Cardiovascular Research 52 (2001) 40–50
www.elsevier.com / locate / cardiores
Electrophysiological features of the mouse sinoatrial node in relation to
connexin distribution
a , ,1
b ,1
a
a
E. Etienne Verheijck * , Marjan J.A. van Kempen , Mike Veereschild , Joost Lurvink ,
Habo J. Jongsma b , Lennart N. Bouman a
a
Academic Medical Center, University of Amsterdam, Task Force Heart Failure and Aging, Department of Physiology, P.O. Box 22700,
1100 DE Amsterdam, The Netherlands
b
University Medical Center Utrecht, Department of Medical Physiology, 3584 CG Utrecht, The Netherlands
Received 29 January 2001; accepted 17 May 2001
Abstract
Objective: The sinoatrial (SA) node consists of a relatively small number of poorly coupled cells. It is not well understood how these
pacemaker cells drive the surrounding atrium and at the same time are protected from its hyperpolarizing influence. To explore this issue
on a small tissue scale we studied the activation pattern of the mouse SA node region and correlated this pattern with the distribution of
different gap junction proteins, connexin (Cx)37, Cx40, Cx43 and Cx45. Methods and Results: The mouse SA node was
electrophysiologically mapped using a conventional microelectrode technique. The primary pacemaker area was located in the corner
between the lateral and medial limb of the crista terminalis. Unifocal pacemaking occurred in a group of pacemaking fibers consisting of
450 cells. In the nodal area transitions of nodal and atrial waveform were observed over small distances (|100 mm). Correlation between
the activation pattern and connexin distribution revealed extensive labeling by anti-Cx45 in the primary and secondary pacemaker area.
Within these nodal areas no gradient in Cx45 labeling was found. A sharp transition was found between Cx40- and Cx43-expressing
myocytes of the crista terminalis and the Cx45-expressing myocytes of the node. In addition, strands of myocytes labeled for Cx43 and
Cx40 protrude into the nodal area. Cx37 labeling was only present between endothelial cells. Furthermore, a band of connective tissue
largely separates the nodal from the atrial tissue. Conclusions: Our results demonstrate strands of Cx43 and Cx40 positive atrial cells
protruding into the Cx45 positive nodal area and a band of connective tissue largely separating the nodal and atrial tissue. This
organization of the mouse SA node provides a structural substrate that both shields the nodal area from the hyperpolarizing influence of
the atrium and allows fast action potential conduction from the nodal area into the surrounding atrium.  2001 Elsevier Science B.V. All
rights reserved.
Keywords: Cell communication; Conduction system; Gap junctions; Sinus node
This article is referred to in the Editorial by T. Opthof
( pages 1 – 4) in this issue.
1. Introduction
The relatively small sinoatrial (SA) node has to drive a
*Corresponding author. Tel.: 131-20-566-4670; fax: 131-20-6919319.
E-mail address: [email protected] (E.E. Verheijck).
1
E. Etienne Verheijck and Marjan J.A. van Kempen contributed equally
to this study.
large mass of surrounding atrial muscle without being
electronically suppressed by its hyperpolarizing influence.
Several different hypotheses concerning the arrangement
between the SA node and the surrounding atrium have
been formulated to explain successful impulse formation
and conduction.
In a series of studies on small isopotential cell groups of
rabbit nodal tissue, the research groups of Kodama and
Boyett [1] investigated regional differences in membrane
properties to explain the complex sino-atrial interaction.
Time for primary review 29 days.
0008-6363 / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved.
PII: S0008-6363( 01 )00364-9
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
Based on the apparent distribution in densities of various
ionic currents, a model was postulated in which a gradual
transition in membrane properties from the SA node
towards the atrium represents part of the substrate to allow
impulse conduction towards the atrium.
In a mathematical study, Joyner and van Capelle [2]
studied the sino-atrial interaction by coupling models of
SA node cells with models of atrial cells. They suggested
that a gradual increase in intercellular coupling might be
required for normal impulse formation together with
impulse conduction from the SA node to the atrium.
In a qualitative ultrastructural study of the rabbit SA
node, it appeared that gap junctions became more frequent
and larger in size in the peripheral zone compared to the
central zone [3]. However, quantitative evidence of a
gradual increase in number or size of gap junctions in the
direction of the surrounding atrium is lacking [4]. In
guinea pig, ten Velde et al. [5] examined the spatial
distribution of connexin (Cx)43 by immunofluorescence to
visualize the structural substrate for impulse propagation
from the SA node to the atrium. They observed no gradient
of Cx43 labeling between the SA node and the atrial
muscle. Instead, they found protrusions of strands of atrial
cells into the primary pacemaker region. A similar interdigitation of nodal cells and atrial cells in the periphery of
the SA node was reported in bovine, rat and human hearts
[6]. In the rabbit SA node, Coppen et al. [7] found some
restricted interdigitation of Cx45-positive nodal cells and
Cx43-positive atrial cells. Furthermore, we found evidence
of a gradual increase in density of atrial cells going from
the primary pacemaker area towards the atrium [8].
The mouse was selected as a model to further explore
the structural substrate that allows a very small mammalian
SA node, located in the vicinity of the atrial tissue, to
overcome the suppressive atrial influence. Therefore, we
determined the spatial distribution of Cx37, Cx40, Cx43
and Cx45 in the nodal and atrial area of the mouse heart.
Electrophysiological mapping experiments were performed
to accurately localize the primary pacemaker region in
relation to subsequent light microscopic and immunohistochemical localization of the different connexins. Our
results demonstrate that: (1) strands of Cx43 and Cx40
labeled atrial cells protrude into the central nodal area of
the mouse heart, (2) a band of connective tissue largely
separates the nodal and atrial area, and (3) transitions of
nodal and atrial waveform occur over small distances
(|100 mm) in the central nodal area. This organization of
the mouse SA node forms a structural substrate for fast
impulse transmission from the SA node into the atrial
tissue, and reduction of the load that the large atrial mass
imposes on the SA nodal tissue.
In addition, this study describes for the first time
function, electrophysiology and connexin distribution in
the mouse SA node and thereby supplies essential information needed to understand the accumulating cardiac
electrophysiological data on transgenic mouse models.
41
2. Methods
2.1. Animals
Twenty-two Swiss mice weighing 41–46 g, of male sex
only, aged 10–12 weeks, were used. Animal care was in
accordance with institutional guidelines and conforms to
the Guide for Care and Use of Laboratory Animals
published by the US National Institutes of Health (NIH
Publication No. 85-23, 1996).
2.2. SA node preparation
2.2.1. Electrophysiological mapping of the SAN
Mice were anaesthetized by ether inhalation. After
cervical dislocation hearts were excised and immersed in a
‘standard’ solution containing (mM): 130.6 NaCl, 24.2
NaHCO 3 , 5.6 KCl, 2.2 CaCl 2 , 0.6 MgCl 2 , 11.1 glucose,
13.2 sucrose, saturated with 95% O 2 and 5% CO 2 at room
temperature. The whole heart was pinned to a perforated
Teflon block to excise the right atrium under a surgical
microscope. The 5-ml tissue bath was perfused with
standard solution at a rate of 10 ml / min.
After removal of the ventricles the right atrium was
opened to expose the crista terminalis, the intercaval area
and the interatrial septum. The preparation was trimmed,
leaving only the SA node region and surrounding atrium.
This preparation was pinned by small stainless steel pins to
the Teflon block with the endocardial side exposed up.
After mounting the preparation the temperature of the
perfusion fluid was slowly raised and kept constant at
3760.38C. Transmembrane potentials were recorded by
conventional glass microelectrodes filled with 2.7 M KCl
and 2 mM K-citrate and connected through an agar-Tyrode
bridge and an AgCl wire to the input of a source follower.
Electrode resistance ranged between 15 and 30 MV. For
the construction of activation maps, impalements were
made 0.2 mm apart and in the primary pacemaker area
down to 0.05 mm apart. A bipolar silver wire electrode
(interelectrode distance 0.3 mm) was placed on the
trabecula of the right auricle, close to the crista terminalis.
This electrode provided a surface electrogram of which the
first deflection served as a time reference for the determination of the sinoatrial conduction time (SACT). The
activation moment of a cell was defined as the moment
that the voltage was halfway between maximum diastolic
potential (MDP) and action potential overshoot, and was
timed in respect to the time reference. A reliable activation
map could be made when the activation moment of at least
50 cells could be measured in an area of 234 mm.
2.3. Data processing
Data were acquired at a sample rate of 5 kHz and were
off-line analyzed using custom-written software (J. Zegers,
University of Amsterdam, The Netherlands). The following
42
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
parameters were measured: cycle length (CLl), SACT,
MDP, duration of action potential at half amplitude
(APD 50 ), duration of action potential between 50% depolarization and 100% repolarization (APD 100 ), maximum
upstroke velocity (V~max ), action potential amplitude (APA),
and diastolic depolarization rate during the first 20 ms after
maximum diastolic potential (DDR). For statistical analysis we used the mean values of the parameters of 10
subsequent action potentials. Results are presented as
mean6S.D.
2.4. Light microscopy
In four experiments the activation map was correlated
with a histological reconstruction. To align the electrical
map and histological staining pattern, the tissue was
marked with 10- to 50-mm-diameter Alcian blue dots after
the activation map was constructed. Therefore, the electrode was backfilled with 1% Alcian blue in 0.5 mM
sodium-acetate (pH 4.0) at the end of the mapping
procedure. The tip of the electrode was broken to lower
resistance and dots were obtained iontophoretically by
applying rectangular 0.5 mA pulses (500 Hz; duration
100–300 ms) for 10 s. Dots were placed at an interval of
200 mm along lines parallel and perpendicular to the crista
terminalis. In total, about 10 dots were placed. The
preparation, together with its Teflon bed was immersed in
buffered formalin for fixation. After dehydration and
removal of the Teflon support two staining methods were
used: (1) two preparations were embedded in paraffin wax
and sectioned (thickness 7 mm) perpendicular to the crista
terminalis. These sections were stained according to the
van Giesson elastic tissue staining method. (2) The other
two preparations were embedded in Tecnovit 7100 plastic
and sectioned in thin sections (2 mm) perpendicular to the
crista terminalis. These sections were stained with hematoxylin and Sirius Red to assess the connective tissue
distribution.
2.5. Immunohistochemistry
After electrophysiological mapping, four other preparations were frozen in liquid nitrogen. The Teflon beds were
removed and serial cryosections of 8-mm thickness were
collected on 3-aminopropyltriethoxysilane- (AAS, Sigma
A3648) coated slides, air-dried and stored at 2808C. Prior
to incubation, slides were equilibrated to room temperature
and rehydrated in phosphate buffered saline (PBS) for 5
min. The unfixed cryosections were incubated in 0.2%
Triton X-100 in PBS for 1 h and subsequently pre-incubated in 2% bovine serum albumin (BSA) in PBS for 30
min. Sections were incubated overnight at room temperature with appropriate dilutions of primary antibodies in
PBS including 10% normal serum from the host of the
secondary antibody.
For labeling of gap-junction proteins, antibodies raised
against synthetic peptides of Cx37, Cx40, Cx43 and Cx45
were used. The anti-Cx37 (3 to 5 mg / ml), specific for
Cx37 only [9], was raised in rabbit and directed against
amino acids 315–331 of rat Cx37. The anti-Cx40 (3 to 5
mg / ml), specific for Cx40 only [10], was raised in rabbit
and directed against amino acids 335–356 of rat Cx40. The
Cx43 specific anti-Cx43 (1 to 3 mg / ml) [11], was raised in
rabbit and directed to amino acids 314–322 of rat Cx43.
The anti-Cx45 (Alpha-Diagnostics) was raised in rabbits
against amino acids 285–298 of dog Cx45 [12]. Myocardial cells were specifically labeled with monoclonal antibodies directed to desmin (MON 3001-clone 33, 1:50,
Monosan).
After about 16 h of incubation with primary antibodies,
the slides were thoroughly rinsed three times for 5 min
with PBS and again incubated with 2% of BSA in PBS for
30 min. Fluorescence labeling was carried out in 2.5 h,
using either fluorescein isothiocyanate (FITC)- or TexasRed (TR)-conjugated secondary antibodies raised against
mouse- (1:100) or rabbit- (1:500) IgGs (Jackson ImmunoResearch Laboratories), diluted in PBS including 10%
serum of the host of the secondary antibody. Finally, slides
were washed in PBS and mounted with Vectashield (Vector
Laboratories). To check the specificity of the labeling,
sections were incubated (1) with the secondary antibody
only, or (2) with pre-immune serum as substitution of the
primary antibody, or (3) with a mixture of pre-incubated
primary antibody and immunogenic peptide (15 to 50
mg / ml). Immunolabeling was examined by epifluorescence
light microscopy (Nikon, Optiphot-X2) and photographs
were taken using Kodak 3200 ASA / TMZ black and white
film.
3. Results
3.1. Activation pattern and action potential
characteristics
In all mammalian SA nodes which we have studied so
far [3], we have always cut the medial limb of the crista
terminalis, which encircles the orifice of the superior caval
veins, without any effect on cardiac rhythm. In the mouse
heart this technique resulted in pronounced irregularities in
interbeat interval (n55). When the muscle bundle was left
intact, the heart showed a steady interbeat interval of
144616 ms (n515) without much change (,5%) in the
time it took to complete an electrophysiological mapping
(|2 h). The final SA node preparation measured about
533 mm and contained the crista terminalis, the superior
and inferior caval veins, the tissue in between and a small
portion of the right atrial wall and the auricular trabecle
(Fig. 1).
The site of the pacemaker fibers could be identified
electrophysiologically in the corner between the lateral and
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
43
Fig. 1. Photomicrograph of the mouse sinoatrial node. Endocardial view on a dissected preparation of the right atrium of the mouse heart, with the right
auricle at the left and the interatrial septum (IAS) at the right. The atrium is cut open to expose the crista terminalis (CT), the IAS and the inlet openings of
the vena cava superior (VCS) and vena cava inferior (VCI). The dashed ellipse indicates the position of the central portion of the sinoatrial node (SAN).
Dashed lines indicate the levels of the cross sections shown in Figs. 3–5.
medial limb of the crista terminalis, which is in the orifice
of the superior caval vein. In this area action potentials
show all characteristics of pacemaker activity comparable
to other mammalian SA node preparations, i.e. pronounced
diastolic depolarization, low upstroke velocity, and long
duration. By a series of successive impalements at short
distance, the size of the electrophysiological SA node, i.e.
the area where pacemaker activity could be delineated,
could be measured as within 300 mm parallel to the crista
terminalis and 150 mm perpendicular to it indicated as
‘SAN’ in Fig. 1.
On the basis of their respective electrical activity a clear
distinction between primary (dominant) and secondary
(latent) pacemaker fibers can be made. Primary pacemakers have a diastolic depolarization rate of at least 85
mV/ s and a maximal upstroke velocity of less than 10 V/ s
(Fig. 2B, second action potential from top, Table 1). Fibers
with these characteristics are limited to an ovally shaped
area of 0.01–0.02 mm 2 (Fig. 2A, oval with asterisk). The
sinoatrial conduction time (SACT) from these fibers to the
atrium was invariably positive up to 5 ms. Furthermore,
pacemaking was unifocally in all experiments (n515).
Around these primary pacemaker fibers there is a zone
of secondary pacemaker fibers, commonly called the
transitional zone [13]. Electrophysiological features of
these fibers (Fig. 2B, third action potential from top, Table
1) are intermediate between the primary pacemaker cells
and the common atrial cell (Fig. 2B, first, fourth and fifth
action potential from top). They have a lower diastolic
depolarization rate and a higher upstroke velocity compared to the primary pacemaker cells (Fig. 2B, Table 1).
Their conduction time varies between a few milliseconds
positive to a few milliseconds negative depending on their
position relative to the preferential conduction path
through the node. In three preparations we were able to
construct a high resolution conduction map of the SA node
(cf. Fig. 2A). In general the impulse travels from the site of
origin preferentially in an oblique caudal direction to the
crista terminalis. In the primary pacemaker area of the SA
node the impulse propagates with a conduction velocity of
about 5.0–6.5 cm / s (117611 mm in 2 ms, n53). In the
periphery of the SA node, conduction was somewhat more
rapid: 8.0–9.0 cm / s (17067 mm in the next 2 ms, n53).
In the transitional zone also fibers could be impaled with
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
44
Fig. 2C shows an enlargement of the central nodal area
of the same experiment displayed in Fig. 2A. Fig. 2D
shows the presence of an atrial type of electrical activity
(second action potential from top) in the vicinity (|100
mm) of fibers showing clear nodal pacemaker activity.
Mark that nodal activity could be recorded between the
atrial like fiber and the crista terminalis (Fig. 2D, first and
third action potential from top). It was not possible to trace
a path of atrial electrical activity from the nodal area
towards the crista terminalis. Similar findings were obtained in one other experiment in which a high resolution
conduction map of the central nodal area was constructed.
3.2. Morphology of the SA node
After electrophysiological mapping, the SA node was
also histologically identified in four experiments by the
following characteristic differences compared to the regular myocardium: (1) the size of the nodal cells is smaller
causing an abundance of nuclei, and (2) the hematoxylin
staining of the cytoplasm is lighter. The transition from
atrial myocardium to nodal tissue is abrupt; on the side of
the free atrial wall a clear separation is formed by broad
bands of connective tissue (Fig. 3, broad arrows). Also
Fig. 2. Activation pattern of a mouse sinoatrial node. (A) The impulse
originated from the site marked by the asterisk. Numbers on isochrones
indicate activation time in ms. CT, crista terminalis; IAS, inter atrial
septum; TRAB, trabecle; VCS, vena cava superior. (B) Action potentials
recorded at five different sites in the sinoatrial node. (C) Shows an
enlargement of the central nodal area of panel (A). (D) Action potentials
recorded at four different sites in the central sinoatrial nodal area. Note
the presence of atrial electrical activity in the central nodal area.
an atrial type of electrical activity. These atrial like fibers,
have a constant diastolic potential (27560.5 mV, n520)
and a high upstroke velocity (.150 V/ s, n520) and were
found both at the endocardial and epicardial site of the
group of cells demonstrating diastolic depolarization.
Table 1
Action potential parameters of mouse sinoatrial fibres
Primary
pacemaker
(n59)
Secondary
pacemaker
(n510)
MDP
(mV)
APA
(mV)
APD 50
(ms)
APD 100
(ms)
DDR
(mV/ s)
V~max
(V/ s)
25164
5662
3161
8265
117620
6.262
24968
5566
3162
8065
70628
27614
MDP, maximum diastolic potential; APA, action potential amplitude;
APD 50 , action potential duration at 50% repolarization; APD 90 , action
potential duration at 90% repolarization; DDR, diastolic depolarization
rate; V~max , maximum upstroke velocity. Mean values6S.D.
Fig. 3. Hematoxylin-stained section through the primary pacemaker area
of the sinoatrial node. Photomicrograph showing a section stained for
hematoxylin. Bottom side of the picture directs towards the auricle, the
upper side towards the interatrial septum (IAS). Left side of the section
shows the endocardial surface, right side shows the epicardial surface of
the atrium. Arrows indicate a broad band of connective tissue between the
atrium and the nodal area. The boxed area indicates the orientation of the
sections of Fig. 4. Arrows indicate connective tissue surrounding clusters
of nodal cells. CT, crista terminalis; IAS, interatrial septum; SAN,
sinoatrial node. Bar represents 50 mm.
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
inside the node the amount of connective tissue is higher
than in the atrium, splitting the nodal cells up into small
clusters (Fig. 3, small arrows). Based on the histological
characteristics the nodal area measured 500 mm in length
and 150 mm in width. This area always coincided with the
electrophysiological SA node.
3.3. Distribution of gap junction proteins
In the next series of experiments we correlated the gap
45
junction distribution to the electrophysiologically mapped
sinoatrial region. Therefore, cryosections of the right atrial
preparation were labeled with fluorescent antibodies raised
against gap junction proteins Cx37, Cx40, Cx43, and Cx45
after electrophysiological mapping.
Myocardial cells were discriminated from non-myocardial cells by an antidesmin antibody that reacts with a
cardiac-specific intermediate filament (Fig. 5E) [14].
Abundant fluorescent labeling of Cx40 and Cx43 protein
could be detected between the cells of the atrial working
Fig. 4. Cross sections through the primary pacemaker of the sinoatrial node. Panel D shows a diagram of the sections presented in panels A to C which are
enlargements of the looped sinoatrial nodal (SAN) area as indicated by the box in Fig. 3. Bottom side of the picture directs towards the auricle, the upper
side towards the interatrial septum (IAS). Left side of the section shows the endocardial (endo) surface, right side shows the epicardial (epi) surface of the
atrium. CT, crista terminalis; SAN, sinoatrial node. Top panels show sections immunostained for (A) Cx43 and (B) Cx45. Gap junction labeling is
represented by the clear white speckles and rods. Panel C is the negative control and shows the specificity of the Cx45 antibody for the peptide against
which it is raised, as tested with an absorption protocol. Bars represent 50 mm.
46
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
Fig. 5. Cross-sections through the secondary pacemaker of the sinoatrial node. Panels A, B, C and E are immunostained for (A) connexin43 (Cx43), (B)
Cx45, (C) Cx40 and (E) desmin. Panel D shows the orientation of these regions (boxed area) within a section that shows an overview of a comparable
region in a hematoxylin-stained heart. Panel F shows a diagram of the sections presented in panels A to E. Bottom side of the picture directs towards the
auricle, the upper side towards the interatrial septum (IAS). Left side of the section shows the endocardial (endo) surface, right side shows the epicardial
(epi) surface of the atrium. CT, crista terminalis; SAN, sinoatrial node. Bars represent 50 mm.
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
myocardium (Figs. 4A, 5A and C). In contrast, between
myocytes in the SA node region, over a large area no
labeling of Cx43 (Figs. 4A and 5A) and Cx40 (Fig. 5C)
could be observed. Although the border between labeled
and non-labeled myocytes appears clear-cut and no gradient in labeling was observed, strands of myocytes
positive for Cx43 (Fig. 5A, arrows) and Cx40 (Fig. 5C,
arrows) protrude into the connexin negative area. As
shown in Figs. 4B and 5B, the SA nodal area showed
Cx45 staining. Whenever sections were incubated with a
mixture of Cx45 antibody that was absorbed against the
peptide to which it was raised, no labeling could be
observed (Fig. 4C), indicating that the Cx45-antibody
recognizes the Cx45 protein. The area devoid of Cx43 and
Cx40 measured about 1250 mm in the direction parallel to
the crista terminalis and 300 mm in the direction perpendicular to it. It extended towards the inflow tract of the
vena cava superior and ended in the roof of the right
atrium, being localized primarily at the epicardial side of
the atrial wall (Fig. 5). At the side of the electrophysiologically determined primary pacemaker area, the area devoid
of Cx40 and Cx43 almost covered the entire myocardial
wall from endocardium to epicardium (Fig. 4). At the side
of the secondary pacemaker area (Fig. 5) broad strands of
myocytes labeling for Cx43 and Cx40 protrude into the
nodal area.
Cx37 labeling was only present between endothelial
cells of arteries embedded in the atrium, where it colocalizes with Cx40 labeling (not shown).
4. Discussion
In the present study, the distribution of gap junctions in
and around the mouse SA node was correlated with the
electrophysiologically determined SA node. The major
Fig. 6. Diagram of the distribution patterns of Cx40, Cx43, and Cx45 in a
cross section through the electrophysiologically determined pacemaker
area of the mouse sinoatrial node.
47
findings as summarized in Fig. 6 are: (1) only Cx45, but
not Cx40 and Cx43, was found in the electrophysiologically determined pacemaker area; (2) strands of myocytes
labeled for Cx40 and Cx43 protrude into the pacemaker
area; (3) no gradient in Cx45 labeling is found in the nodal
area; and (4) a band of connective tissue is localized
between the SA node and the crista terminalis. These
findings substantiate evidence for a structural mechanism
how a small SA node is shielded from the hyperpolarizing
influence of the atrium and still can drive it.
4.1. Activation sequence in the SA node
From this combined electrophysiological and histological study it appears that the structure and function of the
mouse SA node follow the earlier descriptions of the SA
node of other small mammals [3]. The site of origin of the
cardiac impulse differs histologically from the surrounding
myocardial tissue by a greater nuclear density, a lighter
staining of the cytoplasm and the presence of a heavy
mash of connective tissue. In the mouse this area measures
3003150 mm, with a wall thickness of 50 mm or less,
resulting in a volume of about 2.2310 6 mm 3 . Assuming
that the size of the individual nodal cells roughly equals
that of rabbit cells, average cell volume will be in the order
of 2310 3 mm 3 [3]. Assuming a ratio between nodal cells
and connective tissue and interstitial cavity of 2:3, results
in about 450 primary pacemaker cells. This number of
cells is not far off the number of 500, which was estimated
earlier in the guinea-pig heart [15]. It is as yet impossible
to state a minimal number of cells needed for impulse
formation and propagation in the face of the hyperpolarizing load of the atrial tissue.
The preferential conduction pathway in the mouse SA
node runs from the primary center obliquely down to the
crista terminalis. Around the leading pacemaker site the
conduction velocity is about 5.0–6.5 cm / s which is
comparable to the conduction velocity of 2–8 cm / s found
in primary pacemaker area of the rabbit [3]. In the
periphery conduction velocity increased to 8.0–9.0 cm / s,
which is considerably slower compared to the conduction
velocity of 50–80 cm / s found in rabbit [3]. The higher
maximum upstroke velocity found in mouse primary
pacemaker cells compared to rabbit [3] (6.2 vs. 1.2–2 V/ s)
is suggestive for the activation of the fast sodium current
in mouse primary pacemaker fibres. Furthermore, in rabbit
the maximum upstroke velocity is correlated to the conduction velocity.
The shape of the isochrones (Fig. 2A) shows that the
action potentials arrive at the crista terminalis as a broad
wavefront. Recently, it has been suggested that the arrival
of the action potential at the atrial muscle as a broad wave
front facilitates driving of atrial muscle by SA node [1].
Apparently, the crista terminalis acts as conduction path
both to the right auricle, the wall of the right atrium and
the left atrium. The two last structures are reached through
48
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
the medial limb that crosses the orifice of the caval vein to
the left. The fibers in this medial limb show an atrial-like
action potential without diastolic depolarization. Direct
transmission from nodal tissue to the atrial free wall is
limited by coarse strands of connective tissue (Fig. 3),
largely separating the nodal tissue from the atrial myocardium. Connective tissue between the atrium and nodal
tissue has also been described in rabbit (Coppen et al.
[26]), monkey [16] and cow [6]. Oosthoek et al. [6]
showed connective tissue in human but this was not found
by Anderson and Ho [17]. Functionally, such a connective
barrier will help to restrict the electrical load imposed on
the SA node by the atrium.
This preferential obliquely downward conduction pathway in mouse is not common in mammalian SAN. In
rabbit, pig, and guinea-pig the preferential route is obliquely upward, whereas it is transverse in the SA node of the
cat and obliquely downward in the simian [18,19]. Apparently the routing is not related to the size of the node. In
the case of the simian node it was hypothesized that an
obliquely downward routing could be a way to shorten the
route to the AV node. Since the interbeat interval of the
mouse heart is extremely short (144616 ms, n515), this
could be a physiological function in the mouse heart as
well.
4.2. Connexins in the SA node
The presence of gap junctions in the SA node has been
electron microscopically demonstrated in various species
and they appeared to be sparse and smaller in size
compared with those from the surrounding atrium [18].
The connexin composition of gap junctions between sinus
node cells has been the subject of many studies (see
reviews by OptHof [18] and Boyett et al. [1]). The general
concept is that Cx40 and Cx43 are abundantly present in
the atrial myocardium and that nodal cells contain Cx40,
Cx45 or Cx46 [4]. However, some results concerning the
connexin distribution in the SA node have been inconsistent and conflicting. The SA node has been reported to be
either Cx43 positive in rabbit [7], dog [20] and hamster
[21], or Cx43 negative in rabbit [22–24], dog [24], human
[6,24], rat [6,24], bovine [6,24] and mouse (this study). In
this study we neither observe Cx40 labeling in the SA
node of the mouse. Only Cx45 gap junction staining was
positive in the mouse SA node. Discrepancies between the
different studies may be caused by species-dependent
differences in connexin expression. It may also be due to
species-dependent differences in gap junction size.
In the rabbit SAN, the mean number of channels in each
gap junction is reported to be about 90 channels [25],
which is close to the detection threshold of immunofluorescence labeling [5]. Thus, the absence of immunofluorescence labeling does not necessarily implicate the
absence of a specific connexin. Furthermore, the nonspecificity of antibody probes may also influence the
reliability of the data. Recently, Coppen et al. [26] showed
that the Cx45 antibody they used cross-reacts with the
same 4-amino acid sequence of Cx43. Therefore, dual
labeling which shows co-localization of Cx43 and Cx45
immunostaining should be interpreted with great care. It
cannot be ruled out that immunostaining with the antiCx45 antibody in these areas is due to cross-reaction with
Cx43. The anti-Cx45 antibody in our study is from another
source as the one used by Coppen et al. [26]. It may
therefore have a different antigenicity, but because it is
raised against the same peptide sequence, cross-reactivity
of the Cx45 with Cx43 in the atrial working myocardium
cannot be ruled out. However, because the antibody
against the synthetic peptide is a polyclonal isolated from
rabbit serum, it will also contain specific antiCx45 antibodies. Therefore, when no Cx43 labeling is found, sole
Cx45 labeling as observed in the SA nodal region, will be
specific for only Cx45.
4.3. Distribution of connexins and pacemaker function
In a model study, Joyner and van Capelle [2] determined
that a gradual increase of intercellular coupling from the
central part of the SA node towards the periphery may be
important for the SA node to drive the surrounding atrium.
So far, no experimental evidence has been found for such
gradual increase [4]. In this study we did not observe a
gradient in labeling in any of the three connexins studied,
although we cannot exclude subtle gradients in connexin
density below the detection threshold of immunofluorescence labeling. On the other hand we observed strands of
Cx40 and Cx43 positive myocytes that protrude into the
pacemaker area. In the nodal area we observed sharp
transitions in action potential waveform over small distances (|100 mm, Fig. 2C and D). This finding is
suggestive for strands of atrial myocytes running into the
central pacemaker area, which are poorly coupled to the
surrounding nodal tissue. In guinea pig we previously
observed similar sharp transitions in action potential
waveform [5]. Similarly, interdigitation of Cx43 positive
atrial cells and Cx43 negative SA node cells has been
reported in bovine, rat and human hearts [6]. In rabbit,
some interdigitation was present although it was not as
pronounced as in the aforementioned species [7]. The
distinct strands of Cx40 and Cx43 positive cells in the
mouse SA node, instead of a gradient in gap junctions,
might be important to propagate the impulse from the SA
node into the atrial tissue by limiting the contact between
both tissues. In addition, a low gap junction density
together with the low single channel conductance of Cx45
[27] compared to Cx40 [28] and Cx43 [29] may help to
protect the nodal area from the hyperpolarizing influence
of the atrium.
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
For impulse conduction from the SA node to the atrium,
direct electrical contact between these two cell types is
needed. As summarized earlier, nodal cells may contain
Cx40, Cx45, or Cx46, whereas atrial cells contain Cx40
and Cx43. In the mouse SA node we did not find Cx40 in
nodal cells. Coppen et al. [7] found that Cx43 and Cx45
are co-localized at a specific zone at the SA nodal / crista
terminalis border. If the gap junction contains both types of
connexins, it is possible that heterotypic channels of Cx43
and Cx45 are formed between atrial and nodal myocytes. It
has been shown that such a heterotypic channel exhibits
rectifying properties [30]. Recently, Jongsma [4] hypothesized that when these type of heterotypic channels are
inserted in the membrane in the right orientation, upstream
current flow from the atrium to the center of the node tends
to close the gap junctions, whereas downstream current
flow might be unimpeded. Such a specific mechanism
would help to protect the SA node from the hyperpolarizing atrial influence and still allow impulse propagation into
the atrium.
Acknowledgements
This work was supported by the Netherlands Organization for Scientific Research (NWO-902-16-193) and the
Netherlands Heart Foundation (NHS-95.077). We thank
Dr. Ronald Wilders for his invaluable discussion during
the preparation of the manuscript and Dr. D. Gros for
providing us with the antibodies derived against Cx37,
Cx40 and Cx43.
References
[1] Boyett MR, Honjo H, Kodama I. The sinoatrial node, a heterogeneous pacemaker structure. Cardiovasc Res 2000;47:658–687.
[2] Joyner RW, Capelle FJL. Propagation through electrically coupled
cells. How a small SA node drives a large atrium. Biophys J
1986;50:1157–1164.
´
[3] Bleeker WK, Mackaay AJC, Masson-Pevet
MA, Bouman LN,
Becker AE. Functional and morphological organization of the rabbit
sinus node. Circ Res 1980;46:11–22.
[4] Jongsma HJ. Diversity of gap junctional proteins: does it play a role
in cardiac excitation? J Cardiovasc Electrophysiol 2000;11:228–
230.
[5] ten Velde I, de Jonge B, Verheijck EE et al. Spatial distribution of
connexin43, the major cardiac gap junction protein, visualizes the
cellular network for impulse propagation from sinoatrial node to
atrium. Circ Res 1995;76:802–811.
´ S, Mayen AEM et al. Immunohistochemical
[6] Oosthoek PW, Viragh
delineation of the conduction system. I. The sinoatrial node. Circ
Res 1993;73:473–481.
[7] Coppen SR, Kodama I, Boyett MR et al. Connexin45, a major
connexin of the rabbit sinoatrial node, is co-expressed with
connexin43 in a restricted zone at the nodal-crista terminalis border.
J Histochem Cytochem 1999;47:907–918.
49
[8] Verheijck EE, Wessels A, van Ginneken ACG et al. Distribution of
atrial and nodal cells within the rabbit sinoatrial node: models of
sinoatrial transition. Circulation 1998;97:1623–1631.
[9] Delorme B, Dahl E, Jarry-Guichard T et al. Expression pattern of
connexin gene products at the early developmental stages of the
mouse cardiovascular system. Circ Res 1997;81(3):423–437.
[10] Gros D, Jarry-Guichard T, Ten Velde I et al. Restricted distribution
of connexin40, a gap junctional protein, in mammalian heart. Circ
Res 1994;74:839–851.
[11] el Aoumari A, Fromaget C, Dupont E et al. Conservation of a
cytoplasmic carboxy-terminal domain of connexin43, a gap junctional protein, in mammal heart and brain. J Membr Biol
1990;115:229–240.
[12] Kanter HL, Saffitz JE, Beyer EC. Cardiac myocytes express multiple
gap junction proteins. Circ Res 1992;70:438–444.
[13] Bouman LN, Jongsma HJ. Structure and function of the sino-atrial
node: a review. Eur Heart J 1986;7:94–104.
¨
[14] Kjorell
U, Thornell L-E, Lehto V-P, Virtanen I, Whalen RG. A
comparative analysis of intermediate filament proteins in bovine
heart purkinje fibres and gastric smooth muscle. Eur J Cell Biol
1987;44:68–78.
[15] Opthof T, de Jonge B, Mackaay AJC et al. Functional and
morphological organization of the guinea-pig sinoatrial node compared with the rabbit sinoatrial node. J Mol Cell Cardiol
1985;17:549–564.
[16] Alings AMW, Abbas RF, de Jonge B, Bouman LN. Structure and
function of the simian sinoatrial node (Macaca fascicularis). J Mol
Cell Cardiol 1990;22:1453–1466.
[17] Anderson RH, Ho SY. The architecture of the sinus node, the
atrioventricular conduction axis and the internodal atrial myocardium. J Cardiovasc Electrophysiol 1998;9:1233–1248.
[18] Opthof T. Gap junctions in the sinoatrial node: immunohistochemical localization and correlation with activation pattern. J Cardiovasc
Electrophysiol 1994;5:138–143, review; 52 refs.
[19] Bouman LN, Jongsma HJ. The sinoatrial node: structure, inhomogeneity and intercellular interaction. In: Huizinga JD, editor, Pacemaker activity and intercellular communication, Boca Raton, FL:
CRC Press, 1995, pp. 37–50.
[20] Kwong KF, Schuessler RB, Green KG et al. Differential expression
of gap junction proteins in the canine sinus node. Circ Res
1998;82:604–612.
[21] Trabka-Janik E, Coombs W, Lemanski LF, Delmar M, Jalife J.
Immunohistochemical localization of gap junction protein channels
in hamster sinoatrial node in correlation with electrophysiologic
mapping of the pacemaker region. J Cardiovasc Electrophysiol
1993;5:125–137.
[22] Verheule S, van Kempen MJA, Postma S, Rook MB, Jongsma HJ.
Gap junctions in the rabbit sinoatrial node. Am J Physiol 2001; in
press.
[23] Anumonwo JMB, Wang H-Z, Trabka-Janik E, Dunham B, Veenstra
RD. Gap junctional channels in adult mammalian sinus nodal cells:
immunolocalization and electrophysiology. Circ Res 1992;71:229–
239.
[24] Davis LM, Kanter HL, Beyer EC, Saffitz JE. Distinct gap junction
protein phenotypes in cardiac tissues with disparate conduction
properties. J Am Coll Cardiol 1994;24:1124–1132.
´
[25] Masson-Pevet
MA, Bleeker WK, Gros D. The plasma membrane of
leading pacemaker cells in the rabbit sinus node. Circ Res
1979;45:621–629.
[26] Coppen SR, Dupont E, Rothery S, Severs NJ. Connexin45 expression is preferentially associated with the ventricular conduction
system in mouse and rat heart. Circ Res 1998;82:232–243.
[27] Moreno AP, Laing JG, Beyer EC, Spray DC. Properties of gap
junction channels formed of connexin 45 endogenously expressed in
human hepatoma (SKHep1) cells. Am J Physiol 1995;268:C356–
365.
50
E.E. Verheijck et al. / Cardiovascular Research 52 (2001) 40 – 50
[28] Bukauskas FF, Elfgang C, Willecke K, Weingart R. Biophysical
properties of gap junction channels formed by mouse connexin40 in
induced pairs of transfected human HeLa cells. Biophys J
1995;68:2289–2298.
[29] Moreno AP, Rook MB, Fishman GI, Spray DC. Gap junction
channels: distinct voltage-sensitive and -insensitive conductance
states. Biophys J 1994;67:113–119.
[30] White TW, Bruzzone R, Goodenough DA, Paul DL. Voltage gating
of connexins. Nature 1994;371:208–209.