Function and structure of the mouse sinus node: nothing you can see that isn’t shown

See articles by Verheijck et al. [1] (pages 40–50) and Mangoni and Nargeot [2] (pages 51–64) in this issue.

In this issue of Cardiovascular Research two papers are published on the mouse sinus node [1,2]. One is about the activation pattern, the characteristics of the leading pacemaker area and the connexions connecting the sinus node cells with each other and with the surrounding atrial working myocardium [1]. The other paper focuses on the individual membrane currents in single cells isolated from the sinus node including the effects of noradrenaline and acetylcholine [2].

1 Structure

The structure of the mouse heart in late fetal stages was recently described in a paper by Webb et al. [3], but the term ‘sinus node’ was missing in this paper. From this perspective the paper by Verheijck et al. [1] fills a gap in our knowledge on the structure of the murine sinus node. The first very practical piece of information is the observation of the authors that cutting the crista terminalis at the orifice of the superior vena cava, which is normally performed to expose the intercaval region as a flat surface for easier access with microelectrodes, results in marked irregularities. The background of this peculiarity compared with other species remains unclear, but it is useful to know for future studies. We learn from the paper that the area with sinus node type of action potentials is limited to a long axis of 300 μm (parallel to the crista terminalis, albeit that the morphological sinus node cells could be observed over a somewhat longer axis of 500 μm) and a short axis of 150 μm (perpendicular to the crista terminalis). This is very small indeed; the long axis of the rabbit sinus node is 7 mm (see Ref. [4] for structural data in rabbit and several other mammalian species).

The second paper in this issue of Mangoni and Nargeot [2] provides data on membrane currents in the murine sinus node. The mice were different, C57BL6/J mice of 20–23 g rather than Swiss mice of over 40 g [1], and the experiments were unfortunately performed at 26°C. Although this was done to permit comparison with currents expressed in artificial systems, this hampers physiological interpretation.

After cell isolation from the ‘sinus node area’ Mangoni and Nargeot [2] observed 3% ‘spider cells’, 28% ‘spindle cells’, 6% ‘elongated spindle cells’ and an overwhelming 67% ‘atrial cells’. For obvious reasons, there must be a small error in these numbers, but the large number of atrial cells should be looked at with caution. The occurrence of atrial cells in previously electrophysiologically mapped sinus nodes of the rabbit [5] has provoked a debate on their role [6] in impulse conduction from sinus node towards the periphery and the atrium in terms of a gradient model and a mosaic model. As long as we do not exactly know where atrial cells in the very center of the sinus node are located, such a debate is premature. It makes a big difference whether atrial cells are at the epicardial or endocardial side of the sinus node or whether they are really intermingled between nodal cells at large quantities. Mangoni and Nargeot [2] report 67% atrial cells in the sinus node area, but Verheijck et al. [1] have not reported atrial cells within the nodal area. It may be that the 67% atrial cells result purely from the outer zones of the tissue used for cell isolation. Thus, it cannot be excluded that the 67% atrial cells reflect isolation of cells from the working atrial myocardium.

2 Function

The sinoatrial conduction time, the time needed for conduction from the first discharging sinus node cells to the atrial working myocardial cells is only 4 ms. As in other mammalian sinus nodes there is a unifocal type of pacemaking and preferential conduction towards the crista terminalis. Thus, the murine sinus node appears as a miniature ‘normal’ mammalian sinus node in this respect. In Table 1 we have summarized the electrophysiological parameters of the leading pacemaker cells from the mouse sinus node [1] in comparison with sinus nodes from other mammalian species. The cycle length of the isolated murine sinus node is only 144 ms, substantially shorter than in other mammalian species. Fig. 1 shows the relation between body weight and the expected resting cycle length according to previously described allosteric equations (solid line: see references in Ref. [13]). This concerns data on innervated hearts. We have added further data on the cycle length of the isolated sinus node (open circles) and on measured resting cycle length (filled circles). For dog and man we have added the resting cycle length under complete autonomic blockade by propranolol and atropine (open triangles). Interestingly, the data of Verheijck et al., [1] in combination with data on the in vivo cycle length in mice [14,15], indicate that the mouse has a predominant sympathetic tone as other rodents and lagomorphs. In dog and man the opposite is observed: a predominant vagal tone when we accept the resting cycle length under complete autonomic blockade as a surrogate for the cycle length of the isolated sinus node. For monkeys (Macaca fascicularis: different body weights in the two studies) and pigs, Fig. 1 indicates a predominant sympathetic tone. However, the cycle length of the isolated sinus node was measured with a superfused preparation in the former [10]. With a thickness of the preparation of 400 μm, including an endocardium of 77 μm and an epicardium of 136 μm, superfusion may have been borderline. The cycle length of the porcine sinus node was measured with superfusion in combination with perfusion through the sinus node artery [11], but the outlier in the graph casts doubt on this observation.

Table 1

Electrophysiological parameters of the leading pacemaker site in seven mammalian species

Species	Bodyweight	Cycle	Sinoatrial	Action	Upstroke	Maximum	Action	Diastolic	Ref.
	(kg)	length	conduction	potential	velocity	diastolic	potential	depolarization rate
		(ms)	time	amplitude	(V/s)	potential	duration	(mV/s)b
			(ms)	(mV)		(mV)	(ms)a
Mouse	0.043	144	4	56	6.2	−51	82	117	[1]
Guinea-pig	0.540	275	12	67	6.2	−57	137	121	[7]
Rabbit	3.0	386	32	66	1.9	−61	180	83	[7,8]
Cat	3.5	418	29	60	2.8	−59	202	71	[9]
Monkeyc	4	634d	31	65	3.6	−55	286	37	[10]
Pig	17	894	50	63	2.7	−59	393	21	[11]
Man	–	2300e	–	72	5.4	−62	309	–	[12]

a

Action potential duration was measured from halfway of the upstroke of the action potential to the moment of the maximum diastolic potential.

b

Diastolic depolarization rate was measured over the first 200 ms after the moment of the maximum diastolic potential. In the guinea pig this period was 100 ms; in the mouse 20 ms.

c

The monkey species was Macaca fascicularis.

d

The cycle length was calculated from five of the individual sinus nodes in Table 1 from Ref. [10].

e

The cycle length in the four human sinus nodes were measured in superfused preparations. It should be noted that porcine sinus nodes require perfusion through the sinus node artery to maintain normal pacemaker function. It is therefore improbable that the cycle length of 2300 ms as observed in Ref. [12] represents the normal intrinsic human sinus node cycle length.

Fig. 1

Expected relation between body weight and resting cycle length according to allosteric equations (solid line: see for references Ref. [13]). Data on the cycle length of the isolated sinus node (open circles) have been taken from the references indicated in Table 1. Rat: unpublished observations (Opthof). Dog, pup [21]. Data on the resting cycle length (filled circles) have been taken from Ref. [22]. Data on cycle length under complete autonomic blockade (open triangles): man [23], dog [24]. The cycle length of the adult human sinus node [12] has not been included in this graph for reasons explained in the legend of Table 1.

Why do small animals have sinus nodes with shorter cycle length? Table 1 shows that there are hardly any differences in action potential amplitude or maximum diastolic potential. Also, the upstroke velocity of the action potentials are of the same order, suggesting only small differences in density of the L-type Ca²⁺ current which is responsible both for the upstroke of the sinus node action potential and — in part — for the diastolic depolarization phase. The main differences concern action potential duration and diastolic depolarization rate. The relation between cycle length and action potential duration can be regarded as a ‘chicken–egg’ problem and can only be solved by driving the sinus node of different species at the same cycle length. This cannot be performed in practice. One cannot drive a murine sinus node at 500 ms and a rabbit sinus node will not respond to cycle lengths shorter than about 250 ms as is demonstrated very elegantly by its behaviour during atrial fibrillation [16].

Individual membrane currents in the sinus node have been reviewed at the cellular level in detail [17] as well as in relation with the function of the intact node [6]. The large differences in action potential duration between the species are suggestive for substantial differences in density or in kinetic behaviour of the delayed rectifier currents. Mangoni and Nargeot [2] have assessed primarily the density and kinetics of the pacemaker current I_f, which was only observed in cells of the nodal type, not in atrial cells. As stated in the previous section the measurements were performed at 26°C. This explains the very long cycle length of 322 ms and long action potential duration of about 200 ms (deduced from a diastolic interval of 125 ms). The current density of I_f was 18 pA/pF at −120 mV. This is two to three times larger than in the rabbit sinus node as reported previously by Wilders et al. [18] and Honjo et al. [19]. Interestingly, Mangoni and Nargeot did not observe a difference in current density between larger and smaller cells in agreement with Wilders et al. [18], but in contrast to Honjo et al. [19]. The problem with the low temperature should be kept in mind. For this reason we do not compare the kinetics of I_f. The density of I_Kr assessed as a tail current from +20 to −40 mV was 1.3 pA/pF. In rabbit sinus node this density has been estimated at 0.26 pA/pF at the moment of the maximum diastolic potential (Veldkamp, personal communication), but this was measured with simultaneous recording of single channel recordings of I_Kr and action potentials in spontaneously active rabbit sinus node cells. Unfortunately, the data of Mangoni and Nargeot [2] cannot explain the very short action potential in the murine sinus node.

Mangoni and Nargeot [2] have added 10 μM noradrenaline and 5 μM acetylcholine to the cells and observed (at −90 mV) an increase in I_f current by 21% in response to the former and a decrease by 37% in response to the latter. Unfortunately, there are no data on the spontaneous cycle length. Although there seems agreement on the role of I_f as an effector of the chronotropic response to sympathetic stimulation, it is an open question whether or not the decrease in I_f has a physiological meaning. Firstly, the conductance of the membrane of sinus node cells increases, not decreases in response to mild vagal stimulation at least in rabbits [20]. This points to opening of channels rather than closure. Secondly, the I_K-Ach channel was not assessed (see Ref. [6] for details).

3 Intercellular coupling

Verheijck et al. [1] describe the presence of connexin45 in the center of the sinus node. They did not observe a gradient in labeling towards the periphery. However, at the epicardial side of the sinus node they describe an area with both connexin40 and connexin43 labeling. The functional role of this epicardial strand remains obscure (see Fig. 5 in Ref. [1]). Although there seem to be species differences between the types of connexins within the cardiac conduction system (see for review [25]), the observation of absence of connexin43 in the center of the sinus node, which is the major connexin in ventricular working myocardium, fits within a general picture in most species. The problem with absence of connexin43 in the sinus node can be summarized with ‘Nothing you can see that isn’t shown’ as John Lennon sang in ‘All you need is love’. If gap junctions are too small to be effectively labeled by the immunohistochemical technique, we may look at the limitation of a technique rather than at true absence of a structure. At present we still do not know the exact role of different types of connexins. Heteromeric (more than one type of connexins within the same hemichannel) and heterotypic gap junctions (different connexin composition of the two docking hemichannels) do occur [25]. They may play a pivotal role in the electrical communication between sinus node and surrounding atrium, which should within a large frequency domain maintain proper function: successful pacemaking of the heart with avoidance of sinus node standstill by the hyperpolarizing influence of the surrounding atrium. Heterotypic gap junctions might provide the means for differences in centrifugal and centripetal electrical communication as has been suggested previously [26]. In the study of Verheijck et al., there were no indications for cells with either combined connexin45 and connexin40 labeling or combined connexin45 and connexin43 labeling. In the rabbit sinus node there are reports on combined connexin46 and connexin40 labeling within the transitional zone [27], but also on combined connexin45 and connexin43 labeling [28]. In a paper in press Honjo et al. [29] demonstrate that in isolated rabbit sinus node cells larger cells are positive for immunolabeling against connexin43, whereas smaller cells are positive for connexin45. Also, they showed gradual changes in density with cellular dimensions (positive correlation with connexin 43 labeling and negative correlation with connexin45 labeling). Double labeling in larger sinus node cells was observed as well, albeit at limited frequency. Verheijck et al. [1] emphasize the presence of very sharp transitions in connexin labeling. Finally, there is a connective tissue barrier which is suggestive for shielding of a large part of the border zone with the crista terminalis. This makes the role of the epicardial strand with connexin40 and connexin43 at the epicardial side of the sinus node of special interest (see Fig. 5 in Ref. [1]). It might be that in such small sinus nodes extra protection factors are needed for proper function, but this must await further research. Not an original sentence to end this Editorial, but true ……