Anisotropy of electrotonus in the sinoatrial node of the rabbit heart

J Mol

Cell

Cardiol21,407F418

(1989).

Anisotropy

Lennart

N. Bouman,

of Electrotonus the Rabbit Jeroen

J. Duivenvoorden, Jongsma

in the Sinoatrial Heart Feliksas

Node

F. Bukauskas’

of

and Habo

J.

Department of Physiology, Academic Medical Center, University of Amsterdam, The Netherlands and ’ Kaunas Medical Institute, Kaunas, Lithuania, USSR (Received I August 1988, accepted in revisedform 28 December 1988)

L. N. BOWMAN, J. J. DUIVENVOORDEN, F. F. BUKAUSKAS AND H. J. JONGSMA. Anisotropy of Electrotonus in the Sinoatrial Node of the Rabbit Heart. Jour~l of Molecular and Cellular Cardiology (1989) 21, 407-418. In fibers of the sinoatrial node of isolated right atria of rabbits the decay of the electrotonic potential caused by intracellular current injection was measured in two directions: parallel to the crista terminalis and perpendicular to it. A Kf-perfused extracellular suction electrode was used to apply current pulses ( 10e5 A, 60 ms) to fibers located in the primary center of the SA node every fourth cardiac cycle at a fixed moment during diastole. The decay of the electrotonic spread was measured in a series of impalements on a straight line from the current source. Space constants were calculated by fitting single exponential curves to the data. Considerable regional differences in space constant values were found in either direction. Parallel to the crista terminalis the mean value was 529 f 446 pm (s.D., n = 7), perpendicular to it 306 k 295 pm (n = 12); the difference was not significant cf’ < 0.2). However, a significant anisitropy (P < 0.05) of the electrotonic spread was found when measurements were taken from small areas of the node. Large abrupt changes in the electrotonic potential within 200 pm were observed in the center of the node. These data indicate a non-uniformity of electrotonic spread in this part of the SA node. KEY WORDS: Electric coupling; crmtial conduction; Intranodal

Electrotonic interaction; Passive conduction; block; Intercellular coupling; Anisotropy.

Introduction The existence of preferential conduction pathways has been shown to be a common feature of the mammalian sinoatrial node. From mapping experiments in isolated hearts of several species it appears that the impulse travels from its origin in the center of the SA node with a much higher velocity to the crista terminalis than to the myocardium at the septal side of the node. This area is commonly activated via a circuitous route around the mass of nodal cells. There are, however differences in detail in the species which have been studied so far. In rabbit heart (Sano and Yamagishi, 1965; Bleeker et al., 1980), the preferential pathway is directed upwards, parallel to the crista terminalis, which is primarily activated in its cranial part. At the septal side there is a zone of conduction block (Bleeker et al., 1982). To activate the atria1 myocardium beyond this zone the impulse has to travel above and 0022-2828/89/040407

+ 12 $03.00/O

SA node;

Pacemaker

cells: Pret‘-

below the node coming from the lateral (cristal) side. In the guinea-pig (Opthof et al., 1985) the preferential pathway is directed transversely to the crista terminalis. In this species the conduction through the node to the septal side is also slow, but a complete block has not been demonstrated. Nevertheless, the myocardium at the septal side is activated via a circuitous route, which in this species passes the node only caudally. In the cat (Opthof et al., 1986) the directional differences in conduction velocity to the crista terminalis are not large. The crista is invaded over a large area simultaneously. Also in this species there is slow conduction to the septal side of the node, causing the activating impulse to travel around the mass of the node. The shape and orientation of the cells inside the SA node give some clue to the possible cause of this preferential conduction. In all species mentioned the cells have an elongated shape. The length may be up to 30 pm, while IC, 1989 Academic

Press Limited

408

L.N. Bouman er al.

the maximal diameter is 10 pm or less. Only for the rabbit heart (Bleeker et al., 1980) has the spatial orientation of the nodal fibers been described in detail. In light microscopy it can be seen that in the peripheral part the fibers are arranged in bundles with their long axis parallel to the crista terminalis; in the compact central zone of the SA node no preferential orientation can be seen. However in the EM picture also in this area the cells are somewhat elongated, with an orientation of their long axis roughly parallel to the crista terminalis (Masson-P&vet et al., 1979). In the other species fiber orientation has not been studied in detail, but it seems to be less uniform. The aim of the present study is to establish a link between the spatial orientation of the fibers in the rabbit SA node and the direction of the preferential conduction. This was done by measuring the passive distribution of intracellularly injected current pulses in different directions inside the spontaneously beating SA node. A considerable difference could be demonstrated in the decay of the voltage change caused by these current pulses between the long and short axis of fiber arrangment, which may be fundamental to differences in conduction velocity. Although the structure of the SA node differs strongly from a cable or a sheet, we attempted to calculate the space constant for the decay of electrotonic spread from our experimental data for comparison of our results with previous reports (Bonke, 1973a; Seyama, 1976; Bukauskas et al., 1977, 1982; Bleeker et al., 1982). However, although our values fit nicely with those found earlier, the large scatter of our data, especially in the direction perpendicular to the crista terminalis, raises serious doubts about the functional significance of this parameter.

SA node, the intercaval area, the crista terminalis and a sruali part of the right auricle. The preparation was placed in a tissue bath i 5 ml) which was perfused continuously with a Tyrode solution at a rate of 20 ml/min. The composition of the Tyrode was (in mM) : NaCl 130.6, KC1 5.6, NaHCO, 24.2, CaCl, 2, MgCl, 0.6, glucose 11.1 and sucrose 13.2. The solution was equilibrated with 95% O2 and 5% CO2 , the pH was buffered at 7.4 and the temperature was kept constant at 38 f 0.1 “C. Recording of electrical activity Transmembrane potentials were recorded differentially by means of the conventional microelectrode technique. The exploring microelectrode was mounted in a micromanipulator on a mechanical stage, of which the lateral movements could be read on vernier scales with an accuracy of 0.01 mm. The second, extracellular microelectrode was kept in the vicinity of the exploring electrode. The potential differences between the two microelectrodes were recorded ; this differential recording diminished the capacity artefacts of the current pulses. A bipolar surface electrode was positioned on the crista terminalis near the inferior caval vein. The recorded electrogram was used as a time reference. The two coordinates of each impalement were used to construct a two-dimensional activation map, as has been described previously (Bleeker et al., 1980). In short, this was done as follows : recordings of action potentials were made on numerous well-defined spots in the nodal area. The moment at which the transmembrane potential reached the half amplitude of the action potential was timed with respect to the fast deflection of the atria1 electrogram. On the average it took about 30 min to locate the site of earliest activation. Application

Materials

and Methods Preparation

Rabbits (2 to 2.5 kg) of either sex were anaesthetized i.m. with Hypnorm (10 mg fluanison and 0.2 mg fentanyl base per ml). The rabbits were respirated artificially during the operation. The heart was excised rapidly and a preparation was made which consisted of the

of polarizing

current

For the measurement of the decay of electrotonic potential a relatively large amount of current needs to be injected. This was possible by using a suction electrode, as introduced by Bonke (1973a, b). The modification described previously by Sakson et al. ( 1974) was adopted and consisted of perfusion of the lumen of the suction electrode with a potassium-rich (136.2 mM) Tyrode solution (NaCl replaced by

Electrotonus

KCl). In this way the membrane resistance under the electrode was diminished and the membrane was made inexcitable. The suction electrode [Fig. 1 (b)] consisted of a perspex holder (A) in which two concentric glass pipettes (B and C) were fixed with rubber rings. The lumen of this holder was in open connection with the lumen of pipette C ; suction applied at the nozzle drained the fluid out of this pipette. The long narrow ending of the micropipette B fitted into pipette C ; pipette B was slowly perfused (1 drop/s) with a potassium-rich Tyrode solution. For unipolar recording of electrical activity or for application of current an Ag/AgCl wire was led through a silicon rubber seal; this wire extended down into the pipette close to the opening of the tip (closer than shown in the figure). At the start of the experiment the tip of pipette C (inner diameter about 400 pm, outer diameter about 600 pm) was placed on the tissue, at the site of the primary center of the SA node as it was localized previously by mapping of action potentials, and suction of 10 kPa or less was applied. When the contact between the electrode and the tissue was optimal (resistance of about 25 kQ between the tip and the tissue bath) a biphasic action potential of a few millivolts could be recorded using a source follower amplifier (Fig. 1) . The slow perfusion of the inner pipette (B) with an isotonic KC1 solution was then started and within a few minutes the recorded action potential changed from biphasic into monophasic, and increased in amplitude 3 to 15 mV [Fig. l(a)]. The electrode signal was monitored continuously and in successful experiments it remained constant for hours. In those circumstances when apparently the electrode was sealed tightly to the tissue suri‘ace, the start of the potassium perfusion caused in most cases a prolongation of the interval of less than 100 ms [Fig. 1 (a)]. This bradycardia diminished in a few minutes, if the electrode contact was optimal. However, when the contact was leaky, a prolongation of the cycle length with more than 100 ms was observed. This was always reason for repositioning of the electrode. hyperpolarizing pulses were Constant injected into the preparation. These pulses (strength 10 PA, 60 ms duration) were delivered via a stimulus isolation unit by a

in

SA Node

409

ib)

preparation

FIGURE 1. (a) Suction electrode recordings of the electrical activity at the endocardial surface of the sinoatria1 node. When only suction was applied (upper trace) a small biphasic signal was recorded. When the electrode was perfused with a potassium-rich I 136.2 rn~I solution (lower trace) the biphasic signals changed into monophasic action potentials with larger amplitude. Thr increase in cycle length was temporary. (b) Schematic drawing of the construction of the potassium-perfused suction electrode. See text for explanation.

programmable stimulator with a source resistance of 1 Ma. The current pulse was applied about half-way through the diastolic depolarization phase at a coupling interval of 150 to 200 ms after the atria1 electrogram. The actual length of the coupling interval was adjusted according to the intrinsic cycle length. It remained constant throughout the experiment. Ten pulses were given in SUCcession. To avoid cumulative effects of successive pulses, a current pulse was applied only in every fourth cycle. At the end of the experiment the suction electrode was perfused with a 1% solution of Alcian Blue, which stained the site of the electrode as a small blue spot. The coordinates of the position of the

L. N. Bouman

410

ez al.

suction electrode were measured with the exploring microelectrode. In this way we estimated the distance between the center of the suction electrode and the nearest site of impalement (x0 or ya, for measurements parallel or perpendicular to the crista terminalis, respectively). Experimental procedure The suction electrode was placed close to the primary pacemaker area and suction was applied, which caused normally only a slight and temporary disturbance of the cardiac rhythm. Immediately the electrode signal was large enough, perfusion of the electrode with the potassium solution was started. This was subsequently interrupted when the electrode signal was not satisfactory and the electrode replaced. When the electrode signal was optimal a microelectrode impalement was made close to the suction electrode and a few current pulses were applied as a final check to ensure that a sufbciently large amount of current was injected intracellularly to cause a measurable and slowly rising voltage displacement on the transmembrane potential trace. When all conditions were fulfilled recording of intracellular potentials with every fourth cycle superimposed an electrotonic voltage change was started. Data processing The transmembrane action potentials were recorded on magnetic tape (Nagra TI instrumentation recorder, bandwidth 5 kHz), together with the atria1 electrogram. The net electrotonic potential was measured using a PDP-1 l/23 computer (Fig. 2). The average amplitude of ten successive current induced voltage changes was obtained by subtracting the normal diastolic slope in the interval preceding the test interval ( PClcontrol)from the diastolic slope with the current induced voltage change (V&. The extracellular voltage component (O,,,,) was measured during the experiment by withdrawing the exploring microelectrode a few micrometers out of the cell till we recorded the zero line potential including the sometimes considerable artefacts (0 to 15 mV) caused by the extracellular current flow. The net extracellular voltage component (O,,,, - Oeontrot) was subtracted

50 Ins FIGURE 2. Series of signals recorded to calculate the net electrotonic voltage Change (I’.,). l’,,,,, membrane voltage displacement due to a hyperpolarizing current pulse at the end of diastole, superimposed on the intracellularly recorded action potential; VcOn,ro,, action potential prior to the current injection; O,,,,, extracellularly recorded voltage change during current injection; recorded potential prior to the 0 EO,,,lO,, extracellularly current injection. All traces are an average of ten measurements.

from the average intracellular voltage change the net trans( VW,, - ~control) to yield membrane electrotonic potential l’,, . The amplitude of V,, was measured at the end of the voltage signal when it had reached a steady state (i.e. 55 ms after the onset of the current pulse). The electrotonic potential was plotted vs. distance from the current source ;ya ezpin;ntial curves of the format were fitted to the data with a least squares method. Gal~u~ation ofspace constants in the SA node When current is injected from a point source into an isotropic syncytium with a large surface to thickness ratio (two-dimensional syncytium), the steady state decay of the resulting membrane potential displacement with distance can be described by the differential equation (Jack et al., 1975) : a2

a2

1 v, +ygjy

1

v,=o

(1)

in which V, = membrane potential displacement at distance x from point source; L = dmj; d = thickness of syncytium (cm) ; R, = specific membrane resistance (0. cm2) ; Ri = specific internal resistance (lumped cytoplasmatic and junctional resistances) (Q. cm).

Elcctrotonas The solution of equation (Watson, 1966) :

(1)

is known

in SA Node

411

or Ko($?)=e.Ko(+)

(2) in which X,-,( ) = Bessel function of order 0 with imaginary argument; C = constant dependent on magnitude and shape of applied current; when .X is large (X > 2) equation (2) approximates to (Jack et al., 1975) :

From equation (6) I may be calculated when x0 and 1, are measured while K. is known. The thus calculated space constant was called 1LC. Results Intracellular

v,=c.

nl e(- 42) J

2x’

(3)

The slope of this curve approximates the slope of a single exponential curve. Therefore, at large values of x the decay of the electrotonic potential resembles the one-dimensional decay. We were able to measure that decay at a large distance from the current source, since we could inject large current pulses with the suction electrode. A space constant (A,) was then obtained by fitting an exponential function to the measured data. Since the measurements were not always carried out at a large enough distance from the current source, the use of equation (3) was not always justified. Therefore, we applied a correction method (Kukushkin et al., 1974), which is described here briefly. If V, = VO at x = x0, where x0 is the distance from the center of the suction electrode to the nearest point of measurement, equation (2) changes into

v, = c . K, y 0 substitution gives

of equation

(4) in equation

x vo vx= K,(x0/A) . K”0ij

(4) (2)

(5)

The experimentally obtained space constant A, equals the distance across which the electrotonic potential is decreased with a factor e. Substitution of this condition in equation (5) results in

vo vo . K-0x0 - +4 vLc= T = K-,(x,/n) ( 1 >

(6J

current injection into nodalJibers

Of the factors which may interfere with the establishment of a sufficiently tight contact between tissue and electrode both the quality of the electrode and of the tissue seemed to be prominent. The rim of the electrode had to be very smooth and completely free of clefts. This was achieved by fire-polishing of the electrode tip. From the tissue characteristics the thickness of the endocardium seemed to limit success the most. In hearts of rabbits weighing more than 2.5 kg we often failed. The effect of the current pulses on membrane resistance was measured from V/I diagrams. Hyperpolarizing pulses were given varying in strength from 1 to 10 x 10v6 A and the amplitude of the voltage displacement (V,,) was measured in a cell a few hundred micrometers from the current source. If the current pulses caused a progressive decrease of membrane resistance the relation between current strength and V,, would be nonlinear. Figure 3 shows the data from one out of three similar esperiments; in this case the distance of the impaled cell from the center of the current electrode was 310 pm. Current pulses of 60 ms duration were given 150 ms after the upstroke of the action potential. The membrane potential, measured at the end of the from -48.2 mV current pulse, increased (control) to values between -51.8 mV (2 PA pulse) and - 62.1 mV (10 PA pulse). Since hyperpolarizing current pulses were applied, 1, might be activated during the current pulse and hence cause a decrease of membrane resistance (DiFrancesco, 1985; van Ginneken, 1987). However, we considered the occurrence of such an interference unlikely for the following reasons. First, at potentials around -60 mV the amount of totally activated Zr is

L.N.JSouman

412

I 2

0.

I

I 4

1

Current

I 6

I

I 8

I

II IO

(PA)

FIGURE 3. Amplitude of electrotonic voltage change in a cell at a distance of 310 pm from the current source as a function of the strength of the injected current. The amplitude of the passively conducted hyperpolarizing voltage step is linearly related to the current intensity. The calculated function of the regression line is: y = 1.4.x + 0.43 (7 = 1.00).

very small (Irisawa and Noma, 1982 ; DiFrancesco et al., 1986). Secondly, Zr activates very slowly when it is activated on hyperpolarizing voltage clamp steps (DiFrancesco et al., 1986 ; van Ginneken, 1987) ; it reaches a steady state after more than 500 ms. Hence, in our experiments a hyperpolarization of the membrane to a potential even higher than -60 mV would only cause a negligible change of membrane resistance, since the duration of

et al. the applied current pulses was only 60 ms. Seyama (1976) showed in the same tissue that significant changes in membrane resistancr during application of hyperpolarizing current only occurred when the duration of the pulse exceeded 80 ms. Since we found a strong linear correlation between the applied current strength and amplitude of the voltage displacement (Fig. 3), we concluded that, in our experiments up to a current strength of 10 PA, there was no sign of change of the total membrane resistance. Therefore, a current strength of 10 PA was uniformly used for further experiments. Decrement of electrotonic potential in the sinoatrial node In Figure 4 data are presented of three experiments in which it was possible to measure the decay of the electrotonic potential in two directions at right angles : parallel (x) and perpendicular (y) to the crista terminalis while the suction electrode remained at the same position. In the lower panels the maximal rate of rise of the action potentials from the impaled fibers is plotted vs. the distance from the current electrode. A low value indicated that the impaled fiber belonged to the group of typical nodal fibers in the center of the sino-

c 1 (b)

20

'*iA\/

IO 30 L 0

.

&! 200

400

400

8( ‘0 Distance

200 from

400 current

6000 source

200

400

600

801

(pm)

FIGURE 4. Results of the recording of the electrotonic potential (V,,) in three preparations (upper panels) together with the maximal upstroke velocity of the impaled nodal fibers (lower panels). The curves in the upper panels are single exponentials fitted to the data. Perpendicular to the crista terminalis (e,y-direction) the voltage decay is stronger than parallel to it (A, x-direction). The scatter is also larger perpendicular to the crista terminalis. See text.

Electrotonas

TABLE

in

SA Node

1. Values of space constant (I) in two directions

(x andy)

Experiment no.

6 7 8 9 10 11 12 13

413

c.i. (ms) 200 200 150 200 200 200 200 150 150 200 150 150 200 200 150

I (A 300 320 600

446

290

210 470 430 600

116 221 252 438

139 719

23.7 44.7

7 8

138 116 178 199 157 265

9.4 19.8 44.9 45.2 5.1 7.2

8 8 5 8 6 13

121 183 71 140

66.9 33.9 1.7 22.1

7 4 5 5

26.7

500 500 490 450 300 550 510 310 340 180 430

531 608

0.9 33.9 11.5 29.2 13.7 44.6

iyc 1&m! 767 1053

589 156 129 208 239 194 322 128 290 333 833 160 1368

152 216 77 155

Abbreviations: c.i., coupling interval; Xc,, Y,, , distance (pm) between center of suction electrode and nearest point of measurement parallel or perpendicular to the crista terminalis, respectively; I,, , I,, , experimentally measured I parallel or perpendicular to the crista terminalis, respectively; V,, V,, coefficient of variation (lOO*SE/i,, and lOO*SE/&) ; n, number ofobservations; I,, and &, corrected values of I,, and I,,

atria1 node (Bleeker et al., 1980). From these data it can be seen that the measurements shown in Figure 4(b) were made more in the center of the node than those shown in Figure 4(a) and 4(c). In the upper panels the corresponding values of I’,, are plotted against distance as measured from the inner edge of the rim of the suction electrode. Although recordings in either direction were made as close as possible to the suction electrode, we could not get the ranges of recordings equally distant from the current source (i.e., x0 andye were unequal). In Figure 4(a) and 4(b) recordings in the y-direction were made closer to the current source than in the x-direction. Although the ranges of recordings in both directions show little overlap, it is clearly demonstrated that at equal distance from the current source FCXjel was considerably larger than Vtybel+ For instance, Figure 4(b) shows that at a distance of 190 pm from the current source, at which Vfxjel still exceeded 10 mV, V (yre, was already hardly detectable. Despite the large scatter of data, no membrane voltage displacement due to current injection was recorded in fibers impaled at a distance of more than 190 pm from the suction electrode perpendicular to the crista terminalis. Parallel to the crista terminalis however, V,, could still be measured at 350 pm from the current

source. In Figure 4(c) the situation is opposite to that in Figure 4(a) and 4(b) : here most of the measurements in the x-direction are closer to the current source than those in the -ydirection. Moreover, the overlap of ranges of recordings is also greater and the scatter of data perpendicular to the crista terminalis is less than in Figure 4(a) and 4(b). However, this figure also shows that at equal distances from the current source FCxje, is larger than V (rje, and it is even clearer that the spread of electrotonus showed an anisotropy. As we wanted to compare the observed decay of V,, in either direction with previous reports, we calculated space constants from our experimental data according to two models. Using the classic one-dimensional cable model (Bonke, 1973a; Seyama, 1976; Bleeker et al., 1982), we fitted single exponentials to the data to obtain what we called experimental space constants (A,, and A,,), although the large scatter of data, especially in the g-direction, invalidated almost any curve fitting. These space constants are given in Table 1 (the experiments shown in Figure 4(a), (b) and (c) are listed as experiments 10, 9 and 12, respectively). Using a twodimensional model for an anisotropic syncytium (Bukauskas et al., 1982), which reckons with the distance between the center

414

L. N. Bouman

of the suction electrode and the nearest microelectrode impalement [x,-, in equation (4), Methods] we arrived at values that were generally slightly larger (Table 1, A,, and A,,). We also calculated the coefficient of variation V (Table l), whose value depends on the scatter in the experimental data, as an indication of the validity of the space constants. In the direction perpendicular to the crista terminal& V was considerably higher (66.9, 33.9 and 22.1% for exp. 9, 10 and 12, respectively) than parallel to it (33.9, 11.5 and 13.7%). We concluded from this that the one-dimensional model (Bonke, 1973a; Seyama, 1976; Bleeker et al., 1982) as well as the two-dimensional model (Bukauskas et al., 1977, 1982) do not always describe the decay of the electrotonic potential in the direction perpendicular to the crista terminalis appropriately. In one more experiment the voltage decay could be measured in two directions at right angles (experiment 11, Table 1). In all these four experiments we found that the fall in electrotonic potential in the y-direction was much stronger than in the x-direction, yielding a space constant which was significantly shorter in the y-direction than in the xdirection (Mann-Whitney rank-sum test, P < 0.05). Apparently there is a strong anisotropy in the passive conduction within the SA node, which directs the passive current flow mainly parallel to the crista terminalis along the long axis of the node. In other experiments we were not so successful and we only got enough reliable impalements in one direction. The results of these experiments are also included in Table 1, which brings the nuber of calculated space constants in one or the other direction to 19. For each experimental space constant, 1,) we calculated the 1,. The values of the corrected space constant show a large variation, which indicates that within the SA node considerable regional differences in electrotonic spread exist. In the direction parallel to the crista terminalis the average value of the space constant as obtained through exponential curve fitting (A,,) was 333 f 2 15 pm ( &- s.D.), the average of the corrected space constant A,, was 529 + 446 pm. Perpendicular to the crista terminalis the average values were 226 ) 169 pm and 306 + 295 pm, for A,, and I YE, respectively. The mean II,, and ;1,, nor

et al.

-;; IO : II bl ,j t y‘(To 0 Distance

200 from current

1 j

,

400

sowee

(pm1

FIGURE 5. Results of the recording of the electrotonic potential (V,,) perpendicular to the crista terminalis (a), together with the maximal upstroke velocity of the impaled nodal fibers (b). The electrotonic potential does not show an exponential decay, instead it shows an increase with distance from the current source.

the mean A,, and A,, differed significantly (P < 0.2, Mann-Whitney rank-sum test). This is obviously due to the large variation in the measured space contants. In several experiments we observed abrupt changes in the decay of the electrotonic potential Vel within 200 pm in negative or even positive direction. An example of such abrupt V,, drops can be seen at the middle panel of Figure 4, in the measurements perpendicular to the crista terminalis (a). These V,, drops were considerably larger (> 3 mV) than the commonly found spread in the measurements and were observed more frequently in measurements perpendicular to the crista terminalis than parallel to it. Figure 5 shows the results of an experiment in which the decay of the V,, with distance did not even suggest a exponential course. In Figure 5(b), the maximal rate of rise of the impaled fibers did not exceed 5 V/s. From this we concluded that the measurements were done in the region of the primary pacemaker. Not only dropped the V,, abruptly within 200 @m, even at the same position in the preparation the V,, turned out to be different with 3 mV when it was measured in two separate impalements. Moving the microelectrode away from the current electrode, at some sites considerably higher electrotonic potentials could be found than close to the current electrode. In Figure 5 the three data close to the current electrode show this rise, instead of the expected exponential fall in V,, .

Electrotonus A possible technical explanation of these results could be a sudden change in the amount of intracellularly injected current by a loosening of the contact between the suction electrode and the tissue. This is not likely because the measurements were done in random order and we did not observe a continuous fall in electrotonic potential during the experiment. Furthermore we monitored the electrogram from the suction electrode continuously between the periods of current injection and its amplitude did not change during the experiment. Whenever during the experiments the contact of the suction electrode with the tissue deteriorated, the curved onset of V,, (Fig. 2) disappeared immediately. The shape of V,, became rectangular instead. This change of shape of P’,, did not happen either. Therefore we take these irregularities in the voltage decay as serious, being caused by an electrical non-uniformity of the tissue in the SA node. The scatter of data was the strongest in the area of the node in which the fibers had an upstroke velocity of less than 5 V/s, which is the very center of the SA node (Bleeker et al., 1980). Relatively good conducting pathways seem to be alternating with pathways in which there is a steep drop in current flow. The abrupt changes in the electrotonic potential were more often observed in they-direction than in the x-direction; apparently the non-uniformity is the strongest in the center of the node, perpendicular to the crista terminalis. Discussion It has already been shown repeatedly that the voltage displacement (V,,) as in Figure 2 caused by intracellular current pulses in the sinoatrial node diminishes rapidly with distance from the current source (Bonke, 1973a; Seyama, 1976; Bukauskas et al., 1977 ; Bleeker et al., 1982). Bonke (1973a) and Bleeker et al. ( 1982) limited themselves to a fitting of a single exponential only, as they accepted the one-dimensional cable as a suitable model for current spread in the SA node. Only Bukauskas et al. (1977, 1982) adopted as a more likely model a two-dimensional sheet, in which the electrotonic potential decay is described by a Bessel function. By means of a correction method as described by Kukushkin

in SA Node

415

et al. (1974) they calculated a space constant for a two-dimensional sheet without fitting a Bessel function to the data. The method consists of a correction of the space constant obtained by exponential curve fitting [see equation (6), Methods]. This equation contains as a variable the distance between the current source and the nearest point of measurement of the series of impalements for one space constant value. In spite of differences in methods of current injection and calculation the values of the space constant in the rabbit SA node which are reported from our own laboratory and from others are in agreement. The mean value of 529 pm in the direction parallel to the crista terminalis found in the present study is comparable to the previously found values; 465 pm (Bonke, 1973a), 468 pm (Bukauskas et al., 1982)) 520 pm (Bleeker et al., 1982)) 828 pm (Seyama, 1976). Our mean value for the space constant perpendicular to the crista terminalis of 306 pm is also in agreement with previous studies: 205 pm (Bukauskas et uf., 1982) and 310pm (Bleekeretul., 1982). However, these values suggest a homogeneity within the sinoatrial node which is not actually present. First, there is a considerable regional difference in the calculated space constant values. This is the cause of the large standard deviation of the average value of all experiments, which is in agreement with the large standard deviation reported by other authors. Secondly there is a large scatter of data within one small region, especially in the center of the SA node, as can be expected on the basis of the histological structure of the node. When one sees the small clusters of myocardial cells separated by coarse strands of connective tissue as observed by Opthof ( 1988), one would not expect a smooth and gradual fall of electrotonic potential. This lack of homogeneity of the SA node impairs the validity of the use models for intracellular current spread. Hence, in our opinion, the functional significance of the space constants derived from those models is ambiguous. One could even have doubts about the suitability of our method of current injection for this inhomogeneous tissue. The presence of layers of connective tissue could inhibit the current to reach the interior of the nodal cells. Indeed, in many experiments we had to

416

L. N. Bouman

replace the suction electrode one or more times before we succeeded in injecting current intracellularly. The presence of a varying amount of connective tissue under the tip of the suction electrode may be also a reason why we found a large difference in the amount of intracellularly injected current per experiment in spite of equal strength of the applied current. However, the actual voltage change at the site of the current electrode is of no importance for the measurement of the voltage decay and the calculation of the space constant, as long as no changes in membrane resistance are elicited. We think that these did not occur because of the linearity of the voltage-current relationship (Fig. 3). Regardless of the model which is adopted to describe the actual syncytium, our data show clearly that an intracellularly injected current diminishes rapidly with distance and that this decay is much stronger transversely through the SA node than in the direction parallel to the crista terminalis, thus showing a strong anisotropy. As we took many precautions to obtain reliable data (see Methods) we think that the scatter of our data also depicts the real situation. This means that there are different pathways through the SA node with different passive conduction properties, causing a nonuniform electrotonic spread. Strands with a relatively large space constant are intermingled with areas in which there is no electrical continuity. This also seems more to be the case in the direction perpendicular to the crista terminalis than parallel to it. But even for the conducting strands of the SA node it can be concluded that the space constant in the center of the node is lower than in atria1 muscle (Bonke 197313). In nearly all mammalian species studied up to now there is an anisotropy in the velocity of impulse propagation through the SA node showing a preferential conduction along the long axis of the node, parallel to the crista terminalis. This can now be explained as the consequence of the directional differences in electrical coupling described here. In addition, the low degree of coupling towards the septal side of the node may be one cause of the conduction block at this side of the node. We agree however with Bleeker et al. (1982) that this cannot be the only cause, since the same

et al.

degree of weak coupling exists in the direction to the crista, in which direction no block is found. The low degree of excitability of the atria1 fibers at the septal side as described by Bleeker et al. (1982), remains an important additional factor. It may be argued that the creation of a pathway for preferential conduction is not the only functional significance of the remarkable low degree of transverse coupling, especially in the center of the node. From model studies Uoyner and van Capelle, 1986) and experimental evidence (Kodama and Boyett, 1985 ; Opthof et al., 1987; Kirchhof et al., 1987) we know that electrotonic interaction between pacemaker and nonpacemaker tissue may interfere strongly with pacemaker activity. Take as a model a simple pair of cells, one a pacemaker fiber and one a normal atria1 cell. The current flowing between these cells will induce a diastolic depolarization in the atria1 cells but at the same time the rate of diastolic depolarization in the pacemaker cell will be depressed. In the model it was shown that when all cardiac fibers have the same degree of coupling a very large SA node would be necessary to induce spontaneous activity; a small SA node would be clamped to arrest by the large atrium. Returning to the rabbit SA node, we know that it has an ellipsoid shape with its long axis parallel to the crista terminalis (Bleeker et al., 1980). In this direction there is a large number of cells, gradually changing from typically nodal to purely atrial; in this condition a fairly good coupling is possible without the problem of an atria1 hyperpolarizing current. In the other direction the zone of transitional fibers is more narrow, especially at the septal side of the node where there is a rather abrupt transition into the atria1 myocardium. In our view the weak electrical coupling in this direction enables the pacemaker cells to change firing frequency without much interference from the atria1 muscle. The effect of an atria1 electrotonus on the firing rate and electrical activity of the SA node of the rabbit has been demonstrated repeatedly in recent years. In two studies (Kodama and Boyett, 1985; Opthof et al., 1987) it was shown that the intrinsic rate of diastolic depolarization of transitional fibers close to the atrium is higher than of the nor-

Electrotonus mally dominant pacemaker fibers in the center of the intact node. It is assumed that in the intact node diastolic depolarization is depressed in transitional fibers by the atria1 fibers close by. In another study (Kirchhof et al., 1987) it was demonstrated that in an intact rabbit SA node the site of impulse generation shifts to the transitional fibers at the cristal side when the atria1 myocardium is dissected from the node. The normally present weak coupling in the primary center predicts that any intervention which changes coupling resistivity of nodal fibers may induce a pacemaker shift to the transitional fibers. An agent which decreases coupling might allow the transitional fibers to develop their fast intrinsic rhythm by reduction of the hyperpolarizing atria1 current. An agent which increases coupling between fibers might depress the spontaneous activity in the normal primary center, which is relatively close to the septal margin of the

in SA Node

417

SA node and might allow fibers more to the crista to take over pacemaker function. In conclusion, our finding of an anisotropy and non-uniformity of electrotonic interaction in the rabbit SA node is in agreement with the histological structure of the node. The weak electrical coupling of nodal cells along the short axis of the node serves its function as an adjustable pacemaker and is of importance for the localization of the primary pacemaker.

Acknowledgements The authors thank Tobias Opthof and Antoni van Ginneken for assistance and advice during the experiments. Arnold Meijer, Baas Louter and Willem Schreurs are thanked for their valuable technical advice. This study was supported by grant nr. 516-074 from the Netherlands Organization for the Advancement of Research, NWO.

References BLEEKER WK, MACKAAY AJC, MASSON-P&ET M, OPTHOF T, JONGSMA HJ, BOWMAN LN (1982) Asymmetry of thr sinoatrial conduction in the rabbit heart. J Mol Cell Cardiol14: 633-643. BLEEKER WK, MACKAAY AJC, MASSON-P~VET M, OPTHOF T, JONGSMA HJ, BOWMAN LN, BECKER AE (1980) Funrtional and morphological organization ofthe rabbit sinus node. Circ Res 46: 1 l-22. BONKE FIM (1973a) Electrotonic spread in the sinoatrial node ofthe rabbit heart. Pfliigers Arch 339: 17 23. BONKE FIM (197313) Passive electrical properties ofatrial fibers ofthe rabbit heart. Pfliigers Arch 339: 1~ 15. BROWN HF (1982) Electrophysiology of the rabbit sinoatrial node. Physiol Rev 62: 505-530. B~KAUSKAS FF, GUTMAN AM, KISUNAS KJ, VETEIKIS RP (1982) Electrical cell coupling in rabbit sinoatrial node and atrium. Experimental and theoretical evaluation. In: Cardiac Rate andRhythm, edited by LN Bouman, HJ Jongsma, Thr Hague/Boston/London, Martinus Nijhoff Publishers, pp 195-214. BUKAUSKAS FF, VHTEIKIS RP, GUTMAN AM, MUTSKUS KS (1977) Intercellular coupling in the sinus node of the rabbit heart. Biofizika 22: 108-l 12. DIFRANCESCO D (1985) The cardiac hyperpolarizing-activated current I,. Origin and developments. Prog Biophys Mel Biol46: 163-183. DIFRANCESCO D, FERRONI A, MAZZANTI M, TROMBA C (1986) Properties of the hyperpolarizing-activated current ([,I in cells from the rabbit sinoatrial node. J Physio1377: 61-88. IRISAWA H, NOMA A (1982) Pacemaker mechanisms of rabbit sinoatrial node cells. In: Cardiac Rate and Rlpthm. rditrd by LN Bouman, HJ Jongsma, The Hague/Boston/London, Martinus Nijhoff Publishers, pp 33.-51. JACK JJB. NOBLE D, TSIEN RW (1975) E&tic Current Flow in Excitable Cells. Oxford, ClarendonPress. JOYNER RW, VAN CAPELLE FL.1 (1986) Propagation through electrically coupled cells. How a small SA node drives a large atrium. Biophys J 50: 1155-l 164. I K~RCHHOF CIHT. BONKE FIM. ALLEWE MA, LAMMERS WEEP (1987) The influence of the atria1 myocardium on impulse formation in the rabbitsinus node. Pfligers Arch 410: 1981203. KODAMA I, BOYETT MR (1985) Regional differences in the electrical activity of the rabbit sinus node. Pfliigrs Arch 404: 214-226. KUKUSHKIN NI, BUKAUSKAS FF, SAKSON NYE ( 1974) Biophysics of complex systems. Mathematical models 11. ,Model of a two-dimensional anisotropic syncytium. Biofizika 19: 888-893. AM~~~~~-Ph~~~ M, BLEEKER WK, MACKAAY AJC, BOUMAN LN, HOUTKOOPER JM (1979) Sinus node and atrium cells from the rabbit heart: a quantitative electron microscopic description after electrophysiological localization. J Mol Cell Cardiol 11: 555-568. OPTHOF T (1988) The mammalian sinoatrial node. Cardiovasc Drugs Ther 1: 573-597. OPTHOF T, DE JONGE B, MACKAAY AJC, BLEEKER WK, MASSON-P&VET M, JONGSMA HJ, B~UMAN LN [ 1985) Functional and morphological organization of the guinea-pig sinoatrial node compared with the rabbit sinoatrial node. J Mel Cell CardiolI7: 549-564. .,

.,I

418

L. N. Bouman

et al.

OPTHOF T, DE JONGE B, MASSON-P&ET M, JONGSMA HJ, BOUMAN LN (1986) F unctional and morphological orgamzation of the cat sinoatrial node. J Mel Cell Cardiol18: 1015- 1031. OPTHOF T, VAN GINNEKEN ACG, BOWMAN LN, JONCSMA HJ (1987) The intrinsic cycle length in small piews isolatrd from the rabbit sinoatrial I.. de. J Mel Cell Cardioll9: 923-934. SAKSON MY, BUKAUSKAS FF, KUKUSHIN NI, NASONOVA VV ( 1974) Study of electrotonic distribution on the surface of cardiac structures. Biofizika 19: 1043-1049. SANO T, YAMAGISHI S (1965) Spread ofexcitation from the sinus node. Circ Res 26: 423-430. SEYAMA I (1976) Characteristics of the rectifying properties of the sinoatrial node cell of the rabbit. J Physiol 255: 379-397. VAN GINNEKEN ACG (1987) Membrane currents in mammalian cardiac pacemaker cells. Thesis, University of Amsterdam. WATSON GN (1966) A Tre~ieatrseon the Theoheory of Bessel Functions. Cambridge, Cambridge University Press.