BJA Advance Access originally published online on May 14, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
British Journal of Anaesthesia, 2004, Vol. 93, No. 1 34-52
© 2004 The Board of Management and Trustees of the British Journal of Anaesthesia
Cellular mechanisms in sympatho-modulation of the heart
1 Institute of Anaesthesiology, University Hospital Zurich, Switzerland. 2 Institute of Pharmacology and Toxicology, University of Zurich, Switzerland
*Corresponding author: Cardiovascular Anaesthesia Laboratory, Institute of Anaesthesiology, University Hospital Zurich, Rämistrasse 100, 8091 Zurich, Switzerland. E-mail: michael.zaugg{at}usz.ch
 |
Abstract |
|---|
Top
Abstract
IntroductionCardiovascular function relies on complex servo-controlled regulation mechanisms that involve both fast-acting feedback responses and long-lasting adaptations affecting the gene expression. The adrenergic system, with its specific receptor subtypes and intracellular signalling cascades provides the major regulatory system, while the parasympathetic system plays a minor role. At the molecular level, Ca2+ acts as the general signal trigger for the majority of cell activities including contraction, metabolism and growth. During recent years, important new results have emerged allowing an integrated view of how the multifarious Ca2+-signalling mechanisms transmit adrenergic impulses to intracellular target sites. These insights into cellular and molecular mechanisms are pivotal in improving pharmacological control of the sympathetic responses to surgical trauma and perioperative stress. They are examined in detail in this review, with particular emphasis being given to the differences in intracellular signalling between cardiomyocytes and vascular smooth muscle cells.
Keywords: adrenergic receptor signalling; calcium, signalling; myocardial contractility; MAPK signalling; muscle, vascular smooth muscle regulation
|
Introduction |
|---|
Heart function, in terms of contractile force (inotropy), beating frequency (heart rate) and blood supply (vascular tone), relies on a three-tiered control system: (i) immediate and fast feedback responses of the cardiac tissue to the actual mechanical load; (ii) regulation of cardiac performance by the autonomous nervous system involving humoral primary messengers affecting the intracellular signalling systems; and (iii) long-term adaptation to altered physiological and pathological conditions produced by changes in gene expression. The first two modes of regulation are partially overlapping and primarily depend on the sympathetic nervous system. Repetitive peak responses, as they occur perioperatively, are part of the life-supporting adrenergic drive, which, however, may turn into potentially life-threatening maladaptation. Gaining control over sympathetic nervous system activity by blunting the adrenergic responses to the surgical trauma and perioperative stress is an important task in anaesthetic practice. The present review summarizes substantially new experimental results on adrenergic cellular and molecular mechanisms. Because of limitations of space, reviews will often be cited where further references to the primary literature can be found. Unless otherwise stated, the molecular characteristics pertain to the human protein species. Clinical aspects of the individual sympatho-modulatory therapies in perioperative medicine, based on these new experimental findings, will be presented in the article ‘Sympatho-modulatory therapies in perioperative medicine’ by Zaugg and Schaub in this issue.
|
Adrenergic receptor subtype-specific signalling via G-proteins |
|---|
Acute and long-term regulation of myocardial function, including heart rate, systolic and diastolic function and metabolism, are primarily governed by the ß1- and ß2-adrenergic receptor (AR) signalling pathways. A functional role for the
-AR is less well established, although positive and negative inotropic effects have been attributed to specific -AR subtypes, particularly in cardiomyopathies and heart failure.11 24 43 54 69 72 76 The human AR superfamily consists of nine subtypes originating from different genes (with single polypeptide chains varying in length from 408 to 572 amino acid residues): 1A, 1B, 1D, 2A, 2B, 2C, ß1, ß2 and ß3. All these cell surface receptors contain an extracellular N-terminus, seven transmembrane -helices (TM1–TM7) and an intracellular C-terminal region. They are all able to couple to the guanine nucleotide-binding G-proteins and are therefore called G-protein-coupled receptors (GPCRs). The extracellular N-terminal region, together with the extracellular loops between the TM2 and TM7 helices, contributes to the formation of the extracellular ligand binding pocket, whereas the amino acid sequences of the intracellular domains (loops) between the TMs, together with the proximal portion of the C-terminal region, are involved in mediating G-protein coupling. Activated G-proteins transmit signals to specific intracellular targets (Fig. 1). The downstream signalling pathways involve sequential protein phosphorylation cascades that ultimately affect targets in the cytoplasm or operate via transcriptional factors affecting gene expression. A second group of enzymes, the protein phosphatases, are responsible for dephosphorylation and termination of signalling. All intracellular signalling pathways are interconnected and form a robust network with ample redundancy. This complexity allows subtle modulation of individual cellular responses.
|
The GPCRs constitute the third-largest family of genes present in the human genome, and represent a central target structure for drug development. However, their therapeutic use is often limited by unwanted side-effects. Some of these derive from the binding of the drugs to related receptor subtypes.11 72 74 76 90 In the normal human myocardium, ß1-AR is the most abundant species (70–80%),
1 and ß2 account for over 10% each, and ß3 is low (around 1%). However, heart failure caused by either dilated or ischaemic cardiomyopathy is accompanied by selective down-regulation of ß1-AR, entailing a relative increase in both 1A and ß2, which together amount to 
50% of total ARs. In the failing heart, the ß3-AR protein concentration increases by a factor of 3.
ARs couple to the heterotrimeric G-protein complex (G, Gß, G
) on the inner side of the cell membrane.3 12 60 76 94 Upon activation, the G-subunit hydrolyses guanosine triphosphate (GTP) and dissociates from the complex, leaving the ß-subunits as undissociable heterodimer (Fig. 1). Currently there are 20 known G, six Gß and eleven G subunits. When activated, all three 1-ARs interact with the pertussis toxin-insensitive Gq component, which follows the main signalling route via phospholipase C (PLC) leading to diacylglycerol (DAG), producing activation of protein kinase C (PKC) and inositol trisphosphate (IP3) for the liberation of Ca2+ from the sarcoplasmic reticulum (SR). All three 2-ARs couple to the pertussis toxin-sensitive Gi, leading to inhibition of the integral membrane protein adenylyl cyclase (AC), activation of K+-channels, and inhibition of the sarcolemmal L-type Ca2+ entry channels (DHPR). This is in contrast to the ß-ARs, which are more variable in their coupling to G-proteins (Fig. 1). The ß1-AR couples almost exclusively to Gs, inducing positive inotropy (increased contractile amplitude) and positive lusitropy (enhanced relaxation). Its canonical signalling pathways involve activation of AC and protein kinase A (PKA). ß2- and ß3-ARs are both able to signal via Gs, Gi or Gq depending on the physiological or pathophysiological conditions (different states with regard to catecholamines, inflammatory cytokines and angiotensin-II). The Gs and Gq protein families have defined main effector pathways: the AC and the PLC pathways, respectively. The Gi protein family is more amorphous and its signalling flows equally through both the Gi and the Gß complex, affecting several different downstream signalling pathways. Gq-dependent signalling by the heterodimeric Gß to the phosphoinositol-3 kinase (PI3K) pathway was recently established in cardiac hypertrophy.20
AC is an integral membrane protein with 12 transmembrane helices and a molecular weight of 130 kDa.25 26 65 At least nine isoforms of the AC exist (AC1–AC9); AC5 is specifically expressed in cardiomyocytes, whereas AC6 occurs in heart cells other than myocytes. Additional isoforms are also expressed in the heart, but to a lesser degree. In addition to its regulation by G-proteins (stimulatory Gs and inhibitory Gi), PKC further stimulates and PKA inhibits the activity of AC isoforms. Furthermore, high Ca2+ during sustained cell activity inhibits AC5 and AC6, establishing a negative feedback loop. The fine tuning in the regulation of AC is of particular significance as its activity is rate-limiting in adrenergic signal transmission. PLC is a peripheral membrane protein at the cytoplasmic side, hydrolysing phosphatidylinositol 4,5-bisphosphate (PIP2) to DAG and IP3 (Figs 2 and 3). It has an absolute requirement for Ca2+ bound to the active site and comprises three types: PLCß (150 kDa), stimulated by G-proteins; PLC (150 kDa), stimulated by receptor tyrosine kinases; and PLC
(84 kDa), stimulated by transglutaminase-II.29 All three types are expressed in cardiomyocytes.
|
|
|
Regulation of adrenergic receptor signalling |
|---|
The ß2-AR is the most thoroughly studied with regard to signalling. Four paradigms have been delineated from studies of co-expression of wild-type and mutated ß2-AR with various combinations of G-proteins in different in vitro cell systems. One assumes that these regulatory mechanisms also apply to the other GPCRs in different tissues.
(i) Upon agonist activation, the ß2-AR couples to the G-protein heterotrimer, mainly with its third intracellular loop between the transmembrane helices TM5 and TM6 and part of the intracellular C-terminus immediately following the TM6. It couples preferentially to Gs forming a stoichiometric 1:1 complex. However, the ß2-AR may instead also couple to Gi or Gq, activating their corresponding signalling pathways (Fig. 1). The G-protein specificity is, at least in part, determined by the type of agonist.94 This implies that the agonist-induced conformation of the receptor favours the type of G-protein with which it interacts. For instance, salbutamol and ephedrine display a much higher efficacy with Gq than Gs or Gi. On the other hand, isoproterenol is more effective than salbutamol with Gi. It was recently shown that the selective ß2-AR antagonist ICI-118,551 exerts a direct negative inotropic effect by acting as a ß2-AR agonist, directing it away from coupling to Gs towards coupling to Gi.28 Furthermore, upon stimulation by catecholamines the ß2-AR interacts much faster with Gs than with Gi and Gq. This could generate intracellular signals in a timely, ordered fashion. In general, AR signalling is terminated by phosphorylation at multiple sites in the third intracellular loop and in the C-terminal region. However, PKA-dependent phosphorylation in the third loop (serines 261 and 262) and in the proximal C-terminal region (serines 345 and 346) of the ß2-AR switches its predominant coupling from the stimulatory Gs to the inhibitory Gi, thereby inhibiting the AC and promoting activation of the mitogen activated protein kinase (MAPK) signalling cascade (see next section).12 97
(ii) Signal termination is commonly referred to as desensitization, i.e. attenuation of receptor signalling despite the continued presence of a stimulus. Desensitization may occur under physiological or pharmacological stimuli as well as under pathological conditions.12 54 72 Homologous desensitization involves phosphorylation of two adjacent serines (355 and 356) in the C-terminal region of the ß2-AR by a family of G-protein-coupled receptor kinases (GRKs), whose activation does not require the production of second messengers. The six different GRK isoforms are tissue-specific. GRK2 and GRK5 (formerly referred to as ß-ARKs; 68 and 80 kDa respectively) are expressed in cardiomyocytes and display specificity for agonist-activated receptors; non-activated receptors or antagonist-bound receptors are usually not phosphorylated by GRKs. In contrast to ß1- and ß2-AR, the ß3-AR lacks phosphorylation sites and is refractory to desensitization by GRKs or PKA. The Gß subunits seem to play a role in recruiting the cytoplasmic GRKs to the membrane environment of their receptor substrates and the membrane phospholipids required for kinase activation.12 54 Heterologous desensitization occurs in the absence of agonist occupancy and is regulated by the signalling of another receptor via an intermediary second messenger. Induction of cAMP by receptor signalling leads to phosphorylation of several intracellular serines of the ß2-AR by cAMP-dependent PKA. Thus, phosphorylation by PKA turns off signalling through the receptor’s normal partner Gs and, at the same time, facilitates receptor coupling to the inhibitory Gi.12 21 97 Stimulation of PKC via the PLC–PIP2–DAG pathway also leads to phosphorylation of the receptor and its heterologous desensitization.
Desensitization is an acute response involving binding of arrestin to the phosphorylated C-terminus of the ß2-AR. Arrestin, together with the heterotetrameric adapter complex AP2, delivers the receptor to clathrin-coated pits for endocytosis to endosomes or to lysosomes. In the endosomes the ß2-AR is dephosphorylated by specific protein phosphatases. This resensitized receptor then recycles back to the cell membrane. However, fusion of the endosome with lysosomes leads to ß2-AR degradation. Desensitization is not always coupled to internalization but exhibits receptor type-specificity. For instance, under agonist stimulation 50–80% of the ß2-AR internalizes within a few minutes, whereas the ß1-AR does not internalize but remains at the cell surface, even in the desensitized state. The different -ARs internalize only moderately, except for the 2A-AR, which remains at the membrane. In contrast to desensitization, downregulation denotes a chronic process, during which agonist overstimulation promotes increased receptor degradation concomitant with a reduced de novo synthesis rate that does not match the loss of receptors.12 71
(iii) Recent reports on homo- and heterodimerization between receptor subtypes suggest a potential concentration of receptor complexity that could account for previously unexpected pharmacological diversity.3 2-AR and ß2-AR may form homodimers. The ß2-AR was found to form dimers on agonist activation, and the agonist-induced homodimer seems to represent the active ß2-AR species. Dimerization is established between the sixth and seventh transmembrane domains of the two receptors involved.78 Heterodimers have been found between the 2A-AR and ß1-AR, the ß1-AR and ß2-AR, and the 2C and M3-muscarinic receptors, as well as between the ß2-AR and -opioid receptors.3 Formation of the ß1-AR and ß2-AR heterodimer inhibits the agonist-promoted internalization of the ß2-AR and its ability to activate the MAPK cascade46 (see next section).
(iv) A GPCR-associated protein may directly mediate signalling, as in the case of the G-proteins themselves. Alternatively, a GPCR-associated protein may regulate receptor signalling by controlling receptor localization and/or trafficking, e.g. by internalization. Finally, a GPCR-associated protein may act as a scaffold, physically linking the receptor to various effectors. Scaffold proteins are defined as proteins that associate with two or more partners to enhance the efficiency and/or specificity of cellular signalling pathways. The family of PKA-anchoring proteins (AKAPs) was one of the first to be recognized as scaffold proteins.30 85 87 Two AKAPs, AKAP250, also known as gravin, and AKAP79, were found to interact with the C-terminus of the ß2-AR.
|
Relation of adrenergic to global cardiomyocyte signalling |
|---|
The G-protein-stimulated signal pathways continue via distinct classes of protein serine–threonine kinases (PKA, PKB, PKC, PKG), which specifically phosphorylate serine and threonine residues in their target proteins.13 60 61 76 85 87 99 PKA and PKC are the key players in the adrenergic signal relay system. In the absence of cAMP, PKA is an enzymatically inactive tetrameric holoenzyme consisting of two catalytic subunits (PKAC monomer, 40.4 kDa) bound to a regulatory subunit dimer (PKAR monomer, 42.5 kDa). cAMP binds cooperatively to two sites on each PKAR subunit. Upon binding of four molecules of cAMP, the enzyme dissociates into a PKAR dimer with four molecules of cAMP bound, and two free, active PKACs. The PKC exists in a variety of isoforms with functional specificities. There are 10 mammalian PKC isoforms, ranging in molecular weight from 68 to 102 kDa. The four conventional PKCs (
, ß1 with a splice variant ß2, and ) require Ca2+ and DAG for activation; the four novel PKCs (, 
, 
/L and 
) are Ca2+-independent but require DAG for activation; finally, the two atypical PKCs (
and 
) lack both the Ca2+ and the DAG binding domain.
G-proteins are also able to activate the monomeric GTPase Ras via different phosphorelay systems. Ras represents a master switch in transferring extracellular signals for growth and differentiation via the four MAPK cascades to the nucleus (Fig. 2).33 37 53 55 64 Among these four cascades, (i) ERKs (extracellular signal-regulated kinases) are activated by growth factors and regulate cardiac hypertrophy and apoptosis, (ii) the more recently characterized big MAP kinase-1 (BMK1, also called ERK5) pathway transmits oxidative stress signals to the cell nucleus, (iii) the four p38 MAPK isoforms (, ß, and ) are activated by cytokines and environmental stress and are also involved in regulation of apoptosis, and (iv) the JNKs (c-Jun N-terminal kinases) are critical regulators of transcription. Specificity of AR subtype signalling may be achieved by proteins that do not have intrinsic catalytic activity but serve as adapter and anchoring proteins. By keeping the reaction partners of a particular signalling pathway in close proximity to the effector site, they form so-called signalling modules. Such a signalling module of activated PKC with ERK has been demonstrated to be operative in mitochondria, where it induces cardioprotection.5 Another recent example is the conserved sequential MAPK cascade Raf–MEK–ERK, in which the two upstream kinases of the module, Raf and MEK, remain cytoplasmic. In resting cells, ERK is anchored to MEK, whereas upon activation it rapidly detaches and translocates to the nucleus.73 A similarly complex signalling network with redundancies, as revealed between the downstream signalling pathways of ARs (Fig. 1), can also be discerned among the MAPK cascades (Fig. 2).
|
Calcium as a signal transmitter |
|---|
Calcium is highly toxic for any cell, yet it represents the major intracellular messenger regulating most activities, including contractility, metabolism, transport, secretion and transcription. The Ca2+ signal varies in its character depending on the circumstances. One may see a long-lasting cytoplasmic increase in Ca2+ concentration activating metabolism as well as gene expression, or short-lasting (50–200 ms) Ca2+ transients triggering contraction.7 8 14 Calcium signalling for different cellular responses and its consequent requirement for energy production depend on the strictly maintained intracellular Ca2+ homeostasis, which is largely monitored by adrenergic signalling pathways (Fig. 3). The heart makes up less than 0.5% of the body mass, yet it consumes around 11% of the total energy expenditure. The complex scheme in Fig. 3 comprises the explicitly reported interconnections between the intracellular adrenergic and Ca2+ signalling pathways, but in reality probably many more may exist. In the resting myocyte the cytoplasmic Ca2+ concentration is below 10–7 M and increases transiently by almost two orders of magnitude upon stimulation by the action potential (AP). The Ca2+ concentration in the mitochondrial matrix follows closely that of the cytoplasm, involving the regulation of metabolic key enzymes of the tricarboxylic acid cycle.14 99 Several ATP-consuming ion pumps and electrogenically driven ion exchangers are responsible for ensuring a sufficiently low cytoplasmic Ca2+ concentration during diastole. In diastole, the cytoplasmic Ca2+ concentration must be significantly lower than what is found in ultrapure bidistilled water after passage through an ion exchange resin, as it is used in the laboratory. Any cell damage which impairs the tight control over the cytoplasmic Ca2+ concentration leads to abnormal energy metabolism in the mitochondria and ends in cell death.
In contrast to the low cytoplasmic Ca2+ during rest, the Ca2+ concentration outside the myocyte ranges between 1 and 2 mM, thus establishing a gradient of over 10 000-fold. This outside–inside gradient enables the Ca2+ ion to function as a signal, provided there are sufficiently sensitive Ca2+-binding signalling components in the cell.27 Inside the cardiomyocytes, the Ca2+ signal is conveyed to the target sites by the Ca2+-sensing proteins troponin-C (TnC) and calmodulin (CaM), which reversibly bind Ca2+ ions with affinities in the range of 10–5 to 10–6 M. These Ca2+ affinities are just in the range attained by Ca2+ signalling transients, taking into account the intracellular presence of around 1 mM Mg2+ ions, which are competing for the Ca2+ binding sites with an affinity 104 times lower than that of Ca2+. Some Ca2+-regulated targets have their own Ca2+ binding sites and are not dependent on TnC or CaM as an intermediate signal component.
Calcium enters the cardiomyocyte down the concentration gradient via the high voltage-activated (opening at around –20 mV) L-type Ca2+-channels of the sarcolemma (particularly accumulated in the transverse T-tubules) as often as an AP stimulates the cell. The low voltage-activated (opening between –60 and –40 mV) T-type channels are not concentrated in the T-tubules and contribute little to the Ca2+ entry from outside.6 18 34 41 They are primarily found in secretory and smooth muscle cells and in the cardiac nodal cells, where they are involved in rhythm control. The L- and T-type channels belong to a family of cell surface Ca2+-channels composed of four subunits in a 1:1 stoichiometry (1, 2, ß, ). The human 1C (Cav1.2) subunit (2221 amino acids, 249 kDa), occurring in heart (Cav1.2a) and smooth muscle (Cav1.2b sharing 93% homology), contains four domains with six transmembrane helices each (S1–S6, yielding a total of 12 transmembrane helices). The S6s of each domain together form the Ca2+ pore. The four transmembrane S4 helices of 19 amino acids each contain positively charged residues at every third position, together forming the voltage sensor of the pore. The P-loop between the transmembrane helices S5 and S6 is very much conserved among the different Ca2+-channels and provides the filter selectivity for Ca2+. The ß-subunit associates with the -subunit at the inner side of the membrane and determines the kinetics of the channel activities (opening, closing, inactivation). The ß2-subunit (73.5 kDa) is characteristic of the heart while the ß3-subunit (54.5 kDa) is more typical of smooth muscle. The Cav1.2 channel activity is enhanced by phosphorylation by PKA, PKC and CaMK, but is inhibited by Ca2+ when its cytoplasmic concentration is increased during sustained cell activity (negative feedback control).34 39 41 93 Calcium binds directly to the C-terminal intracellular part of the Cav1.2, which contains a Ca2+-binding EF-hand domain (see below), but nearby is also a binding site for CaM, which contributes to sensing Ca2+ signalling. For prolonged elevation of cytoplasmic Ca2+, ill-defined ligand-gated channels (also called store-operated channels) are thought to be involved in Ca2+ entry after the Ca2+ release from internal stores.14 22
For excitation–contraction coupling, Ca2+ is primarily released from intracellular stores in response to extracellular stimuli (Fig. 3). The endoplasmic reticulum (ER) in non-muscle cells and its derivative, the SR in striated muscles (cardiac and skeletal) as well as in smooth muscle, are able to accumulate Ca2+ up to a concentration of 30 mM and to store it bound to proteins with multiple low-affinity Ca2+-binding sites, such as calsequestrin and calreticulin. Calsequestrin (46.4 kDa) is found in cardiac and skeletal SR while calreticulin (45.0 kDa) is mainly present in the SR of smooth muscle and ER of non-muscle cells. Both the SR and ER contain two ligand-gated Ca2+ release channels, the ryanodine binding receptor (RYR) and the IP3 binding receptor (IP3R). The Ca2+ entering the cell on stimulation induces a far larger (5- to 20-fold) release of Ca2+ from the closely positioned intracellular SR into the cytoplasm via the RYR, a process called Ca2+-induced Ca2+ release (CICR).7 14 16
The RYR channel forms a tetramer with four equal subunits of 565 kDa each, which combine with four regulatory proteins, called FKBP12.6. FKBP, with a mass of 12.6 kDa belongs to the group of cyclophilins, which accelerate protein folding, acting as peptidyl-prolyl cis-trans isomerases or rotamases. They are inhibited by the immunosuppressor drugs FK506 (tacrolimus) and rapamycin (sirolimus), but not by cyclosporin A. Of the three isoforms, RYR1 (with a total molecular weight of around 2300 kDa) occurs in skeletal muscle, RYR2 in cardiomyocytes and RYR3 in non-muscle cells.49 88 In addition, the RYR2 serves as scaffold protein, combining with numerous key regulatory components in the junctional SR complex, including CaM, PKA, type-1 and type-2 phosphatases and, at the luminal SR surface, triadin and calsequestrin. Calcium can also be released from the SR via the IP3R channel, which is composed of four equal subunits of 313 kDa each. The IP3R channel also exists in three isoforms, with IP3R2 in cardiomyocytes. The IP3R2 is activated by IP3 produced by PLC. The rate and extent of Ca2+ liberation by IP3 is much lower than for CICR via the RYR2, and thus hardly contributes to excitation–contraction coupling in cardiomyocytes. However, intracellular Ca2+ release by IP3 is important in the slow motion of smooth muscle contraction and in fine-tuning the activity of atrial myocytes, where the SR has more IP3R2 than in the ventricular myocytes. RYR2 and IP3R2 share structural and functional similarities and have some sequence similarity in their C-terminal domains, although the latter is about half the size of the former.77 The two Ca2+ release channel types interact and are inhibited by high Ca2+ and CaM, but become activated by phosphorylation by PKA, PKC or a Ca2+-CaM-dependent protein kinase-II (CaMK-II) (Fig. 3). Phosphorylation of RYR2 induces dissociation of the FKBP regulatory protein, which inhibits the RYR2 channel when bound to it.
|
Intracellular calcium handling in cardiomyocytes |
|---|
Cardiomyocyte contraction and cell activity in general are terminated by removing Ca2+ from its regulatory sites on TnC, CaM and other proteins. This is achieved by quickly and efficiently lowering the cytoplasmic Ca2+ concentration.8 14 27 45 Two proteins are involved in moving Ca2+ out of the myocyte, the sarcolemmal Ca2+ pump (PMCA) and the Na+/Ca2+-exchanger (or antiporter, NCX1), and one in taking it up into the intracellular store (the Ca2+ pump of the SR, SERCA2a) (Fig. 3). The immediate removal of Ca2+ from the regulatory sites is effected by the SERCA2a pump, which has high affinity but low transport capacity. The NCX1 extrudes Ca2+ out of the cell during diastole with relatively low affinity but high transport capacity, thus ensuring a sufficiently low Ca2+ concentration during cell rest. While the NCX1 couples the outward transport of one Ca2+ to the entry of three Na+ down its outside–inside gradient, the SR and the sarcolemmal Ca2+ pumps depend on ATP consumption. The outside–inside Na+ gradient (established by the Na+/K+ pump, which also uses ATP) allows the electrogenically driven exchange of Ca2+ against Na+ by the NCX1. Stimulation of the Na+/K+ pump by phosphorylation by PKA and/or PKC is under the control of the ARs, in particular the ß1-AR (Figs 2 and 3). In mice and rats, about 92% of the Ca2+ ions are pumped back into the SR by the SERCA2a and only 7% are extruded out of the cell by the NCX1. In larger mammals, such as rabbits, cats, dogs and humans, SERCA2a and NCX1 are able to lower the Ca2+ concentration to about 70 and 28% respectively. The slow sarcolemmal PMCA pump contributes little (1–2%) to the process of lowering the cytoplasmic Ca2+ concentrations in the heart, whereas its function is decisive for Ca2+ homeostasis in smooth and non-muscle cells.
The amount of AP-induced Ca2+ entry is primarily dictated by the duration of the AP and the open probability of the L-type Ca2+-channel. The Ca2+ extruded after the heart-beat must match the amount of Ca2+ that enters just before the beat, otherwise the cell would not be in steady state but would either lose or gain Ca2+. This provides a quantitative framework for the dynamic Ca2+ fluxes in the cardiomyocytes. The SERCA2a pump is one of the main players in terminating contraction and restoring resting cytoplasmic Ca2+ concentrations. It is known that during heart failure in humans, as well as in a rabbit model, the functional expression of SERCA2a is reduced and NCX1 is increased.8 32 This results in a loss of SR Ca2+ concomitant with enhanced Ca2+ extrusion, leading to a net loss of intracellular Ca2+. Consequently, less Ca2+ is available in the SR for the subsequent heart-beats, which is the central cause of systolic contractile deficit in heart failure.
The NCX1 (splice variant NCX1.1) is the cardiac isoform with 938 amino acids (104 kDa), whereas the NCX2 is preferentially expressed in the brain and NCX3 in skeletal muscle.62 70 82 100 Its topology is not yet clear, suggesting either 11 or nine transmembrane helices. Interestingly, the NCX1 may function in reverse mode during the plateau of the AP (Ca2+ influx coupled with Na+ outflux). This may be of physiological significance as the NCX1, like the L-type Ca2+-channel, is concentrated in the transverse T-tubular system and may be able to elicit Ca2+-induced Ca2+ release via the RYR2 from the SR. This would reinforce the Ca2+ release process from the RYR2 ascribed to the interaction of the L-type Ca2+-channel with the RYR2 (as mentioned above). The NCX1 can be phosphorylated and stimulated by both PKA and PKC, and is therefore under the control of ARs. PMCA and SERCA belong to a subfamily of the P-type ATPases. The sarcolemmal Ca2+ pump (PMCA) transports one Ca2+ per ATP out of the cell, while SERCA2a takes two Ca2+ per ATP up into the SR. From the isoforms of the four genes giving rise to PMCA1-PMCA4 and over 20 splice variants, PMCA1c (1249 amino acids, 138 kDa, with 10 putative transmembrane helices) seems to be the main cardiac species, but other isoforms are also expressed in cardiomyocytes. PMCA activity depends on binding CaM, and may be further stimulated by phosphorylation via PKA and/or PKC (Fig. 3).
The cardiac SR Ca2+ pump SERCA2a, together with its regulatory protein phospholamban (PLN), is the most important component linking adrenergic control to inotropy and rhythmicity.4 8 14 22 48 Three genes were identified for the SR Ca2+ pump, SERCA1, SERCA2 and SERCA3, which are spliced into several isoforms. SERCA1a is mainly expressed in fast skeletal muscle, while SERCA1b is abundant in fetal and neonatal tissues. The SERCA2 gene encodes four splice variants: SERCA2a, expressed in the heart and in slow skeletal muscle, SERCA2b, expressed in smooth muscle and with variants (types 2 and 3) in non-muscle cells and (type 4) in neuronal cells. The recently solved crystal structure of the SERCA1a pump in the Ca2+-bound state reveals (besides the 10 transmembrane helices) three large cytoplasmic domain structures constituting the nucleotide binding site, the catalytic site and the phosphorylation site.91 The cardiac SERCA2a is 997 amino acids long (110 kDa) and is under the direct control of a CaM-dependent kinase-II (CaMK-II), which enhances its transport capacity by phosphorylation of Ser38. The major regulator of SERCA2a, however, is PLN, with 52 amino acids (6.1 kDa) and one transmembrane helix (C-terminal amino acids 32–52). It is predominantly expressed in ventricular cardiac muscle, but also in small amounts in slow-twitch skeletal muscle, smooth muscle and endothelial cells. As a monomer, it associates with and efficiently inhibits the SERCA2a pump, by interaction of its transmembrane helix with helices of the pump and its cytoplasmic domain with the cytoplasmic domain of the pump. Its inhibitory effect is delicately regulated by phosphorylation induced by different signalling pathways. The cytoplasmic domain is amenable to regulation by phosphorylation of Ser16 by PKA, Thr17 by CaMK-II, and Ser10 by PKC (Fig. 3). Phosphorylation to various degrees causes gradual dissociation of PLN from SERCA2a, relieving the inhibition of the Ca2+ pump and thus increasing the rate of relaxation (lusitropic effect). When not associated with SERCA2a, PLN polymerizes into pentamers.
To further increase the complexity, sarcolipin (SLN), which is a homologue of PLN with 31 amino acids (3.8 kDa), forming just one transmembrane helix and lacking most of the cytoplasmic portion present in PLN, was discovered recently.4 48 It is highly expressed in human fast-twitch skeletal muscle and in small amounts predominantly in atrial muscle. SLN inhibits the SERCA1a (fast skeletal muscle) and SERCA2a (cardiomyocytes) by lowering the Ca2+-binding affinity and slowing the ATPase turnover rate. In addition, SLN is able to induce a superinhibitory effect apparently by binding to PLN and thus preventing PLN from polymerizing into pentamers. A small amount of SLN may be sufficiently potent to shift the equilibrium of pentameric PLN towards the monomer, inhibiting SERCA2a. In view of the decisive power of the SR-SERCA2a pump in regulating contractility, both SERCA2a and PLN represent potential targets for new therapeutic approaches.
|
Adrenergic fine-tuning of contractility |
|---|
TnC and CaM are the main transmitters of the intracellular Ca2+ signal. Both belong to the EF-hand protein superfamily, which contains 66 distinct subfamilies with a total of almost 600 known proteins.59 The characteristic EF-hand, with 30 amino acids, consists of a flexible loop (eight amino acids) with five residues, providing oxygen for reversibly complexing with Ca2+, which is flanked on both sides by a short
-helix of approximately 11 amino acids. Reversible binding of Ca2+ to the loop induces a conformational change that moves the two rigid helices relative to one another, and this movement is transmitted to the associated target proteins. Both TnC and CaM contain four such Ca2+-binding domains, which have arisen from a common single Ca2+ binding ancestor protein by several steps of gene duplication and fusion. The expression of TnC is restricted to striated muscles (skeletal and cardiac), where it regulates contraction and is located in the heterotrimeric troponin complex on the actin filament in the sarcomeres. This complex consists, besides TnC, of troponin-I (inhibitory component, TnI) and troponin-T (tropomyosin binding component, TnT).51 66 67 TnC contains either two ions or two ions bound to the two C-terminal binding sites for protein stabilization, whereas in the fast skeletal muscle isoform (159 amino acids, 18.0 kDa) the two N-terminal sites reversibly bind signalling Ca2+.27 59 79 In the cardiac TnC (TnC) isoform (161 amino acids, 18.4 kDa) only one of the two N-terminal binding sites is functional for regulation. Binding of Ca2+ to the regulatory site of cTnC modulates its interaction with TnI and, as a consequence, it also affects the interaction of TnI with TnT and tropomyosin. In the absence of Ca2+, TnI inhibits the interaction of myosin-II with actin (relaxation). Binding of Ca2+ to cTnC relieves the inhibitory effect of TnI, and contraction ensues. The troponin complex is positioned along the actin filament at every seventh actin monomer, and the Ca2+-induced conformational changes of TnC are imparted to all actin monomers via TnI and TnT (around 32 kDa) acting through the two tropomyosin threads that run on both sides along the entire actin filament (Fig. 4). Tropomyosin is an elongated dimeric protein about 40 nm long, comprising two -helical subunits (284 amino acids each, 32.8 kDa) arranged in parallel orientation. The tropomyosin molecules are strung together head-to-tail, thus forming threads along the actin filaments with a length of 600–1000 nm. Cardiac TnI (210 amino acids, 24.0 kDa) is somewhat larger than its skeletal muscle counterparts and includes a unique 32-residue N-terminal extension with two adjacent residues, Ser23 and Ser24, that can be phosphorylated by either PKA or PKC. TnI phosphorylation under the control of distinct AR subtypes (Fig. 3) lowers the affinity of TnC for calcium ions. Calcium then more readily dissociates from TnC at the end of systole, thus enhancing relaxation (lusitropic effect).
|
In contrast to TnC, CaM (148 amino acids, 16.7 kDa) is an ubiquitously expressed Ca2+ signalling protein found in all eukaryotic cells. Its amino acid sequence is identical in the human, rat, mouse, rabbit, chicken, frog, sea urchin and torpedo fish.59 77 This makes it one of the most highly conserved proteins known. With four Ca2+ bound, it regulates over 40 enzymes, channels and structural proteins. In the so-called apo form, the Ca2+ binding sites are empty or may be occupied by Mg2+. It operates by (i) reversible binding to the target protein when Ca2+ is bound, (ii) remaining combined with some proteins in both apo and Ca2+-bound form, as in the heteromeric phosphorylase complex, or (iii) forming part of the target protein structure as heterochimera, the gene encoding EF-hands being fused to the gene of the target protein. While Ca2+-CaM interacts with PMCA, L-type Ca2+-channel and RYR2 (as mentioned above) for direct regulation, most of its regulatory activities are mediated by CaM-dependent kinases (CaMKs) and subsequent phosphorylation reactions (Fig. 3). CaMKs are activated by phosphorylation, like most protein kinases, either by autophosphorylation triggered by binding of CaM or by another CaM-dependent kinase, called CaMK kinase (CaMKK). The role of this CaMKK is reminiscent of that of MAP kinase kinase (MAPKK) in the MAP kinase cascade (Fig. 2).38 44 Several types of CaMK are multisubstrate kinases that phosphorylate different target proteins, while others are specific for one particular protein; for instance, the myosin light chain kinase (MLCK). The multisubstrate CaMK-II is a multimeric enzyme composed of up to 12 subunits, whereas the skeletal and cardiac muscle MLCK (65 kDa) is monomeric. The MLCK phosphorylates specifically the regulatory light-chain subunit (RLC) of cardiac myosin-II at Ser14, rendering the contractile apparatus more sensitive to the Ca2+ trigger (positive inotropy). The same can also be achieved by RLC phosphorylation by PKC (Fig. 3).
|
Frank–Starling and negative feedback mechanisms |
|---|
Physiological contraction generates both isometric force (ventricular pressure) and rapid shortening to eject the blood. As discussed above, there are two main ways to change the strength of contraction: by altering the amplitude and/or duration of the Ca2+ transient, and by altering the sensitivity of the contractile filaments to Ca2+ (by phosphorylation of the myosin RLC and/or TnI). Calcium-sensitivity also increases by mechanical stretching as the heart fills with blood, resulting in stronger subsequent contraction.23 This is due to the compression of the transverse filament lattice that occurs on stretching and brings the contractile filaments in the sarcomeres closer together. This facilitates actin–myosin interaction and increases the affinity of TnC for Ca2+. It represents an important autoregulatory mechanism by which the heart adjusts to increased diastolic filling, and is called the Frank–Starling response.
In general, the positive inotropic and lusitropic effects are mainly elicited by ß1- and ß2-AR stimulation and mediated through cAMP-dependent PKA phosphorylation of PLN, TnI, the L-type Ca2+-channel and RYR2. The degree of positive inotropy is directly related to the amount of cellular Ca2+, which steadily increases with time under these conditions. Without safety precautions, it would end in cardiac arrest in systolic contracture. This is, however, normally not observed (except with digitalis glycoside overdoses), as the cardiac phosphodiesterase (PDE) isoforms degrade cAMP, thus keeping the overall ß1-AR stimulation at bay. At least four types of PDE isoforms (PDE1–PDE4, varying in size from 61 to 124 kDa), with over 20 splice variants, most of them membrane-associated, have been identified in cardiomyocytes.25 35 52 83 Via cyclic nucleotide metabolism, PDEs are also involved in the regulation of the L-type Ca2+-channels. The activity of PDE1 depends on the association of the four Ca2+ bound form of CaM, which increases concomitantly with the increase in cytoplasmic Ca2+; PDE2 is stimulated by cAMP; PDE3 is inhibited by cAMP; and PDE4 is insensitive to cAMP. The PDEs are operative as homodimers and contribute decisively to the contractile responsiveness. The fact that PDE activity reduces inotropy by degradation of cAMP prompted the development of PDE inhibitors, such as amrinone, milrinone and enoximone (bipyridines inhibit PDE enzymes, as do methylxanthines), which are in clinical use for the acute treatment of congestive heart failure. Despite its complexity, the PDE system provides a powerful negative feedback control system of great physiological significance (Fig. 3).
In conclusion, although Ca2+ is traditionally described as a second messenger that is liberated from intracellular stores (SR), Ca2+ entering the cell may activate a number of processes acting directly as a first messenger.14 In particular, it amplifies its own signalling capacity by the Ca2+-induced Ca2+ release from the SR via the RYR2, thus also acting as a second messenger. Furthermore, the second messenger IP3, liberated by PLC from PIP2 (as mentioned above), provokes the release of Ca2+ via the IP3R of the SR, acting then as a third messenger. In addition, adrenergic signalling pathways exert their effects (balancing between positive or negative inotropy and lusitropy) almost exclusively by modulating cytoplasmic Ca2+ concentrations and signalling transients in amplitude and frequency, thus adding yet another step to the Ca2+ signalling cascade. Consequently, Ca2+ may well be viewed as operating at the same time as a first, second and third messenger. In the healthy heart, the combined signalling circuits coordinate contractility and energy production in a concerted way. The cellular and molecular basis explains how adrenergic stimulation induces positive inotropy together with positive lusitropy (faster relaxation) under increased haemodynamic load. This combination of stronger but shorter contraction twitches makes it possible to accommodate more beats per time interval, thus increasing cardiac output under increased workload. However, upon transition from compensated haemodynamics to overt heart failure or during ischaemia/reperfusion injuries, the integrated signalling network may become unbalanced and dysfunctional, eventually ending in collapse.32 49 71 79 99
|
Adrenergic regulation of heart rate |
|---|
Maintaining adequate perfusion of vital organs is the main task of the myocardium. Adaptation to increased workload is effected by coordination of two principal mechanisms, which both are under sympathetic and parasympathetic control: (i) positive inotropy as described above, and (ii) increase in heart rate. Recently a voltage-gated cation channel family was cloned comprising four members: HCN1–HCN4, in the 89–129 kDa range.9 40 56 HCN channels open upon hyperpolarization and close at positive potentials. cAMP and cGMP increase channel activity by shifting the activation curve of the channels to more positive voltages. The stimulatory effect of cyclic nucleotides is not dependent on protein phosphorylation but is due to a direct interaction with the HCN channel protein. HCN channels contain six transmembrane helices (S1–S6) and are believed to assemble in tetramers. The S4 segment is positively charged and serves as voltage sensor. The C-terminal region of all HCN channels contains a cyclic nucleotide binding domain that confers regulation. The channels are activated during membrane hyperpolarization after the termination of an action potential and provide an inward Na+ current that slowly depolarizes the plasma membrane. Sodium permeates the channels five times more readily than K+. HCN channels are found in neurons and heart cells. HCN4 represents the predominant species in the sino-atrial node (SAN) controlling heart rate and rhythm. HCN2 is the most abundant neuronal channel and is found almost ubiquitously in the brain. Sympathetic stimulation of the SAN raises cAMP concentrations, thus accelerating diastolic depolarization and heart rate. Stimulation of muscarinic acetylcholine receptors slows the heart rate by the opposite action (Fig. 1). Given the key role of HCN channels in cardiac pacemaking, these channels represent promising pharmacological targets for the development of novel drugs for the treatment of cardiac arrhythmias and ischaemic heart disease.
|
Regulation of vascular smooth muscle contraction |
|---|