyl}benzylguanidinium acetate)-codivinylbenzene, a tin pre-cursor anchored to a polystyrene matrix, was obtained fromMolecular Insight Pharmaceuticals. High-performance liquid chro-matography gamma was performed using an Agilent 1100 HPLCequipped with a Raytest (Germany) Model Gabi radioisotope de-tector. Analytical reversed-phase HPLC (RP-HPLC) analyses wereperformed using a Chromolith Performance RP-18 endcappedcolumn (3 × 100 mm). The column was eluted in a linear gradient(0–100%) of acetonitrile [0.1% trifluoroacetic acid (TFA)] in water(0.1% TFA) for 5 min; flow rate of 2 mL/min; and room temper-ature absorbance λ = 214 nm. [124I]Metaiodobenzylguanidine(MIBG) synthesis was conducted as follows: 3.2 mg of the resinwas swollen for 5 min in 50 μL ethanol, and 25 μL of water wasadded; 120 MBq [124I] in a volume of 35 μL was added, followedby 25 μL of a solution of hydrogen peroxide in acetic acid. ThepH value of the reaction mixture was pH 4. The reaction mixturewas stirred at a temperature of 85 °C for 20 min. At this time,HPLC analysis revealed a complete conversion, and 20 μL ofa saturated solution of methionine in water was added to quenchthe reaction. The resin was separated by filtration over a sterilefilter and washed with 0.9% NaCl. The mixture was neutralizedwith 35 μL of 1 M NaOH, and the pH was adjusted to 7.4 withphosphate buffer. Radio-HPLC revealed the purity (>95%) ofthe product. Mice were measured 7 wk after TAC surgery 1 hafter i.v. injection of 10.73 ± 0.69 MBeq of [124I]-MIBG. Theanesthetized animals [2% (vol/vol) sevoflurane; Abbott] wereplaced in an Inveon small animal PET scanner (Siemens). Re-gion of interest (ROI) analysis of the acquired images was per-formed using Siemens imaging software. The resulting imagedata were normalized to the administered activity to parame-terize the microPET images in terms of %ID/g (percentage ofinjected dose per gram of heart corrected for radioactive decayto the time of injection). The uptake profiles were fitted toa one-compartment model (realized with the PMOD software)to estimate the influx rate constant (K1, milliliters of medium pergrams of cells per min), and the volume of distribution (Vd).

Heart Tissue Preparation. Mice were killed by cervical dislocation,and hearts were removed and weighed promptly. Relative heartweight was calculated as ratio to body weight. The left ventricleswere transversely dissected, and parts of the left ventricles werequickly frozen in liquid nitrogen for protein and RNA isolationand analysis. Some hearts were arrested in diastole and com-pletely fixed for 48 h in 4% (vol/vol) paraformaldehyde dissolvedin 0.1 M PBS (pH 7.4), embedded in paraffin, longitudinally cutinto 5-μm sections, and stained for histomorphometry.

Primer for RNAAnalysis.The primers and probes that were used areas follows—rat: myh6 sense 5′-tgcagaagaaactgaaggaaaa-3′ andantisense 5′-gctcggcctctagctcct-3′; myh7 sense 5′-caccaacaacccctacgatt-3′ and antisense 5′-agcacatcaaaggcgctatc-3′; th sense5′-tctccctgaggggtacaaaa-3′ and antisense 5′-gaattttggcttcaaatgtctca-3′; mouse: ednra sense 5′-tgtgagcaagaaattcaaaaattg-3′and antisense 5′-atgaggcttttggactggtg-3′; nppa sense 5′-cacagatctgatggatttcaaga-3′ and antisense 5′-cctcatcttctaccggcatc-3′;col3a1 sense 5′-tcccctggaatctgtgaatc-3′ and antisense 5′-tgagtcgaattggggagaat-3′; myh7 sense 5′-cgcatcaaggagctcacc-3′ and an-tisense 5′-ctgcagccgcagtaggtt-3′; gapdh sense 5′-gggttcctataaa-tacggactgc-3′ and antisense 5′-ccattttgtctacgggacga-3′; th sense 5′-

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 1 of 9

cccaagggcttcagaagag-3′ and antisense 5′-gggcatcctcgatgagact-3′. ForSERCA (atp2a2) we used a taqman universal PCR mastermix(Applied Biosystems).

Isolated Heart Perfusion. Wistar rats were anesthetized with thio-pental (100 mg/kg, i.p.). The hearts were rapidly cut out andrinsed in ice-cold buffer, and the aorta was cannulated for per-fusion according to the Langendorff method. Within one ex-periment, 8–12 spontaneously beating hearts were perfusedsimultaneously at a constant coronary flow and a constant tem-perature of 37.5 °C. The perfusion medium was a modified Krebs-Henseleit solution (125 mmol/L NaCl, 16.9 mmol/L NaHCO3,0.2 mmol/L Na2HPO4, 4.0 mmol/L KCl, 1.85 mmol/L CaCl2, 1.0mmol/L MgCl2, 11 mmol/L glucose, and 0.027 mmol/L EDTA).The buffer was gassed with 95% (vol/vol) O2 and 5% (vol/vol)CO2, and the pH was adjusted to 7.4. Cardiac [3H]-norepi-nephrine (NE) uptake was determined as described previously(9). Briefly, a bolus of [3H]-NE (1 mL, 3 μCi, 100 pmol NE;Amersham-Buchler) was injected into the perfusion system andproportionally distributed to the hearts and blank channels.Radioactivity was measured in the effluent. The amount of [3H]-NE extracted by the hearts (uptake) was expressed as the per-centage of radioactivity measured in the blank channels. Over-flow of endogenous NE was evoked by electrical field stimulationas described previously (10). Briefly, isolated perfused heartswere stimulated with bipolar platinum electrodes in the presenceof atropine (1 μmol/L) to prevent vagal activation (S1; 1 min,4 Hz, 5 V). To distinguish between exocytotic and net release ofNE, a second stimulation was performed in the presence of theNE reuptake inhibitor desipramine (DMI). Endogenous NE in thecoronary venous effluate was determined by HPLC and electro-chemical detection.

Determination of Neurohormones. The tissue concentration of NEin mice was determined by a radioenzymatic assay as previouslydescribed (10). Plasma concentrations of the stable N-terminalpart (amino acids 1–98) of the prohormone form of atrial na-triuretic peptide (proANP) were determined in venous bloodsamples taken from the femoral vein. ProANP was determinedusing enzyme immunoassays according to the manufacturer’sinstructions (Biomedica). For determination of NE tissue levelsin rats, the left ventricle was rinsed in ice-cold 0.9% saline andfrozen in liquid nitrogen until HPLC and electrochemical de-tection were performed as previously described in detail (11).

Histology. H&E and Masson�s trichrome stainings were per-formed as previously described (12). Cardiomyocyte size wasassessed on H&E-stained sections by using National Institutes ofHealth Image J software (http://rsb.info.nih.gov/ij/). More than80 randomly chosen cardiomyocytes from each group wereanalyzed to measure cross-sectional cardiomyocyte area. Toquantify cardiac fibrosis, 20 trichrome-stained sections (magni-fication 20×) from the left ventricle were randomly selected, andmorphometric analysis using Image J was performed. Photographswhere acquired with an Olympus SZH zoom stereo dissection scopewith an Optronics DEI-750 CCD digital camera. All data wereanalyzed by a single observer blinded to the mouse genotypes.

Survival Analysis. DSs were treated as described above (DS, DS +DA). Ten rats per group were monitored, and deaths were recordedevery day. Survival was described by standardKaplan–Meier analysis.

Isolation of CMs. Rat CMs were isolated from newborn Sprague–Dawley rats as previously described (13). After isolation, CMswere maintained in DMEM containing 10% (vol/vol) FCS, 2%(vol/vol) antibiotics (penicillin/streptomycin), and 2 mM L-Gln.For the assessment of CM hypertrophy, cells were cultured for3–5 d and then starved (0% FCS) for 24 h and stimulated with 10

nM ET1 for 18–24 h. Murine CMs were isolated from neonatalETAfl/fl mice as described previously (14). Pups were decapi-tated, and ventricular tissue was carefully dissected and disso-ciated in 0.41 mg/mL collagenase (Sigma) and 0.30 mg/mLpancreatin (Sigma) in Ads buffer (116 mM NaCl, 20 mMHepes, 0.8 mM NaH2PO4, 1 g/liter glucose, 5.4 mM KCl, and0.8 mM MgSO4, pH 7.35). CMs were plated on laminin/poly-L-lysine–coated plates, and cells were treated with adenovirusexpressing Cre-recombinase (a kind gift by Oliver Müller, Uni-versity Hospital, Heidelberg, Germany) for 12 h with 12 multi-plicity of infection (MOI). For CM-SN coculture, neonatal ratSNs were plated 1 d after preparation of CMs. After 3 d, mediumwas changed to serum-free medium for 24 h. For hypertrophicexperiments, cells were then treated with 10 nM ET1 for 18–24 hbefore measurement of fixation.

Plasmids and Reagents. CMs were transduced with adenovirusesharboring FLAG-HDAC4 (Ad-HDAC4), and FLAG-HDAC5(Ad-HDAC5) were infected for 24 h as described previously (13),starved for 4 h, and stimulated with 10 nM ET1 for 12 h. Adenovirusharboring Cre recombinase was a kind gift by Oliver Müller and wasgenerated according to the manufacturer’s instructions (ViraPowerAdenoviral Expression System; Invitrogen). Prazosin (αARi; Sigma)and propranolol (βARi; Sigma) were used at a concentrationof 1 μM. Darusentan (kind gift of Klaus Muenter, Knoll AG,Mannheim, Germany) was used at a concentration of 1 μM.

Adenovirus Production. A luciferase reporter construct harboringthree binding sites for the myocyte enhancer factor 2 [Ad-3xmyocyteenhancer factor 2 (MEF2)-Luc] was obtained from Seven HillsBioreagents. After generation, the adenoviruses were amplifiedand purified with the Adeno-X Purification Kit (BD), and itsinfectious units per microliter were determined with the Adeno-XRapid Titer Kit (BD).

RNA Quantification. Total RNA was isolated from ventriculartissue or from cell culture experiments by using TRIzol (Invi-trogen). Total RNA was digested with DNase, and cDNA syn-thesis of 500 ng of RNA was carried out by using a SuperScriptfirst-strand synthesis system for RT-PCR (Invitrogen). Quanti-tative real-time PCR (qPCR) was performed with UniversalProbeLibrary (Roche) by using TaqMan Universal PCR Mas-termix (Applied Biosystems) and detection on a 7500 Fast Cycler(Applied Biosystems) as described previously (5).

Coimmunoprecipitation. Cells were harvested in Tris (50 mM, pH7.4), NaCl (150–900 mM), EDTA (1 mM), and Triton X-100 (1%)supplemented with protease inhibitors (Complete; Roche Diag-nostics) and PMSF (1 mM). Cells were further disrupted by passagethrough a 25-gauge needle, and cell debris were removed by cen-trifugation. FLAG-tagged proteins were immunoprecipitated withM2-agarose conjugate (Sigma-Aldrich) and thoroughly washed withlysis buffer. Bound proteins were resolved by SDS/PAGE, trans-ferred to PVDF membranes, and immunoblotted as indicated witheither a monoclonal anti-FLAG antibody (M2; Sigma-Aldrich) or apolyclonal anti–14-3-3 antibody (Abcam). Proteins were visualizedwith a chemiluminescence system (Santa Cruz Biotechnology).

Western Blot Analysis. Cultured cells were harvested in Tris(50 mM, pH 7.4), NaCl (150 mM), EDTA (1 mM), and TritonX-100 (1%) supplemented with protease inhibitors (Complete;Roche) and PMSF (1 mM). The cells were further disrupted bypassage through a 25-gauge needle, and cell debris were removed bycentrifugation. Proteins were resolved by SDS/PAGE, transferred toPVDF membranes, and immunoblotted as indicated in the figureswith tyrosine hydroxylase (TH) antibody (rabbit; Santa Cruz), ETAantibody (rabbit; Santa Cruz) or GAPDH (mouse; Santa Cruz),

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 2 of 9

followed by the appropriate secondary HRP-labeled antibodiesand visualized with a chemiluminescence system (Santa Cruz).

[3H]-Leucine Incorporation. Protein synthesis was examined by anal-ysis of [3H]-leucine incorporation.CMsandSN-CMcoculturesweremaintained for 4 d in DMEM containing 10% (vol/vol) FCS, 2%(vol/vol) antibiotics (penicillin/streptomycin), and 2 mM L-Gln,starved (0%FCS) for 24 h, and stimulatedwith 10 nMET1.Cultureswere treated with inhibitors 1 h before stimulation. After stimula-tion, 3 μL Ci of [3H]-leucine was added, and the cells were culturedfor an additional 18 h. After two washes with ice-cold PBS, neonatalrat ventricular myocytes (NRVMs) were precipitated with 10%(vol/vol) trichloroacetic acid for 30minat 4 °C.Theprecipitateswerethen solubilized in 0.2 N NaOH for more than 4 h. Radioactivitywas measured in a liquid scintillation counter.

MEF2 Reporter Assay.CMs were transduced with Ad-3xMEF2-Luc(10 MOI) construct, expressing three copies of an MEF2 binding

site to detect endogenous MEF2 activity. CMs were infected 4 hbefore washing and adding SN to avoid infection of neuronalcells. After 24-h culturing, cells were starved for 24 h. Aftertreatment with 10 nM ET1 for 24 h, cells were harvested, andluciferase levels were determined.

Indirect Immunofluorescence. For immunofluorescence analysis,cocultures and monocultures were grown on laminin-coated cov-erslips (BD Bioscience), fixed in paraformaldehyde [4% (vol/vol)],permeabilized in0.3%TritonX-100, andblocked inPBS containinggoat serum [5% (vol/vol)]. Primary antibodies against FLAG(monoclonal rabbit; Sigma), α-actinin (monoclonal mouse;Sigma), and TH (polyclonal rabbit; Cell Signaling) were used ata dilution of 1:200. Secondary antibodies conjugated to eitherfluorescein or Texas red were used at a dilution of 1:200. Imageswere captured at a magnification of 20–60× with an OlympusBX-51 fluorescence microscope.

1. Kedzierski RM, et al. (2003) Cardiomyocyte-specific endothelin A receptor knockoutmice have normal cardiac function and an unaltered hypertrophic response toangiotensin II and isoproterenol. Mol Cell Biol 23(22):8226–8232.

2. Parlato R, Otto C, Begus Y, Stotz S, Schütz G (2007) Specific ablation of thetranscription factor CREB in sympathetic neurons surprisingly protects againstdevelopmentally regulated apoptosis. Development 134(9):1663–1670.

3. Agah R, et al. (1997) Gene recombination in postmitotic cells. Targeted expression ofCre recombinase provokes cardiac-restricted, site-specific rearrangement in adultventricular muscle in vivo. J Clin Invest 100(1):169–179.

4. Hill JA, et al. (2000) Cardiac hypertrophy is not a required compensatory response toshort-term pressure overload. Circulation 101(24):2863–2869.

5. Backs J, et al. (2009) The delta isoform of CaM kinase II is required for pathologicalcardiac hypertrophy and remodeling after pressure overload. Proc Natl Acad Sci USA106(7):2342–2347.

6. Kong Y, et al. (2006) Suppression of class I and II histone deacetylases blunts pressure-overload cardiac hypertrophy. Circulation 113(22):2579–2588.

7. Buss SJ, et al. (2006) Spironolactone preserves cardiac norepinephrine reuptake insalt-sensitive Dahl rats. Endocrinology 147(5):2526–2534.

8. Hardt SE, et al. (2002) Accelerated cardiomyopathy in mice with overexpression ofcardiac G(s)alpha and a missense mutation in the alpha-myosin heavy chain.Circulation 105(5):614–620.

9. Backs J, et al. (2001) The neuronal norepinephrine transporter in experimental heartfailure: Evidence for a posttranscriptional downregulation. J Mol Cell Cardiol 33(3):461–472.

10. Backs J, Bresch E, Lutz M, Kristen AV, Haass M (2005) Endothelin-1 inhibits the neuronalnorepinephrine transporter in hearts of male rats. Cardiovasc Res 67(2):283–290.

11. Touyz RMSE, Schiffrin EL (2003) Role of endothelin in human hypertension. Can JPhysiol Pharmacol 81(6):533–541.

12. Song K, et al. (2006) The transcriptional coactivator CAMTA2 stimulates cardiacgrowth by opposing class II histone deacetylases. Cell 125(3):453–466.

13. Backs J, Song K, Bezprozvannaya S, Chang S, Olson EN (2006) CaM kinase II selectivelysignals to histone deacetylase 4 during cardiomyocyte hypertrophy. J Clin Invest116(7):1853–1864.

14. Balza RO, Jr, Misra RP (2006) Role of the serum response factor in regulatingcontractile apparatus gene expression and sarcomeric integrity in cardiomyocytes.J Biol Chem 281(10):6498–6510.

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 3 of 9

Fig. S1. Generation of pre- and postsynaptic ETA KO. (A) Breeding scheme to obtain presynaptic (SN, sympathetic neuron specific) and postsynaptic (CM,cardiomyocyte specific) KO of the ETA (αMHC, myosin heavy chain; DBH, dopamin β hydroxylase). (B) Transcript levels in the stellate ganglion and the leftventricle of the indicated genes were detected by quantitative real-time PCR (n = 5–10 per group). Values indicate relative expression level normalized tocontrol littermates (without Cre expression or floxed allele) of each strain.

Fig. S2. Quantification of aortic stenosis. We quantified the aortic stenosis of TAC and sham-operated animals. Values indicate the flow rate within thestenosis. No significant difference was found between the groups. Values are presented as mean ± SEM; *P < 0.05 sham vs. TAC (n ≥ 5 per group).

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 4 of 9

Fig. S3. Cardiomyocyte ETA is not involved in pathological cardiac remodeling. (A) Representative four-chamber view of CM-KO 7 wk after sham or TACsurgery. (Scale bar, 1 mm.) (B) Representative trichrome-stained sections, shown as indicated. (Scale bar, 100 μm.) (C) Representative transthoracic echocar-diography of the left ventricle (M-Mode, midventricular). Groups are shown as indicated. (D) Heart weight/body weight ratios of control littermates and KO(n = 7–11 per group). (E) Quantification of CM size (>80 cells per sample from more than five fields of view; n = 3). (F) Quantification of fibrosis normalized tocontrol littermates (n ≥ 3 per group). (G) Fractional shortening of CM-KO and control littermates 3 wk after TAC (n = 6–11 per group). (H) Lung weight/bodyweight ratios of control littermates and CM-KO animals 7 wk after TAC surgery (n = 7–11 per group). (I) Transcript levels in hearts from CM-KO 7 wk after TACsurgery from genes as indicated, normalized to control littermates (n = 4–8 per group). Values presented as mean ± SEM. *P < 0.05; **P < 0.01.

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 5 of 9

Fig. S4. Reduced cardiac fibrosis in SN-KO after TAC. Representative trichrome-stained sections from SN-KO (A) and CM-KO (B), shown as indicated. (Scalebar, 50 μm.)

Fig. S5. Sympathetic neurons regulate cardiomyocyte hypertrophy through sympathetic ETA. (A) Treatment with 10 nM ET1 for 1 h leads to an increase in NErelease into the cell culture medium of sympathetic neuron-cardiomyocyte (SN-CM) coculture, which was attenuated by the pretreatment with darusentan(1 μM; ETAi). (B) Western blot analysis of the SN-specific marker tyrosine hydroxylase (TH) confirmed the presence of SNs in CM-SN coculture. (C) Representativeimage in CM surface area after 24-h ET1 treatment (10 nM). (D) Neonatal rat ventricular myocytes (NRVMs, CMWT) were cocultured with SNs from ETA

fl/fl miceand treated with Ad-Cre (SNKO). (Top) Representative α-actinin stained CMs. (Middle) Anti-ETA immunocytochemistry, confirming reduced ETA expression(magnification in white rectangle). Please also note the higher expression of ETA in SNs in relation to CMs. Gray-scaled panels show transmitted light images tovisualize the presence of sympathetic neurons. (White bar, 50 μm.) Cultures were treated as indicated (ET1 10 nM, 18 h). (E) Cell surface area was quantified(n = 50–70 cells). Values are shown as mean ± SEM. *P < 0.05 as indicated. (F) CMs from ETAfl/fl mice were cocultured with WT rat SNs (SNWT). For ex vivo KO ofETA, cells were transduced with Ad-Cre (CMKO) and then treated with a high dosage (100 nM) of ET1 for 24 h as indicated; 100 nM ET1 resulted in substantialCM hypertrophy even in CM monoculture, which was attenuated in CMKO but gained again by the addition of SNWT. Immunocytochemistry was performedusing antibodies recognizing sarcomeric α-actinin (red; Upper) and TH (green; Lower). (G) CM cell surface area was then quantified (n = 24–40 cells per group).*P < 0.05 as indicated. Values are presented as mean ± SEM.

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 6 of 9

Fig. S6. In vitro deletion of ETA in NMVM from ETA floxed mice. (A and B) Protein expression was detected by Western blot analysis 5 d after Ad-Cretransduction (n = 3 per group). IB, immunoblot.

Fig. S7. Selective β1-adrenergic inhibition blocks exaggerated cardiomyocyte hypertrophy. CM cell surface area was quantified in CM and SN-CM culturesafter treatment with ET1 and pretreatment with selective (bisoprolol) and nonselective inhibitors of β1-adrenergic receptors. #P < 0.05 vs. ET1-treated; *P <0.05 ET1-treated SN-CM vs. CM (n = 100–150 per group) (ET1: 10 nM 24 h; 1 μM propranolol; 1 μM bisoprolol; 10 μM atenolol added 1 h before ET1 treatment).

Fig. S8. β-adrenergic activation leads to increased 14-3-3 binding to HDAC4 but reduced 14-3-3 binding to HDAC5. (A and B) Coimmunoprecipitation assayusing cellular extracts from NRVMs expressing FLAG-HDAC4 and FLAG-HDAC5. The FLAG-tagged proteins were immunprecipitated with an antibody againstFLAG. Associated endogenous 14-3-3 was detected with an antibody against 14-3-3. The experiment was performed with and without stimulation, withisoproteronol as indicated. IB, immunoblot; IP, immunoprecipitation; ISO, isoproteronol.

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 7 of 9

Table

S1.

Summaryofclinical

endothelin

receptoran

tagonisttrials

Timeof

trea

tmen

tDate

Acronym

Referen

ceDrug

Selective

for

Design

Patien

tsnumber

Trea

tedwith

adrenergic

receptor

antagonists

Ben

eficial

effects?

Result

24wk

2004

EARTH

(1)

DET

ARan

domized

,double-blin

d,

placebo-controlle

d

642

73%

No

Noben

eficialeffectsonLV

ESV

(MRI),ch

anges

inLV

EF,

neu

rohorm

onal

mea

suremen

t,6-min

walktest,quality

oflife,

NYHA

class,global

assessmen

t,an

dco

mposite

clinical

status.

3wk

2002

HEA

T(2)

DET

ARan

domized

,double-blin

d,

placeboco

ntrolle

d

157

46%

Yes

Positive

forcardiacindex

after3wk,

noch

angein

pulm

onarywed

ge,

pulm

onaryarterial

pressure,

pulm

onaryva

scularresistan

ce,rightarterial

pressure,

hea

rtrate,mea

nartery

pressure

andplasm

acatech

olamines.

2wk

1998

(3)

BET

Aan

dET

BRan

domized

,double-blin

d,

placeboco

ntrolle

d

360%

Yes

After

2wk:

improve

dcardiacoutput,reducedsystem

ican

dpulm

onaryva

scularresistan

ce;noch

angein

hea

rtrate.

3wk

2001

(4)

DET

ARan

domized

,placeboco

ntrolle

d21

33%

Yes

Improve

men

tofflow-m

ediatedva

sodila

tations(FMD).

26wk

2005

Rea

ch-1

(5)

BET

Aan

dET

BRan

domized

,double-blin

d,

placebo-controlle

d

370

14–26

%No/short

term

;ye

sfor17

6patients

inthelong-term

follo

w-up

Term

inationbecau

seofsafety

concerns.17

6patients

werefollo

wed

uplongterm

(26wk):26

%im

prove

din

NYHA

class(compared

with19

%placebo);28

%showed

adeclin

ein

NYHA

class(compared

with43

%placebo)P=

0.04

5.Noch

angein

systolic

pressure,

hea

rtrate

orQTtime.

Noch

angewhen

allpatients

areincluded

into

thean

alysis,suggestingabiphasic

response

withea

rlyworsen

ingan

dlaterim

prove

men

tin

bosentan-treated

patients.

3wk

2004

(6)

DET

ARan

domized

,double-blin

d,

placebo-controlle

d

2334

%Yes

Significan

tben

efitsin

hem

odyn

amic

variab

les,mea

npulm

onaryartery

pressure,an

dhea

rtrate.A

significan

treductionin

mea

narterial

bloodpressure,

BNPdecreased

significan

tlyin

patients

with

darusentan,wherea

sbig-endothelin

remained

unchan

ged

.Pro-A

NPtended

todecreasein

theactive

trea

tmen

tgroup,butdid

notreachstatisticalsignifican

ce.

6h

2004

(7)

TET

Aan

dET

B

Ran

domized

,double-blin

d,

placebo-controlle

d

130

17–31

%Yes

Increa

sein

cardiacindex

,decreasein

BNPleve

ls,positive

tren

din

patients

subjectivedyspnea

score,reduce

urineoutputin

highdosage.

6h

2003

(8)

TET

Aan

dET

B

Ran

domized

,double-blin

d,

placebo-controlle

d

292

56%

Yes

Improve

men

tin

hem

odyn

amics,im

prove

dcardiacindex

,in

tenden

cyim

prove

ddyspnea

,decreased

risk

of

clinical

worsen

ing

upto

72h

2003

RITZ-4

(9)

TET

Aan

dET

B

Ran

domized

,double-blin

d,

placebo-controlle

d

193

71–79

%No

Noch

angein

dea

th,worsen

ingHF,

recu

rren

tisch

emia

or

new

myo

cardialinfarction

upto

24h

2003

Ritz-5

(10)

TET

Aan

dET

B

Ran

domized

,double-blin

d,

placebo-controlle

d

8445

–60

%No

Nodifference

inSO

(2),dea

th,recu

rren

tpulm

onary

edem

a,mechan

ical

ventilation,myo

cardialinfarction

duringthefirst24

h.

Instudiesin

whichpatients

did

notreceiveoronly

asm

allpercentagereceived

antiad

renergic

therap

ies,read

outs

tended

tobepositive

.B,bosentan;D,darusentan;T,

tezo

sentan.

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 8 of 9

1.Anan

dI,et

al.(20

04)Lo

ng-term

effectsofdarusentanonleft-ven

tricularremodellin

gan

dclinical

outcomes

intheEn

dothelin(A

)Recep

torAntagonistTrialin

Hea

rtFa

ilure

(EARTH

):Ran

domised

,double-blin

d,p

lacebo-controlle

dtrial.La

ncet

364(94

31):34

7–35

4.2.

Luscher

TF,et

al.(200

2)Hem

odyn

amic

andneu

rohumoraleffectsofselectiveen

dothelin

A(ET(A))receptorblockad

ein

chronic

hea

rtfailu

re:Th

eHea

rtFa

ilure

ET(A

)Recep

torBlockad

eTrial(H

EAT).Circu

lation10

6(21

):26

66–26

72.

3.Su

tsch

G,et

al.(199

8)Sh

ort-term

oralen

dothelin-recep

toran

tagonisttherap

yin

conve

ntionally

trea

tedpatients

withsymptomatic

seve

rech

ronic

hea

rtfailu

re.Circu

lation98

(21):226

2–22

68.

4.Berger

R,et

al.(200

1)Effectsofen

dothelin

areceptorblockad

eonen

dothelialfunctionin

patients

withch

ronic

hea

rtfailu

re.Circu

lation10

3(7):981

–98

6.5.

Packer

M,et

al.(200

5)Clin

ical

effectsofen

dothelin

receptoran

tagonism

withbosentanin

patients

withseve

rech

ronic

hea

rtfailu

re:Resultsofapilo

tstudy.

JCardFa

il11

(1):12

–20

.6.

Bergler-Klein

J,Pa

cher

R,Berger

R,Bojic

A,Stan

ekB(200

4)Neu

rohumoralan

dhem

odyn

amic

effectsoftheselectiveen

dothelin

antagonistdarusentanin

adva

ncedch

ronic

hea

rtfailu

re.JHea

rtLu

ngTran

splant23

(1):20

–27

.7.

Cotter

G,et

al.(200

4)Th

ehem

odyn

amic

andneu

rohorm

onal

effectsoflow

dosesoftezo

sentan(anen

dothelin

A/B

receptoran

tagonist)

inpatients

withacute

hea

rtfailu

re.Eu

rJHea

rtFa

il6(5):601

–60

9.8.

Torre-AmioneG,et

al.(200

3)Hem

odyn

amic

andclinical

effectsoftezo

sentan,an

intrav

enousdual

endothelin

receptoran

tagonist,in

patients

hospitalized

foracute

decompen

satedhea

rtfailu

re.JAm

CollCardiol42

(1):14

0–14

7.9.

O’ConnorCM,e

tal.;Ran

domized

Intrav

enousTe

ZosentanStudy-4Inve

stigators

(200

3)Te

zosentanin

patients

withacute

hea

rtfailu

rean

dacute

coronarysyndromes:R

esultsoftheRan

domized

Intrav

enousTe

ZosentanStudy(RITZ-4).J

Am

Coll

Cardiol41

(9):14

52–14

57.

10.KaluskiE,

etal.(20

03)RITZ-5:

Ran

domized

intrav

enousTe

Zosentan(anen

dothelin-A

/Ban

tagonist)

forthetrea

tmen

tofpulm

onaryed

ema:

Aprospective

,multicen

ter,double-blin

d,placebo-controlle

dstudy.

JAm

CollCardiol41

(2):20

4–21

0.

Lehmann et al. www.pnas.org/cgi/content/short/1409026111 9 of 9