The sinus venosus myocardium contributes to the atrioventricular canal: potential role during atrioventricular node development?

The presence of distinct electrophysiological pathways within the atrioventricular node (AVN) is a prerequisite for atrioventricular nodal reentrant tachycardia to occur. In this study, the different cell contributions that may account for the anatomical and functional heterogeneity of the AVN were investigated. To study the temporal development of the AVN, the expression pattern of ISL1, expressed in cardiac progenitor cells, was studied in sequential stages performing co-staining with myocardial markers (TNNI2 and NKX2-5) and HCN4 (cardiac conduction system marker). An ISL1+/TNNI2+/HCN4+ continuity between the myocardium of the sinus venosus and atrioventricular canal was identified in the region of the putative AVN, which showed a pacemaker-like phenotype based on single cell patch-clamp experiments. Furthermore, qPCR analysis showed that even during early development, different cell populations can be identified in the region of the putative AVN. Fate mapping was performed by in ovo vital dye microinjection. Embryos were harvested and analysed 24 and 48 hrs post-injection. These experiments showed incorporation of sinus venosus myocardium in the posterior region of the atrioventricular canal. The myocardium of the sinus venosus contributes to the atrioventricular canal. It is postulated that the myocardium of the sinus venosus contributes to nodal extensions or transitional cells of the AVN since these cells are located in the posterior region of the AVN. This finding may help to understand the origin of atrioventricular nodal reentrant tachycardia.


Introduction
The cardiac conduction system (CCS) initiates electrical activation of the heart and ensures proper electrical propagation resulting in coordinated mechanical activation of the myocardium. Disturbances in the normal electrical activation pattern of the heart resulting in cardiac arrhythmias are an important cause of mortality and morbidity in the general population [1].
During early cardiac development fusion of the bilateral cardiogenic plates establishes the primary heart tube (PHT). During further development a subpopulation of cells from the splanchnic mesoderm continues to differentiate towards a cardiac fate and is recruited to the PHT at the venous and arterial poles. The splanchnic mesoderm at the venous pole gives rise to the myocardium of the sinus venosus. This myocardium surrounds the cardinal and pulmonary veins and shows distinct expression patterns as compared with the atrial working myocardium [2]. In contrast to the atrial myocardium, the sinus venosus myocardium expresses Tbx18 [3], ISL1 [2,4], Shox2 [5], Pdgf-receptor alpha [6], Tbx3 [7,8] and RHOA [2], while it is negative for NKX2-5 during early stages of development [3,9]. The sinus venosus myocardium includes the definitive right-sided sinoatrial node (SAN) as well as a transient left-sided SAN, and will differentiate into a working myocardial phenotype during further development, with exception of the future right-sided SAN. The SAN will retain a primitive phenotype as characterized by relatively high automaticity, slow conduction and poor coupling of cells and becomes the dominant pacemaker of the heart [10]. The right and left venous valves are also components of the sinus venosus myocardium and are known substrates for atrial fibrillation and tachycardia in the fully developed heart [1]. They extend along the posterior wall of the right atrium towards another important structure of the CCS, the atrioventricular node (AVN). The AVN delays conduction of the electrical impulse generated by the SAN, thereby ensuring adequate filling of the ventricles. The AVN is part of the myocardium of the atrioventricular (AV) canal and consists of a compact node, situated in the posterior part of the AV canal, covered by transitional cells as well as nodal extensions running towards the vestibules of the mitral and tricuspid valves [11]. The exact origin of the individual components of the AVN is still unclear. Studies indicate that at least the compact node derives from the myocardium of the AV canal [12,13], suggesting that the majority of cells of the AVN is formed from one source of cells. However, the presence of electrophysiologically and morphologically distinct cells (i.e. compact nodal cells, transitional cells and nodal extensions) in the region of the AVN indicates contributions from different sources, and/or with different transcriptional programmes. A dual contribution would be a possible explanation for the occurrence of these different cell types and may help to understand the occurrence of atrioventricular nodal reentrant tachycardia (AVNRT). A prerequisite for AVNRT to occur is the presence of electrophysiologically distinct pathways within the AVN [14], resulting in slow conduction in the so-called slow pathway, unidirectional conduction in the fast pathways and a central area of block. The slow pathway is a common target for catheter ablation of AVNRT, and has been described anatomically as the inferior nodal extension [15]. It has been postulated that the transitional cells and nodal extensions are an atrial contribution to the AVN [16] or are derived from the myocardium of the sinus venosus [2]. Interestingly, the inferior nodal extension has a similar gene expression profile as the SAN [17], and pacemaker activity from the AV junction has been attributed to the inferior nodal extension [18], thereby suggesting a developmental relation between the AVN and the sinus venosus myocardium.
ISL1, a LIM homeodomain transcription factor that plays an important role during cell proliferation, differentiation and survival [19], is expressed in the sinus venosus myocardium during early stages of development. Genetic lineage tracing with an inducible Isl1-Cre revealed important differences in timing of addition of cells to the developing heart and AVN. When Cre-recombinase was induced with Tamoxifen at embryonic day (E)7, a large number of Isl1-positive cells were found in the heart at E9. Induction of Cre at E9 and analysis at E11 revealed far less Isl1-positive cells, with a cluster of cells in the region of the developing AVN [20]. This could indicate that the AVN develops from multiple sources, with contributions at different time-points during development. Immunohistochemical analysis of ISL1 expression in chick embryos revealed a region of ISL1+ cells in the dorsal atrial wall, just cranial to the AV junction, which is known to be the location of the future AVN, leading to the hypothesis of a contribution of the sinus venosus myocardium to the AVN [2]. However, no cell lineage tracing studies independent of gene expression to prove a contribution from the sinus venosus myocardium have been performed.
In this study, vital dye labelling experiments aimed at studying different contributions to the developing AVN were performed, to determine the origin of cells contributing to the AVN. First, expression patterns marked by double or triple staining with ISL1 (expressed in cardiac progenitor cells), the myocardial markers TNNI2 and NKX2-5 and the CCS marker HCN4, were studied in sequential stages of chick and mouse development. These results were used as a roadmap for the targeted vital dye labelling experiments.

Immunohistochemistry
To obtain a developmental series of chicken heart development, fertilized eggs of the White Leghorn chicken were incubated at 37°C and 80% humidity. Embryos (HH7-31) were excised, staged according to Hamburger and Hamilton criteria [21], and fixed in 4% paraformaldehyde for 24 hrs. Subsequently, they were embedded in paraffin and sectioned serially (5 lm) for immunohistochemical analysis.
To confirm the results obtained in chick embryos, immunohistochemical analysis was extended to wild-type mouse embryos with a mixed genetic background of E10.5, E11.5 and E12.5 (mouse line has been described previously [22]). The morning of the vaginal plug was considered E0.5. Pregnant dams were euthanized with CO 2 exposure followed by cervical dislocation. Processing of embryos after excision was similar to the chicken embryos. Animal care was in accordance with national and institutional guidelines and approved by the animal experiments committee of the Leiden University Medical Center (Permit Number 13067).

In situ hybridizations
For detection of HCN4 mRNA expression in the developing CCS, in situ hybridizations in chick embryos were performed (HH15-23) and analysed as described previously [23].

Laser capture microdissection procedure
Laser capture microdissection (LCM) was performed to specifically isolate tissue of three distinct structures: 1. Myocardial continuity sinus venosus and AV canal (SV-AVC); 2. More caudal region of posterior AV canal (AVC); 3 Lateral wall right ventricle (RV). Embryos of HH21 were extracted and fixated for 24 hrs in 4% paraformaldehyde, embedded in paraffin and serially sectioned at 5 lm. Sections were mounted on membrane-covered slides (MembraneSlide 1.0 PEN; Carl Zeiss Microscopy, Thornwood, NY, USA). Tissue from the different structures of interest were dissected using the PALM microbeam (Carl Zeiss Microscopy) and RoboSoftware package and subsequently transferred to AdhesiveCaps (Carl Zeiss Microscopy). After microdissection, the tissue was stored at À80°C. The outline of the dissected tissue was photographed in all embryos and examples of the different regions before and after dissection are shown in Figure S1.

RNA isolation, cDNA synthesis and qPCR of laser microdissection tissue
After microdissection, RNA was isolated using the RecoverAll TM Total Nucleic Acid Isolation Kit for FFPE (Ambion, Life Technologies). cDNA synthesis was performed with the iScript TM cDNA synthesis kit (Bio-Rad, Hercules, CA, USA). After cDNA synthesis, qPCR was performed for TNNT2, NKX2-5, HCN4 and ISL1. For ISL1 a nested approach was necessary to detect mRNA expression. After cDNA synthesis, for ISL1, 15 cycles of pre-amplification were performed with 10 ll Sybr-Green Mastermix (Bio-Rad), 8.5 ll nuclease free water, 0.25 ll forward and 0.25 ll reverse primer and 1 ll of cDNA (final volume of 20 ll). For the other three genes, no pre-amplification was needed. The qPCR was performed with 10 ll of SybrGreen Mastermix (Bio-Rad), 1 ll of forward primer, 1 ll of reverse primer, 1 ll of cDNA (or 1 ll nuclease free water as negative control) and 7 ll nuclease free water per reaction (final volume 20 ll). Reactions were carried out in triplicate for each sample. qPCR was performed on a Bio-Rad CFX96 real-time system. Melting curve analysis was performed to verify single PCR product amplification. Chicken GAPDH was used as the reference standard for normalization, and relative differences in mRNA expression were determined. The primers used are described in Table 1.
Whole mount ISL1 staining HH21 embryos were extracted and the heart was carefully removed from the thorax under the preparation microscope. The heart was fixated in DMSO:100% Methanol (ratio 4:1) for 2 hrs at 4°C, after which the tissue was rehydrated in 75%, 50% and 25% methanol in PBS. After rinsing thoroughly with PBS for 30 min. at room temperature, the tissue was permeabilized with PBS-Tween for 1 hr. Blocking was performed with PBS-Tween-20 with 1% BSA for 1 hr at room temperature. The ISL1 primary antibody (Clone 40.2d6, 1/100; Developmental Studies Hybridoma Bank), was dissolved in PBS-Tween-20 with 1% BSA and the tissue was incubated with the antibody overnight at 4°C in a shaking plate. The next morning, the tissue was rinsed with PBS and incubation with Vectastain ABC staining kit (PK-6100; Vector Labs) was performed for 90 min. at room temperature. After rinsing with PBS, the tissue was incubated for 30 min. in 3-30diaminobenzidinetetrahydrochloride (DAB) (D5637; Sigma-Aldrich) dissolved in Tris/Maleate (pH 7.6). Finally, a fresh solution of DAB with 5 ll of H 2 O 2 was added for a brief moment (AE5 sec.) to the tissue, after which the tissue was thoroughly rinsed in PBS to stop the visualization reaction.

Single cell patch-clamp recordings
Electrophysiological data were obtained by single cell patch-clamp of HH21-22 chick embryonic cardiomyocytes of the continuity between the sinus venosus and AV canal and compared to cardiomyocytes from the RV. Whole mount ISL1 stained embryos were used to identify the ISL1+ continuity between the sinus venosus and AV canal, which aided in specifically dissecting this tissue under the preparation microscope. As control tissue, a piece of RV was collected. The dissected tissue was collected in 1.5 ml tubes and 800 ll of 0.05% Table 1 qPCR primers

Primer Sequence
Trypsin (Life Technologies) was added. The tube was placed in a heated (37°C), shaking plate for 10 min. After this, the tubes were vortexed until all tissue was dissociated and centrifuged at 114 g for 5 min. The supernatant was removed and the cells were resuspended in medium containing DMEM high glucose (Life Technologies), 10% foetal calf serum (Sigma-Aldrich), 1% penicillin-streptomycin (Life Technologies), 1% non-essential amino acids (Life Technologies) and 1% L-Glutamine (Life Technologies). Cells were plated on glass coverslips and used for single cell patch-clamp recording.
Recordings were performed on single cardiomyocytes, 1-2 days after cell dissociation. Action potential measurements of spontaneously contracting cells were recorded. Action potentials were recorded with the perforated patch-clamp technique using an Axopatch 200B amplifier (Molecular Devices Sunnyvale, CA, USA) and were filtered (5 kHz) and digitized (40 kHz). Data were acquired with pClamp10.1 (Molecular Devices, Sunnyvale, CA, USA) and analysis was performed with custom-made software. Potentials were corrected for the estimated change in liquid junction potential.

Micro-injection procedure
Medial labelling was performed by opening the coelomic cavity with subsequent injection in the ISL1+/TNNI2+ sinus venosus myocardium ( Fig. 1A-E). To calculate the volume of labelling, all microscopic sections that contained DiI/5-TAMRA were photographed and ImageJ was used to calculate the volume of labelling. The total volume of labelling and the volume of labelling in the ISL1-/TNNI2+ atrial myocardium directly bordering the ISL1+/TNNI2+ sinus venosus myocardium were calculated for the 14 embryos that were analysed in the medial group 1-2 hrs after labelling.
Lateral, right-sided labelling of the mesenchyme and vessel wall of the right cardinal vein (RCV) was performed just caudal to the entrance of the RCV in the sinus venosus ( Fig. 1F-H). Injection in this region was performed through the developing lateral body wall leaving the coelomic cavity intact. This injection site was chosen to assure lateral labelling of mesenchymal structures and to minimize leakage to other cardiac components. Embryos were collected at selected time-points (HH15-25) to assess the location of labelling.

3D reconstruction
AMIRA (Template Graphics Software, Inc. Houston, TX, USA) reconstructions aided in clarifying the three-dimensional relationship between different structures. Embryos were serially sectioned and immunofluorescently stained for both TNNI2 and ISL1. The slides were photographed and processed with Adobe Photoshop CS6. Finally, the slides were stacked and aligned and structures of interest were labelled with the AMIRA software. A subset of reconstructions was converted to 3D interactive PDFs, which are available online. Different functions (e.g. buttons to show, hide or make different structures transparent) were included in the interactive PDFs, using Javascript. The scripts that were used were previously written and published [24] and slightly modified for the current PDFs.

Statistical analysis
Results are expressed as mean AE SEM. Comparisons were made using unpaired Student's t-test (normal distribution) or Mann-Whitney U-test when data were not normally distributed. P < 0.05 was considered statistically significant.

Results
Identification of an ISL1+/TNNI2+/HCN4+ myocardial continuity between the sinus venosus myocardium and posterior portion of the AV canal This first phase of the study was aimed at analysis of expression patterns of ISL1, performing immunostaining at stages HH7-31. The first expression of TNNI2 was seen by HH8, within the medial borders of the bilateral cardiogenic plates, co-localizing with ISL1 ( Fig. 2A-D). Strong expression of ISL1 was also found in the foregut endoderm ( Fig. 2A-C). At HH9, the bilateral plates of mesoderm met in the midline, and started to fuse (Fig. 2F). TNNI2 expression increased and co-localization with ISL1 was still found in the majority of cells (Fig. 2E-H). By HH10, the bilateral plates had fused and formed a semitubular structure, with the ventral side formed by the TNNI2+ myocardium, and the dorsal side bordered by the ISL1+ foregut endoderm. By HH11-12, the heart is almost closed posteriorly, forming the PHT. Most cardiomyocytes of the tube have lost ISL1 expression ( Fig. 2I and J), except for populations of ISL1+/TNNI2+ cells at the arterial (Fig. 2I-L) and venous pole.
Analysis of ISL1 expression in subsequent developmental stages revealed ISL1/TNNI2 co-expression in the myocardium of the sinus venosus. At HH14-15, the myocardium connecting the atria to the AV canal is negative for ISL1 (Fig. 3A-F). The AV canal was defined based on morphological criteria, i.e. the presence of thick endocardial cushions and less densely organized myocardium underlying the cushions (Fig. 3E, I and K). More caudally the sinus venosus myocardium is positive for ISL1 and TNNI2 (Fig. 3G and H). Thus, there is initially a region of ISL1-/TNNI2+ myocardium between the sinus venosus myocardium and the AV canal myocardium (Fig. 3B and C). To further clarify the relation between the different structures, an interactive PDF of the three-dimensional reconstruction shown in Figure 3 is available online (Data S1). At this stage, the RVV and LVV are not evident, yet.
During further development, the RVV becomes recognizable as a double myocardial leaflet, of which the inner leaflet (SV luminal side) shows co-expression of ISL1 and TNNI2. The outer leaflet is ISL1-/ TNNI2+ (atrial luminal side; Fig. 3I and J). An ISL1+/TNNI2+ continuity between the myocardium of the sinus venosus and the posterior part of the AV canal becomes apparent at HH16-17 ( Fig. 3K and L). At HH 21, this continuity still shows ISL1 and TNNI2 expression and is located between the posterior portion of the AV canal (region of the (putative) AVN) and the cranial portion of the sinus venosus at the level of the entrance of the cardinal veins in the sinus venosus (Fig. 4E). The AV canal is again characterized by the presence of thick endocardial cushions with spongious myocardium underlying the cushions (Fig. 4G-J). The ISL1+/TNNI2+ area includes the myocardium of the SAN, RCV, RVV and LCV, including the left SAN (Fig. 4A-H). The central part of the sinus venosus myocardium is connected to the myocardium of the posterior AV canal (Fig. 4C, E and G), where ISL1 expression is seen in the AV canal myocardium (Fig. 4J and K). An interactive PDF of the three-dimensional reconstruction is available online, which shows in further detail the continuity between the ISL1+/TNNI2+ sinus venosus myocardium and the posterior AV canal (Data S2). An ISL1/NKX2-5/TNNI2 triple staining was performed and showed a gradient in ISL1 expression from the SV myocardium (strongest expression) towards the AV canal myocardium (decrease in expression) (Fig. 4I-L). NKX2-5 expression showed a similar but opposite gradient (Fig. 4I-L), suggesting differentiation of ISL1+/ TNNI2+/NKX2-5-sinus venosus cardiomyocytes towards NKX2-5+/ TNNI2+/ISL1-AV canal cardiomyocytes.
During further development, ISL1 expression became confined to the myocardial sleeve of the RCV at the location of the future SAN ( Fig. 4M-P), with no discernible expression of ISL1 in the continuity between the myocardium of the sinus venosus and AV canal any more.
To investigate whether the myocardial continuity between the sinus venosus and posterior AV canal indeed had a 'pacemaker-like' phenotype, in situ hybridizations were performed in chick embryos for HCN4, which is used as a marker for the (developing) CCS [23]. The connection between the sinus venosus myocardium and the myocardium of the posterior portion of the AV canal (Fig. 5A) showed HCN4 expression (Fig. 5B), shown at HH19. To verify the co-localization of ISL1, TNNI2 and HCN4 in this region in mammals, a triple staining for ISL1, TNNI2 and HCN4 was performed in wild-type mouse embryos. ISL1+/TNNI2+/HCN4+ cells were found in the region of the myocardial continuity between the sinus venosus and AV canal (shown at E11.5 in Fig. 5C-F). These results indicate that the cells of the myocardial continuity between the sinus venosus and posterior AV canal have a pacemaker-like phenotype.
Characterization of the myocardial continuity between the sinus venosus and AV canal confirms the pacemaker-like phenotype Laser capture microdissection was performed to characterize the sinus venosus-AV canal continuity (SV-AVC in Fig. 6) and to quantify difference in expression levels of TNNT2, NKX2-5, ISL1 and HCN4 in this region. These results were compared to the more caudal tissue of the posterior AV canal and tissue from the lateral wall of the RV to better identify the phenotype of the SV-AVC region.
Expression of TNNT2 and NKX2-5 was comparable between the different samples, confirming the myocardial identity of all tissue samples that were collected (Fig. 6B and C). Analysis of HCN4 expression showed the highest relative expression in the SV-AVC continuity, with a lower level of expression in the AV canal and RV (Fig. 6D). The difference in expression levels between the myocardial continuity and the RV was significant (P = 0.008). There was a trend towards significance in the difference between the continuity and the AV canal sample (P = 0.056). ISL1 expression was only detectable in the samples taken from the SV-AVC continuity (Fig. 6E).
Single cell patch-clamp was performed to functionally characterize the cells of the myocardial continuity between the sinus venosus and AV canal. Whole mount ISL1 staining was performed to identify the region of the continuity between the sinus venosus myocardium and posterior portion of the AV canal (Fig. 6A). Cells from this region were compared to cells from the RV (Fig. 6A). The action potential of cells from the SV-AVC continuity showed a pacemaker-like phenotype (Fig. 6F), while the action potential of cells from the RV had a ventricular-like phenotype (Fig. 6J). The SV-AVC cells had a higher spontaneous beating frequency (Fig. 6H), a lower amplitude (Fig. 6I), a shorter APD50% (Fig. 6K) and APD90% (Fig. 6L), in line with what is expected from pacemaker-like cells. The results are summarized in Table 2.
Together these results confirm the pacemaker-like phenotype of the cells from the myocardial continuity between the sinus venosus and AV canal. Furthermore, results show that even during early development, it is possible to distinguish different cell populations within the region of the putative AVN.

Physical lineage tracing shows incorporation of sinus venosus myocardium in the posterior region of the AV canal
The abovementioned results show the identification of an ISL1+/ HCN4+/TNNI2+ myocardial continuity between the sinus venosus and posterior AV canal. At early developmental stages, ISL1-/TNNI2+ atrial myocardium is positioned between the ISL1+/TNNI2+ sinus venosus myocardium and the ISL1-/TNNI2+ AV canal myocardium. During further development, ISL1+/TNNI2+ myocardium is continuous with the ISL1-/TNNI2+ AV canal myocardium. To further study, the possible contribution from the ISL1+/TNNI2+ sinus venosus myocardium to the posterior portion of the AV canal, lineage tracing experiments were performed. In ovo microinjection experiments were performed in chicken embryos with vital dyes (DiI-5-TAMRA) aimed at labelling the ISL1+/TNNI2+ sinus venosus myocardium at stages when this myocardium is not yet in contact with the AV canal (HH14-15). This myocardial labelling was termed medial labelling, since the labelling was applied more medial than the second group of labelling performed in the mesenchyme outside the sinus venosus myocardium (termed lateral labelling). First, the results of the medial myocardial labelling will be discussed.
To confirm the location of the medial labelling, a group of embryos was harvested 1-2 hrs after microinjection (n = 14). DiI/5-TAMRA was found in the myocardium of the sinus venosus and the myocardium surrounding the RCV, both positive for ISL1 and TNNI2 (Fig. 1A-C and E, 14/14 embryos, 100%; in Fig. 1B all 14 separate microinjections are superimposed to clarify the extent of the total labelling). Labelling was also present in the mesenchyme surrounding the RCV (Fig. 1D, 14/14 embryos, 100%). In 9/14 embryos (64%), a small area of labelling (on average 5.9% of total labelling) was also found in the ISL1-/TNNI2+ atrial myocardium directly bordering the ISL1+/TNNI2+ sinus venosus myocardium, but no leakage to the ISL1-/TNNI2+ connection between the inferior region of the atrium/ sinus venosus and AV canal myocardium or the AV canal itself was seen. Limited leakage to other cardiac structures, such as the posterior OFT and ventricles was restricted to the upper layer of myocardial cells, while the labelling at the sinus venosus myocardium was applied to the full thickness of the myocardium. This first group of embryos therefore verified the correct location of injection. An interactive PDF of the three-dimensional reconstruction is available online, which shows the location of all separate initial microinjections (Data S1).
Analysis of the embryos after 24 hrs of re-incubation (n = 10, HH19-22) revealed DiI/5-TAMRA in the sinus venosus myocardium (9/10 embryos, 90%), putative SAN (10/10 embryos, 100%), RVV (10/10 embryos, 100%), right atrium (7/10 embryos, 70%) and the continuity between the sinus venosus myocardium and the posterior AV canal myocardium (8/10 embryos, 80%; Fig. 7A-G). The labelled area was largely positive for ISL1 and TNNI2 (Fig. 7D-G). Two embryos did not show DiI/5-TAMRA in the continuity between the myocardium of the sinus venosus and AV canal. In one of these embryos, initial labelling was applied more cranially, only resulting in labelling of the atrial myocardium. The other embryo showed DiI/5-TAMRA in the caudal portion of the sinus venosus myocardium. Here, initial injection was applied more caudally than in the eight embryos that did show labelling of the continuity.
In one embryo, labelling was not found in the continuity. This is explained by the more caudal initial injection that was performed, as was seen in one embryo in the second group.
Lateral labelling of the RCV mesenchyme does not result in labelling of the myocardium of the sinus venosus or AV canal The above-described results showed that in all embryos, DiI/5-TAM-RA was found in the mesenchyme of the RCV. The labelled cells that were found in the myocardial continuity between sinus venosus and AV canal can therefore also be derived from this mesenchyme. To exclude this possibility, a second group of more lateral labelling of the RCV mesenchyme was performed at HH15-17 and the first group of embryos was harvested 1-2 hrs after injection (n = 6). Analysis showed DiI/5-TAMRA in the RCV mesenchyme and lateral body wall (6/6 embryos, 100%; Fig. 1F-H). None of the analysed embryos showed leakage to myocardial structures.
Together, the medial and lateral labelling experiments show that ISL1+/TNNI2+ myocardium of the sinus venosus is incorporated in the posterior region of the AV canal.

Discussion
The adult AVN consists of multiple cell types (i.e. compact nodal cells, transitional cells and nodal extensions) with different morphological and electrophysiological characteristics. The initiation of AVNRT requires the presence of pathways with different electrophysiological characteristics, resulting in slow conduction in the so-called slow pathway, unidirectional conduction in the fast pathway and a central area of block. Cellular and functional heterogeneity of the node is thus a prerequisite for AVNRT to occur. A key finding of this study was the demonstration of incorporation of sinus venosus myocardium in the AV canal in the posterior region of the AVN. It has been suggested that the transitional cells of the AVN are an atrial contribution to the AVN [25]. The inferior nodal extension, which acts as the slow pathway in AVNRT, has been described as atrial myocardium, based on histological criteria [16]. The CCS-LacZ [26,27] and HCN4nLacZ [28] mice, which were used to visualize the developing CCS, demonstrated that the sinus venosus myocardiumthat includes the SAN at its proximal end as well as the left and right venous valves -is connected to the AVN. Analysis of protein expression of HNK-1 in human embryos [29], mRNA expression of HCN4 in chicken embryos [23] and Tbx3 in mice [30] showed the same connection between the sinus venosus myocardium and the AVN. More recently, it was established that Isl1+ cardiogenic mesoderm contributed to the AVN [20]. These results suggest that the sinus venosus myocardium and venous valves are connected with the AVN through specialized myocardium of the CCS. However, these studies did not show a direct embryonic contribution from the sinus venosus myocardium to the developing AV canal. To overcome these limitations, it is necessary to directly label the cells from the sinus venosus myocardium and trace these cells during development.
Recently, two separate research groups performed vital dye labelling experiments and showed a very early contribution (4-6 somite stage in mouse [31] and HH8 in chick [32]) from a region posterior to the cardiac crescent to the AV canal myocardium. Dom ınguez et al. showed that the majority of cells comprising the murine AV canal are derived from an Isl1+ and Mlc2a-pool of cells [31]. Bressan et al. showed a contribution to the developing AV canal from an even more posteriorly located region, which is Isl1-and Nkx2-5- [32]. These experiments suggest that the AV canal myocardium, and thus the AVN, is derived from a larger population of cells than cardiac crescent cells alone. However, these results still do not prove the addition of cells from the sinus venosus myocardium to the developing AV canal.
This study is aimed at the area of myocardial continuity between the sinus venosus and AV canal. Expression data show that in this tissue, ISL1 is strongly expressed in the sinus venosus myocardium and decreases towards the AV canal. NKX2-5 expression shows a similar but opposite pattern, with an increase in expression towards the AV canal myocardium, suggesting that ISL1+ cells are differentiating towards a more restricted cardiac fate in this region. The continuity between sinus venosus myocardium and the AV canal also showed HCN4 mRNA expression and was identified in murine embryos co-expressing ISL1, TNNI2 and HCN4. HCN4 is responsible for the 'funny current', enabling cardiomyocytes to depolarize

Conclusion
Incorporation of sinus venosus myocardium in the posterior part of the AV canal myocardium was demonstrated. It is postulated that the sinus venosus myocardium contributes to the nodal extensions or transitional cells of the AVN. Multiple sources of cells contributing to the AVN may explain the heterogeneity of the node, and form the substrate underlying AV nodal reentrant tachycardia.