- Email: [email protected]
Molecular architecture of the human specialised atrioventricular conduction axis
Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Contents lists available at ScienceDirect
Journal of Molecular and Cellular Cardiology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y j m c c
Original article
Molecular architecture of the human specialised atrioventricular conduction axis I.D. Greener a,1, O. Monfredi a,1, S. Inada a, N.J. Chandler a, J.O. Tellez a, A. Atkinson a, M.-A. Taube a, R. Billeter b, R.H. Anderson a, I.R. Efimov c, P. Molenaar d, D.C. Sigg e, V. Sharma e, M.R. Boyett a,2, H. Dobrzynski a,⁎,2 a
University of Manchester, UK University of Nottingham, UK Washington University, St Louis, USA d Queensland University of Technology & University of Queensland, Australia e Medtronic Inc., USA b c
a r t i c l e
i n f o
Article history: Received 28 August 2010 Received in revised form 16 December 2010 Accepted 17 December 2010 Available online 21 January 2011 Keywords: Atrioventricular node Ion channels Gap junctions Arrhythmias
a b s t r a c t The atrioventricular conduction axis, located in the septal component of the atrioventricular junctions, is arguably the most complex structure in the heart. It fulfils a multitude of functions, including the introduction of a delay between atrial and ventricular systole and backup pacemaking. Like any other multifunctional tissue, complexity is a key feature of this specialised tissue in the heart, and this complexity is both anatomical and electrophysiological, with the two being inextricably linked. We used quantitative PCR, histology and immunohistochemistry to analyse the axis from six human subjects. mRNAs for ~ 50 ion and gap junction channels, Ca2+-handling proteins and markers were measured in the atrial muscle (AM), a transitional area (TA), inferior nodal extension (INE), compact node (CN), penetrating bundle (PB) and ventricular muscle (VM). When compared to the AM, we found a lower expression of Nav1.5, Kir2.1, Cx43 and ANP mRNAs in the CN for example, but a higher expression of HCN1, HCN4, Cav1.3, Cav3.1, Kir3.4, Cx40 and Tbx3 mRNAs. Expression of some related proteins was in agreement with the expression of the corresponding mRNAs. There is a complex and heterogeneous pattern of expression of ion and gap junction channels and Ca2+handling proteins in the human atrioventricular conduction axis that explains the function of this crucial pathway. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction The atrioventricular conduction axis (or atrioventricular node) is located in the septal component of the atrioventricular junctions, and is arguably the most complex structure in the heart. It has a multitude of functions, such as providing a critical delay between atrial and ventricular systole, protecting against life-threatening rapid supraventricular rates, and providing contingent pacemaking in the event of failure of the sinus node [1]. It is also the site of one of the most common cardiac arrhythmias, namely atrioventricular nodal reentrant tachycardia. Like any other multifunctional tissue, complexity is one of its key features. This complexity is both structural and functional. Histological analysis of the specialised muscular axis responsible for conduction of the cardiac impulse across the plane of atrioventricular insulation reveals various components common to all mammals, including the human [2].
⁎ Corresponding author at: Cardiovascular Medicine, Faculty of Medical and Human Sciences, University of Manchester, Core Technology Facility, 46 Grafton Street, Manchester M13 9NT, UK. Tel.: + 44 161 275 1182. E-mail address: [email protected] (H. Dobrzynski). 1 Joint first authors. 2 Joint senior authors. 0022-2828/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.yjmcc.2010.12.017
Inferiorly, between the coronary sinus and tricuspid valve annulus, is found the inferior nodal extension (INE). The INE connects to the compact node (CN). The CN occupies the apex of the triangle of Koch, becoming the penetrating bundle (PB), or bundle of His, when it penetrates into the central fibrous body. Also connecting to the CN is a transitional area (TA) made up of transitional atrial myocytes. Electrophysiological data relating to the function of the axis has largely been derived from rabbit preparations. Amongst the most important findings in terms of function is the presence of dual inputs to the axis [3,4]. The fast pathway is located antero-superiorly, whereas the slow pathway is located between the orifice of the coronary sinus and the attachment of the septal leaflet of the tricuspid valve. The fast pathway is thought to correspond to the TA and the slow pathway to the INE [5,6]. In the rabbit, three cell types are widely accepted as making up the atrioventricular conduction axis. As compared to atrial myocytes with a fast Na+-dependent action potential upstroke and negative resting potential, N (typical nodal) cells have a slow Ca2+-dependent upstroke and a more positive diastolic potential, whereas the transitional AN and NH cells have an intermediate upstroke and diastolic potential [7]. N cells are thought to make up the INE (slow pathway) and CN, the transitional AN cells are thought to make up the TA (fast pathway), and the transitional NH cells are thought to be present in the PB [7,8].
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Electrophysiological data regarding the human atrioventricular conduction axis have largely been restricted to electrode recordings during physiological testing and modification of arrhythmic substrates. Clearly, fast and slow pathways exist in humans [9]. More recently, Efimov and colleagues have provided further insights into the electrophysiology of the axis in the human with optical action potential recordings [10]. Due to limited access to live human tissue, there is no information regarding specific ionic currents or cell–cell coupling properties in the human. To remedy this deficiency, we have undertaken a molecular analysis of transcripts and proteins governing electrical excitability, hoping to provide further insight into the behaviour of the human atrioventricular conduction axis. 2. Materials and methods We obtained specimens including the area of the triangle of Koch and the adjacent membranous part of the septum from six diseased human hearts acquired from the Prince Charles Hospital District, Chermside, Australia (ethics approval, EC2565; work also ethically approved by University of Manchester). A clinical profile of the patients is shown in the Data Supplement. We used quantitative PCR (qPCR), histology and immunohistochemistry to investigate the molecular make-up of the components of the atrioventricular conduction axis. For qPCR, we sampled the components of the axis, along with the atrial muscle (AM) and ventricular muscle (VM), as outlined in Fig. 1A, from 30 to 50 (60 μm thick) frozen sections, which underwent a novel and rapid haematoxylin and eosin (H&E) staining protocol [11]. In brief, from multiple quick H&E stained sections, tissue from each region (TA, INE, CN, PB) were micro-dissected by hand under a dissecting microscope with a fine surgical needle. Each region was identified by its histological appearance. The identification was guided by Masson's trichrome stained sections as well as sections immunolabelled for Cx43
643
such as those in Fig. 1. Total RNA was extracted from a given region from all relevant tissue sections. This was repeated for all six hearts. Following extraction of total RNA from the samples and reverse transcription to produce cDNA, the amounts of ~50 cDNAs for selected cell type-markers, connexins, ion channels, Ca2+-handling proteins and receptors were measured using qPCR. The data were analysed using the ‘ΔCT technique’: to correct for variations in the amount of total RNA in the different samples, the amounts of the different cDNAs measured were normalised to the amount of cDNA for a housekeeper (ribosomal protein, 28S). The amount of 28S cDNA was similar in each tissue investigated (data not shown). For a given ion channel etc., the amount of mRNA is proportional to the amount of cDNA and, therefore, this method allows the abundance of the same mRNA in different tissues to be compared. Tentative conclusions can also be drawn about the relative amounts of different mRNAs from the relative amounts of the corresponding cDNAs. However, the conclusions must be tentative, because the efficiency of the reverse transcription step (to produce cDNA) varies for different mRNAs. If the difference between two cDNAs was greater than 10-fold, the difference is likely to reflect to a difference in the abundance of the two corresponding mRNAs (rather than difference in efficiency of reverse transcription). However, if the difference between two cDNAs was less than 10-fold, the result must be interpreted cautiously. Data for individual transcripts are plotted here, whereas the data for groups of related transcripts are plotted in the Data Supplement (Figs. S1–S7) to facilitate comparison. All transcripts in each sample were measured in triplicate. In general, the three measurements were close. The average of these three measurements was then taken and the mean ± SEM of the average values for each heart was calculated – the n number for the calculation is, therefore, the number of hearts. In figures, the mean± SEM is shown; in addition, the individual averages are shown as open circles. The SEM (as well as the open circles) shows the variability between hearts. Significant
Fig. 1. Histological and immunohistochemical characteristics of human atrioventricular conduction axis. A, Masson's trichrome stained sections through INE (left), CN (middle) and PB (right). Myocytes stained purple, connective tissue blue. Dotted lines highlight areas (INE, CN and PB) sampled for qPCR. B, high magnification images of triple immunolabelling of Cx43 (green signal; at gap junctions), caveolin3 (red signal; within cell membrane) and vimentin (blue signal; within fibroblasts) in AM (left), INE, CN and PB. CFB, central fibrous body.
644
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
differences in the abundance of mRNAs in different tissues were identified using one-way ANOVA. A difference was assumed to be significant when P b 0.05. The data are shown in an alternative format in Table 1 (see also Fig. S8 in the Data Supplement), which shows mean values of the relative abundance of all mRNAs normalised to that in the AM. A more detailed description of the methods can be found in the Data Supplement. A list of the principal abbreviations used is in the Glossary. 3. Results 3.1. Morphological characteristics of components of atrioventricular conduction axis Staining with Masson's trichrome technique of sections obtained through the triangle of Koch and the membranous septum revealed the location of the components of the specialised muscular axis responsible for atrioventricular conduction (Fig. 1A). With this technique, the
Table 1 Relative abundance (mean value) of mRNA normalised to mRNA abundance in AM (%).
ADBR1 ADBR2 ANP Ca2+v1.2 Ca2+v1.3 Ca2+v3.1 Ca2+v3.3 Cx31.9 Cx40 Cx43 Cx45 DPP6 FREQ HCN1 HCN2 HCN3 HCN4 ERG IP3R1 KchAP KchIP2 Kir2.1 Kir2.2 Kir2.3 Kir3.1 Kir3.4 Kv1.4 Kv1.5 Kv4.2 Kv4.3 KvB1 KvB2 KvLQT1 minK Na+v1.1 Na+v1.3 Na+v1.4 Nav1.5 Nav1.6 Nav1.7 NavB1 NavB2 NavB3 NavB4 NCX1 PLN RyR2 RyR3 SERCA2a Tbx3
AM
TA
INE
CN
PB
VM
100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100
89 113 30 84 356 210 95 103 54 64 83 103 129 164 70 92 261 87 85 94 80 37 57 106 93 79 38 69 285 122 107 78 87 118 138 147 152 66 92 159 83 55 105 129 95 58 89 269 115 175
59 38 3 79 649 282 15 Not investigated 49 21 60 286 106 187 86 366 480 82 32 65 80 22 59 32 20 129 63 22 256 52 83 48 33 85 784 111 51 17 156 135 58 68 33 37 35 0 26 302 60 263
77 108 4 105 1762 559 110 92 174 21 118 103 607 527 142 86 863 164 113 164 144 44 141 131 178 414 40 61 1044 98 74 70 121 191 151 165 42 49 195 181 69 112 103 53 173 61 32 116 48 498
35 93 14 56 547 318 87 104 250 53 61 1949 688 51 67 107 505 172 96 142 88 91 50 112 68 370 651 41 643 140 46 82 78 102 435 82 52 126 188 101 76 113 114 64 126 99 82 118 54 320
63 70 4 112 18 8 58 106 27 112 88 Not investigated 74 7 73 83 84 77 101 79 140 337 3 253 6 66 345 39 106 85 152 166 91 133 213 118 43 116 118 77 59 141 153 94 115 120 110 82 66 73
cardiac myocytes are stained purple, and the connective tissue is stained blue. Adjacent tissue sections were immunolabelled for Cx43 (major connexin in working myocardium), caveolin3 (membrane-bound protein expressed by all cardiac myocytes) and vimentin (expressed by fibroblasts). As shown in Fig. 1B, whereas there was labelling for caveolin3 in all cardiac myocytes, labelling for Cx43 was absent from the INE and CN, but present in the AM, PB and VM. Labelling of vimentin was primarily observed in the INE and the CN. In small mammals, Cx43 has been shown to be absent from these components of the axis, but abundantly expressed in the working myocardium [e.g., 8]. Our immunofluorescence analysis revealed this also to be the case in the human (Fig. 1B). 3.2. Nodal markers ANP (atrial natriuretic peptide), a hormone expressed by atrial muscle [e.g., 8,11,12], and Tbx3, a transcription factor expressed in the cardiac conduction system [8,11–13], were used to characterise the tissues studied by qPCR. qPCR showed ANP mRNA to be highly abundant in the AM, but poorly expressed in the INE, CN, PB and VM (Fig. 2). An intermediate level of expression of ANP mRNA was noted in the TA (Fig. 2). Tbx3 mRNA was most abundant in the CN and it was poorly expressed in the AM and VM (Fig. 2). The level of Tbx3 in the working myocardium is presumably a basal level of this transcription factor. In our previous work on the human sinus node [11], we also detected a ‘basal level’ of Tbx3 in the working myocardium. There was an intermediate level of Tbx3 expression in the TA, INE and PB (Fig. 2). The significantly lower expression of ANP mRNA, but higher expression of Tbx3 mRNA, demonstrated by qPCR in the components of the atrioventricular conduction axis compared to the working myocardium confirms that, during dissection, there was little or no contamination of these tissues with working myocardium. 3.3. Connexins Cx31.9, Cx40, Cx43 and Cx45 form ~ 9, ~ 200, ~ 80 and ~ 30 pS gap junction channels, respectively. The levels of Cx40, Cx43 and Cx45 cDNAs were generally N10-fold greater than that of Cx31.9 (Fig. 2). There was significantly higher expression of Cx40 mRNA in the CN and PB compared to the other tissues (Fig. 2). Immunohistochemistry showed that Cx40 protein was also highly abundant in the CN and AM, but absent from the VM (Fig. 3). As expected, immunohistochemistry showed that Cx40 protein was more abundant in the PB than in the working myocardium (Fig. S9, Data Supplement). There was significantly lower expression of Cx43 mRNA in the INE, CN and PB as compared to the working myocardium (Fig. 2). There was an intermediate level of expression of Cx43 mRNA in the TA. Cx43 protein was distributed in a similar manner in the different tissues (Fig. 1B). The levels of Cx31.9 and Cx45 mRNAs were similar in each of the six tissues (Fig. 2). However, the amount of Cx31.9 mRNA in comparison with other connexin mRNAs was negligible in the human heart (Fig. S2), and this is in agreement with the study of Kreuzberg et al. at the protein level [14]. 3.4. HCN channels The HCN channels (HCN1–HCN4) are responsible for the hyperpolarization-activated or ‘funny’ current, If [15]. It is well known that If plays a key role in pacemaker activity in the heart including the human sinus node [15,16]. In the CN, for example, HCN4 cDNA was more abundant than HCN2 cDNA (but not more than N10fold more abundant), N10-fold more abundant than HCN1 cDNA, which was N100-fold more abundant than HCN3 cDNA (Fig. 2). There was a significantly higher expression of HCN1 mRNA in the CN compared to the PB and VM (Fig. 2). There was also a significantly higher expression of HCN4 mRNA in the INE, CN and PB compared to
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
645
Fig. 2. Relative abundance of mRNA for cell type-markers, connexins and HCN channel subunits. Means + SEM (n = 6) shown. a–fsignificantly different from appropriately lettered bars (one-way ANOVA). Individual data for all six subjects shown by open circles.
the working myocardium (Fig. 2). Because HCN4 was the most abundant isoform, it was investigated at the protein level using immunohistochemistry. Fig. 3 shows that HCN4 protein was expressed in the sarcolemma of the CN cells, although the signal was below the detection threshold in the working myocardium. The level of HCN2 and HCN3 mRNAs was similar in most of the six tissues (Fig. 2). 3.5. Na+ channels Voltage-gated Na+ channels are responsible for the inward Na+ current, INa. INa is responsible for the rapid action potential upstroke in the working myocardium. The cardiac Na+ channel, Nav1.5, is primarily responsible for INa. There was a significantly lower expression of Nav1.5 mRNA in the INE and CN as compared to the AM, PB and VM (Fig. 4). Immunohistochemistry also showed that, although Nav1.5 protein was detectable in the working myocardium, it was not detectable in the CN (Fig. 3). In small mammals, other Na+ channels have also been shown to be expressed in cardiac myocytes [17]. These Na+ channels (Nav1.1, Nav1.3, Nav1.4, Nav1.6 and Nav1.7) at the mRNA level are also expressed in human cardiac tissues (Fig. 4; Nav1.2 not studied as we have previously shown it to be poorly expressed in human heart [11]). However, their cDNA levels were N10-fold lower than that of Nav1.5 cDNA (Fig. 4), suggesting that these Na+ channels may play a minor role in human. The levels of the mRNAs were similar in each of the six tissues (Fig. 4). Of the Na+ channel β-subunits (Navβ1–Navβ4), Navβ1 was the most abundant β-subunit (at cDNA level at least) and the levels of the mRNAs were again similar in each of the six tissues (Fig. 4). 3.6. Ca2+ channels Two separate Ca2+ currents have been recorded from cardiac myocytes, the L-type (ICa,L) and the T-type (ICa,T) [18]. Cav1.2 and Cav1.3
are responsible for ICa,L and Cav3.1–Cav3.3 are responsible for ICa,T. All channels except Cav3.2 were studied since previously we have shown Cav3.2 to be undetectable in human heart) [11]. In the working myocardium, Cav1.2 cDNA was N10-fold more abundant than Cav1.3 cDNA (Fig. 5), whereas in the CN they were of comparable abundance (Fig. 5). The expression of Cav1.2 mRNA was similar in each of the six tissues (Fig. 5). On the other hand, there was a significantly higher expression of Cav1.3 mRNA in the INE, CN and PB compared to the working myocardium (Fig. 5). Cav3.1 cDNA was generally N10-fold more abundant than Cav3.3 cDNA (Fig. 5). Cav3.1 mRNA was distributed in a similar manner to Cav1.3 mRNA (Fig. 5). Consistent with this, Cav3.1 protein was detected in the CN, but not in the AM (Fig. 3). Cav3.3 mRNA did not vary significantly among the tissues (Fig. 5).
3.7. Ca2+-handling proteins Intracellular Ca2+ has been suggested to play an important role in pacemaking in small mammals [19]. The expression level of the Na+– Ca2+ exchanger, NCX1, at the mRNA level, was similar in most tissues, however it was significantly more abundant in the CN compared to the INE (Fig. 5; highest NCX1 data point does not account for this and when removed from statistical analysis, the difference remained significant). Consistent with this, immunohistochemistry showed that NCX1 protein was expressed both in the PB and working myocardium (Fig. S9, Data Supplement). mRNA for the sarcoplasmic reticulum (SR) Ca2+ pump, SERCA2a, did not vary significantly among the tissues (Fig. 5). RyR2 and RyR3 are SR Ca2+ release channels; RyR2 cDNA was N10-fold more abundant than RyR3 cDNA (Fig. 5). RyR2 mRNA tended to be less abundant in the INE and CN (Fig. 5). Consistent with this, immunohistochemistry showed that RyR2 protein was detectable in the working myocardium, but not in the CN (Fig. S9, Data Supplement). RyR3 mRNA was significantly more abundant in the INE as compared to the VM
646
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Fig. 3. Expression of Cx40, caveolin3, HCN4, Nav1.5 and Cav3.1 proteins in CN and working myocardium. High magnification images of immunolabelling of Cx40 and caveolin3 (top; Cx40, green signal; caveolin3, red signal), HCN4 (middle top; green signal), Nav1.5 (middle bottom; red signal) and Cav3.1 (bottom; green signal) in CN (left) and AM/VM (right) shown. Scale bar for each panel in shown in the bottom right panel.
(Fig. 5). IP3 receptor (IP3R1) mRNA tended to be less abundant in the INE (Fig. 5).
3.8. Transient outward K+ channels The transient outward K+ current (Ito) is responsible for the early phase of repolarization (phase 1) of the action potential. The ion channels responsible for Ito are the voltage-dependent K+ channels, Kv1.4, Kv4.2 and Kv4.3. Kv4.3 is regarded as the major ion channel underlying Ito in the human and although Kv4.3 cDNA was more abundant than Kv1.4 and Kv4.2 cDNAs, it was not necessarily N10-fold greater, especially in the CN (Fig. 6). Kv1.4 mRNA was significantly more abundant in the CN as compared to the AM, TA and INE and Kv4.2 mRNA was significantly more abundant in the CN as compared to all other tissues (Fig. 6). The level of Kv4.3 mRNA was similar in all tissues (although expression in PB was significantly greater than in INE; Fig. 6).
3.9. Delayed rectifier K+ channels There are three delayed rectifier K+ currents, ultra-rapid (IK,ur), rapid (IK,r) and slow (IK,s). The ion channels responsible for IK,ur, IK,r and IK,s are Kv1.5, ERG and KvLQT1, respectively. ERG cDNA N KvLQT1 cDNA N Kv1.5 cDNA (but not necessarily by N10-fold; Fig. 6). The expression level of Kv1.5 mRNA was significantly higher in the AM compared to the INE, PB and VM while the level of ERG mRNA was similar in each of the six tissues, whereas there was significantly lower expression of KvLQT1 mRNA in the INE as compared to all other tissues (Fig. 6). 3.10. Voltage-gated K+ channel regulatory proteins We have also studied accessory proteins for voltage-gated K+ channels. FREQ (frequenin; Ca2+-binding protein known to regulate Kv4 channels), DPP6 (dipeptidyl aminopeptidase-like protein 6; proposed auxiliary subunit of Kv4 channels) and KChAP (chaperone for Kv4.3)
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Fig. 4. Relative abundance of mRNA for Na+ channel subunits. See Fig. 2 legend for details.
Fig. 5. Relative abundance of mRNA for Ca2+ channel subunits and Ca2+-handling proteins. See Fig. 2 legend for details.
647
648
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Fig. 6. Relative abundance of mRNA for voltage-gated K+ channel subunits. See Fig. 2 legend for details.
regulate transient outward K+ channels. Both FREQ and DPP6 mRNAs were significantly more abundant in the CN and PB compared to the VM (similar to distribution of Kv1.4 and Kv4.2; Fig. 6). KChAP mRNA tended to be distributed in a similar manner (Fig. 6). Together with KvLQT1, minK is responsible for IK,s and was distributed in a similar manner to KvLQT1 mRNA (Fig. 6). Three β-subunits are thought to regulate delayed rectifier K+ channels, Kvβ1–Kvβ3; we studied Kvβ1 and Kvβ2 mRNAs as previously we have shown Kvβ3 to be poorly expressed in human heart [11]. Kvβ1 and Kvβ2 mRNAs did not vary among the tissues studied (Fig. 6). 3.11. Inward rectifier K+ channels The background inward rectifier K+ current, IK1, is generated by the ion channels, Kir2.1–Kir2.4. mRNA levels for Kir2.1–Kir2.3 only were measured since previously we have shown Kir2.4 to be poorly expressed in human heart [11]. In the VM, Kir2.1 cDNA N Kir2.3 cDNA N K ir 2.2 cDNA, whereas in the AM, K ir 2.2 cDNA N K ir 2.1 cDNA N Kir2.3 cDNA (but not necessarily by N10-fold; Fig. 7). The level of Kir2.1 mRNA was significantly higher in the VM compared to all other tissues and it tended to be lower in the TA, INE and CN than in the AM (Fig. 7). The level of Kir2.2 mRNA was similar in most tissues (although it was significantly higher in AM and CN compared to VM; Fig. 7). Kir2.3 mRNA did not vary significantly among the tissues (Fig. 7). The ACh-activated K+ current, IKACh, is generated by a heteromultimer of Kir3.1 and Kir3.4 [20]. Kir3.1 and Kir3.4 mRNAs were similarly abundant as expected (Fig. 7). The level of Kir3.1 mRNA was significantly higher in the AM, TA, CN and PB compared to the VM, whereas the level of Kir3.4 mRNA was significantly higher in the CN and PB compared to the TA and VM (Fig. 7).
3.12. Adrenergic receptors The atrioventricular conduction axis is under rigorous autonomic regulation. mRNAs for β1 (ADβR1) and β2 (ADβR2) receptors were detectable in all tissues investigated. The levels of the two receptors were similar (Fig. 7). β1 receptor (ADβR1) mRNA tended to be lower in the PB, whereas β2 receptor (ADβR2) mRNA tended to be lower in the INE (Fig. 7; see Discussion). 4. Discussion We have shown a complex portrait of transcripts, and whenever possible proteins, underlying the electrical excitability of the specialised muscular axis responsible for atrioventricular conduction in the human heart. Our data provide novel insights into the electromolecular properties of the atrioventricular conduction axis, in particular within the major sub-compartments of this complex structure. 4.1. Pacemaking Various factors are responsible for pacemaking in the heart (for example in sinus node), but two important ones are the absence of IK,1 and the presence of the pacemaker current, If. In the rabbit at least, the working myocardium has a resting potential generated by IK,1 (although density of IK,1 in AM is less than in VM, despite resting potentials being similar), whereas the diastolic potential of N cells (present in INE and CN) is more positive, because of a lack of IK,1 [8]. Consistent with this, in the human, Kir2.1 mRNA was VMN AMN INE≈CN (Kir2.1 is one of the principal channels responsible for IK,1; Fig. 7). In the rabbit, the resting potential of the AN cells in the TA is also known to be more positive [7]
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
649
Fig. 7. Relative abundance of mRNA for inward rectifier K+ channel subunits and adrenergic receptors. See Fig. 2 legend for details.
and, in the human, the abundance of Kir2.1 mRNA was also low in this region (Fig. 7). However, unlike Kir2.1 mRNA, Kir2.2 and Kir2.3 mRNAs were more evenly expressed (Fig. 7). To explore the consequences of the pattern of expression of ion channels, the mathematical model of the human atrial action potential developed by Courtemanche et al. [21] was modified based on ion channel gene expression in the different compartments of the atrioventricular conduction axis (see Chandler et al. [11] for details). For example, the conductance determining IK,1 was multiplied by the sum of Kir2.1, Kir2.2 and Kir2.3 mRNAs (corrected for differences in single channel conductance) expressed as a ratio of the sum for AM. Other ionic conductances were adjusted in an analogous manner. The modelling suggests that the diastolic potential of the nodal tissues in the human will be more positive than the resting potential of the working myocardium (data not shown). The absence of Kir2.1 and presumably of IK,1 along the human atrioventricular conduction axis will facilitate pacemaking, because IK,1 will not oppose inward pacemaker currents. If is an important pacemaker current and in small mammals it is present in the cardiac pacemaker tissues and generally not in the working myocardium [15]. Consistent with this, the major If channel, HCN4, at the mRNA level (and protein level), was abundantly expressed in all the major compartments of the atrioventricular conduction axis, but not in the working myocardium (Figs. 2 and 3). In contrast, HCN1 mRNA was only more highly expressed in the CN and HCN2 was uniformly expressed in all tissues. In the mathematical modelling described above, If was introduced into the model, because it is not present in the Courtemanche et al. [21] model of the human atrial action potential. In the CN, the conductance determining If was set to be one-third of that in the human sinus node [16] and the conductance in the other compartments of the atrioventricular conduction axis was adjusted based on HCN mRNA expression. The modelling predicted that the INE and CN show pacemaker activity in the human (data not shown). The CN is predicted to show the fastest pacemaker activity (cycle length, 825 ms; data not shown) and this is consistent with the finding that, in the human heart, the leading pacemaker site at the atrioventricular junction is the CN and its junction with the PB [22]. In the rabbit heart, however, the INE is the leading pacemaker site [23]. 4.2. Action potential upstroke The fast upstroke of the action potential in the working myocardium is due to the presence of INa, carried by Nav1.5. In contrast, in the rabbit,
the cells of the INE and CN typically have an action potential with a slow upstroke [7]. The same may be true in the human since there was a high level of expression of Nav1.5 mRNA (and protein) in the AM, PB and VM, an intermediate level in the TA and CN and a low level in the INE (Figs. 3 and 4). In the mathematical modelling described above, the conductance determining INa was adjusted based on the Nav1.5 mRNA expression. The modelling predicted the maximum upstroke velocity, dV/dtmax, to be 192 V/s in the AM, 84 V/s in the TA, 2 V/s in the INE and CN, 191 V/s in the PB and 149 V/s in the VM. This compares favourably to dV/dtmax along the atrioventricular conduction axis in the rabbit: ~150 V/s in the AM, ~45 V/s in the TA, ~13 V/s in the INE and CN, ~30 V/s in the PB, ~400 V/s in the Purkinje fibres and ~200 V/s in the VM [4,7,24]. Loss-of-function mutations in Nav1.5 result in disturbances of atrioventricular conduction, including atrioventricular block [25]. Because Nav1.5 is poorly expressed in the INE and CN, this is presumably the result of a decrease in INa in the TA (normal input into AVN) and PB (output from AVN). The characteristic slow upstroke of nodal myocytes is largely driven by ICa,L, at least in small mammals [7]. The major L-type Ca2+ channel, Cav1.2, was uniformly distributed in the different tissues (Fig. 5). However, theory suggests that a Ca2+ channel with a more negative activation threshold than Cav1.2 may be necessary in the absence of INa to maintain excitability [26]. Two such Ca2+ channels, the L-type Ca2+ channel, Cav1.3, and the T-type Ca2+ channel, Cav3.1, were more highly expressed in the nodal tissues (especially in CN) than in the working myocardium (Figs. 3 and 5). Data from Cav1.3 knockout mice indicate the pivotal functional role of the Cav1.3 channel [27]. Furthermore, autoimmune antibodies to Cav1.3 result in atrioventricular block [28]. Cav3.1 knockout mice display marked disturbances of atrioventricular conduction, suggesting that this channel too plays a pivotal role [29]. In the CN, the three Ca2+ channels, Cav1.2, Cav1.3 and Cav3.1, were present in comparable amounts (Fig. 5). Although all Ca2+ channels may be present in all cells, an alternative explanation is the presence of multiple specialised cells within the CN with different Ca2+ channels. 4.3. Repolarization Ito has two major cardiac components, Ito,f and Ito,s, which show fast and slow, respectively, recovery from inactivation. Kv4.2 and Kv4.3 are responsible for Ito,f, whereas Kv1.4 is responsible for Ito,s. In the working myocardium of the human heart, Kv4.3 and Ito,f are believed to be the major isoforms. However, in the atrioventricular conduction axis of the
650
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
rabbit, both Ito,f and Ito,s can be present [7,30]. The same may be true in the human, because in the CN and PB, the expression levels of Kv1.4, Kv4.2 and Kv4.3 mRNAs were comparable (Fig. 6). It is interesting that in the nodal tissues the expression levels of Kv1.4, Kv4.2, frequenin, DPP6 and possibly KChAP were greater in the CN and PB than in the working myocardium (Fig. 6). In the rabbit, Ito is small or absent in the atrioventricular conduction axis [7], but this is unlikely to be the case in the human, because there was no downregulation of any of the Ito genes (Fig. 6). IK,ur is considered to be a major repolarizing current in the human atrium. Consistent with this, Kv1.5 was abundantly expressed in the AM (Fig. 6). The low levels of Kv1.5 mRNA measured in the compartments of the atrioventricular conduction axis, as well as the VM, suggest a negligible role for this current in these tissues. The delayed rectifier currents, IK,r and IK,s, are crucial for repolarization in the working myocardium in various species including the human. However, along the rabbit atrioventricular conduction axis only IK,r is present (IK,s is absent) [8]. Along the human atrioventricular conduction axis, although ERG (IK,r) cDNAN KvLQT1 (IK,s) cDNA, the level of KvLQT1 cDNA was still substantial and, therefore, IK,s as well as IK,r may be important. 4.4. Autonomic regulation Autonomic regulation of the atrioventricular conduction axis is important. Mediating the sympathetic effects are the β-adrenergic receptors, of which β1 (ADβR1) and β2 (ADβR2) are the main isoforms. We observed a uniform distribution of β1 receptor mRNA in the various tissues studied, albeit a trend towards a lower abundance in the PB (Fig. 7). This trend reflects a previous quantitative receptor autoradiography study, which showed that, in the human, the bundle of His had a lower (one fifth) β1 adrenergic receptor expression compared to the axis and surrounding working myocardium [31]. In the same study, β2 adrenergic receptor expression was reported to be uniform [31] which is consistent with the result in Fig. 7 (with the exception of INE, which was not investigated in the earlier study). Thus activation of both β1 and β2 adrenergic receptors in the INE and CN could affect conduction, whilst a lesser role is envisaged for the β1 adrenergic receptor in the bundle of His [31]. Our data are in agreement with optical mapping of the human atrioventricular conduction axis [22]. Parasympathetic effects are mediated by Kir3.1 and Kir3.4 (responsible for IK,ACh). In keeping with tight parasympathetic control of the human atrioventricular conduction axis, Kir3.4 at least showed greater expression in the CN and PB, suggesting these compartments are the main targets for slowing of atrioventricular conduction by increased parasympathetic tone. 4.5. Intracellular Ca2+-handling Within the working myocardium, Ca2+-handling plays a pivotal role in determining cardiac contractile performance. The same cannot be said regarding the cardiac conduction system. On the other hand, the atrioventricular conduction axis is well known to have inherent pacemaking properties and recently a “Ca2+-clock” has been shown to play an important role in pacemaking [19]. It is possible, therefore, that there is a unique signature of Ca2+-handling proteins in the axis. Our data show a trend towards reduced SERCA2a mRNA and RYR2 mRNA (and protein) in the INE and CN (Fig. 5; Fig. S9, Data Supplement), as occurs in the rabbit atrioventricular conduction axis [8] and human sinus node [11]. In the human, there was also a tendency for increased expression of the Na+-Ca2+ exchanger (NCX1 mRNA) in the CN and this could be important for pacemaking (Fig. 5). 4.6. Conduction velocity The major role of the atrioventricular conduction axis is to slow conduction from the atria to the ventricles to allow for ventricular
filling. The INE is held by some to be the slow pathway into the CN [5] and in the rabbit conduction along the INE is indeed slow [7]. In contrast, conduction through the His-Purkinje system (of which PB is the most proximal component) is fast to ensure synchronous activation of the VM. A major factor determining the conduction velocity of the action potential is the coupling conductance, which in turn is determined by the expression of connexins. mRNAs for the small conductance connexins, Cx31.9 and Cx45, were uniformly expressed in the various tissues (Fig. 2). However, Cx43 mRNA and protein, responsible for medium conductance channels, were poorly expressed in the INE and CN compared to the working myocardium (Figs. 1 and 2). Cx40 mRNA, responsible for large conductance channels, also tended to be more poorly expressed in the INE (Fig. 2). Conversely, Cx40 mRNA (and protein) was highly expressed in the CN and PB and this may be correlated with the high conduction velocity of the His–Purkinje system (Fig. 2; Fig. S9, Data Supplement). The high expression of Cx40 in the CN is not a surprise. We have previously reported that Cx40 protein is present in the CN of the rat heart in the ‘lower nodal cells’, although it is absent in the upper nodal cells [32]. In the rat heart, as in the human heart, Cx40 protein is also expressed in the bundle of His as expected [32]. In the mouse, knockout of Cx40 results in varying degrees of atrioventricular block [33]. Another major factor determining the conduction velocity is the upstroke velocity of the action potential, which in turn is determined by the expression of the cardiac Na+ channel, Nav1.5. Nav1.5 mRNA and protein were poorly expressed in the INE and CN (Figs. 3 and 4). Together the expression of connexins and Nav1.5 can explain the variation in the conduction velocity along the atrioventricular conduction axis. The mathematical modelling described above was used to predict the conduction velocity along the axis. The conduction velocity was measured along a string of 300 electrically coupled model cells; the coupling conductance between cells was scaled according to the expression of connexins (corrected for differences in single channel conductance). The computed conduction velocity was 53 cm/s in the AM; the conduction velocity was slowest in the INE (~4 cm/s) and highest in the PB (~100 cm/s) as expected (data not shown). 4.7. Dual pathway electrophysiology The INE is thought to constitute the slow pathway in the human, whereas the TA is thought to constitute the fast pathway. The INE has a similar (but not identical) gene expression profile to the CN consistent with the hypothesis that both regions contain true nodal or N cells. The mathematical modelling described above predicted the INE to show pacemaker activity like the CN and have an action potential profile (including slow upstroke) similar to that of the CN (data not shown). However, the INE is predicted to have a slower conduction velocity than the CN (data not shown). In contrast, the TA has a transitional gene expression profile consistent with the hypothesis that it is comprised of transitional AN cells. It is predicted not to show pacemaker activity and to have an action potential with an intermediate upstroke velocity and that conducts at an intermediate speed (data not shown). These results are consistent with the hypothesis that the fast and slow pathways have distinct electrophysiological behaviours. 5. Conclusions and clinical implications In summary, we have described a unique and complex expression of various transcripts in the human atrioventricular conduction axis. The data can explain (i) the slow conduction along much of the axis, (ii) the fast conduction along the bundle of His, (iii) the subsidiary pacemaker activity of the axis and (iv) the dual pathway electrophysiology of the axis.
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Disclosures None. Acknowledgments This study was supported by Medtronic Inc. Appendix A. Supplementary data Supplementary data to this article can be found online at doi:10.1016/ j.yjmcc.2010.12.017. References [1] Hucker WJ, Dobrzynski H, Efimov IR. Mechanisms of AV nodal excitability and propagation. In: Zipes DP, Jalife J, editors. From Cell to Bedside. 5th Edition. Saunders: Elsevier Science; 2009. [2] Anderson RH, Yanni J, Boyett MR, Chandler NJ, Dobrzynski H. The anatomy of the cardiac conduction system. Clin Anat 2009;22:99–113. [3] Moe GK, Preston JB, Burlington H. Demonstration of a dual A–V nodal conduction system in the isolated rabbit heart. Circ Res 2008;19:378–93. [4] Medkour D, Becker AE, Khalife K, Billette J. Anatomic and functional characteristics of a slow posterior AV nodal pathway: role in dual-pathway physiology and reentry. Circulation 1998;98:164–74. [5] Inoue S, Becker AE. Posterior extensions of the human compact atrioventricular node: a neglected anatomic feature of potential clinical significance. Circulation 1998;97:188–93. [6] Li J, Greener ID, Inada S, Nikolski VP, Yamamoto M, Hancox J, et al. Computer threedimensional reconstruction of the atrioventricular node. Circ Res 2008;102: 975–85. [7] Inada S, Hancox J, Zhang H, Boyett MR. One-dimensional mathematical model of the atrioventricular node including atrio-nodal, nodal, and nodal-his cells. Biophys J 2009;97:2117–27. [8] Greener ID, Tellez JO, Dobrzynski H, Yamamoto M, Graham GM, Billeter R, et al. Ion channel transcript expression at the rabbit atrioventricular conduction axis. Circ Arrhythmia Electrophysiol 2009;2:305–15. [9] Jackman WM, Beckman KJ, McClelland JH, Wang X, Friday KJ, Roman CA, et al. Treatment of supraventricular tachycardia due to atrioventricular nodal reentry, by rediofrequency catheter ablation of slow-pathway conduction. N Engl J Med 1992;327:313–8. [10] Hucker WJ, Fedorov VG, Foyil KV, Moazami N, Efimov IR. Optical mapping of the human atrioventricular junction. Circulation 2008;117:1474–7. [11] Chandler NJ, Greener ID, Tellez JO, Inada S, Musa M, Molenaar M, et al. Molecular architecture of the human sinus node: insights into the function of the cardiac pacemaker. Circulation 2009;119:1562–75. [12] Tellez JO, Dobrzynski H, Greener ID, Graham GM, Laing E, Honjo H, et al. Differential expression of ion channel transcripts in atrial muscle and sinoatrial node in rabbit. Circ Res 2006;99:1384–93. [13] Hoogaars VM, Tessari A, Moorman AF, de Boer PA, Hagoort J, Soufan AT, et al. The transcriptional repressor Tbx3 delineates developing central conduction system of the heart. Circ Res 2004;62:489–99. [14] Kreuzberg MM, Liebermann M, Segschneider S, Dobrowolski R, Dobrzynski H, Kaba R, et al. Human connexin31.9, unlike its orthologous protein connexin30.2 in the mouse, is not detectable in the human cardiac conduction system. J Mol Cell Cardiol 2009;46:553–9. [15] DiFrancesco D. Pacemaker mechanisms in cardiac tissue. Annu Rev Physiol 1993;55:455–72. [16] Verkerk AO, Wilders R, van Borren MM, Peters RJ, Broekhuis E, Lam K, et al. Pacemaker current (If) in the human sinoatrial node. Eur Heart J 2007;28:2472–8.
651
[17] Maier SKG, Westenbroek RE, Yamanushi TT, Dobrzynski H, Boyett MR, Catterall WA, et al. An unexpected requirement for brain-type sodium channels for control of heart rate in the mouse sinoatrial node. Proc Natl Acad Sci USA 2003;100: 3507–12. [18] Dobrzynski H, Boyett MR, Anderson RH. New insights into pacemaker activity. Promoting understanding of sick sinus syndrome. Circulation 2007;115:1921–32. [19] Bogdanov KY, Vinogradova TM, Lakatta EG. Sinoatrial nodal cell ryanodine receptor and Na+-Ca2+ exchanger: molecular partners in pacemaker regulation. Circ Res 2001;88:1254–8. [20] Krapivinsky G, Gordon EA, Wickman K, Velimirovic B, Krapivinsky L, Clapham DE. The G-protein-gated atrial K+ channel IKACh is a heteromultimer of two inwardly rectifying K+-channel proteins. Nature 1995;374:135–41. [21] Courtemanche M, Ramirez RJ, Nattel S. Ionic mechanisms underlying human atrial action potential properties: insights from a mathematical model. Am J Physiol 1998;275:H301–21. [22] Fedorov VV, Ambrosi CM, Hucker WJ, Glukhov AV, Foyil KV, Wuskell J, et al. Human AV junctional pacemaker shift due to cholinergic and adrenergic stimulations: optical imaging with a novel long wavelength voltage-sensitive dye. Circulation 2008;118:S520. [23] Dobrzynski H, Nikolski VP, Sambelashvili AT, Greener ID, Yamamoto M, Boyett MR, et al. The site of origin and molecular substrate of AV junctional rhythm in the rabbit heart. Circ Res 2003;93:1102–10. [24] Aslanidi OV, Sleiman RN, Boyett MR, Hancox JC, Zhang H. Ionic mechanisms for electrical heterogeneity between rabbit Purkinje and ventricular cells and tissue. Biophys J 2010;98:2410–31. [25] Wang DW, Viswanathan PC, Balser JR, George Jr AL, Benson DW. Clinical, genetic, and biophysical characterization of SCN5A mutations associated with atrioventricular conduction block. Circulation 2002;105:341–6. [26] Inada S, Mitsui K, Honjo H, Boyett MR. Why is Cav1.3 expressed in the sinoatrial node? Biophys J 2005. [27] Mangoni ME, Couette B, Bourinet E, Platzer J, Reimer D, Striessnig J, et al. Functional role of L-type Cav1.3 Ca2+ channels in cardiac pacemaker activity. Proc Natl Acad Sci USA 2003;100:5543–8. [28] Qu Y, Baroudi G, Yue Y, Boutjdir M. Novel molecular mechanism involving alpha1D (Cav1.3) L-type calcium channel in autoimmune-associated sinus bradycardia. Circulation 2005;111:3034–41. [29] Mangoni ME, Traboulsie A, Leoni AL, Couette B, Marger L, Le Quang K, et al. Bradycardia and slowing of the atrioventricular conduction in mice lacking Cav3.1/ a1G T-type calcium channels. Circ Res 2006;98:1422–30. [30] Mitcheson JS, Hancox JC. Characteristics of a transient outward current (sensitive to 4-aminopyridine) in Ca2+-tolerant myocytes isolated from the rabbit atrioventricular node. Pflugers Arch 1999;438:68–78. [31] Elnatan J, Molenaar P, Rosenfeldt FL, Summers RS. Autoradiographic localisation and quantitation of beta1- and beta2-adrenoceptors in the human atrioventricular conducting system: a comparison of patients with idiopathic dilated cardiomyopathy and ischemic heart disease. J Mol Cell Cardiol 1994;26:313–23. [32] Yoo S, Dobrzynski H, Fedorov VV, Xu SZ, Yamanushi TT, Jones SA, et al. Localization of Na+ channel isoforms at the atrioventricular junction, including the atrioventricular node. Circulation 2006;114:1360–71. [33] Hagendorff A, Schumacher B, Kirchhoff S, Luderitz B, Willecke K. Conduction disturbances and increased atrial vulnerability in connexin40-deficient mice analyzed by transesophageal stimulation. Circulation 1999;99:1508–15.
Glossary AM: Atrial muscle AVN: Atrioventricular node CN: Compact node INE: Inferior nodal extension PB: Penetrating bundle/bundle of His TA: Transitional area VM: Ventricular muscle
Contents lists available at ScienceDirect
Journal of Molecular and Cellular Cardiology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y j m c c
Original article
Molecular architecture of the human specialised atrioventricular conduction axis I.D. Greener a,1, O. Monfredi a,1, S. Inada a, N.J. Chandler a, J.O. Tellez a, A. Atkinson a, M.-A. Taube a, R. Billeter b, R.H. Anderson a, I.R. Efimov c, P. Molenaar d, D.C. Sigg e, V. Sharma e, M.R. Boyett a,2, H. Dobrzynski a,⁎,2 a
University of Manchester, UK University of Nottingham, UK Washington University, St Louis, USA d Queensland University of Technology & University of Queensland, Australia e Medtronic Inc., USA b c
a r t i c l e
i n f o
Article history: Received 28 August 2010 Received in revised form 16 December 2010 Accepted 17 December 2010 Available online 21 January 2011 Keywords: Atrioventricular node Ion channels Gap junctions Arrhythmias
a b s t r a c t The atrioventricular conduction axis, located in the septal component of the atrioventricular junctions, is arguably the most complex structure in the heart. It fulfils a multitude of functions, including the introduction of a delay between atrial and ventricular systole and backup pacemaking. Like any other multifunctional tissue, complexity is a key feature of this specialised tissue in the heart, and this complexity is both anatomical and electrophysiological, with the two being inextricably linked. We used quantitative PCR, histology and immunohistochemistry to analyse the axis from six human subjects. mRNAs for ~ 50 ion and gap junction channels, Ca2+-handling proteins and markers were measured in the atrial muscle (AM), a transitional area (TA), inferior nodal extension (INE), compact node (CN), penetrating bundle (PB) and ventricular muscle (VM). When compared to the AM, we found a lower expression of Nav1.5, Kir2.1, Cx43 and ANP mRNAs in the CN for example, but a higher expression of HCN1, HCN4, Cav1.3, Cav3.1, Kir3.4, Cx40 and Tbx3 mRNAs. Expression of some related proteins was in agreement with the expression of the corresponding mRNAs. There is a complex and heterogeneous pattern of expression of ion and gap junction channels and Ca2+handling proteins in the human atrioventricular conduction axis that explains the function of this crucial pathway. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction The atrioventricular conduction axis (or atrioventricular node) is located in the septal component of the atrioventricular junctions, and is arguably the most complex structure in the heart. It has a multitude of functions, such as providing a critical delay between atrial and ventricular systole, protecting against life-threatening rapid supraventricular rates, and providing contingent pacemaking in the event of failure of the sinus node [1]. It is also the site of one of the most common cardiac arrhythmias, namely atrioventricular nodal reentrant tachycardia. Like any other multifunctional tissue, complexity is one of its key features. This complexity is both structural and functional. Histological analysis of the specialised muscular axis responsible for conduction of the cardiac impulse across the plane of atrioventricular insulation reveals various components common to all mammals, including the human [2].
⁎ Corresponding author at: Cardiovascular Medicine, Faculty of Medical and Human Sciences, University of Manchester, Core Technology Facility, 46 Grafton Street, Manchester M13 9NT, UK. Tel.: + 44 161 275 1182. E-mail address: [email protected] (H. Dobrzynski). 1 Joint first authors. 2 Joint senior authors. 0022-2828/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.yjmcc.2010.12.017
Inferiorly, between the coronary sinus and tricuspid valve annulus, is found the inferior nodal extension (INE). The INE connects to the compact node (CN). The CN occupies the apex of the triangle of Koch, becoming the penetrating bundle (PB), or bundle of His, when it penetrates into the central fibrous body. Also connecting to the CN is a transitional area (TA) made up of transitional atrial myocytes. Electrophysiological data relating to the function of the axis has largely been derived from rabbit preparations. Amongst the most important findings in terms of function is the presence of dual inputs to the axis [3,4]. The fast pathway is located antero-superiorly, whereas the slow pathway is located between the orifice of the coronary sinus and the attachment of the septal leaflet of the tricuspid valve. The fast pathway is thought to correspond to the TA and the slow pathway to the INE [5,6]. In the rabbit, three cell types are widely accepted as making up the atrioventricular conduction axis. As compared to atrial myocytes with a fast Na+-dependent action potential upstroke and negative resting potential, N (typical nodal) cells have a slow Ca2+-dependent upstroke and a more positive diastolic potential, whereas the transitional AN and NH cells have an intermediate upstroke and diastolic potential [7]. N cells are thought to make up the INE (slow pathway) and CN, the transitional AN cells are thought to make up the TA (fast pathway), and the transitional NH cells are thought to be present in the PB [7,8].
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Electrophysiological data regarding the human atrioventricular conduction axis have largely been restricted to electrode recordings during physiological testing and modification of arrhythmic substrates. Clearly, fast and slow pathways exist in humans [9]. More recently, Efimov and colleagues have provided further insights into the electrophysiology of the axis in the human with optical action potential recordings [10]. Due to limited access to live human tissue, there is no information regarding specific ionic currents or cell–cell coupling properties in the human. To remedy this deficiency, we have undertaken a molecular analysis of transcripts and proteins governing electrical excitability, hoping to provide further insight into the behaviour of the human atrioventricular conduction axis. 2. Materials and methods We obtained specimens including the area of the triangle of Koch and the adjacent membranous part of the septum from six diseased human hearts acquired from the Prince Charles Hospital District, Chermside, Australia (ethics approval, EC2565; work also ethically approved by University of Manchester). A clinical profile of the patients is shown in the Data Supplement. We used quantitative PCR (qPCR), histology and immunohistochemistry to investigate the molecular make-up of the components of the atrioventricular conduction axis. For qPCR, we sampled the components of the axis, along with the atrial muscle (AM) and ventricular muscle (VM), as outlined in Fig. 1A, from 30 to 50 (60 μm thick) frozen sections, which underwent a novel and rapid haematoxylin and eosin (H&E) staining protocol [11]. In brief, from multiple quick H&E stained sections, tissue from each region (TA, INE, CN, PB) were micro-dissected by hand under a dissecting microscope with a fine surgical needle. Each region was identified by its histological appearance. The identification was guided by Masson's trichrome stained sections as well as sections immunolabelled for Cx43
643
such as those in Fig. 1. Total RNA was extracted from a given region from all relevant tissue sections. This was repeated for all six hearts. Following extraction of total RNA from the samples and reverse transcription to produce cDNA, the amounts of ~50 cDNAs for selected cell type-markers, connexins, ion channels, Ca2+-handling proteins and receptors were measured using qPCR. The data were analysed using the ‘ΔCT technique’: to correct for variations in the amount of total RNA in the different samples, the amounts of the different cDNAs measured were normalised to the amount of cDNA for a housekeeper (ribosomal protein, 28S). The amount of 28S cDNA was similar in each tissue investigated (data not shown). For a given ion channel etc., the amount of mRNA is proportional to the amount of cDNA and, therefore, this method allows the abundance of the same mRNA in different tissues to be compared. Tentative conclusions can also be drawn about the relative amounts of different mRNAs from the relative amounts of the corresponding cDNAs. However, the conclusions must be tentative, because the efficiency of the reverse transcription step (to produce cDNA) varies for different mRNAs. If the difference between two cDNAs was greater than 10-fold, the difference is likely to reflect to a difference in the abundance of the two corresponding mRNAs (rather than difference in efficiency of reverse transcription). However, if the difference between two cDNAs was less than 10-fold, the result must be interpreted cautiously. Data for individual transcripts are plotted here, whereas the data for groups of related transcripts are plotted in the Data Supplement (Figs. S1–S7) to facilitate comparison. All transcripts in each sample were measured in triplicate. In general, the three measurements were close. The average of these three measurements was then taken and the mean ± SEM of the average values for each heart was calculated – the n number for the calculation is, therefore, the number of hearts. In figures, the mean± SEM is shown; in addition, the individual averages are shown as open circles. The SEM (as well as the open circles) shows the variability between hearts. Significant
Fig. 1. Histological and immunohistochemical characteristics of human atrioventricular conduction axis. A, Masson's trichrome stained sections through INE (left), CN (middle) and PB (right). Myocytes stained purple, connective tissue blue. Dotted lines highlight areas (INE, CN and PB) sampled for qPCR. B, high magnification images of triple immunolabelling of Cx43 (green signal; at gap junctions), caveolin3 (red signal; within cell membrane) and vimentin (blue signal; within fibroblasts) in AM (left), INE, CN and PB. CFB, central fibrous body.
644
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
differences in the abundance of mRNAs in different tissues were identified using one-way ANOVA. A difference was assumed to be significant when P b 0.05. The data are shown in an alternative format in Table 1 (see also Fig. S8 in the Data Supplement), which shows mean values of the relative abundance of all mRNAs normalised to that in the AM. A more detailed description of the methods can be found in the Data Supplement. A list of the principal abbreviations used is in the Glossary. 3. Results 3.1. Morphological characteristics of components of atrioventricular conduction axis Staining with Masson's trichrome technique of sections obtained through the triangle of Koch and the membranous septum revealed the location of the components of the specialised muscular axis responsible for atrioventricular conduction (Fig. 1A). With this technique, the
Table 1 Relative abundance (mean value) of mRNA normalised to mRNA abundance in AM (%).
ADBR1 ADBR2 ANP Ca2+v1.2 Ca2+v1.3 Ca2+v3.1 Ca2+v3.3 Cx31.9 Cx40 Cx43 Cx45 DPP6 FREQ HCN1 HCN2 HCN3 HCN4 ERG IP3R1 KchAP KchIP2 Kir2.1 Kir2.2 Kir2.3 Kir3.1 Kir3.4 Kv1.4 Kv1.5 Kv4.2 Kv4.3 KvB1 KvB2 KvLQT1 minK Na+v1.1 Na+v1.3 Na+v1.4 Nav1.5 Nav1.6 Nav1.7 NavB1 NavB2 NavB3 NavB4 NCX1 PLN RyR2 RyR3 SERCA2a Tbx3
AM
TA
INE
CN
PB
VM
100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100
89 113 30 84 356 210 95 103 54 64 83 103 129 164 70 92 261 87 85 94 80 37 57 106 93 79 38 69 285 122 107 78 87 118 138 147 152 66 92 159 83 55 105 129 95 58 89 269 115 175
59 38 3 79 649 282 15 Not investigated 49 21 60 286 106 187 86 366 480 82 32 65 80 22 59 32 20 129 63 22 256 52 83 48 33 85 784 111 51 17 156 135 58 68 33 37 35 0 26 302 60 263
77 108 4 105 1762 559 110 92 174 21 118 103 607 527 142 86 863 164 113 164 144 44 141 131 178 414 40 61 1044 98 74 70 121 191 151 165 42 49 195 181 69 112 103 53 173 61 32 116 48 498
35 93 14 56 547 318 87 104 250 53 61 1949 688 51 67 107 505 172 96 142 88 91 50 112 68 370 651 41 643 140 46 82 78 102 435 82 52 126 188 101 76 113 114 64 126 99 82 118 54 320
63 70 4 112 18 8 58 106 27 112 88 Not investigated 74 7 73 83 84 77 101 79 140 337 3 253 6 66 345 39 106 85 152 166 91 133 213 118 43 116 118 77 59 141 153 94 115 120 110 82 66 73
cardiac myocytes are stained purple, and the connective tissue is stained blue. Adjacent tissue sections were immunolabelled for Cx43 (major connexin in working myocardium), caveolin3 (membrane-bound protein expressed by all cardiac myocytes) and vimentin (expressed by fibroblasts). As shown in Fig. 1B, whereas there was labelling for caveolin3 in all cardiac myocytes, labelling for Cx43 was absent from the INE and CN, but present in the AM, PB and VM. Labelling of vimentin was primarily observed in the INE and the CN. In small mammals, Cx43 has been shown to be absent from these components of the axis, but abundantly expressed in the working myocardium [e.g., 8]. Our immunofluorescence analysis revealed this also to be the case in the human (Fig. 1B). 3.2. Nodal markers ANP (atrial natriuretic peptide), a hormone expressed by atrial muscle [e.g., 8,11,12], and Tbx3, a transcription factor expressed in the cardiac conduction system [8,11–13], were used to characterise the tissues studied by qPCR. qPCR showed ANP mRNA to be highly abundant in the AM, but poorly expressed in the INE, CN, PB and VM (Fig. 2). An intermediate level of expression of ANP mRNA was noted in the TA (Fig. 2). Tbx3 mRNA was most abundant in the CN and it was poorly expressed in the AM and VM (Fig. 2). The level of Tbx3 in the working myocardium is presumably a basal level of this transcription factor. In our previous work on the human sinus node [11], we also detected a ‘basal level’ of Tbx3 in the working myocardium. There was an intermediate level of Tbx3 expression in the TA, INE and PB (Fig. 2). The significantly lower expression of ANP mRNA, but higher expression of Tbx3 mRNA, demonstrated by qPCR in the components of the atrioventricular conduction axis compared to the working myocardium confirms that, during dissection, there was little or no contamination of these tissues with working myocardium. 3.3. Connexins Cx31.9, Cx40, Cx43 and Cx45 form ~ 9, ~ 200, ~ 80 and ~ 30 pS gap junction channels, respectively. The levels of Cx40, Cx43 and Cx45 cDNAs were generally N10-fold greater than that of Cx31.9 (Fig. 2). There was significantly higher expression of Cx40 mRNA in the CN and PB compared to the other tissues (Fig. 2). Immunohistochemistry showed that Cx40 protein was also highly abundant in the CN and AM, but absent from the VM (Fig. 3). As expected, immunohistochemistry showed that Cx40 protein was more abundant in the PB than in the working myocardium (Fig. S9, Data Supplement). There was significantly lower expression of Cx43 mRNA in the INE, CN and PB as compared to the working myocardium (Fig. 2). There was an intermediate level of expression of Cx43 mRNA in the TA. Cx43 protein was distributed in a similar manner in the different tissues (Fig. 1B). The levels of Cx31.9 and Cx45 mRNAs were similar in each of the six tissues (Fig. 2). However, the amount of Cx31.9 mRNA in comparison with other connexin mRNAs was negligible in the human heart (Fig. S2), and this is in agreement with the study of Kreuzberg et al. at the protein level [14]. 3.4. HCN channels The HCN channels (HCN1–HCN4) are responsible for the hyperpolarization-activated or ‘funny’ current, If [15]. It is well known that If plays a key role in pacemaker activity in the heart including the human sinus node [15,16]. In the CN, for example, HCN4 cDNA was more abundant than HCN2 cDNA (but not more than N10fold more abundant), N10-fold more abundant than HCN1 cDNA, which was N100-fold more abundant than HCN3 cDNA (Fig. 2). There was a significantly higher expression of HCN1 mRNA in the CN compared to the PB and VM (Fig. 2). There was also a significantly higher expression of HCN4 mRNA in the INE, CN and PB compared to
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
645
Fig. 2. Relative abundance of mRNA for cell type-markers, connexins and HCN channel subunits. Means + SEM (n = 6) shown. a–fsignificantly different from appropriately lettered bars (one-way ANOVA). Individual data for all six subjects shown by open circles.
the working myocardium (Fig. 2). Because HCN4 was the most abundant isoform, it was investigated at the protein level using immunohistochemistry. Fig. 3 shows that HCN4 protein was expressed in the sarcolemma of the CN cells, although the signal was below the detection threshold in the working myocardium. The level of HCN2 and HCN3 mRNAs was similar in most of the six tissues (Fig. 2). 3.5. Na+ channels Voltage-gated Na+ channels are responsible for the inward Na+ current, INa. INa is responsible for the rapid action potential upstroke in the working myocardium. The cardiac Na+ channel, Nav1.5, is primarily responsible for INa. There was a significantly lower expression of Nav1.5 mRNA in the INE and CN as compared to the AM, PB and VM (Fig. 4). Immunohistochemistry also showed that, although Nav1.5 protein was detectable in the working myocardium, it was not detectable in the CN (Fig. 3). In small mammals, other Na+ channels have also been shown to be expressed in cardiac myocytes [17]. These Na+ channels (Nav1.1, Nav1.3, Nav1.4, Nav1.6 and Nav1.7) at the mRNA level are also expressed in human cardiac tissues (Fig. 4; Nav1.2 not studied as we have previously shown it to be poorly expressed in human heart [11]). However, their cDNA levels were N10-fold lower than that of Nav1.5 cDNA (Fig. 4), suggesting that these Na+ channels may play a minor role in human. The levels of the mRNAs were similar in each of the six tissues (Fig. 4). Of the Na+ channel β-subunits (Navβ1–Navβ4), Navβ1 was the most abundant β-subunit (at cDNA level at least) and the levels of the mRNAs were again similar in each of the six tissues (Fig. 4). 3.6. Ca2+ channels Two separate Ca2+ currents have been recorded from cardiac myocytes, the L-type (ICa,L) and the T-type (ICa,T) [18]. Cav1.2 and Cav1.3
are responsible for ICa,L and Cav3.1–Cav3.3 are responsible for ICa,T. All channels except Cav3.2 were studied since previously we have shown Cav3.2 to be undetectable in human heart) [11]. In the working myocardium, Cav1.2 cDNA was N10-fold more abundant than Cav1.3 cDNA (Fig. 5), whereas in the CN they were of comparable abundance (Fig. 5). The expression of Cav1.2 mRNA was similar in each of the six tissues (Fig. 5). On the other hand, there was a significantly higher expression of Cav1.3 mRNA in the INE, CN and PB compared to the working myocardium (Fig. 5). Cav3.1 cDNA was generally N10-fold more abundant than Cav3.3 cDNA (Fig. 5). Cav3.1 mRNA was distributed in a similar manner to Cav1.3 mRNA (Fig. 5). Consistent with this, Cav3.1 protein was detected in the CN, but not in the AM (Fig. 3). Cav3.3 mRNA did not vary significantly among the tissues (Fig. 5).
3.7. Ca2+-handling proteins Intracellular Ca2+ has been suggested to play an important role in pacemaking in small mammals [19]. The expression level of the Na+– Ca2+ exchanger, NCX1, at the mRNA level, was similar in most tissues, however it was significantly more abundant in the CN compared to the INE (Fig. 5; highest NCX1 data point does not account for this and when removed from statistical analysis, the difference remained significant). Consistent with this, immunohistochemistry showed that NCX1 protein was expressed both in the PB and working myocardium (Fig. S9, Data Supplement). mRNA for the sarcoplasmic reticulum (SR) Ca2+ pump, SERCA2a, did not vary significantly among the tissues (Fig. 5). RyR2 and RyR3 are SR Ca2+ release channels; RyR2 cDNA was N10-fold more abundant than RyR3 cDNA (Fig. 5). RyR2 mRNA tended to be less abundant in the INE and CN (Fig. 5). Consistent with this, immunohistochemistry showed that RyR2 protein was detectable in the working myocardium, but not in the CN (Fig. S9, Data Supplement). RyR3 mRNA was significantly more abundant in the INE as compared to the VM
646
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Fig. 3. Expression of Cx40, caveolin3, HCN4, Nav1.5 and Cav3.1 proteins in CN and working myocardium. High magnification images of immunolabelling of Cx40 and caveolin3 (top; Cx40, green signal; caveolin3, red signal), HCN4 (middle top; green signal), Nav1.5 (middle bottom; red signal) and Cav3.1 (bottom; green signal) in CN (left) and AM/VM (right) shown. Scale bar for each panel in shown in the bottom right panel.
(Fig. 5). IP3 receptor (IP3R1) mRNA tended to be less abundant in the INE (Fig. 5).
3.8. Transient outward K+ channels The transient outward K+ current (Ito) is responsible for the early phase of repolarization (phase 1) of the action potential. The ion channels responsible for Ito are the voltage-dependent K+ channels, Kv1.4, Kv4.2 and Kv4.3. Kv4.3 is regarded as the major ion channel underlying Ito in the human and although Kv4.3 cDNA was more abundant than Kv1.4 and Kv4.2 cDNAs, it was not necessarily N10-fold greater, especially in the CN (Fig. 6). Kv1.4 mRNA was significantly more abundant in the CN as compared to the AM, TA and INE and Kv4.2 mRNA was significantly more abundant in the CN as compared to all other tissues (Fig. 6). The level of Kv4.3 mRNA was similar in all tissues (although expression in PB was significantly greater than in INE; Fig. 6).
3.9. Delayed rectifier K+ channels There are three delayed rectifier K+ currents, ultra-rapid (IK,ur), rapid (IK,r) and slow (IK,s). The ion channels responsible for IK,ur, IK,r and IK,s are Kv1.5, ERG and KvLQT1, respectively. ERG cDNA N KvLQT1 cDNA N Kv1.5 cDNA (but not necessarily by N10-fold; Fig. 6). The expression level of Kv1.5 mRNA was significantly higher in the AM compared to the INE, PB and VM while the level of ERG mRNA was similar in each of the six tissues, whereas there was significantly lower expression of KvLQT1 mRNA in the INE as compared to all other tissues (Fig. 6). 3.10. Voltage-gated K+ channel regulatory proteins We have also studied accessory proteins for voltage-gated K+ channels. FREQ (frequenin; Ca2+-binding protein known to regulate Kv4 channels), DPP6 (dipeptidyl aminopeptidase-like protein 6; proposed auxiliary subunit of Kv4 channels) and KChAP (chaperone for Kv4.3)
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Fig. 4. Relative abundance of mRNA for Na+ channel subunits. See Fig. 2 legend for details.
Fig. 5. Relative abundance of mRNA for Ca2+ channel subunits and Ca2+-handling proteins. See Fig. 2 legend for details.
647
648
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Fig. 6. Relative abundance of mRNA for voltage-gated K+ channel subunits. See Fig. 2 legend for details.
regulate transient outward K+ channels. Both FREQ and DPP6 mRNAs were significantly more abundant in the CN and PB compared to the VM (similar to distribution of Kv1.4 and Kv4.2; Fig. 6). KChAP mRNA tended to be distributed in a similar manner (Fig. 6). Together with KvLQT1, minK is responsible for IK,s and was distributed in a similar manner to KvLQT1 mRNA (Fig. 6). Three β-subunits are thought to regulate delayed rectifier K+ channels, Kvβ1–Kvβ3; we studied Kvβ1 and Kvβ2 mRNAs as previously we have shown Kvβ3 to be poorly expressed in human heart [11]. Kvβ1 and Kvβ2 mRNAs did not vary among the tissues studied (Fig. 6). 3.11. Inward rectifier K+ channels The background inward rectifier K+ current, IK1, is generated by the ion channels, Kir2.1–Kir2.4. mRNA levels for Kir2.1–Kir2.3 only were measured since previously we have shown Kir2.4 to be poorly expressed in human heart [11]. In the VM, Kir2.1 cDNA N Kir2.3 cDNA N K ir 2.2 cDNA, whereas in the AM, K ir 2.2 cDNA N K ir 2.1 cDNA N Kir2.3 cDNA (but not necessarily by N10-fold; Fig. 7). The level of Kir2.1 mRNA was significantly higher in the VM compared to all other tissues and it tended to be lower in the TA, INE and CN than in the AM (Fig. 7). The level of Kir2.2 mRNA was similar in most tissues (although it was significantly higher in AM and CN compared to VM; Fig. 7). Kir2.3 mRNA did not vary significantly among the tissues (Fig. 7). The ACh-activated K+ current, IKACh, is generated by a heteromultimer of Kir3.1 and Kir3.4 [20]. Kir3.1 and Kir3.4 mRNAs were similarly abundant as expected (Fig. 7). The level of Kir3.1 mRNA was significantly higher in the AM, TA, CN and PB compared to the VM, whereas the level of Kir3.4 mRNA was significantly higher in the CN and PB compared to the TA and VM (Fig. 7).
3.12. Adrenergic receptors The atrioventricular conduction axis is under rigorous autonomic regulation. mRNAs for β1 (ADβR1) and β2 (ADβR2) receptors were detectable in all tissues investigated. The levels of the two receptors were similar (Fig. 7). β1 receptor (ADβR1) mRNA tended to be lower in the PB, whereas β2 receptor (ADβR2) mRNA tended to be lower in the INE (Fig. 7; see Discussion). 4. Discussion We have shown a complex portrait of transcripts, and whenever possible proteins, underlying the electrical excitability of the specialised muscular axis responsible for atrioventricular conduction in the human heart. Our data provide novel insights into the electromolecular properties of the atrioventricular conduction axis, in particular within the major sub-compartments of this complex structure. 4.1. Pacemaking Various factors are responsible for pacemaking in the heart (for example in sinus node), but two important ones are the absence of IK,1 and the presence of the pacemaker current, If. In the rabbit at least, the working myocardium has a resting potential generated by IK,1 (although density of IK,1 in AM is less than in VM, despite resting potentials being similar), whereas the diastolic potential of N cells (present in INE and CN) is more positive, because of a lack of IK,1 [8]. Consistent with this, in the human, Kir2.1 mRNA was VMN AMN INE≈CN (Kir2.1 is one of the principal channels responsible for IK,1; Fig. 7). In the rabbit, the resting potential of the AN cells in the TA is also known to be more positive [7]
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
649
Fig. 7. Relative abundance of mRNA for inward rectifier K+ channel subunits and adrenergic receptors. See Fig. 2 legend for details.
and, in the human, the abundance of Kir2.1 mRNA was also low in this region (Fig. 7). However, unlike Kir2.1 mRNA, Kir2.2 and Kir2.3 mRNAs were more evenly expressed (Fig. 7). To explore the consequences of the pattern of expression of ion channels, the mathematical model of the human atrial action potential developed by Courtemanche et al. [21] was modified based on ion channel gene expression in the different compartments of the atrioventricular conduction axis (see Chandler et al. [11] for details). For example, the conductance determining IK,1 was multiplied by the sum of Kir2.1, Kir2.2 and Kir2.3 mRNAs (corrected for differences in single channel conductance) expressed as a ratio of the sum for AM. Other ionic conductances were adjusted in an analogous manner. The modelling suggests that the diastolic potential of the nodal tissues in the human will be more positive than the resting potential of the working myocardium (data not shown). The absence of Kir2.1 and presumably of IK,1 along the human atrioventricular conduction axis will facilitate pacemaking, because IK,1 will not oppose inward pacemaker currents. If is an important pacemaker current and in small mammals it is present in the cardiac pacemaker tissues and generally not in the working myocardium [15]. Consistent with this, the major If channel, HCN4, at the mRNA level (and protein level), was abundantly expressed in all the major compartments of the atrioventricular conduction axis, but not in the working myocardium (Figs. 2 and 3). In contrast, HCN1 mRNA was only more highly expressed in the CN and HCN2 was uniformly expressed in all tissues. In the mathematical modelling described above, If was introduced into the model, because it is not present in the Courtemanche et al. [21] model of the human atrial action potential. In the CN, the conductance determining If was set to be one-third of that in the human sinus node [16] and the conductance in the other compartments of the atrioventricular conduction axis was adjusted based on HCN mRNA expression. The modelling predicted that the INE and CN show pacemaker activity in the human (data not shown). The CN is predicted to show the fastest pacemaker activity (cycle length, 825 ms; data not shown) and this is consistent with the finding that, in the human heart, the leading pacemaker site at the atrioventricular junction is the CN and its junction with the PB [22]. In the rabbit heart, however, the INE is the leading pacemaker site [23]. 4.2. Action potential upstroke The fast upstroke of the action potential in the working myocardium is due to the presence of INa, carried by Nav1.5. In contrast, in the rabbit,
the cells of the INE and CN typically have an action potential with a slow upstroke [7]. The same may be true in the human since there was a high level of expression of Nav1.5 mRNA (and protein) in the AM, PB and VM, an intermediate level in the TA and CN and a low level in the INE (Figs. 3 and 4). In the mathematical modelling described above, the conductance determining INa was adjusted based on the Nav1.5 mRNA expression. The modelling predicted the maximum upstroke velocity, dV/dtmax, to be 192 V/s in the AM, 84 V/s in the TA, 2 V/s in the INE and CN, 191 V/s in the PB and 149 V/s in the VM. This compares favourably to dV/dtmax along the atrioventricular conduction axis in the rabbit: ~150 V/s in the AM, ~45 V/s in the TA, ~13 V/s in the INE and CN, ~30 V/s in the PB, ~400 V/s in the Purkinje fibres and ~200 V/s in the VM [4,7,24]. Loss-of-function mutations in Nav1.5 result in disturbances of atrioventricular conduction, including atrioventricular block [25]. Because Nav1.5 is poorly expressed in the INE and CN, this is presumably the result of a decrease in INa in the TA (normal input into AVN) and PB (output from AVN). The characteristic slow upstroke of nodal myocytes is largely driven by ICa,L, at least in small mammals [7]. The major L-type Ca2+ channel, Cav1.2, was uniformly distributed in the different tissues (Fig. 5). However, theory suggests that a Ca2+ channel with a more negative activation threshold than Cav1.2 may be necessary in the absence of INa to maintain excitability [26]. Two such Ca2+ channels, the L-type Ca2+ channel, Cav1.3, and the T-type Ca2+ channel, Cav3.1, were more highly expressed in the nodal tissues (especially in CN) than in the working myocardium (Figs. 3 and 5). Data from Cav1.3 knockout mice indicate the pivotal functional role of the Cav1.3 channel [27]. Furthermore, autoimmune antibodies to Cav1.3 result in atrioventricular block [28]. Cav3.1 knockout mice display marked disturbances of atrioventricular conduction, suggesting that this channel too plays a pivotal role [29]. In the CN, the three Ca2+ channels, Cav1.2, Cav1.3 and Cav3.1, were present in comparable amounts (Fig. 5). Although all Ca2+ channels may be present in all cells, an alternative explanation is the presence of multiple specialised cells within the CN with different Ca2+ channels. 4.3. Repolarization Ito has two major cardiac components, Ito,f and Ito,s, which show fast and slow, respectively, recovery from inactivation. Kv4.2 and Kv4.3 are responsible for Ito,f, whereas Kv1.4 is responsible for Ito,s. In the working myocardium of the human heart, Kv4.3 and Ito,f are believed to be the major isoforms. However, in the atrioventricular conduction axis of the
650
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
rabbit, both Ito,f and Ito,s can be present [7,30]. The same may be true in the human, because in the CN and PB, the expression levels of Kv1.4, Kv4.2 and Kv4.3 mRNAs were comparable (Fig. 6). It is interesting that in the nodal tissues the expression levels of Kv1.4, Kv4.2, frequenin, DPP6 and possibly KChAP were greater in the CN and PB than in the working myocardium (Fig. 6). In the rabbit, Ito is small or absent in the atrioventricular conduction axis [7], but this is unlikely to be the case in the human, because there was no downregulation of any of the Ito genes (Fig. 6). IK,ur is considered to be a major repolarizing current in the human atrium. Consistent with this, Kv1.5 was abundantly expressed in the AM (Fig. 6). The low levels of Kv1.5 mRNA measured in the compartments of the atrioventricular conduction axis, as well as the VM, suggest a negligible role for this current in these tissues. The delayed rectifier currents, IK,r and IK,s, are crucial for repolarization in the working myocardium in various species including the human. However, along the rabbit atrioventricular conduction axis only IK,r is present (IK,s is absent) [8]. Along the human atrioventricular conduction axis, although ERG (IK,r) cDNAN KvLQT1 (IK,s) cDNA, the level of KvLQT1 cDNA was still substantial and, therefore, IK,s as well as IK,r may be important. 4.4. Autonomic regulation Autonomic regulation of the atrioventricular conduction axis is important. Mediating the sympathetic effects are the β-adrenergic receptors, of which β1 (ADβR1) and β2 (ADβR2) are the main isoforms. We observed a uniform distribution of β1 receptor mRNA in the various tissues studied, albeit a trend towards a lower abundance in the PB (Fig. 7). This trend reflects a previous quantitative receptor autoradiography study, which showed that, in the human, the bundle of His had a lower (one fifth) β1 adrenergic receptor expression compared to the axis and surrounding working myocardium [31]. In the same study, β2 adrenergic receptor expression was reported to be uniform [31] which is consistent with the result in Fig. 7 (with the exception of INE, which was not investigated in the earlier study). Thus activation of both β1 and β2 adrenergic receptors in the INE and CN could affect conduction, whilst a lesser role is envisaged for the β1 adrenergic receptor in the bundle of His [31]. Our data are in agreement with optical mapping of the human atrioventricular conduction axis [22]. Parasympathetic effects are mediated by Kir3.1 and Kir3.4 (responsible for IK,ACh). In keeping with tight parasympathetic control of the human atrioventricular conduction axis, Kir3.4 at least showed greater expression in the CN and PB, suggesting these compartments are the main targets for slowing of atrioventricular conduction by increased parasympathetic tone. 4.5. Intracellular Ca2+-handling Within the working myocardium, Ca2+-handling plays a pivotal role in determining cardiac contractile performance. The same cannot be said regarding the cardiac conduction system. On the other hand, the atrioventricular conduction axis is well known to have inherent pacemaking properties and recently a “Ca2+-clock” has been shown to play an important role in pacemaking [19]. It is possible, therefore, that there is a unique signature of Ca2+-handling proteins in the axis. Our data show a trend towards reduced SERCA2a mRNA and RYR2 mRNA (and protein) in the INE and CN (Fig. 5; Fig. S9, Data Supplement), as occurs in the rabbit atrioventricular conduction axis [8] and human sinus node [11]. In the human, there was also a tendency for increased expression of the Na+-Ca2+ exchanger (NCX1 mRNA) in the CN and this could be important for pacemaking (Fig. 5). 4.6. Conduction velocity The major role of the atrioventricular conduction axis is to slow conduction from the atria to the ventricles to allow for ventricular
filling. The INE is held by some to be the slow pathway into the CN [5] and in the rabbit conduction along the INE is indeed slow [7]. In contrast, conduction through the His-Purkinje system (of which PB is the most proximal component) is fast to ensure synchronous activation of the VM. A major factor determining the conduction velocity of the action potential is the coupling conductance, which in turn is determined by the expression of connexins. mRNAs for the small conductance connexins, Cx31.9 and Cx45, were uniformly expressed in the various tissues (Fig. 2). However, Cx43 mRNA and protein, responsible for medium conductance channels, were poorly expressed in the INE and CN compared to the working myocardium (Figs. 1 and 2). Cx40 mRNA, responsible for large conductance channels, also tended to be more poorly expressed in the INE (Fig. 2). Conversely, Cx40 mRNA (and protein) was highly expressed in the CN and PB and this may be correlated with the high conduction velocity of the His–Purkinje system (Fig. 2; Fig. S9, Data Supplement). The high expression of Cx40 in the CN is not a surprise. We have previously reported that Cx40 protein is present in the CN of the rat heart in the ‘lower nodal cells’, although it is absent in the upper nodal cells [32]. In the rat heart, as in the human heart, Cx40 protein is also expressed in the bundle of His as expected [32]. In the mouse, knockout of Cx40 results in varying degrees of atrioventricular block [33]. Another major factor determining the conduction velocity is the upstroke velocity of the action potential, which in turn is determined by the expression of the cardiac Na+ channel, Nav1.5. Nav1.5 mRNA and protein were poorly expressed in the INE and CN (Figs. 3 and 4). Together the expression of connexins and Nav1.5 can explain the variation in the conduction velocity along the atrioventricular conduction axis. The mathematical modelling described above was used to predict the conduction velocity along the axis. The conduction velocity was measured along a string of 300 electrically coupled model cells; the coupling conductance between cells was scaled according to the expression of connexins (corrected for differences in single channel conductance). The computed conduction velocity was 53 cm/s in the AM; the conduction velocity was slowest in the INE (~4 cm/s) and highest in the PB (~100 cm/s) as expected (data not shown). 4.7. Dual pathway electrophysiology The INE is thought to constitute the slow pathway in the human, whereas the TA is thought to constitute the fast pathway. The INE has a similar (but not identical) gene expression profile to the CN consistent with the hypothesis that both regions contain true nodal or N cells. The mathematical modelling described above predicted the INE to show pacemaker activity like the CN and have an action potential profile (including slow upstroke) similar to that of the CN (data not shown). However, the INE is predicted to have a slower conduction velocity than the CN (data not shown). In contrast, the TA has a transitional gene expression profile consistent with the hypothesis that it is comprised of transitional AN cells. It is predicted not to show pacemaker activity and to have an action potential with an intermediate upstroke velocity and that conducts at an intermediate speed (data not shown). These results are consistent with the hypothesis that the fast and slow pathways have distinct electrophysiological behaviours. 5. Conclusions and clinical implications In summary, we have described a unique and complex expression of various transcripts in the human atrioventricular conduction axis. The data can explain (i) the slow conduction along much of the axis, (ii) the fast conduction along the bundle of His, (iii) the subsidiary pacemaker activity of the axis and (iv) the dual pathway electrophysiology of the axis.
I.D. Greener et al. / Journal of Molecular and Cellular Cardiology 50 (2011) 642–651
Disclosures None. Acknowledgments This study was supported by Medtronic Inc. Appendix A. Supplementary data Supplementary data to this article can be found online at doi:10.1016/ j.yjmcc.2010.12.017. References [1] Hucker WJ, Dobrzynski H, Efimov IR. Mechanisms of AV nodal excitability and propagation. In: Zipes DP, Jalife J, editors. From Cell to Bedside. 5th Edition. Saunders: Elsevier Science; 2009. [2] Anderson RH, Yanni J, Boyett MR, Chandler NJ, Dobrzynski H. The anatomy of the cardiac conduction system. Clin Anat 2009;22:99–113. [3] Moe GK, Preston JB, Burlington H. Demonstration of a dual A–V nodal conduction system in the isolated rabbit heart. Circ Res 2008;19:378–93. [4] Medkour D, Becker AE, Khalife K, Billette J. Anatomic and functional characteristics of a slow posterior AV nodal pathway: role in dual-pathway physiology and reentry. Circulation 1998;98:164–74. [5] Inoue S, Becker AE. Posterior extensions of the human compact atrioventricular node: a neglected anatomic feature of potential clinical significance. Circulation 1998;97:188–93. [6] Li J, Greener ID, Inada S, Nikolski VP, Yamamoto M, Hancox J, et al. Computer threedimensional reconstruction of the atrioventricular node. Circ Res 2008;102: 975–85. [7] Inada S, Hancox J, Zhang H, Boyett MR. One-dimensional mathematical model of the atrioventricular node including atrio-nodal, nodal, and nodal-his cells. Biophys J 2009;97:2117–27. [8] Greener ID, Tellez JO, Dobrzynski H, Yamamoto M, Graham GM, Billeter R, et al. Ion channel transcript expression at the rabbit atrioventricular conduction axis. Circ Arrhythmia Electrophysiol 2009;2:305–15. [9] Jackman WM, Beckman KJ, McClelland JH, Wang X, Friday KJ, Roman CA, et al. Treatment of supraventricular tachycardia due to atrioventricular nodal reentry, by rediofrequency catheter ablation of slow-pathway conduction. N Engl J Med 1992;327:313–8. [10] Hucker WJ, Fedorov VG, Foyil KV, Moazami N, Efimov IR. Optical mapping of the human atrioventricular junction. Circulation 2008;117:1474–7. [11] Chandler NJ, Greener ID, Tellez JO, Inada S, Musa M, Molenaar M, et al. Molecular architecture of the human sinus node: insights into the function of the cardiac pacemaker. Circulation 2009;119:1562–75. [12] Tellez JO, Dobrzynski H, Greener ID, Graham GM, Laing E, Honjo H, et al. Differential expression of ion channel transcripts in atrial muscle and sinoatrial node in rabbit. Circ Res 2006;99:1384–93. [13] Hoogaars VM, Tessari A, Moorman AF, de Boer PA, Hagoort J, Soufan AT, et al. The transcriptional repressor Tbx3 delineates developing central conduction system of the heart. Circ Res 2004;62:489–99. [14] Kreuzberg MM, Liebermann M, Segschneider S, Dobrowolski R, Dobrzynski H, Kaba R, et al. Human connexin31.9, unlike its orthologous protein connexin30.2 in the mouse, is not detectable in the human cardiac conduction system. J Mol Cell Cardiol 2009;46:553–9. [15] DiFrancesco D. Pacemaker mechanisms in cardiac tissue. Annu Rev Physiol 1993;55:455–72. [16] Verkerk AO, Wilders R, van Borren MM, Peters RJ, Broekhuis E, Lam K, et al. Pacemaker current (If) in the human sinoatrial node. Eur Heart J 2007;28:2472–8.
651
[17] Maier SKG, Westenbroek RE, Yamanushi TT, Dobrzynski H, Boyett MR, Catterall WA, et al. An unexpected requirement for brain-type sodium channels for control of heart rate in the mouse sinoatrial node. Proc Natl Acad Sci USA 2003;100: 3507–12. [18] Dobrzynski H, Boyett MR, Anderson RH. New insights into pacemaker activity. Promoting understanding of sick sinus syndrome. Circulation 2007;115:1921–32. [19] Bogdanov KY, Vinogradova TM, Lakatta EG. Sinoatrial nodal cell ryanodine receptor and Na+-Ca2+ exchanger: molecular partners in pacemaker regulation. Circ Res 2001;88:1254–8. [20] Krapivinsky G, Gordon EA, Wickman K, Velimirovic B, Krapivinsky L, Clapham DE. The G-protein-gated atrial K+ channel IKACh is a heteromultimer of two inwardly rectifying K+-channel proteins. Nature 1995;374:135–41. [21] Courtemanche M, Ramirez RJ, Nattel S. Ionic mechanisms underlying human atrial action potential properties: insights from a mathematical model. Am J Physiol 1998;275:H301–21. [22] Fedorov VV, Ambrosi CM, Hucker WJ, Glukhov AV, Foyil KV, Wuskell J, et al. Human AV junctional pacemaker shift due to cholinergic and adrenergic stimulations: optical imaging with a novel long wavelength voltage-sensitive dye. Circulation 2008;118:S520. [23] Dobrzynski H, Nikolski VP, Sambelashvili AT, Greener ID, Yamamoto M, Boyett MR, et al. The site of origin and molecular substrate of AV junctional rhythm in the rabbit heart. Circ Res 2003;93:1102–10. [24] Aslanidi OV, Sleiman RN, Boyett MR, Hancox JC, Zhang H. Ionic mechanisms for electrical heterogeneity between rabbit Purkinje and ventricular cells and tissue. Biophys J 2010;98:2410–31. [25] Wang DW, Viswanathan PC, Balser JR, George Jr AL, Benson DW. Clinical, genetic, and biophysical characterization of SCN5A mutations associated with atrioventricular conduction block. Circulation 2002;105:341–6. [26] Inada S, Mitsui K, Honjo H, Boyett MR. Why is Cav1.3 expressed in the sinoatrial node? Biophys J 2005. [27] Mangoni ME, Couette B, Bourinet E, Platzer J, Reimer D, Striessnig J, et al. Functional role of L-type Cav1.3 Ca2+ channels in cardiac pacemaker activity. Proc Natl Acad Sci USA 2003;100:5543–8. [28] Qu Y, Baroudi G, Yue Y, Boutjdir M. Novel molecular mechanism involving alpha1D (Cav1.3) L-type calcium channel in autoimmune-associated sinus bradycardia. Circulation 2005;111:3034–41. [29] Mangoni ME, Traboulsie A, Leoni AL, Couette B, Marger L, Le Quang K, et al. Bradycardia and slowing of the atrioventricular conduction in mice lacking Cav3.1/ a1G T-type calcium channels. Circ Res 2006;98:1422–30. [30] Mitcheson JS, Hancox JC. Characteristics of a transient outward current (sensitive to 4-aminopyridine) in Ca2+-tolerant myocytes isolated from the rabbit atrioventricular node. Pflugers Arch 1999;438:68–78. [31] Elnatan J, Molenaar P, Rosenfeldt FL, Summers RS. Autoradiographic localisation and quantitation of beta1- and beta2-adrenoceptors in the human atrioventricular conducting system: a comparison of patients with idiopathic dilated cardiomyopathy and ischemic heart disease. J Mol Cell Cardiol 1994;26:313–23. [32] Yoo S, Dobrzynski H, Fedorov VV, Xu SZ, Yamanushi TT, Jones SA, et al. Localization of Na+ channel isoforms at the atrioventricular junction, including the atrioventricular node. Circulation 2006;114:1360–71. [33] Hagendorff A, Schumacher B, Kirchhoff S, Luderitz B, Willecke K. Conduction disturbances and increased atrial vulnerability in connexin40-deficient mice analyzed by transesophageal stimulation. Circulation 1999;99:1508–15.
Glossary AM: Atrial muscle AVN: Atrioventricular node CN: Compact node INE: Inferior nodal extension PB: Penetrating bundle/bundle of His TA: Transitional area VM: Ventricular muscle