Stretch-stimulated ANF secretion: Are the ATP-sensitive potassium channels the missing link?

See article by Saegusa et al.[9](pages 60–68) in this issue. In 1981, de Bold et al. reported that the administration of atrial extracts to rats induced diuresis and natriuresis [1]. Soon afterwards, the substance with diuretic and natriuretic properties present in the atrial extracts was identified as the atrial natriuretic peptide (ANF) [2]. This important discovery led to the new concept that the heart was not only a mechanical pump but also an endocrine organ. Since the characterization of ANF, many reports appeared in the literature describing a wide spectrum of biological actions displayed by this peptide supported by the wide distribution of the receptors in different tissues and cell types. Other natriuretic peptides were also characterized: BNP and CNP. The family of natriuretic peptides plays a relevant role in the regulation of sodium and water as well as in cardiovascular homeostasis. ANF is synthesized, stored and secreted by mammalian atrial cardiocytes. The peptide is continuously released from the heart but appropriate mechanical (atrial stretch) or neuroendocrine (endothelin-1 or α1 adrenergic stimulation) stimuli enhance the rate of release with or without a concomitant increase in the synthesis [3]. Although many factors modulate the release of ANF, the major physiological stimulus is atrial stretch evoked by hemodynamic changes. Stretch secretion coupling results in an immediate increase of ANF secretion, but this output rapidly diminishes although the hormone stores are not significantly depleted. Even if the stimulus is maintained, the output of the hormone decreases. The secretion of ANF following stretch derives from a depletable pool of newly synthesized hormone with no effect on the synthesis [4]. The chronic elevation of ANF is maintained mainly by neuroendocrine mechanisms and involves increased synthesis [3]. Signaling for ANF release in response to endothelin-1 (ET-1) and norepinephrine has been widely studied and established, but the mechanisms underlying transduction of the mechanical stretch stimulus into a secretory response by atrial myocytes are poorly understood. An extensive literature exists on this issue, diverse mechanisms have been proposed, and various candidates have emerged as mediators between stretch and ANF secretion, including ion channels and second messengers as well as ET-1 and norepinephrine. Although ET-1 is a potent stimulator of ANF secretion, locally released ET-1 is not involved in stretch-induced ANF output. In fact, the secretion of ANF evoked by either ET-1 or α1 adrenergic stimulation requires several minutes to become fully established, whereas that induced by stretch is almost instantaneous [5]. Among the second messengers, phospholipase C (PLC) and cAMP were reported to be involved in ANF release, but current evidence seems not to support the idea that these second messenger mediate the relationship between stretch and ANF output. PLC activity is enhanced by cardiac stretch and protein kinase C (PKC) stimulates ANF secretion. However, PKC inhibitors have been shown not to affect stretch-induced ANF release [6], although PKC activators positively modulate it [7]. The extensive literature on the role of cAMP in ANF secretion evoked by stretch is controversial. Nevertheless, these second messengers have been shown to modulate ATP-sensitive potassium channels (KATP channels), a mechanosensitive ion channel present not only in the heart but also in other tissues [8]. Furthermore, in the present issue of Cardiovascular Research, Saegusa et al. report by using a novel approach that ATKATP channels are involved in stretch-evoked ANF release, as discussed further below [9]. The possible involvement of ion channels is supported by the observation that a rapid increase in ANF secretion occurs following stretch. The role of Ca2+ channels has not yet been well defined, although most evidence favors the participation of L-type voltage-gated Ca2+ channels in the negative modulation of ANF release. The existence of stretch-activated ion channels in the atria appears to be a promising clue to solve the enigma. In the atria, various stretch-activated channels exist, including KATP, chloride and cation channels. Although experiments with ion substitution have remained inconclusive thus far, the pharmacological manipulation of KATP channels has provided promising but different results according to the preparation used. Thus, KATP channels openers inhibit ANF output evoked by stretch in isolated hearts but have no effect in isolated atria [10,11]. Moreover, KATP channel blockers increase stretch-induced ANF release in isolated hearts but inhibit the secretion in isolated atria [10,11]. Basal ANF release is not affected by KATP channel modulators. In agreement with the results found in isolated hearts, it was reported that in cultured atrial appendage myocytes, where the effects of endothelial cells, nerve endings and mechanical effects of the ventricles are excluded, distinct putative KATP channel openers (diazoxide, pinacidil and 2-deoxyglucose) exert inhibitory effects on stretch-stimulated ANF release [12]. Further, the inhibition of stimulated ANF release correlates with the magnitude of the stretch-activated KATP current. These findings clearly indicate that atrial stretch activates KATP channels and supports the notion that stretch-induced potassium efflux appears to be involved in ANF secretion. However, the concern is whether results from neonate atrial appendage myocytes can be extended to adult myocytes and to in vivo studies. A further step in elucidating this issue was the recent finding that Gi/o proteins are involved in stretch-stimulated ANF secretion [13]. The release of ANF induced by stretch is abolished in the presence of pertussis toxin, suggesting that Gi/o proteins are involved in the intracellular pathway responsible for the stretch coupling mechanism. Other authors reported that ET-1 inhibits cardiac KATP channels via pertussis toxin-sensitive G proteins [14]. Whether Gi/o proteins and KATP channels are related components of the complex mechanism underlying stretch-stimulated ANF secretion remains to be elucidated. KATP channels also play a relevant role in glucose-stimulated insulin release. Similarities as well as discrepancies exist regarding their functioning and regulation in the heart and pancreas [15]. In the present issue, Saegusa et al. advance our knowledge regarding the role of the KATP channels in stretch-induced ANF release by using mutant mice lacking the Kir6.2 subunit of these channels [9]. This is a different and interesting experimental approach in the study of the issue. The authors further strengthen the view that KATP channels are involved in stretch-stimulated ANF release. The mechanical stimulus of atrial myocytes would activate KATP channels in association with ANF secretion. Intracellular calcium would be negatively modulated by KATP channels, which would in turn result in a negative feedback mechanism controlling ANF output. The authors suggest that an impaired KATP channel function plays a protective role in pathophysiological states, whereas under physiological conditions these channels would act as a negative feedback mechanism for ANF output. Although a step ahead, the search for the missing link between stretch and ANF release still continues.