OUP user menu

Appearance of a ventricular 5-HT4 receptor-mediated inotropic response to serotonin in heart failure

Eirik Qvigstad , Trond Brattelid , Ivar Sjaastad , Kjetil Wessel Andressen , Kurt A. Krobert , Jon Arne Birkeland , Ole M. Sejersted , Alberto J. Kaumann , Tor Skomedal , Jan-Bjørn Osnes , Finn Olav Levy
DOI: http://dx.doi.org/10.1016/j.cardiores.2004.11.017 869-878 First published online: 1 March 2005


Background: Current pharmacological treatment of congestive heart failure (CHF) addresses changes in neurohumoral stimulation or cardiac responsiveness to such stimulation. Yet, undiscovered neurohumoral changes, adaptive or maladaptive, may occur in CHF and suggest novel pharmacological treatment. Serotonin [5-hydroxytryptamine (5-HT)] enhances contractility and causes arrhythmias through 5-HT4 receptors in human atrium and ventricle but not through rat ventricular 5-HT4 receptors.

Objective: We investigated whether CHF could induce ventricular responsiveness to serotonin.

Methods: Postinfarction CHF was induced in male Wistar rats by coronary artery ligation. Contractility was measured in left ventricular papillary muscles 6 weeks after infarction. Messenger RNA was quantified by RT-PCR and cAMP by RIA.

Results: Serotonin caused positive inotropic (−logEC50=7.5) and lusitropic effects in CHF but not Sham papillary muscles. The inotropic effect of 10 μM serotonin in CHF (31.3 ± 2.2%) was of similar size as the effect of 10 μM isoproterenol (34.0 ± 1.7%). The effects of serotonin were antagonised by GR113808 (0.5–5 nM), consistent with mediation through 5-HT4 receptors. This was further supported by positive inotropic effects of the 5-HT4-selective partial agonist RS67506. Carbachol blunted the serotonin responses and serotonin increased ventricular and cardiomyocyte cAMP, consistent with coupling to Gs and adenylyl cyclase. Quantitative RT-PCR revealed fourfold increased 5-HT4(b) mRNA expression in CHF vs. Sham ventricles.

Conclusion: Functional ventricular 5-HT4 receptors are induced by myocardial infarction and CHF of the rat heart. We propose that they are a model for ventricular 5-HT4 receptors of human failing heart and may play a pathophysiological role in heart failure.

  • Serotonin (5-HT)
  • Heart failure
  • 5-HT4 receptor
  • Ventricle
  • Contractile function

1. Introduction

Our understanding of congestive heart failure (CHF) pathophysiology has changed dramatically from a hemodynamic model with treatment aimed at correcting hemodynamic defects, to a neurohumoral model with treatment aimed at preventing maladaptive, biological changes [1,2]. In addition to the changes in, e.g., the adrenergic system, there may be changes, adaptive or maladaptive, in other neurohumoral systems during CHF. Exploration of such possible changes may increase our understanding of CHF.

The neurohumoral system related to the neurotransmitter and vasoactive indoleamine serotonin is one candidate, and increased plasma levels and activity of serotonin have been reported in human CHF [3,4]. Circulating serotonin, mainly derived from enterochromaffin cells of the gastrointestinal tract, is taken up, stored, transported by and released from platelets [5]. Serotonin can also be captured and released by sympathetic nerve endings [6] and activate 5-hydroxytryptamine (5-HT) receptors in the human atrium [7,8]. Serotonin exerts diverse effects through at least 14 different receptors, named 5-HT1 to 5-HT7 with subgroups, of which 5-HT4, 5-HT6 and 5-HT7 receptors increase cAMP-formation through the G-protein Gs, similar to β-adrenoceptors [9].

Serotonin causes cardiostimulation either directly or indirectly through modulation of norephinephrine release from sympathetic nerve terminals [6]. Direct inotropic effects of serotonin mediated through 5-HT4 receptors have been observed in atrium [6–8] and ventricle [10] of man and pig, and through atrial 5-HT2A receptors in rat [11], and 5-HT4 receptors have been implicated in arrhythmia [8,12,13]. In addition, 5-HT4 receptor mRNA was detected in rat atrium, but not ventricle [14]. Early reports indicated lack of inotropic effects of serotonin in human [15,16] and porcine [17,18] ventricle. However, 5-HT4(a) and 5-HT4(b) receptor mRNA have been detected in human left and right ventricle [19,20]. Moreover, we recently demonstrated that phosphodiesterase inhibition uncovers functional ventricular 5-HT4 receptors in porcine and failing human ventricle and 5-HT4(b) mRNA was increased fourfold in failing human hearts compared to donor hearts [10]. The aim of the present study was to explore the possibility that serotonergic signalling through 5-HT4 receptors may become apparent in the remaining ventricular myocardium of the well-characterised model of postinfarction CHF in rats, following coronary artery ligation [21–23]. Our results revealed the appearance of 5-HT4 receptor-mediated inotropic and lusitropic effects of serotonin in CHF rat papillary muscle, accompanied by a fourfold increase in 5-HT4 receptor mRNA.

2. Materials and methods

2.1. CHF model

The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH Publication No. 85-23, revised 1996). Two animals per cage in a temperature-regulated room on a 12:12-h day/night cycle were given access to food and water ad libitum. As described [24], an extensive myocardial infarction (MI) was induced in 320 g male Wistar rats under anaesthesia (68% N2O/29% O2/2–3% Isofluran) by proximal ligation of the left coronary artery. Six weeks later, CHF rats were included in the study if left ventricular end-diastolic pressure (LVEDP, measured by catheterisation as described [24]) was ≥ 15 mm Hg and the rats had clinical signs of CHF. Rats with LVEDP<15 mm Hg without symptoms of CHF were included in the myocardial infarction nonfailing (MInf) group [23]. Sham-operated animals (Sham) underwent identical surgical procedure without coronary artery ligation. The endocardial surface of CHF and MInf hearts was digitally photographed and infarct size (% of inner surface) was traced on screen (KS100, Kontron electronics, Germany).

2.2. Isolated papillary muscles

Posterior left ventricular papillary muscles (diameter<1.0 mm) were prepared, mounted in 31 °C organ baths at 1.8 mM Ca2+, equilibrated and field-stimulated at 1 Hz [24] and the contraction–relaxation cycles (CRCs) were recorded and analysed as described [24,25] with respect to maximal developed force (Fmax, mN), maximal development of force [(dF/dt)max], time to peak force (TPF), time to 80% relaxation (TR80) and relaxation time (t80=TR80−TPF). Inotropic responses were expressed as increase of Fmax (% and mN) and of (dF/dt)max (%). Lusitropic effects were expressed as decrease of t80. Blockers of α1-adrenoceptors (prazosin 1 μM), β-adrenoceptors (timolol 1 μM), muscarinic cholinergic receptors (atropine 1 μM) and, where indicated, 5-HT2A receptors (ketanserin 0.1 μM) were added 90 min before agonist. Serotonin was added to the organ bath cumulatively (concentration–response curves) or as a bolus (10 μM). The apparent inhibition constant Kb for the selective 5-HT4 blocker GR113808 was calculated from ratios of serotonin EC50 values with and without GR113808. GR113808 did not influence basal CRC characteristics or electrical stimulation threshold (not shown). Concentration–response curves were constructed by estimating centiles (EC10 to EC100) and calculating the corresponding means, and the horizontal positioning expressed as −logEC50 values [24].

2.3. cAMP content in perfused hearts

Hearts were excised and retrogradely perfused [24] at 31 °C with the additional presence of the 5-HT7 receptor antagonist SB269970 (0.1 μM) to eliminate stimulation of cAMP production by vascular 5-HT7 receptors [26]. Spontaneously beating hearts were equilibrated for 30 min before 1-min perfusion with 5-HT (10 μM). Infarcted area was quickly removed and remaining heart was freeze-clamped. Frozen heart powder (300 mg) was homogenised at 4 °C in 5 ml 5% trichloroacetic acid and cAMP content was determined by radioimmunoassay [27].

2.4. cAMP accumulation in cardiomyocytes

Hearts were perfused (Langendorff setup) with Buffer-1 (mM): NaCl–130, Hepes–25, d-glucose–22, KCl–5.4, MgCl2–0.5, NaH2PO4–0.4, insulin–0.01 μg/ml (pH 7.4), then with Buffer-1 containing 200 U/ml collagenase Type-II and 0.1 mM Ca2+ until the aortic valves were digested. Left ventricle free wall and septum were minced and gently shaken at 37 °C for 3–4 min in the same solution plus 1% BSA. Following filtration (nylon mesh–200 μm) and sedimentation (37 °C), the cell pellet was washed three times in Buffer-1/1% BSA/0.1 mM Ca2+, resedimented in Buffer-1/1% BSA with first 0.2 mM, then 0.5 mM Ca2+and incubated 2 h (37 °C) in laminin-coated six-well plates. Plated cells were washed in Buffer-1/1 mM Ca2+ and incubated 2 h. Receptor blockers (as in contractility experiments) were added 10 min prior to 500 μM 3-isobutyl-1-methylxanthine, 10 min prior to 15-min incubation with 5-HT (10 μM). Incubations were stopped by trichloroacetic acid (5% final), and intracellular plus extracellular cAMP (expressed as % increase above basal) was determined by radioimmunoassay [27].

2.5. Qualitative and quantitative RT-PCR

Fresh noninfarcted left ventricle tissue or papillary muscle, collected after completion of functional analyses, was stored in RNAlater (Ambion) until use, then prepared and qualitative and quantitative RT-PCR performed as described [20]. Sets of primers (targeted to intron/exon boundaries to avoid genomic DNA signals) and probes (Double-Dye Oligonucleotide, 5′-labeled with the fluorescent reporter dyes FAM (5-HT4(b) and 5-HT2A), JOE [glyceraldehyde-3-phosphate dehydrogenase (GAPDH)] or Yakima Yellow [YY; atrial natriuretic peptide (ANP)] and quenched with TAMRA (5-HT4(b), 5-HT2A and GAPDH) or Dark Quencher (DQ; ANP) for quantitative PCR were designed as described [28]. The names and sequences of upper (U) and lower (L) primers and probes (P) used were (5′-3′): 5-HT4(b): ON283(U), CATGTGCTAAGGTATACAGTGGAATGT; ON284(L), GCAGCCACCAAAGGAGAAGTT; TM14(P), FAM-CTGTGAGGTGACACCGACTCTCCCATT-TAMRA; 5-HT2A: ON273(U), TTCACCACAGCCGCTTCAA; ON274(L), ATCCTGTAGTCCAAAGACTGGGATT; TM9(P), FAM-ATGGATATACCTACAGATATGGTCGTCCACACGGCAAT-TAMRA; ANP: ON285(U), ATCTGATGGATTTCAAGAACC; ON286(L), CTCTGAGACGGGTTGACTTC; TM16(P) YY-CGCTTCATCGGTCTGCTCGCTCA-DQ; GAPDH: ON279(U), CCTGCACCACCAACTGCTTA; ON290(L), GGCATGGACTGTGGTCATGA; TM12(P), JOE-TGGCCAAGGTCATCCATGACAACTTTG-TAMRA.

2.6. Statistics

All results are expressed as mean ± S.E.M. unless otherwise indicated and statistical significance assessed with unpaired and paired Student's t-tests or nonparametric Mann–Whitney test.

3. Results

3.1. Characteristics of CHF animals

All rats in the CHF group had large anterolateral infarcts and signs of CHF including tachypnea, pleural effusion and pulmonary congestion, while the rats in the MInf group had infarcts of variable size and no signs of CHF. For animal characteristics and hemodynamic data, see Table 1.

View this table:
Table 1

Animal and papillary muscle characteristics

3.2. Inotropic effects of serotonin

Serotonin (10 μM) elicited a monophasic positive inotropic response in CHF papillary muscles in contrast to the complete lack of mechanical response in Sham (n=6; Figs. 1 and 2A). With a diffusion delay of about 5 s, the time from serotonin addition to 50% and maximal inotropic response was 22 ± 2 s and 1–2 min, respectively (Fig. 1–inset). Serotonin (10 μM) increased Fmax in CHF by 31.3 ± 2.2% (n=10, p<0.0001; basal: 5.7 ± 0.2 mN; 5-HT: 7.3 ± 0.3 mN) and (dF/dt)max by 44.5 ± 2.9%, similar to the effect of 10 μM isoproterenol in CHF [Fmax increase: 34.0 ± 1.7%, n=10, p<0.0001; basal: 5.4 ± 0.4 mN; Iso: 7.3 ± 0.5 mN; Fig. 2A; (dF/dt)max increase: 58.5 ± 7.0%]. Isoproterenol increased Fmax in Sham by 113.4 ± 4.8% [n=6, p<0.0001; basal: 5.7 ± 0.6 mN; Iso: 12.2 ± 1.3 mN; Fig. 2A; (dF/dt)max increase: 111.6 ± 8.6%]. After stabilisation of the serotonin response in CHF, isoproterenol did not cause a significant further increase in Fmax (Iso: 7.7 ± 0.4 mN; n=10, nonsignificant vs. 5-HT) or (dF/dt)max (60.4 ± 5.8%). The concentration–response curve for serotonin exhibited a −logEC50M value of 7.49 ± 0.08 (n=5) with 12.1 ± 1.2% (p=0.0018) increase in Fmax [basal: 5.6 ± 0.5 mN; 5-HTmax: 6.3 ± 0.5 mN obtained at about 1 μM; (dF/dt)max increase:18.1 ± 3.5%]. The lower maximal inotropic effect in these experiments compared to single-dose (10 μM serotonin) experiments, e.g., Fig. 2A, may reflect desensitisation as well as more complete activation of relaxing components during cumulative agonist addition in the concentration–response experiments.

Fig. 2

Inotropic response to serotonin and to isoproterenol (Iso) in CHF (A–C) and Sham (A) papillary muscles. (B,C) Representative average contraction relaxation cycles in a CHF papillary muscle before addition of agonist (–) and at maximal steady state inotropic response to 10 μM serotonin (---) or 100 μM isoproterenol (----), expressed as contractile force (mN) (B) and normalised to maximal force (100%) (C).

Fig. 1

Inotropic effect of serotonin only in CHF. Original recordings in representative Sham (upper) and CHF (lower) papillary muscles. Serotonin (10 μM) induced a positive inotropic response only in the CHF rat. Inset: Development of the inotropic response, expressed as increase in Fmax in percent of individual maximal response (n=10).

3.3. Lusitropic response to serotonin

Serotonin reduced TR80 more than TPF, resulting in reduced t80 in papillary muscles from CHF but not Sham rats (Table 2, Fig. 2B and C). Similar to β-adrenoceptor stimulation [25], these contraction–relaxation cycle changes demonstrate lusitropic effects in addition to inotropic effects of serotonin, consistent with common signalling pathways.

View this table:
Table 2

Characteristics of contraction–relaxation cycles (CRCs) in Sham and CHF papillary muscles before and after subsequent addition of serotonin (10 μM) and isoproterenol (100 μM, Iso)

3.4. The inotropic response to serotonin is inhibited by cholinergic stimulation and associated with increased cAMP levels

In the absence of atropine, stimulation of muscarinic acetylcholine receptors by carbachol (30 μM) reversed the positive inotropic response to 10 μM serotonin (n=5, p=0.002; Fig. 3A) and partially reversed TPF, TR80 and t80 after 1–2 min (not shown). In contrast, carbachol alone did not cause a reduction in basal; instead, on occasion, a small increase in basal force was observed (not shown). Atropine (1 μM) restored the serotonin response in the presence of carbachol and reestablished the contraction–relaxation cycle characteristics. This functional antagonism by cholinergic stimulation, as expected from stimulation of Gi-coupled M2 muscarinic acetylcholine receptors, is consistent with involvement of the Gs-cAMP pathway. Accordingly, cAMP levels in perfused hearts increased by 14.8 ± 4.6% (n=5, p=0.02; Fig. 3B) after 1 min serotonin (10 μM) stimulation (basal: 0.415 ± 0.017; serotonin: 0.476 ± 0.019 pmol/mg wet weight), compared to about 100% increase by 0.1 μM isoproterenol [29]. To eliminate a possible influence of noncardiomyocytes, cardiomyocytes isolated from failing hearts were stimulated for 15 min with 10 μM serotonin, resulting in increased cAMP levels by 43.1 ± 23.6% (n=8, p=0.05; Fig. 3B), which was 5% of that obtained with 10 μM isoproterenol (not shown).

Fig. 3

The inotropic response to serotonin is reversible by cholinergic stimulation and mediated via 5-HT4 receptors. (A) Changes in Fmax in CHF papillary muscles by sequential addition of 10 μM serotonin (5-HT), 30 μM carbachol (5-HT+CCh) and 1 μM atropine (5-HT+CCh+Atropine). Carbachol reversed and atropine restored the inotropic response to serotonin. (B) Increased cAMP levels in CHF hearts perfused for 1 min with 10 μM serotonin and in cardiomyocytes from CHF hearts following 15-min stimulation with 10 μM serotonin in the presence of IBMX. (C) Representative contraction–relaxation cycles in a CHF papillary muscle before addition of agonist (–), at maximal steady-state inotropic response to 10 μM serotonin (---) and following reversal by 1 μM GR113808 (---). (D,E) Concentration–response curves for serotonin in CHF papillary muscles without and with 0.5 or 5 nM GR113808 (D) or 0.1 μM ketanserin (E). Ketanserin (0.1 μM) was present in (D). Inotropic response is expressed in percent of maximum in each experiment. (F) Inotropic effect of the 5-HT4-selective partial agonist RS67506 (RS67) in CHF papillary muscle, as well as attenuation of the agonist effect of subsequently added serotonin. *p<0.05 vs. Basal; **p<0.05 vs. RS67506 (10 μM); ***p<0.05 vs. 10 μM 5-HT. (G) Antagonistic effect of RS67506 of the inotropic effect of the full agonist serotonin. The paradoxical apparent reduction of the inotropic response, measured as maximum developed force, below basal after RS67506 addition is caused by the hastened relaxation induced by serotonin and not reversed within the time frame of the experiment, as illustrated for GR113808 in (C). *p<0.05 vs. Basal; **p<0.05 vs. 5-HT.

3.5. Evidence for involvement of 5-HT4 receptors

Given subsequently to serotonin, the 5-HT4-selective antagonist GR113808 (1 μM) completely reversed the inotropic response and partially reversed the lusitropic effect, when measured after 1–5 min (Fig. 3C). The apparent incomplete reversal of the lusitropic effect most likely reflects slow reversal kinetics explained by slow dephosphorylation of some proteins involved in the lusitropic effect, e.g., troponin-I [30–32]. Serotonin concentration–response curves without and with 0.5 or 5 nM GR113808 were nearly parallel with −logEC50M values of 7.62 ± 0.06 (n=6), 7.32 ± 0.06 (n=6) and 6.37 ± 0.10 (n=6), respectively (Fig. 3D), and no difference in basal and maximal Fmax. The pooled average values for basal and maximal Fmax were 5.4 ± 0.4 and 6.1 ± 0.4 mN (n=18), respectively. The average apparent GR113808 inhibition constant (Kb) was 0.4 nM (−logKb=9.4), consistent with the reported affinity of GR113808 at 5-HT4 receptors [33]. Since GR113808 displays at least 1000-fold selectivity for the 5-HT4 receptor over all other receptors examined [33], we conclude that the effect is 5-HT4 receptor mediated. We also tested the ability of the 5-HT4 selective partial agonist RS67506 [34] to elicit a positive inotropic effect. RS67506 (1 μM) increased Fmax by 3.9 ± 0.4% (n=3, p=0.013, Fig. 3F) in CHF in contrast to no effect in Sham (data not shown). An increase to 10 μM RS67506 did not cause any additional augmentation of Fmax (Fig. 3F). Subsequent addition of 10 and 100 μM of 5-HT increased Fmax further to 9.9 ± 0.3% (n=3) and 16.8 ± 0.9% (n=3), respectively, above basal. The low effects of serotonin at these concentrations reflect the antagonist activity of the partial agonist RS67506 towards the full agonist serotonin in the failing hearts (Fig. 3F). When added subsequently to 5-HT (10 μM), RS67506 (10 μM) effectively reversed the developed positive inotropic effect (5-HT: 36.3 ± 6.5%, n=3 vs. RS67506: −3.8 ± 2.6%, n=3; Fig. 3E). Accordingly, RS67506 acted as a partial 5-HT4 agonist with an intrinsic activity around 10% compared to the full agonist serotonin in ventricle from failing hearts.

A 5-HT2A receptor-mediated inotropic effect of serotonin was demonstrated in rat atrium [11]. However, in CHF rat papillary muscles, ketanserin (0.1 μM) did not shift the concentration–response curve for serotonin to higher agonist concentrations [−logEC50M value with ketanserin 7.62 ± 0.06 (n=6) vs. 7.49 ± 0.08 (n=5) without, nonsignificant], demonstrating that the 5-HT2A receptor is not involved (Fig. 3E). The maximal inotropic responses with and without ketanserin were similar [Fmax increase without: 12.1 ± 1.2%; basal: 5.6 ± 0.5 mN, 5-HTmax: 6.3 ± 0.5 mN (n=5); with: 12.2 ± 1.2%; basal: 5.0 ± 0.4 mN, 5-HTmax: 5.7 ± 0.5 mN (n=6), nonsignificant; (dF/dt)max increase: 18.1 ± 3.5% vs. 17.0 ± 2.3%; nonsignificant].

3.6. Induction of 5-HT4(b) receptor mRNA in CHF

Quantitative RT-PCR was used to determine the mRNA level for 5-HT4(b) receptor, 5-HT2A receptor, and the heart failure-marker ANP normalised to GAPDH mRNA. In left ventricle and papillary muscle, respectively, 5-HT4(b) mRNA levels were 4- and 18-fold higher in CHF vs. Sham, whereas 5-HT2A mRNA levels were unchanged and ANP mRNA levels increased (Fig. 4). The ratio between normalised 5-HT4(b) and 5-HT2A mRNA levels was increased in CHF in both left ventricle and papillary muscle, confirming increased 5-HT4(b) mRNA level relative to 5-HT2A, independent of GAPDH (Table 3).

Fig. 4

Expression of 5-HT4(b), 5-HT2A and ANP mRNA in left ventricle and papillary muscle of Sham and CHF rats. Messenger RNA for 5-HT4(b) receptor (upper panels), 5-HT2A receptor (middle panels) and ANP (lower panels) was quantified by real-time quantitative RT-PCR in Sham and CHF left ventricle (left panels) and papillary muscle (right panels). The results were normalised to GAPDH and Sham was assigned a value of 1. *CHF (n=11) vs. Sham (n=7) p<0.01; **CHF (n=5) vs. Sham (n=4) p<0.05.

View this table:
Table 3

Ratio of 5-HT4(b) receptor mRNA to 5-HT2A receptor mRNA

3.7. Maximum inotropic response to serotonin and 5-HT4(b) receptor mRNA related to infarction size

In papillary muscles from infarcted nonfailing (MInf) rats, serotonin (10 μM) elicited an inotropic response qualitatively similar to the response observed in CHF animals. The magnitude of the response correlated positively (r2=0.62, p<0.0001) with infarction size up to 30–40% of the inner myocardial surface in MInf animals (Fig. 5A). In MInf hearts with infarct size 30–40%, the inotropic effect was of the same magnitude as in CHF hearts which all had infarct size >40%. An increase with infarction size was also observed for 5-HT4(b) and ANP mRNA expression, whereas 5-HT2A mRNA levels were unchanged (Fig. 5B). 5-HT4(b) and ANP mRNA levels correlated positively (r2=0.37, p<0.01). These findings are consistent with a gradual transition of phenotype related to the extent of myocardial changes secondary to the infarction.

Fig. 5

Induction of serotonin responsiveness and 5-HT4 mRNA in nonfailing infarcted hearts, depending on infarction size. (A) Inotropic response to 10 μM serotonin in papillary muscle and (B) 5-HT4, 5-HT2A and ANP mRNA levels in remote (noninfarcted) tissue from left ventricle are shown as a function of infarction size. Nonfailing infarcted (MInf) hearts were grouped according to infarction size. Levels of mRNA were normalised to GAPDH and the group with MI size ≥ 10% was assigned a value of 1. The infarction size of CHF hearts was 44.4 ± 3.5%. For MI size ≥ 10% in (B), the error bars indicate ± half range of n=2.

4. Discussion

We demonstrate for the first time the induction of a ventricular inotropic response to serotonin in infarcted hearts and CHF. The mechanical response, absent in Sham, is of similar size as the maximal inotropic response to β-adrenoceptor stimulation in CHF. The response to serotonin is mediated by Gs-coupled 5-HT4 receptors and accompanied by small increases of cAMP levels. The 5-HT4 receptor mRNA expression is also increased in CHF.

We recently reported functional 5-HT4 receptors in failing human ventricle and an increase in 5-HT4 mRNA in failing human hearts [10]. The appearance of functional 5-HT4 receptors in failing rat ventricle, together with the evidence in human heart, adds new dimensions to neurohumoral changes in CHF and serotonergic signalling in the heart. Since no inotropic effect of serotonin was previously observed in human [15,16] or porcine [17,18] ventricles, administration of 5-HT4 receptor antagonists to prevent arrhythmia [12] was thought of as an atrial-specific intervention [13]. We now propose that the appearance of ventricular 5-HT4 receptor function after myocardial infarction and increased ventricular 5-HT4 mRNA in rats with developing and manifest heart failure may provide an experimental model for 5-HT4 receptor function in human ventricle. 5-HT4 receptor antagonists may therefore target not only atrial 5-HT4 receptors [8,12], but also ventricular 5-HT4 receptors [10].

5-HT4 receptor mRNA was detectable in the Sham rat ventricles (this study) and in rat atria [11,14], but functional responses could not be elicited in these tissues, possibly due to low or absent receptor protein expression. Robust inotropic responses have previously been detected in human and porcine atria despite a very low 5-HT4 receptor density [35,36]. This is apparently also the case in the failing rat ventricle, as we found the 5-HT4 receptor density to be below the limit of detection with both [3H]GR113808 and [125I]SB207710 as radioligands.

Although more than one mechanism of action may be involved, we hypothesise, based on several lines of evidence, that the inotropic effect of serotonin in CHF is at least partly cAMP mediated: (1) Serotonin increased cAMP levels, although to a small extent, in perfused hearts and isolated cardiomyocytes. (2) Serotonin elicited fast-developing inotropic and lusitropic effects–characteristics similar to responses to isoproterenol. (3) The pharmacological criteria for the involvement of Gs-coupled 5-HT4 receptors were fulfilled, and 5-HT4 receptor mRNA was induced in CHF. (4) The inotropic effect of serotonin was attenuated by muscarinic receptor activation. (5) In isoproterenol-stimulated muscle, serotonin did not cause additional response, as would be expected if different signalling pathways were used. Thus, in addition to the measured increase in cAMP levels, all the functional and pharmacological characteristics observed in relation to the serotonin response are to be expected for cAMP-mediated mechanical responses in mammalian heart muscle [25]. Furthermore, activation of 5-HT4 receptors in human atria increases cAMP levels, accompanied by increased PKA activity and phosphorylation of L-type calcium channels, phospholamban and troponin-I, similar to β-adrenoceptor activation [7,8].

Although the lack of an additional effect of serotonin in the presence of isoproterenol and the similarity of the time courses strongly indicate that both agonists fully activate a common signalling pathway in a functionally important compartment, the increase in total cAMP content induced by serotonin was small compared to that induced by isoproterenol. Compartmentation of cAMP may result in differing responses (mechanical and biochemical) to activation of cAMP production by different agonists and different receptor types [37–39]. Correlating increases of cAMP in different subcellular compartments with the functional response indicated that cAMP bound in a particulate fraction was of special relevance [39]. This fraction amounted to about 20% of the total cAMP content in rat heart. Thus, a small increase in total tissue level of cAMP may reflect a sufficient increase in a functionally important compartment. The present results do not, however, exclude that other mechanisms of action may be involved in the response to serotonin.

The mechanism of induction of the 5-HT4-mediated inotropic response is unknown. The results from MInf rats indicate that manifest CHF is not a prerequisite. The lack of a further increase in inotropic response from MInf rats with infarction size 30–40% to CHF rats may reflect that the inducing mechanism(s) are fully activated in this situation or that there may be a desensitisation of the Gs-mediated responses in CHF (as for β-adrenoceptor signalling) superimposed upon the induction. The trigger mechanism for induction of the serotonin response should be sought among changes related to the extent of myocardial damage. Similar studies with different CHF models will be helpful to further address this question. Previous studies did not detect 5-HT4 mRNA in rat ventricle [11,14,40]. Although 5-HT4 mRNA is present in atria from normal rats [11,14,40], no functional response was detected [11]. Although speculative, it is tempting to suggest that 5-HT4 receptor-expression in CHF ventricle, like ANP, reflects some kind of “atrialisation” of compromised ventricular myocytes. The positive correlation between 5-HT4 mRNA and ANP mRNA in MInf hearts is in line with this.

A crucial question is whether the endogenous serotonin level is sufficient to activate myocardial 5-HT4 receptors in vivo. Plasma serotonin concentration in rats was estimated to approximately 0.3 μM [11], about 10-fold higher than the EC50 (0.02–0.03 μM) for serotonin in the present study, at recombinant 5-HT4 receptors [19] and in human atrium [7,8]. In addition, sympathetic nerve endings have been shown to capture and release 5-HT analogous to norepinephrine, thus stimulating myocardial 5-HT4 receptors [6], consistent with an activation of myocardial 5-HT4 receptors in vivo.

The induction of the ventricular serotonin effect in infarcted hearts and CHF may be considered compensatory, e.g., to rescue contractile function. In CHF, sympathetic nerve activity and plasma levels of norepinephrine are increased to levels shown to cause desensitisation in the β-adrenoceptor signalling cascade and to induce cardiomyocyte injury [41]. Overexpression of cardiac β1-adrenoceptors or Gαs in transgenic mice also produces cardiomyopathy [42,43]. In this light, chronic induction of a cAMP-mediated serotonin response may appear maladaptive, considering that although β-adrenoceptor signalling is desensitised in CHF, treatment with β-blockers is beneficial [2]. The plasma norepinephrine level in CHF is positively correlated to mortality. Corresponding data for serotonin are lacking. There are reports of increased plasma levels and activity of serotonin in human CHF [3,4], and increased left ventricular serotonin concentrations were reported in hypertensive heart disease [44]. We have recently found functional 5-HT4 receptors in the ventricular myocardium of patients with terminal heart failure, including ischaemic heart disease and reported that serotonin can elicit experimental ventricular arrhythmias [10]. It is therefore conceivable that 5-HT4-selective antagonists may prevent life-threatening arrhythmias in terminal heart failure. Although speculative, if serotonin stimulates the ventricular 5-HT4 receptors in vivo in CHF, blockade of these receptors may have a beneficial effect, in analogy with β-adrenoceptor-blockade.

5. Conclusion

A 5-HT4 receptor-mediated inotropic response to serotonin, associated with increased 5-HT4 receptor mRNA levels, is induced in ventricular myocardium of infarcted and subsequently failing rat heart. The 5-HT4 receptor and β-adrenoceptors apparently share a common signalling pathway in CHF rats. Future challenges comprise exploring the trigger mechanisms responsible for induction of the serotonin response and exploring the pathophysiological and potential therapeutic importance of this novel neurohumoral change in CHF. We propose that the infarcted failing rat heart is an experimental model for investigation of ventricular myocardial 5-HT4 receptors, present also in the failing human heart.


Supported by The Norwegian Council on Cardiovascular Diseases, The Research Council of Norway, Anders Jahre's Foundation for the Promotion of Science, The Novo Nordisk Foundation and The Family Blix foundation.


  • 1 EQ and TB contributed equally to the present work.

  • Time for primary review 30 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
View Abstract