OUP user menu

Cardiomyocytes as effectors of nitric oxide signalling

Mike Seddon , Ajay M. Shah , Barbara Casadei
DOI: http://dx.doi.org/10.1016/j.cardiores.2007.04.031 315-326 First published online: 15 July 2007


Nitric oxide (NO) generated constitutively within the heart has long been known to influence myocardial function; however, the precise nature of these effects has been controversial – at least in part – because of the experimental use of non-isoform-selective inhibitors of NO synthases (NOS) and unwarranted extrapolation from results obtained with NO donors. Recent studies using NOS-selective inhibitors and genetically modified models are beginning to redress the balance. It is well established that agonist-stimulated release of NO from eNOS in the coronary endothelium exerts paracrine effects on cardiomyocytes, predominantly affecting the timing of relaxation as well as myocardial oxygen consumption. A significant recent advance has been the finding that both eNOS and nNOS are constitutively expressed in distinct subcellular locations within cardiomyocytes. The relative autocrine role of these isoforms in the cardiomyocyte remains to be fully clarified but evidence suggests that the autocrine effects of nNOS may include the modulation of basal inotropy and relaxation, β-adrenergic responsiveness, and the force-frequency relationship. Myocardial eNOS, on the other hand, may be involved in mediating the inotropic response to sustained stretch. These effects may change significantly in the diseased heart where the expression, activity and/or coupling of NOS isoforms to downstream effectors may be altered.

In this article, we review the current understanding of this important but complex field, focussing particularly on contractile function and on recent advances in knowledge regarding the autocrine functions of nNOS-derived NO.

  • Nitric oxide
  • Cardiac function
  • Cardiomyocyte
  • Calcium
  • Mice

Time for primary review 32 days

1 Introduction

NO is a ubiquitous intra- and inter-cellular signalling molecule generated by a family of NO synthases (NOSs), which catalyse the conversion of the amino acid l-arginine to l-citrulline in a reaction that requires O2 and cofactors. NO is involved a wide range of physiological and pathophysiological processes, with its best-established action in the cardiovascular system being the regulation of vascular homoeostasis. Over the last 15 years or more, it has been established that NO also has important effects on myocardial function which include the modulation of contractile function, energetics, substrate metabolism, cell growth and survival [1–3]. Whilst there is a general consensus that NO influences several aspects of myocardial function, the precise nature of these effects has been controversial. Part of this controversy almost certainly relates to the widespread use of various NO donors as experimental tools; it is now understood that these agents do not necessarily mimic the myocardial effects of endogenous NO, which is generated in a temporally and spatially restricted manner that is also dependent upon the stimulus for NO release. Secondly, the interpretation of studies undertaken in the intact heart or in multicellular preparations is complicated by the fact that several cardiac cell types generate NO and that observed effects on contractile function may in some cases reflect changes in coronary flow or autonomic transmission [4]. Thirdly, a large number of studies have been undertaken using non-selective inhibitors of NO synthases (NOS). Although initial interpretation of these studies was predicated on the assumption that the effects were attributable to the “endothelial” NOS (eNOS or NOS3), the mammalian myocardium has been found to express both eNOS and a “neuronal” NOS (nNOS or NOS1); eNOS is found in coronary and endocardial endothelial cells and cardiomyocytes [5] whereas nNOS has been localised to cardiac autonomic nerves and ganglia and cardiomyocytes [4,6]. A major recent advance has been the discovery that eNOS and nNOS are expressed in distinct subcellular compartments in the cardiomyocyte [5–7] where they are expected to couple to distinct effector molecules and to elicit quite different effects following enzyme activation. This is because the diffusion distance of NO within cardiac myocytes is likely to be limited to a local environment by both a high cytoplasmic concentration of myoglobin (which has a high affinity for NO and acts to scavenge it) and – particularly in disease states – by an abundance of superoxide anions, which can react with NO to limit its bioavailability. Therefore the local functional impact of NO may differ significantly depending on the NOS isoform involved, making the interpretation of experiments using non-isoform selective NOS inhibitors difficult and potentially misleading. Furthermore, expression, intracellular localisation, and activity of constitutive NOS isoforms may change in the presence of cardiac disease, leading to altered functional effects [8–10]. Finally, the high output iNOS isoform may also be expressed in many different cells (e.g., endothelial cells, cardiomyocytes, and inflammatory cells) in pathological settings [3]. In view of this complexity, elucidation of the physiological and pathophysiological effects of NO is likely to require a combination of approaches including single cell, whole heart and in vivo experiments, isoform-specific pharmacological probes and the use of state-of-the-art molecular approaches. In this article, we review some of the emerging data from such approaches, focussing mainly on the effects of eNOS and nNOS on myocardial function. The effects of iNOS-derived NO are beyond the scope of this article.

2 Paracrine effects of eNOS-derived NO

2.1 Basal myocardial function

Early studies in multicellular non-perfused preparations showed that stimulation of NO release from coronary microvascular endothelium affected the functional properties of adjacent cardiomyocytes [11]. The most significant effect was an earlier onset of relaxation. Subsequent studies that administered substance P, bradykinin, angiotensin converting enzyme (ACE) inhibitors or sodium nitroprusside in isolated whole hearts [12,13] and in human subjects studied during cardiac catheterization [14–16] showed a similar enhancement of left ventricular (LV) relaxation, independent of changes in coronary flow (Fig. 1A). These effects, which are mimicked by the application of NO donors or cGMP-analogues to isolated cardiomyocytes (Fig. 1B), have been attributed to a cGMP-dependent, protein kinase (PKG)-mediated phosphorylation of troponin I (TnI), leading to a reduction in myofilament Ca2+ sensitivity [17,18] (Fig. 1C). In humans, stimulation of eNOS activity by intra-coronary infusion of substance P was also associated with changes in the LV end-diastolic pressure–volume relationship suggestive of an acute reduction in LV stiffness [15] (Fig. 1A), which was analogous to the increase in myocyte diastolic length observed after the application of cGMP-analogues, PKG or NO donors [17]. Interestingly, these effects on relaxation and diastolic stiffness are only observed after stimulation of NO release from vascular eNOS since the inhibition of basal NO production with the non-selective NOS inhibitor L-NMMA has not been associated with changes in LV relaxation [19]. Similarly, studies in the hearts of eNOS−/− mice have not shown significant changes in myocyte relaxation [20] or LV diastolic function [21].

Fig. 1

A. Left ventricular pressure (LVP)-volume (LVV) loops obtained before (square) and after (circle) intracoronary infusion of substance P (20 pmol/min) in humans. Stimulation of NO release from the coronary endothelium by substance P infusion increases LV compliance and slightly decreases the end-systolic LV pressure. B. Application of the cGMP-analogue, 8-Bromo-cGMP, to isolated rat cardiomyocytes hastens relaxation and decreases cell shortening (lower panel) without affecting the amplitude and rate of decay of the intracellular Ca2+ transient (upper panel). C. Immunoblots showing that the troponin I phosphorylated fraction (TnI-P) is increased in rat hearts treated with the NO donor DEA-NO (10 μmol/L) compared with vehicle-treated hearts. Total troponin I (TnI) is not affected. Taken together these findings suggest that stimulation of NO release from the coronary endothelium may hasten myocardial relaxation and enhance LV compliance via a cGMP-dependent reduction in myofilament Ca2+ sensitivity. Modified from Paulus et al. [15], Shah et al. [17] and Layland et al. [18] respectively.

2.2 β-adrenergic responses

NO donors have been consistently shown to modulate β-adrenergic inotropic responsiveness in many models, with low doses reducing and high doses enhancing the response [3]. The mechanism underlying these effects may involve both cGMP-dependent and independent mechanisms [22]. However, the effects of endogenously generated NO on β-adrenergic responses have proved more difficult to unravel. Initial studies reported that non-selective NOS inhibitors enhanced the inotropic response to β-adrenergic stimulation both in vitro and in vivo, suggesting that constitutive NO release reduces β-adrenergic responsiveness; however, others failed to observe such effects (reviewed in [3]). A more specific approach has been to study eNOS−/− mice. Whereas studies in isolated hearts or in vivo in these animals demonstrated enhanced LV β-adrenergic inotropic response [21,23,24], all studies [23,25–27] but one [24] found no such effect in eNOS−/− cardiomyocytes or muscle preparations. Furthermore, no differences in the basal or isoproterenol-stimulated L-type Ca2+ current were found between eNOS−/− and control cardiomyocytes [23,25–27], suggesting that the chronic absence of eNOS-derived NO does not affect PKG or phosphodiesterase activity in the murine LV myocardium.

A possible explanation for the inconsistency between results in isolated cardiomyocytes and in whole heart/in vivo preparations may be that the predominant physiological effect of eNOS-derived NO on myocardial β-adrenergic responsiveness is paracrine and necessitates the production of NO from the coronary microvascular endothelium. Definitive proof of such a possibility would probably require studies in tissue-specific eNOS−/− mice. Although cardiomyocyte-specific adenoviral gene transfer of eNOS in eNOS−/− mice has been shown to normalize the in vivo LV inotropic response to β-adrenergic receptor stimulation [28], this experimental approach may not necessarily reproduce the physiological effects of constitutively expressed eNOS.

2.3 The Frank–Starling response

Prendergast et al. [29] reported that the inhibition of endogenous NO in isolated ejecting guinea pig hearts either with L-NMMA or with haemoglobin attenuated preload-induced increases in cardiac output in this preparation, independent of changes in coronary flow, suggesting that NO may be implicated in the positive inotropic response to increased preload, i.e., the Frank–Starling response (Fig. 2A). Neither the relevant cellular source of NO, the NOS isoform involved nor the mechanism of action of NO were directly addressed in this study, although the effectiveness of haemoglobin may suggest that a paracrine action of (endothelial eNOS-derived) NO was involved. This idea is supported by data from Pinsky et al. [30] who showed that the increase in myocardial NO production in response to mechanical stimuli in the rabbit heart ex vivo was abolished after denuding cardiac endothelial and endocardial cells (Fig. 2B–D). A potential subcellular mechanism responsible for this finding could be that NO-induced phosphorylation of TnI in its N-terminal [18] augments the positive inotropic effects of stretch through increased cross-bridge cycling [31], analogous to the recent data indicating that cardiac TnI phosphorylation in the N-terminal contributes significantly to the positive inotropic effects of isoproterenol, even though it reduces myofilament Ca2+ sensitivity [32].

Fig. 2

A. In isolated guinea-pig hearts, the increase in cardiac output in response to increases in preload (i.e., the Frank–Starling response, open circles) is decreased by NOS inhibition with L-NMMA (closed circles). † denotes a significant interaction between preload and L-NMMA vs. time-control group; * denotes a significant effect of L-NMMA vs. time-control group at equivalent preload. Modified from Prendergast et al. [29]. B. Intra-myocardial measurements of NO release in response to compression in the intact ex-vivo (non-beating) rabbit heart, C. after cardiac endothelial cells were partially denuded by intracoronary perfusion of Triton X-100, and D. after incubation with the NOS inhibitor, L-NMMA. The magnitude and time of application of the applied force are indicated by the vertical arrows. These data suggest that the heart releases NO in response to mechanical stimuli and that the source of this NO is likely to be the vascular endothelium. Modified from Pinsky et al. [30].

2.4 Myocardial oxygen consumption

Another paracrine action of endothelial eNOS-derived NO supported by several studies is a reversible inhibition of myocardial O2 consumption, independent of contractile function, which has been best characterized by Hintze and colleagues (reviewed in [33]). These authors have shown that the stimulation of NO release by agonists such as bradykinin and ACE inhibitors reduces mitochondrial respiration, possibly through inhibition of the electron transport chain at complexes I and II. It was also found that bradykinin-induced reduction in myocardial O2 consumption in vitro was abolished in eNOS−/− mice. Several studies have demonstrated changes in myocardial O2 consumption in vivo compatible with the above mechanism, using either inhibition of NOS or stimulation of NO release (e.g. [34,35]), although not all studies have confirmed these data [36]. These NO-dependent reductions in myocardial O2 consumption may serve to improve mechanical efficiency.

3 Autocrine effects of NO in cardiomyocytes: eNOS vs. nNOS

An important reason for apparently conflicting findings on the effects of endogenous NO on myocardial contractile function has been a lack of appreciation, until relatively recently, that cardiomyocytes express both eNOS and nNOS. The eNOS isoform is mostly localized to caveolae, a site where caveolin-3 is also localized and where several signal transduction pathways have been shown to be modulated by NO (reviewed in [37]). Myocardial nNOS was first localized to the sarcoplasmic reticulum (SR) [6], although later studies also found this isoform bound to sarcolemmal membrane proteins [7,38], especially in the diseased heart [8–10] (Fig. 3).

Fig. 3

LV lysates were immunoprecipated (IP) with antibodies raised against Caveolin 3 (Cav 3) (A), nNOS (B) or the ryanodine receptor Ca2+ release channels (RyR) (C) and immunoblotted (IB) for eNOS, Cav-3 and nNOS, respectively. In sham-operated rats (Sh), eNOS co-immunoprecipitated with Cav-3 (A) whereas nNOS appeared to interact preferentially with RyR (B–C). This pattern changed in failing rat hearts (HF), where a greater proportion of nNOS co-immunoprecipitate with Cav-3 (B). Both eNOS-Cav-3 and nNOS-RyR complexes appeared to be reduced in the presence of HF (A–C). Modified from Bendall et al. [10].

4 Autocrine effects of myocardial eNOS-derived NO

4.1 Basal and β-adrenergic myocardial function

As discussed in the previous section, most studies in isolated cardiomyocytes or muscle preparations from eNOS−/− mice found no difference in basal contraction or β-adrenergic response compared to control mice. Martin et al. [20] conducted a systematic investigation into the potential contribution of eNOS to the regulation of inotropy in isolated cardiomyocytes, under a wide range of experimental conditions. These experiments confirmed that eNOS gene deletion had no significant effect on basal or â-adrenergic cell shortening and relaxation; however, they also found that the inotropic response to â-adrenergic stimulation differed significantly depending on the age of animals and the choice of control mice (e.g., C57BL6/j mice vs. true littermate controls), suggesting that these factors may have contributed to previous inconsistencies in the literature. Other studies have reported that moderate myocyte-specific overexpression of eNOS leads to an inhibition of â-adrenergic responses [28,39,40], although this was not confirmed in mice with higher levels of cardiomyocyte-specific eNOS expression [41], where eNOS activity may be dysfunctional due to insufficient availability of the critical NOS co-factor, tetrahydrobiopterin [42]. Nevertheless, these studies suggest that pathophysiological conditions leading to increased expression of myocardial eNOS could result in inhibition of β-adrenergic responsiveness.

4.2 The Anrep effect

An alternative autocrine mechanism for stretch-induced augmentation of contractile function was suggested by Petroff et al. [43], who found that myocardial NO release contributed significantly to the increase in the frequency and amplitude of Ca2+ sparks and peak intracellular Ca2+ transient that subtend the positive inotropic response to sustained stretch (i.e., the Anrep effect). These authors provided evidence that this effect involved eNOS activation through Akt-mediated phosphorylation, was independent of cGMP, and was abolished in myocytes from eNOS−/− mice. This work implies that eNOS-derived NO may modify the ryanodine receptor Ca2+ release channels (RyR), leading to an increase in their open probability, in the absence of changes in SR Ca2+ content. As there is evidence indicating that an isolated effect on RyR open probability would only cause a transient increase in inotropy [44], it seems likely that eNOS-derived NO may target other proteins involved in excitation-contraction coupling under conditions of sustained stretch.

5 Autocrine effects of myocardial nNOS-derived NO

A major new paradigm in NO biology in the last few years has been the gradual realization that nNOS-derived NO may play an important role in the physiological regulation of myocardial contraction and Ca2+ fluxes. nNOS was initially found in the SR in both human and mouse myocardium, where it was initially found to inhibit Ca2+ uptake through the SR Ca2+ pump (SERCA2a) in SR microvesicles [6]. Subsequently, nNOS was shown to co-immunoprecipitate with the RyR [8–10,24]. From this location in the SR, nNOS-derived NO may influence both the open probability of RyRs [45] and L-type Ca2+ channels [46], which are situated in close proximity to RyRs at dyadic sarcolemmal-SR junctions. However, myocardial nNOS has also been localized to the sarcolemma [7,38], particularly in the LV myocardium of remodelled and failing hearts [8–10]. The subcellular localization of nNOS is dependent on interactions between its PDZ domain and scaffold adaptor proteins, which may include dystrophin, α-syntrophin [7] and caveolin-3 [47]. With regard to its regulation, it is of interest that phosphorylation of nNOS causes a reduction in enzymatic activity [48,49] rather than an increase, as shown for eNOS [50].

5.1 Basal and β-adrenergic myocardial function and Ca2+ fluxes

Ashley et al. [51] reported that field-stimulated (1–6 Hz) cardiomyocytes from nNOS−/− mice had a greater basal cell shortening and slower relaxation compared to wild-type littermates. Importantly, the nNOS inhibitor L-VNIO induced similar effects in wild-type cells but no further changes in nNOS−/− myocytes. This was reproduced in another study [46] which showed that the increased myocyte contraction was associated with a greater Ca2+ influx through L-type Ca2+ channels and a larger SR Ca2+ content and [Ca2+]i transient amplitude in nNOS−/− myocytes (Fig. 4A–C) or in wild-type myocytes exposed to selective nNOS inhibitors. The rate of decay of [Ca2+]i was found to be prolonged in nNOS−/− myocytes, suggesting that mechanisms responsible for Ca2+ reuptake or extrusion may account for the impaired myocardial relaxation observed in these animals.

Fig. 4

A. The current–voltage relationship shows that the Ca2+ current density is greater in nNOS−/−myocytes (filled squares) than in controls (open circles) over the voltage range −30 to +20 mV (P<0.05, n=16 and 21, respectively). * P<0.05. B. Average raw data trace showing the indo-1 fluorescence ratio (410/495 nm) in control (nNOS+/+) and nNOS−/−myocytes (n=17 and 19, respectively). Transients recorded from nNOS−/−myocytes had greater peak fluorescence and were slower to decay. C. The amplitude of cell shortening (elicited by a 200 ms depolarising step from −40 to 0 mV and expressed as a percentage of the resting cell length) was greater in LV myocytes from nNOS−/− mice compared with their wild type littermates (nNOS+/+). D. Similarly, the LV preload recruitable stroke work (PRSW, a chamber size and load-independent index of LV inotropy in vivo) was significantly increased in nNOS−/− mice. Taken together, these findings suggest that an increase in Ca2+ influx via the L-type Ca2+ channels, leading to an increase in the intracellular Ca2+, may be the main mechanism responsible for the increased myocardial inotropy in nNOS−/− mice. Modified from Sears et al. [46] and Dawson et al. [52].

Consistent with these findings, nNOS−/− mice also showed a modest increase in LV contractile function in vivo compared to wild-type littermates, whether assessed by echocardiography [46] or by LV pressure–volume analyses [52] (Fig. 4D) and a prolongation of the time constant of LV isovolumic relaxation, both under basal conditions and in the presence of β-adrenergic stimulation with dobutamine. These data are in partial agreement with Barouch et al. [24] who reported a small increase in basal LV contractile function in nNOS−/− mice in vivo compared with C57BL/6j mice. However, this group only found a non-significant trend towards an increase in basal contraction or in the L-type Ca2+ current in cardiomyocytes from nNOS−/− mice [24,53] and a suppressed lusitropic response to higher stimulation frequencies [53]. More recently, adenoviral-mediated overexpression of nNOS has been found to cause a reduction in Ca2+ current in sinoatrial node cells [54]; similarly, conditional, myocardial-specific nNOS overexpression has been associated with a decrease in Ca2+ current density, [Ca2+]i transient amplitude and cell shortening in isolated myocytes and in vivo [55].

These data suggest that under basal conditions nNOS-derived NO may exert an inhibitory effect on Ca2+ influx and myocardial contraction whilst promoting relaxation. Nevertheless, findings cannot be easily compared between groups due to significant differences in experimental conditions (e.g., choice of controls, level of nNOS expression, stimulation frequency and temperature).

nNOS may also modulate the inotropic response to β-adrenergic stimulation, although the mechanism responsible for this action remains controversial. For instance, the inotropic effect of low concentrations (<10 nmol/L) of isoproterenol was found to be enhanced in cardiomyocytes from nNOS−/− mice [24,51] or after acute nNOS inhibition [51]. However, with higher concentrations of isoproterenol, nNOS−/− cardiomyocytes showed a similar increase in cell shortening (vs. wild type littermates) in one study [20] and no further increase in inotropy in another (vs. C57BL/6j) [24]. In vivo, pharmacological β-adrenergic stimulation has consistently been shown to elicit a significantly smaller LV inotropic response in nNOS−/− mice compared with control mice [24,52].

However, it should be noted that the assessment of the role of myocardial nNOS in the regulation of LV function and β-adrenergic responsiveness in conventional nNOS−/− models in vivo may be confounded by the effects of nNOS gene deletion on central sympathetic outflow and on noradrenaline release from cardiac sympathetic nerves [4], both of which may indirectly influence basal and β-adrenergic LV function in the whole animal.

5.2 Intracellular signalling and targets of nNOS-derived NO

The intracellular signalling pathways involved in mediating the autocrine effects of nNOS-derived NO are still poorly understood. Generally speaking, it is known that several actions of NO are mediated through stimulation of soluble guanylate cyclase and elevation of cGMP levels, which in turn may regulate contraction through two main pathways (a) modulation of cAMP phosphodiesterases and cAMP levels, and (b) activation of the cGMP-dependent protein kinases (PKG) (reviewed in [22]). In addition, NO can interact directly with reactive thiols in many proteins, leading to post-translational modifications that induce significant functional changes. Notably, NO can increase the RyR open probability through this mechanism [43,45]. Similarly, S-nitrosylation of thiols on L-type Ca2+ channels has been shown to induce stimulatory or inhibitory effects on their function [56–58]. Evidence suggests that such effects could be preferentially mediated by nNOS-derived NO [59]; however, the relative contribution of cGMP-dependent effects and S-nytrosylation in mediating the myocardial effects of nNOS-derived NO remains to be evaluated.

The location of nNOS in the SR and/or sarcolemma suggests that ion channels and transporters involved in the regulation of Ca2+ cycling in the myocyte may be obvious targets for NO downstream signalling — e.g., SERCA2A, phospholamban, the RyR, the sarcolemma Ca2+ ATPase, and L-type channels. The documented nNOS-mediated inhibition of Ca2+ entry through L-type Ca2+ channels [46] is a likely mechanism through which nNOS-derived NO regulates contraction and intracellular Ca2+ stores, suggesting that NO produced by this isoform may exert a negative feedback regulation on Ca2+ influx (since increases in intracellular Ca2+ would stimulate nNOS synthesis of NO which in turn attenuates L-type Ca2+ current).

The reported effects of nNOS on SR function are more difficult to reconcile with the observed effects of nNOS inhibition or gene deletion on contraction and relaxation in intact myocytes. nNOS immunoprecipitates with the RyR [8–10,24] and exogenous NO has been shown to reversibly increase the open probability of purified RyR by S-nitrosylating free thiols [45,60]. If the opening of the RyRs were increased during diastole, the resulting leak of Ca2+ would be expected to lead to a lower SR Ca2+ content in the presence of nNOS (consistent with the higher SR Ca2+ content in nNOS−/− myocytes).

Interestingly, eNOS has also been reported to co-purify with RyR [61] and to increase RyR open probability and the amplitude of the intracellular Ca2+ transient under conditions of sustained myocardial stretch via a cGMP-independent mechanism [43].

Other data have suggested that nNOS-derived NO may inhibit SERCA2A activity [6], consistent with experimental findings showing an increase SR Ca2+ load in nNOS−/− myocytes [46]. However, an nNOS-induced reduction in SERCA2A activity would be expected to be associated with faster twitch relaxation and rate of decay of the intracellular Ca2+ transient in nNOS−/− myocytes, which is exactly the opposite of what has been reported [46,51–53]. The possibility that faster myocardial relaxation in the presence of nNOS might be due to reduced myofilament Ca2+ sensitivity (as previously reported for NO donors or cGMP/PKG in isolated rat myocytes [17,18]) was excluded by Sears et al. [46]. However, Khan et al. [62] showed that, in the presence of a high field-stimulation frequency (6 Hz), myofilament sensitivity might be decreased in cardiomyocytes from nNOS−/− mice; although this does not explain the absence of a lusitropic response to increasing stimulation frequencies they had previously reported in these animals [53].

An alternative explanation for the slower relaxation in the absence of nNOS may be provided by the observation that phospholamban (PLN) phosphorylation is reduced in the LV myocardium of nNOS−/− mice compared with their wild type littermates [63]. Although these findings could account for a slower rate of SR Ca2+ uptake in nNOS−/− myocytes [64], they would be expected to lead to a decreased contraction by decreasing SR Ca2+ stores — contrary to the experimental results. Taken together, these findings suggest that the observed effects of nNOS deletion or inhibition on contraction may reflect a dominant effect of increased Ca2+ influx via the L-type Ca2+ channel over the consequences of a reduction in SERCA2A activity. This would seem plausible given the parallel results of Brittsan et al. [65] who demonstrated impaired relaxation but a trend towards increased basal contraction in a transgenic model expressing a non-phosphorylatable mutant of PLN. In these animals, despite abolition of isoproterenol-induced effects on the rate of decline of intracellular Ca2+ transients, the overall inotropic response to isoproterenol was unchanged, due to a compensatory 25% increase in Ca2+ current density. Nevertheless, it seems likely that the net result of these contrasting influences of nNOS-derived NO on inotropy may change in pathophysiological conditions (e.g. in the presence of heart failure) where the balance between Ca2+ handling mechanism is altered or in humans where the inhibitory action of phospholamban on SERCA2A activity is much greater.

5.3 Myocardial redox state

A further layer of complication is added by the fact that the bioavailability of NO is markedly influenced by the free radical superoxide, which reacts avidly with NO and inactivates it. As such, any setting in which there is increased generation of superoxide may lead to a reduced NO bioavailability. Recent studies have suggested the presence of a specific interaction between nNOS and the activity of xanthine oxydoreductase (XOR) in the murine LV myocardium. In particular, Khan et al. [62] have shown that nNOS and XOR are co-localized in the SR of murine cardiomyocytes and that absence or inhibition of nNOS is associated with an increased XOR-dependent superoxide generation. This leads to the possibility that at least part of the cardiac phenotype of nNOS−/− mice may be a consequence of an increase in myocardial oxidative stress. Indeed, Kinugawa et al. [66] showed that the bradykinin-or carbachol-induced reduction in myocardial O2 consumption was attenuated in nNOS−/− mice. This finding was reversed upon application of superoxide scavengers or XOR inhibitors, suggesting that nNOS gene deletion may lead to a reduction in the bioactivity of eNOS-derived NO secondary to increased superoxide generation.

It should also be noted that the product of the reaction between NO and superoxide, i.e., peroxynitrite, has been implicated in cellular responses that span from a subtle modulation of cell signalling to cytotoxicity and cell death. In living cells, the biological chemistry of NO, superoxide, and peroxynitrite is complex and difficult to predict with implications for several disease states, which we are only beginning to unravel (reviewed in [67]). In particular, peroxynitrite oxidation of the NOS co-factor tetrahydrobiopterin appears to be a key mechanism underlying NOS ‘uncoupling’ both in the vascular endothelium [68] and, as shown more recently, in the myocardium [69,70]. In the presence of tetrahydrobiopterin deficiency NOSs synthesize superoxide rather than NO, thereby perpetrating a vicious circle leading to further oxidative damage and reduced NO bioavailability. The effects of this phenomenon on myocardial function remain largely unexplored.

5.4 The force–frequency relationship (FFR)

The myocardial positive FFR is largely due to an increase in the intracellular Ca2+ available for contraction at higher stimulation frequencies. Early studies suggested that endogenous NO may depress the myocardial positive FFR. For example, the positive FFR or shortening–frequency relationship was enhanced by non-selective NOS inhibitors in isolated hamster papillary muscles [71] and rat cardiac myocytes [72], respectively. However, a subsequent study in rat papillary muscle did not confirm these findings [73]. More recently, the FFR has been studied in nNOS−/− mice. Khan et al. [53] reported that the in vivo positive FFR was attenuated in nNOS−/− mice compared to C57BL/6j mice, suggesting that nNOS-derived NO may enhance the FFR. They hypothesized that an increased diastolic Ca2+ leak from RyR, rather than a change in SERCA2a activity, might explain the combination of the lower FFR and the reduced frequency-induced rise in SR Ca2+ content in nNOS−/− mice. However, it is difficult to reconcile this mechanism with their observation that the main difference between nNOS−/− and control myocytes occurred at the highest frequency, when diastolic time – and therefore any diastolic leak – should be less. Subsequently, the same authors have shown that inhibition of XOR with allopurinol normalises FFR in nNOS−/− myocytes, in the absence of changes in intracellular Ca2+ (Fig. 5), implying that a superoxide-mediated reduction in myofilament Ca2+ sensitivity may be the mechanism underlying the reduced FFR in nNOS−/− mice [62].

Fig. 5

Inhibition of xanthine oxidoreductase activity (XOR) with allopurinol (dashed lines in A, grey bars in B, and open symbols in C) restores the depressed force–frequency relationship in nNOS−/−cardiomyocytes (A, top panel and B, left panel) without affecting the amplitude of the intracellular Ca2+ transient (A, lower panel and B, right panel) or the intracellular Ca2+ stores (C, expressed as the % increase in Fura-2 fluorescence elicited by caffeine); C57BL/6j mice were used as wild type controls (WT). B *, P<0.05, nNOS−/− vs. WT and eNOS−/−, P<0.05 vs. nNOS−/− without allopurinol. C *, P<0.05 vs. 1 Hz; P<0.05, NOS1−/− vs. WT and eNOS−/−. These data suggest that production of reactive oxygen species by XOR in the nNOS−/− myocardium causes a reduction in myofilament Ca2+ sensitivity at 6 Hz; which accounts for the suppressed force–frequency relationship in these animals. eNOS gene deletion does not affect sarcomere length (SL) or intracellular Ca2+ under these conditions. Modified from Khan et al. [62].

Given the importance of the FFR as a homeostatic control mechanism in normal human physiology and the knowledge that an impaired FFR is a major component of the phenotype of the failing human heart, any potential role of endogenous NO in modulating the FFR in humans is of interest. Only one study has attempted to address this question to date. Cotton et al. [19] studied the effects of intracoronary L-NMMA on the response to atrial pacing in humans with normal LV function and patients with heart failure but found no effect on FFR in either group. However, this study clearly did not address a potential NOS isoform-specific effect on FFR.

6 Altered eNOS and nNOS function in heart failure

Although many published reports have suggested a relationship between altered eNOS expression/activity and contractile dysfunction in various cardiac pathological settings, definitive evidence linking eNOS with contractile dysfunction is limited (reviewed in [3]). In particular, selective gene deletion of eNOS has not produced consistent effects on LV remodelling after myocardial infarction in mice [74,75] whereas moderate myocardial specific eNOS overexpression has been shown to be beneficial [76]. The lack of eNOS-selective inhibitors has no doubt contributed to this paucity of information and to date it remains unclear whether constitutive myocardial eNOS activity plays a role in regulating myocardial function in remodelled or failing hearts. In contrast, recent studies have indicated that the effects of nNOS inhibition may differ between healthy and failing hearts. Myocardial nNOS expression and activity have been reported to be increased following experimental MI in rats [8,10] or mice [52], in human failing hearts [9] and in spontaneously hypertensive rats [77]. In these circumstances, increased nNOS expression was accompanied by a translocation from the SR to the sarcolemma, where nNOS associated with caveolin-3 (Fig. 3) and the regulatory protein HSP90. These findings were associated with a reduction in the level of RyR/nNOS complexes, suggesting that both activity and target proteins of nNOS-derived NO may differ in the remodelled myocardium.

Bendall et al. [10] have studied the effects of acute nNOS inhibition (using SMTC in vivo or L-VNIO ex vivo) on LV function in failing rat hearts post-MI. They found that, under basal conditions, nNOS inhibition increased basal LV inotropy and prolonged the time constant of LV isovolumic relaxation in sham-operated rat hearts whereas in failing hearts these effects were significantly blunted. In contrast, inhibition of nNOS enhanced the inotropic and lusitropic response to β-adrenergic stimulation in failing hearts but had no significant effect in sham-operated rats, suggesting that myocardial nNOS overexpression may contribute to the depressed β-adrenergic inotropic responsiveness observed in heart failure. Such an effect could be construed as a beneficial adaptive mechanism to protect the diseased heart from the harmful effects of excessive catecholamine stimulation. Consistent with an adaptive role for myocardial nNOS overexpression in the remodelled myocardium, Saraiva et al. [78] and Dawson et al. [52] found that nNOS−/− mice developed more severe LV remodelling (Fig. 6) and impaired β-adrenergic reserve after MI compared with control mice with similar infarct size. The subcellular mechanisms underlying these in vivo findings remain to be elucidated; however, Saraiva et al. [78] suggested that increased XOR activity in the myocardium of nNOS−/− mice may contribute to the maladaptive phenotype.

Fig. 6

nNOS gene deletion exacerbates adverse LV remodelling after myocardial infarction. LV volumes were evaluated by 3D-echocardiography in sham-operated and infarcted (MI) nNOS−/− mice (n=20) and wild type littermates (nNOS+/+, n=25) over 8 weeks of follow-up. ESVI and EDVI: end-systolic and end-diastolic LV volumes normalized for body weight. ESVI was slightly smaller in sham-operated nNOS−/−mice compared with their WT littermates (§ P<0.01, A). LV dilatation was significantly greater in infarcted nNOS−/− mice compared to their WT littermates, as indicated by their greater ESVI (*P<0.02) and EDVI (**P<0.05) (A–B). The relative increase over time in ESV and EDV in infarcted mice is shown in (C) and (D). Adverse LV remodelling was more accentuated in infarcted nNOS−/− mice at all time points (*P<0.01 for ESV and **P<0.05 for EDV) compared with infarcted WT mice with the same infarct size. Modified from Dawson et al. [52].

In the context of disease, it may also be important to consider possible NOS “uncoupling” and consequent generation of superoxide [69,70] as well as changes in the myocardial redox state due to increased superoxide generation by other sources, such as mitochondria and NADPH oxidases. In addition, it is possible that altered nNOS activity in the autonomic nervous system might impact on cardiac contractile function (reviewed in [4]).

7 Conclusions

It is now generally accepted that the constitutive NOS isoforms, eNOS and nNOS, subserve distinct functions in the myocardium as a consequence of their different cellular and subcellular localization, regulation and coupling to downstream targets (summarised in Fig. 7). eNOS expressed in coronary microvascular endothelial cells exerts paracrine effects on myocardial contraction, which include modulation of myocardial relaxation, oxygen consumption and probably β-adrenergic inotropic responsiveness. The effects of eNOS expressed within LV myocytes remain somewhat controversial; there is evidence implicating eNOS-derived NO in the increased inotropy elicited by sustained stretch and myocardial-specific overexpression of eNOS has been shown to play a role in inhibiting β-adrenergic inotropic responsiveness and post-MI LV dysfunction and remodelling.

Fig. 7

Paracrine and autocrine regulation of cardiomyocyte function by NO is dependent on the distinct subcellular locations of nNOS and eNOS. nNOS is preferentially localized to the SR. Evidence so far suggests that nNOS-derived NO may inhibit Ca2+ influx through the L-type Ca2+ channels and stimulate SR Ca2+ re-uptake by promoting PLN phosphorylation. An effect on RyR Ca2+ release channels has been hypothesised but not directly proven. nNOS-derived NO may also modulate the inotropic response to β-adrenergic stimulation (possibly in a bi-phasic fashion) and inhibit XOR activity, thereby limiting myocardial oxidative stress and, indirectly, increasing NO availability within the myocardium. eNOS exerts its main effects through the release of NO from nearby coronary microvascular endothelium. The paracrine effects of eNOS-derived NO are believed to be cGMP-dependent and include hastening relaxation and increasing myocardial distensibility (via a PKG-dependent reduction in myofilament Ca2+ sensitivity), inhibiting β-adrenergic inotropy, and reducing mitochondrial respiration and O2 consumption. eNOS is also found in cardiomyocytes (mostly in caveolae), where it is involved in mediating the positive inotropic response to sustained stretch by increasing the RyR open probability.

nNOS must now be considered the constitutive NOS isoform chiefly responsible for physiological NO-mediated autocrine regulation of cardiomyocyte contraction and relaxation (at least in rodents), mainly through modulation of excitation-contraction coupling. Emerging data also suggest that nNOS-derived NO may have important functions in the pathophysiology of adverse cardiac remodelling.

It should be noted that although eNOS and nNOS are known to be expressed in the human heart and in some cases to be dysregulated in disease, data on the functional relevance of these isoforms for human contractile function is extremely limited. This is especially important in light of the known differences in the contribution of different transporters and pumps to Ca2+ cycling between rodents and humans, and is therefore an important area for future study.


The authors' work is supported by the British Heart Foundation.


  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]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]
  71. [71]
  72. [72]
  73. [73]
  74. [74]
  75. [75]
  76. [76]
  77. [77]
  78. [78]
View Abstract