| Mapping the cellular mechanisms regulating atrial natriuretic peptide secretion: | ||
|---|---|---|
| Prev | Chapter 2. Review of literature | Next |
Although mechanical stretch activates multiple second messenger systems, it is unknown which molecules are directly activated by stretch and which molecules are indirectly activated by other upstream modulators. A mechanosensitive molecule is assumed to have some interaction with the plasma membrane in order to sense tension of the membrane. One candidate for mechanosensor are stretch-activated (SA) ion channels, which are suggested to interact directly with cytoskeleton and thus can sense the cell stretch (Hamill & McBride 1992, Sackin 1995). SA channels have been identified e.g. in vascular endothelial (Lansman et al. 1987, for review see Davies 1995) and smooth muscle (Davis et al. 1992b) cells as well as in atrial (Sadoshima et al. 1992, Kim 1993) and ventricular myocytes (Craelius 1993, Ruknudin et al. 1993). These include potassium channels, nonselective cation channels, and stretch-inactivated channels (Sadoshima et al. 1992, Sigurdson et al. 1992, Ruknudin et al. 1993, Sackin 1995). Opening of SA channels causes an increase in intracellular Ca2+ concentration because some of these channels are permeable to Ca2+. Nonselective SA channels in the heart have been shown to be blocked by Gd3+ (Sadoshima et al. 1992) and stretch-induced rise in the intracellular Ca2+ concentration has been shown to be decreased by Gd3+ in cardiac cells (Sigurdson et al. 1992). Gd3+ has also been shown to block the stretch-induced increase in BNP mRNA levels in isolated superfused atrium, suggesting that Gd3+ -sensitive stretch-activated channels may be involved in stretch-induced hypertrophic responses (Laine et al. 1996). In contrast, Sadoshima and co-workers have shown (1992) that treatment of myocytes with Gd3+ does not affect stretch-induced immediate-early gene experssion or stretch-induced increase in the rate of protein synthesis. However, these results do not deny the possibility that a mechano-sensitive channel, which is Gd3+ insensitive, may work as a mechanotransducer.
Another candidate for mechanosensors are integrins, transmembrane receptors that couple components of the extracellular matrix with the actin cytoskeleton (Wang et al. 1993, Parson 1996). Cytoskeleton may form a complex with membrane proteins such as ion channels, adenylyl cyclase, and Na+/H+ exchanger and regulate their responsiveness to external forces (Watson 1991). Multiple integrins are expressed in the heart (Baldwin & Buck 1994). Integrins are suggested to interact e.g. with FAK, which further interacts with various signaling molecules (for review see Sadoshima & Izumo 1997). Studies by Sadoshima and co-workers (1992) argue against the possibility that SA cation channels, actin microfilaments, microtubules, integrins or contractile activity are necessary for the stretch-induced immediate-early gene induction.
Also G protein-linked receptors are likely to be involved in mechanotransduction. Activation may be secondary to flow-mediated binding of a known ligand, or the receptor may itself be uniquely mechanosensitive. Strong evidence for G protein responses to shear stress implicates endothelial receptors as putative flow sensors (for review see Davies 1995).
In cardiac cells mechanical stimuli-induced signal transduction is characterized by activation of multiple second messenger systems (Vandenburgh 1992, Sadoshima & Izumo 1993a, for review see Sadoshima & Izumo 1997). In cultured neonatal cardiac myocytes mechanical stretch causes activation of phospholipases C, D, and A2; tyrosine kinases; p21ras; Raf-1; MAPKs and their activators; SAPK; 90-kDa S6; PKC; and probably other molecules as well (Sadoshima & Izumo 1993a, Sadoshima et al. 1993, Yamazaki et al. 1995, Komuro et al. 1995). Phosphatidylinositol turnover has been shown to be increased by mechanical stimuli in cardiac myocytes (von Harsdorf et al. 1988, 1989, Sadoshima & Izumo 1993a) and mechanical stretch also activates phospholipases within minutes (Sadoshima & Izumo 1993a) leading to the generation of various lipid-derived second messengers such as IP3, DAG, arachidonic acids, and phosphatidic acid. PKC activity and concentration has been reported to increase during development of left ventricular hypertrophy induced by pressure overload in rat hearts (Gu & Bishop 1994). In addition, mechanical stretch could increase Ang II secretion leading to the activation of phospholipases (Sadoshima & Izumo 1993b). In vitro mechanical stress stimulates the activation of all components of the Raf-MEK-ERK signaling cascade in neonatal cardiac myocytes (Sadoshima & Izumo 1993a, Yamazaki et al. 1995). When cardiac myocytes of neonatal rats cultured on a deformable silicone dish were stretched, activity of SAPK was increased (Komuro et al. 1996). Activation of SAPK by stretch is relatively slow (15min) compared with ERK activation (5 min)(Komuro et al. 1996, Sadoshima & Izumo 1997). Recently, hypotonic cell swelling of cardiac myocytes has also been shown to activate ERKs and SAPK (Sadoshima et al. 1996). Thus, the activation of these pathways may be a common signaling mechanism in response to different types of increased membrane tension in cardiac myocytes.
RPTKs have transmembrane segments, and some of the nonreceptor-type PTKs are anchored to the inner surface of cell membranes. Thus it is possible that membrane stretch directly causes conformational changes of tyrosine kinases, thereby activating them. Sadoshima & Izumo (1993a) have shown that mechanical stretch of cardiac myocytes causes a rapid increase in phosphotyrosine content of protein such as p42, p44, p60, p70, p85, p120, and p170 within 1 minute. A significant increase in tyrosine phosphorylation can be observed in 5 s (Sadoshima et al. 1996). Thus, the activation of PTKs appears to be one of the earliest cellular responses observed in response to mechanical stretch. Hypotonic swelling-induced immediate-early gene expression was abolished by PTK inhibitors but not by inhibitors of PKC, PLC, or Ang II antagonists (Sadoshima et al. 1996). In contrast, stretch-induced immediate-early gene induction was inhibited by inhibitors of PTK, PKC and PLC, as well as by Ang II antagonists (Sadoshima & Izumo 1993a, Sadoshima et al. 1993). Thus although hypotonic cell swelling and linear stretch activate separate signaling mechanisms, tyrosine kinase activation is required for both stimuli (Sadoshima & Izumo 1997). However, the mechanism of tyrosine kinase activation by mechanical stress is still far from clear.
The major determinant of ANP secretion is myocyte stretch (Ruskoaho 1992, de Bold et al. 1996). A rapid and complete inhibition of stretch-induced ANP secretion was observed after cellular ATP depletion (Page et al. 1991), thus emphasizing the energy-dependence of the secretory event. When considering the stretch-dependency of ANP secretory response the SA cation channels could be the most obvious candidates initiating the stretch-secretion coupling. In support of this, Laine and co-workers (1994) found that Gd3+, a known blocker of SA channels, dose-dependently inhibited stretch-induced ANP secretion in superfused atrium.
Several lines of evidence support the concept that PKC activation may promote ANP secretion from the heart. In the atrial and ventricular myocytes PKC is present in both membrane and cytosolic fractions (Kuo et al. 1984, Yuan & Sen 1986). Stimulation of phosphatidylinositol turnover and formation of IP3 was noted when right atria were dilated (von Harsdorf et al. 1988, 1989). Tumor promoting phorbol esters, which activate intracellular PKC, have been shown to stimulate basal ANP secretion in cultured myocytes (Matsubara et al. 1988, Iida and Page 1988, 1989, Shields & Glembotski 1989) and in perfused hearts (Ruskoaho et al. 1985, 1986b, for review see Ruskoaho 1992).
Ruskoaho et al. (1990) have shown that activity of PKC appears to positively regulate stretch-induced ANP release. When 12-O-tetradecanoyl-phorbol-13-acetate (TPA), tumor promoting phorbol ester known to stimulate PKC, was added to the perfusion fluid, a dose-dependent augmentation of stretch-induced ANP release was observed. In addition, inactive phorbol ester 4α-PDD had no effect (Ruskoaho et al. 1990). It is expected that hormones and neurotransmitters that activate phosphoinositide hydrolysis in heart cells would influence stretch-mediated ANP secretion. In fact, ET-1, α-adrenergic agonists, Ang II and vasopressin have been shown to augment the atrial stretch-induced ANP release both in vivo and in vitro (Ruskoaho 1992). PKC inhibitor H-7 has been reported to decrease secretagogue-induced release of ANP in isolated atria (Ishida et al. 1988). In addition, H-7 decreased atrial stretch-induced ANP secretion (Page et al. 1990) but conflicting results exists (Ishida et al. 1988, Ruskoaho 1992). ANP secretion produced by passive left ventricular wall stretch was inhibited by PKC inhibitor staurosporine (Kinnunen et al. 1993). These results indicate that PKC may mediate a part of the stretch-secretion coupling of ANP.
The role of other second messengers in the regulation of stretch-dependent ANP secretion is still far from clear. In the isolated perfused rat heart infusion of forskolin, a compound that increases cAMP concentration and then intracellular Ca2+, dose-dependently inhibited stretch-induced ANP release (Ruskoaho et al. 1990). Stretch-dependent ANP secretion can also be inhibited by 8-chlorophenylthio-cAMP and caffeine in noncontracting rat atria (Page et al. 1990) and by isoprenaline pretreatment in contracting rat atria (Agnoletti et al. 1992), further suggesting that cAMP may be a negative modulator of ANP secretion.
Several studies have addressed the importance of cytosolic Ca2+ as a possible intracellular messenger mediating stretch-induced ANP secretion. Atrial stretch and the rate of contractions, factors known to modulate ANP secretion (Ruskoaho 1992), are linked to changes in intracellular Ca2+. Processing of ANP to the released hormone seems to be dependent on Ca2+ (Ito et al. 1988). In addition, ANP granules inside the cells contain high concentration of Ca2+ (Somlyo et al. 1988, Thibault & Doubell 1992), further supporting the possibility of a connection between the intracellular Ca2+ level and ANP secretion. ANP secretion is also known to be dependent on several neurotransmitters and humoral agents, which induce phosphoinositide hydrolysis and the formation of IP3, eventually leading to Ca2+ mobilization from intracellular stores. ANP secretion has been shown to be dependent on several protein kinase activities and many of these kinases are Ca2+-dependent (Ruskoaho 1992).
Experiments have revealed both positive and negative modulation of stretch-induced ANP secretion by cellular Ca2+ (for review see Ruskoaho 1992). In Ca2+-depleted hearts in the absence of spontaneous contractility total ANP release by stretch was partially suppressed (Ito et al. 1988). Left atrial pressure-induced increase in ANP release in the isolated perfused heart preparation was significantly attenuated by low Ca2+, nifedipine, CaM antagonist W-7 and ryanodine (Katoh et al. 1990). Stretch-induced ANP secretion was also inhibited by CaM-binding drug in isolated rat atria (Page et al. 1990). Pretreatment by ryanodine, which inhibits the release of Ca2+ from the SR, inhibited stretch-induced ANP secretion in isolated rat atria (Kuroski-de Bold & de Bold, 1991). In addition, Laine and co-workers (1994) reported that ryanodine inhibited stretch-induced ANP secretion both in contracting and noncontracting superfused atrium. These result shows that internal Ca2+ stores are involved in the stretch-secretion coupling of ANP release.
In contrast, Ruskoaho and co-workers (1990) have shown that in the perfused rat heart preparation Bay K8644, a compound that increases the concentration of intracellular Ca2+, dose-dependently inhibited stretch-stimulated ANP release. In addition, a dose-dependent decrease in the rate of ANP secretion in isolated rat atria was noted at higher extracellular Ca2+ levels (Page et al. 1991). Ca2+ also inhibited ANP secretion from osmotically stretched neonatal atrial myocytes (Greenwald et al. 1988). Stretch-induced ANP release in isolated rat atrial preparation was independent of extracellular Ca2+ (Kuroski-de Bold & de Bold, 1991, de Bold et al. 1996) or ryanodine-sensitive Ca2+-release (Page et al. 1990). Similarly, ANP release from isolated rat atria induced by stretching was not inhibited by depolarization with KCl or a low concentration of external Ca2+ (Agnoletti et al. 1992). L-type Ca2+ channel blockers nifedipine (Deng & Lang 1992) and diltiazem (Laine et al. 1994) did not decrease stretch-induced ANP secretion, suggesting that the mechanotransduction may be mediated by other than voltage-gated Ca2+ channel. Thus, experiments by using various Ca2+ modulating agents suggest both negative and positive modulation of stretch-stimulated ANP release by intracellular Ca2+ concentration. These conflicting results concerning the effect of Ca2+ on stretch-induced ANP secretion may be explained by methodological differences including the species used, the age of the myocytes and the preparation technique of isolated atrial preparation. The rate of contractions, which is affected by changes in Ca2+ homeostasis, may also modulate the ANP secretory responses. This may explain the differences between the observations of studies where isolated spontaneously beating or paced organ preparations or arrested or slowly beating cell cultures are used. Furthermore, even under identical experimental conditions Ca2+ may have a dual effect on cellular ANP release by stretch; an initial transient stimulation followed by more marked inhibition was noted as intracellular Ca2+ was increased (Page et al. 1991).
Most of the ET produced by endothelial cells is released into the basolateral site (Wagner et al. 1992) and the circulating levels of the ETs are very low (Simonson 1993). Therefore, ET must be considered rather as a local autocrine or paracrine factor than as a circulating hormone. Recently it has been suggested that endothelial cells to some extent store ET-1 (Macarthur et al. 1994, McClellan et al. 1994), and when endothelial cells in culture are stretched, ET-1 can be released rapidly (Macarthur et al. 1994). Thus, it has been speculated that ET-1 may be the mediator of stretch-induced ANP secretion response. Mäntymaa and co-workers (1990) used a modification of the perfused heart preparation which permitted distension of right atrium. They found that ET-1 dose-dependently increased stretch-induced ANP secretion. In addition, it has been reported that in neonatal atrial cell cultures the response to ET administration in the presence of cyclic stretch was significantly greater than either cyclic stretch or ET alone (Gardner et al. 1991). Leskinen and co-workers (1997b) determined the effects of selective ET receptor antagonists on both baseline and atrial stretch-induced ANP and NT-ANP release in conscious rats. Volume load-induced ANP and NT-ANP release was reduced by the ETA receptor antagonist BQ-123 and the mixed ETA/B receptor antagonist bosentan. These results agree with the observation that passive immunization with ET-1 antiserum decreases volume load-induced ANP secretion in anesthetized rats (Fyhrquist et al. 1993). The effect of ET on ANP secretion seems to be mediated by ETA receptors since ETA receptor antagonists inhibit ET-stimulated secretion of ANP in cultured atrial myocytes (Irons et al. 1993, Leite et al. 1994, Thibault et al. 1994).
In addition to ET, vascular endothelium as well as endocardial cells have been shown to produce endothelium-derived relaxing factor. Several actions of NO on myocardial contractile performance have been reported, including mediation of cytokine (Finkel et al. 1992) and cholinergic responses (Balligand et al. 1993), modulation of β -adrenergic inotropic responses (Balligand et al. 1993) and changes in myocardial relaxation, force-frequency relationship and Frank-Starling response (Smith et al. 1991, Brady et al. 1993, Prendengast 1997). The role of NO in the regulation of stretch-induced cardiac hormone secretion is not clear. A potent inhibitor of NOS, L-NAME (Rees et al. 1990), has been recently reported to increase volume expansion-induced ANP secretion in conscious rats (Leskinen et al. 1995), suggesting that NO may negatively influence ANP secretion in vivo. In perfused rat atria, the inhibition of NO activity has been shown to restore the normal response to stretch in the presence of acetylcholine (Skvorak & Dietz 1997).