In this study, we aimed to characterize in the pig the inotropic and lusitropic effects of 5-HT, and the 5-HT4 receptor agonists prucalopride, tegaserod, R149402 and R199715 in the left atria, and their chronotropic effects in spontaneously beating right atria; the role of PDE enzymes in the regulation of the inotropic response was studied by using the PDE enzyme inhibitor IBMX. To verify whether the responses are developmentally related, we also studied the left atrial myocardial and sinoatrial 5-HT4 receptors of newborn piglets.
Inotropic effects in the porcine left atrium
Involvement of the 5-HT4receptor: 5-HT, as well as the 5-HT4 receptor agonists prucalopride, tegaserod, R149402 and R199715, evoked positive inotropic effects in the porcine left atrium. 5-HT4 receptors have been shown to be the only functional receptor for 5-HT in porcine right and left atrium (Lorrain et al., 1992; Medhurst & Kaumann, 1993; Parker et al., 1995). The 5-HT4 receptor antagonist GR113808 was able to prevent and/or reverse the agonist-induced inotropic responses, supporting the idea that the effects are 5-HT4 receptor-mediated. The increased contraction force upon 5-HT4 receptor activation is probably due to an increase in L-type Ca2+ current, an effect that has been described in human atrial myocytes for 5-HT (Ouadid et al., 1992; Pau et al., 2003), and very recently also for prucalopride (Pau et al., 2005). A mechanism involving the phosphorylation of these Ca2+ channels by cAMP-dependent protein kinase was proposed, similar to the well-documented cardiac effects of β-adrenergic receptor activation by agonists (Ouadid et al., 1992). This is in accordance with the 5-HT4 receptor-induced stimulation of adenylyl cyclase and PKA in human atrial strips (Kaumann et al., 1990). The increased L-type Ca2+ current and the subsequent Ca2+-induced release of Ca2+ from the sarcoplasmatic reticulum, through the activation of ryanodine channels, is thus a plausible explanation of the observed positive inotropic action of the 5-HT4 receptor agonists.
Fading of 5-HT4receptor-induced inotropic response: The positive inotropic response observed after the administration of a single concentration of agonist was transient. This time dependency was most clear for 5-HT, where the response returned to basal values within 15 min. For the higher concentrations, the response stabilized below the basal contraction level. The same time dependency was observed for the increase in contraction rate (+dF/dt). This excludes the possibility that a hastened relaxation causes the fading of the contraction force. For experiments with 5-HT, the contraction rate stabilized at a level that was not significantly different from basal for all concentrations of 5-HT tested. The other agonists showed slower kinetics in the rising as well as in the descending phase of the inotropic response, that could also stabilize below basal. Furthermore, there is a clear influence of the administration method on the concentration–response relationship. In a cumulative administration protocol, 5-HT appeared less efficacious compared to a noncumulative administration, while for the other agonists no positive inotropic response was observed in the cumulative protocol. Moreover, the cumulative concentration–response curve of 5-HT shows a typical bell shape. These observations suggest the involvement of a mechanism inducing fading of the inotropic response. In the cumulative protocol, the administration of the highest 5-HT concentration was not able to overcome the progressive fading of the response. Therefore, we can exclude the possibility that degradation of the ligand contributes to the fading response.
As in this study, in the human paced left atrium, the cumulative administration of 5-HT generates a bell-shaped concentration–response curve (Sanders & Kaumann, 1992). On the other hand, the response to a single concentration of 5-HT did not fade in the paced human right atrium, while the response to 5-CT and renzapride faded after a prolonged exposure to the agonists (Kaumann et al., 1991). It is noteworthy that the response following a cumulative administration of renzapride appeared bell shaped, while the response to 5-HT was not. This indicates that the mechanisms regulating the generated response can be different between different species, between both atria or even between different agonists.
Different mechanisms can be involved in the observed fading of the inotropic response. Some splice variants of the 5-HT4 receptor have been shown to couple to Gi proteins in transfected cells as well as in rat cardiac myocytes (Pindon et al., 2002; Castro et al., 2005). We consider it unlikely that the fading response is caused by a time-dependent switch in G protein selectivity from Gs to Gi, which has been described for β2-adrenergic receptors (Heubach et al., 2004; Martin et al., 2004) since GR113808 would also block this Gi-mediated effect and would thus not decrease the contraction force once it has stabilized.
5-HT4 receptors have been shown to desensitize upon activation by their ligands (Ansanay et al., 1992; Ronde et al., 1995; Mialet et al., 2003). The 5-HT4(a) receptor undergoes rapid (t(1/2)∼2 min) concentration-dependent phosphorylation after stimulation with 5-HT (Ponimaskin et al., 2005), which is the initial trigger for desensitization leading to uncoupling of the receptor from the G proteins (Ferguson, 2001). Both in the presence and absence of IBMX a maximal inotropic response for 5-HT was reached within 1–2 min after the administration of a single concentration of 5-HT, which is in accordance with the phosphorylation time profile described for the 5-HT4(a) receptor. The occurrence of desensitization tallies with the difference in maximal effect of 5-HT between noncumulative and cumulative concentration–response curves (and the absence of positive inotropic effects upon cumulative administration of prucalopride, tegaserod, R149402 and R199715) since in a cumulative protocol, previously administered agonist concentrations, causing (partial) receptor desensitization, will interfere with the equilibrium between the additional agonist concentrations and the remaining receptors. When the inotropic response to a 5-HT4 receptor agonist stabilized below basal levels, GR113808 was still able to induce a further small reduction of the contraction force. The contraction velocity, stabilized near basal, was also further reduced to a level below basal by GR113808. Similar to what has been observed in the human ventricle after stimulation of IBMX-treated ventricular trabeculae with 5-HT (Brattelid et al., 2004b), GR113808 also reversed the stable increased contraction force upon stimulation with prucalopride or tegaserod in the presence of IBMX. Although 5-HT4 receptors have been shown to possess a large, splice variant-dependent, constitutive activity (Claeysen et al., 1999), it seems unlikely that GR113808 acted as an inverse agonist at a constitutively activated 5-HT4 receptor, as GR113808 did not show any effect on the basal contraction force, both in the absence or presence of IBMX (data not shown). It can thus be deduced that the 5-HT4 receptors are at least still partially active and thus certainly not completely desensitized, as they remain accessible for GR113808 when the response has stabilized.
Role of PDE enzymes
PDE enzymes, which catalyse the breakdown of cAMP, obviously play a role in the observed fade. Indeed, the inhibition of these enzymes with IBMX results in a marked increased responsiveness to single concentrations of prucalopride and tegaserod with no fading for 30 or 60 min, respectively (contraction force as well as +dF/dt). Furthermore, IBMX greatly potentiates the 5-HT-induced contractions, and unravels the contractions upon the cumulative administration of prucalopride, tegaserod, R149402 and R199715. This is in line with a recent study, in which the inhibition of PDE led to the discovery of functional porcine and human ventricular 5-HT4 receptors (Brattelid et al., 2004b), hereby contradicting previous reports, where no 5-HT4 receptor-mediated effects in human and porcine ventricles could be shown, probably because no PDE inhibitor was used (Jahnel et al., 1992; Schoemaker et al., 1992; 1993). Furthermore, in the presence of IBMX, the concentration–effect relation for prucalopride was the same for cumulative and noncumulative administrations. This suggests that PDEs are the major cause of the observed contrast between the effect of the agonists with both administration methods in the absence of IBMX. This does not exclude the possibility of the involvement of concentration-dependent (partial) rapid desensitization since this might be masked by IBMX in the cumulative protocol.
For the β2-adrenoreceptor a mechanism has been demonstrated by which β-arrestins, which are also involved in 5-HT4 receptor desensitization (Ponimaskin et al., 2005), contribute to the degradation of cAMP by recruiting cAMP-specific PDEs to ligand-activated receptors (Perry et al., 2002). Both β2-adrenergic and 5-HT4 receptors are expressed on the same cardiac cells and can even form heterodimers (Ouadid et al., 1992; Berthouze et al., 2005). Therefore, in analogy with β-adrenoreceptors, β-arrestins could also be involved in the recruitment of PDEs upon 5-HT4 receptor activation, and the observed fading in our experiments could thus well be caused by an integrated cascade of events that involve the coordinated activation of receptor-desensitizing mechanisms and the quenching of the receptor-induced activity.
Furthermore, PKA anchor proteins (AKAPs) serve to sequester PKA to distinct subcellular compartments for local activation, and the AKAP–PKA complex scaffolds and activates PDEs, which maintain compartmentation by limiting the diffusion of the second messenger (Dodge et al., 2001; Baillie & Houslay, 2005). This mechanism generates local PDE activity and therefore controls the spatial gradient of cyclic nucleotides (Zaccolo & Pozzan, 2002; Carlisle Michel et al., 2004; Rochais et al., 2004) and provides a molecular framework for a negative feedback mechanism by which cAMP may regulate its own levels (Rochais et al., 2004). It is possible that, also for the 5-HT4 receptor, PDEs play a role in the targeting of the response, or, as Brattelid et al. (2004b) suggested, that they have a protective role against 5-HT-induced cardiostimulation. The cardiac compartmentation has been shown to be greatly reduced by IBMX (Jurevicius & Fischmeister, 1996). The increased inotropic effect of the 5-HT4 receptor agonists in the presence of IBMX could therefore be a direct consequence of the inhibition of PDE action and/or, since compartmentation is lost, result from the more diffuse activation of cAMP substrates.
The reduction of the contraction force below basal after stimulation with an agonist is in line with the loss of the positive inotropic response by PDE action. The observation that the contraction force stabilized below basal, while the +dF/dt stabilized at the basal level, can be explained by the decrease of R2 at this moment. This indicates that there is still a lusitropic effect present, which is thus less sensitive to the negative influence of the PDEs than the positive inotropic response. This prolonged relaxant effect might be related to the lusitropic role of phospholamban and troponin I (see below), as they have slow dephosphorylation kinetics (Garvey et al., 1988).
The same argumentation could explain the observations with GR113808. Just as PDEs, GR113808 preferentially blocks the positive inotropic response over the lusitropic one, since R2 remains decreased. This is in line with the study of Brattelid et al. (2004b), who found that the positive inotropic effects of 5-HT in human ventricle were completely reversed by GR113808, whereas the lusitropic effects were only partially reversed. However, in the presence of IBMX, R2 was unaffected by the stimulation with prucalopride, yet the maximal contraction force and velocity (independent of lusitropic responses) was also reduced below its basal level by GR113808 (which cannot be accounted for by PDEs in this condition). Therefore, some negative inotropic mechanism other than PDE activation seems to be involved here.
Comparison of 5-HT4 receptor agonists: In the presence of IBMX, prucalopride, R149402 and R199715 acted as partial agonists in the paced left atrium, while no significant difference in efficacy was found between tegaserod and 5-HT. These results confirm previous findings with piglet sinoatrial and human right atrial tissue, in which substituted benzamides behaved as partial agonists (Kaumann et al., 1991; Villalon et al., 1991; Medhurst & Kaumann, 1993). Langlois & Fischmeister (2003) already stated that the stimulation of 5-HT4 receptors in piglet isolated right atrium is a useful method for evaluating the pharmacological profile of 5-HT4 receptor ligands. Prucalopride also acted as a partial agonist on the L-type Ca2+ current through 5-HT4 receptors in human atrial cells (Pau et al., 2005). This study together with our current findings show the partial agonistic character of prucalopride in a cardiac assay. It is interesting that R149402 and R199715, which are both structurally related to prucalopride, are more potent but less efficacious than their parent molecule. Tegaserod results in slow inotropic responses, and, for the higher concentrations, displays an additional inotropic response following a lag phase. In the literature, tegaserod is described as a high-affinity, potent 5-HT4 receptor agonist. Compared to these reports, the potency by which tegaserod increased left atrial contraction force in the present study was considerably lower (Beattie et al., 2004). But, since in our experiments, the curve to tegaserod has a low Hill slope, these findings might be related to a very difficult penetration of the chemical into the tissue. Although tegaserod has been shown to interact with 5-HT1 and 5-HT2 receptors (Briejer et al., 2001; Beattie et al., 2004), we consider it doubtful that these interactions occur in porcine atrium as 5-HT4 receptors are the only ones described in pig atria. However, the possibility that tegaserod interacts with non-5-HT4 receptors cannot be excluded. Interestingly, a low Hill slope is also observed for the noncumulative concentration–response curve to 5-HT, which is structurally related to tegaserod as they both are indolic compounds. Another alternative explanation could be that these indoles stabilize a specific receptor state that activates multiple G proteins, a mechanism that is called agonist-directed trafficking (Kenakin & Morgan, 1989; Kenakin, 1995).
Influence of development: In newborn piglets, the 5-HT4 receptor agonists were less potent and/or less efficacious in inducing left atrial inotropic responses than in young pigs; results with R149402 and R199715 could even not be adequately fitted. A similar age dependency was recently shown for 5-HT4 receptor-associated inotropic responses induced by 5-HT in porcine ventricular trabeculae carnae. 5-HT was 15-fold more potent and twice more efficacious in the trabeculae of 3 months old pigs than of newborn piglets. Furthermore, the increase of PKA activity caused by 10 μM 5-HT was twice as big in young pigs than in newborns (Brattelid et al., 2004b). In our left atrial experiments, a similar potency shift, but a smaller efficacy difference for 5-HT, is found. The developmental changes that cause the development-related increase in response to 5-HT4 receptor activation are not yet understood. One possibility is that the 5-HT4 receptor expression increases during development. The expression level of 5-HT4 receptors in the atria of piglets is indeed very low, being less than 1% of that of β1-adrenoreceptors (Kaumann et al., 1995). In comparison, in human (adult) atria the expression of 5-HT4 receptors equals 10 and 20% of that of β1- and β2-adrenoreceptors, respectively (Kaumann et al., 1996). In accordance with this higher receptor density, the 5-HT-induced inotropic responses on human atria are bigger than on piglet atria (Kaumann & Sanders, 1998).
Another possibility is that the 5-HT4 receptors become more efficiently coupled to their effectors during development. It has indeed been shown that chronic treatment with β-blockers causes an increased response to 5-HT at 5-HT4 receptors (Sanders et al., 1995; Pau et al., 2003), while it is documented that treatment with β-blockers does not change 5-HT4 receptor mRNA (Grammer et al., 2001). The increased sensitivity could be induced by an increased expression of G-protein. An age dependency of the expression, as well as the receptor-mediated activation of G-proteins, has been shown in human atrium (Kilts et al., 2003).
Lusitropic effects in the porcine left atrium
It has previously been shown that 5-HT changes the time to reach peak force in human left and right atria and ventricles (Kaumann et al., 1991; Sanders & Kaumann, 1992; Brattelid et al., 2004b). This correlates with the following sequence of events, proposed by Kaumann et al. (1990): Activated PKA catalyses the phosphorylation of the inhibitory regulatory protein phospholamban. This results in liberation of Ca2+ ATPase activity, resulting in an accelerated Ca2+ re-uptake in the sarcoplasmatic reticulum. PKA also phosphorylates troponin I, which causes a reduction in the affinity of Ca2+ for troponin C. This mechanism has been observed with catecholamines through β1- and β2-adrenoreceptors (Kaumann et al., 1999). We observed a hastening of the onset of relaxation (tpf) as well as a shortening of the relaxation time (t50) for 5-HT and all 5-HT4 receptor agonists. Interestingly, prucalopride, tegaserod, R149402 and R199715 induced lusitropic effects when administered cumulatively, while they had no clear positive inotropic effect. This can be explained by the preferential inhibition of the inotropic response by PDEs as discussed above. The potencies observed for both lusitropic effects of all agonists are comparable with each other. As observed for the inotropic response in the presence of IBMX, R149402 and R199715 were more potent than the other agonists.
An interesting observation is that prucalopride, R149402 and R199715, which are benzofuran compounds, are half as efficacious in reducing the tpf when administered noncumulatively than when administered cumulatively, while this is not the case for 5-HT and tegaserod, which possess an indolic structure. This phenomenon is not consistently observed for the reduction in t50. As a consequence, these benzamide-derived compounds induce a smaller reduction in the tpf than 5-HT in a noncumulative administration regime. A structural differentiation between 5-HT4 receptor agonists has been made previously by Pindon et al. (2002). They showed a cAMP-independent Ca2+ influx through 5-HT4(a), but not 5-HT4(b), receptor activation that was 2–3 times higher for benzamide-like structures than for indolamines in HEK 293 cells. Supposed this mechanism also occurs in the porcine atrial system, the additional inward Ca2+ transient would promote systolic contraction but hamper diastolic relaxation. Indeed, during diastole, the activation of 5-HT4 receptors results in a decreased Ca2+ concentration at the contractile proteins through a PKA-dependent phosphorylation of phospholamban and the linked increase in Ca2+ ATPase activity. The activation of the receptor with benzofurans, however, would then at the same time induce an inward Ca2+ flux and thus counteracting relaxation. The more efficacious relaxation with the benzofurans in the cumulative protocol could then be explained by the time-related inactivation of this additional Ca2+ effect or by 5-HT4 receptor desensitization. The latter should then reduce the Ca2+ influx, while phospholamban and/or troponin I remain phosphorylated for a longer time. The slow dephosphorylation kinetics for these proteins have indeed been shown, in agreement with our hypothesis (Garvey et al., 1988). The lower lusitropic efficacy of 5-HT in the cumulative protocol, compared to the noncumulative protocol, supports the idea of 5-HT4 receptor desensitization. The different coupling of benzamide-derived compounds fits into the concept of agonist-directed trafficking mentioned before (Kenakin & Morgan, 1989).
Chronotropic effects in the porcine right atrium
Porcine sinoatrial 5-HT4 receptors gained a lot of attention because of the lack of a human model to study 5-HT4-receptor-mediated tachycardia (Kaumann, 1990; Medhurst & Kaumann, 1993; Kaumann, 1994). 5-HT and all 5-HT4 receptor agonists tested caused positive chronotropic effects in the isolated right atrium of young pigs. An interesting difference with the left atrial inotropic response is that all agonists induced their effect when added cumulatively in the absence of IBMX. It is possible that the receptor is more efficiently coupled to the transducing mechanism in the sinus node than in the myocardium, and/or that PDEs are less involved in the 5-HT4-mediated response in the sinus node. Receptor-induced cAMP binds indeed directly to HCN channels to provoke the chronotropic effect, thus without the involvement of PKA, and steps further downstream in the cascade (Ulens & Tytgat, 2001), which might contribute to more efficient coupling. Furthermore, preliminary data also showed no fading of the response upon the administration of a single concentration of 5-HT, corresponding to a less pronounced influence of PDEs on the sinus rhythm. This may be explained by a different transduction mechanism in the compartments responsible for the frequency coupling, in the sinus node. It is possible that the influence of PDE in these compartments is limited. An interesting thought would be that the 5-HT4 receptor splice variants responsible for frequency coupling are different from those responsible for contraction coupling. If these splice variants show a different affinity for arrestin, AKAP or other members of the cascade complex, this could result in a lack of PDE recruiting. For example, it has been shown that the Na+/H+ exchanger regulatory factor (NHERF) interacts with AKAP. For β2-adrenoreceptors, this interaction has been shown to be functionally involved in the tight regulation of the signalling molecules into microdomains (Xiang & Kobilka, 2003). Furthermore, NHERF interacts with PDZ motifs on the 5-HT4(a) receptor, but not on the 5-HT4(e) receptor or with the 5-HT4(b) receptor (Joubert et al., 2004). Thus, there may exist different compartments for the different 5-HT4 receptor splice variants.
The benzamide-derived compounds prucalopride, R149402 and R199715 behaved as partial agonists, while tegaserod reached a maximal effect that was not different from 5-HT. The occurrence of arrhythmia was also most prominent in experiments with tegaserod. As for the inotropic response, the very low Hill slope associated with tegaserod suggests a very difficult penetration of the chemical into the tissue or an additional transducing mechanism associated with a specific receptor state.
5-HT and the 5-HT4 receptor agonists also increased the right atrial frequency in newborn piglets; the potency was similar as in young pigs. Results for 5-HT were in good agreement with the literature on sinoatrial 5-HT4 receptors in newborn piglets (Kaumann, 1990; Medhurst & Kaumann, 1993). Prucalopride, tegaserod, R149402 and R199715, but not 5-HT, are less efficacious in newborns. The decrease in efficacy was also observed for the inotropic responses in left atrium. It is likely that a general developmental change in receptor density and/or coupling efficiency is responsible for a bigger 5-HT4 receptor-mediated left atrial inotropic and right atrial chronotropic response. It is thus surprising that 5-HT, as well as isoprenaline, induced the same absolute increase in heart rate in newborns as in young pigs, irrespective of the higher basal beating rate in the former.