Podoplanin deficient mice show a rhoa-related hypoplasia of the sinus venosus myocardium including the sinoatrial node

Authors


Abstract

We investigated the role of podoplanin in development of the sinus venosus myocardium comprising the sinoatrial node, dorsal atrial wall, and primary atrial septum as well as the myocardium of the cardinal and pulmonary veins. We analyzed podoplanin wild-type and knockout mouse embryos between embryonic day 9.5–15.5 using immunohistochemical marker podoplanin; sinoatrial-node marker HCN4; myocardial markers MLC-2a, Nkx2.5, as well as Cx43; coelomic marker WT-1; and epithelial-to-mesenchymal transformation markers E-cadherin and RhoA. Three-dimensional reconstructions were made and myocardial morphometry was performed. Podoplanin mutants showed hypoplasia of the sinoatrial node, primary atrial septum, and dorsal atrial wall. Myocardium lining the wall of the cardinal and pulmonary veins was thin and perforated. Impaired myocardial formation is correlated with abnormal epithelial-to-mesenchymal transformation of the coelomic epithelium due to up-regulated E-cadherin and down-regulated RhoA, which are controlled by podoplanin. Our results demonstrate an important role for podoplanin in development of sinus venosus myocardium. Developmental Dynamics 238:183–193, 2009. © 2008 Wiley-Liss, Inc.

INTRODUCTION

During early embryogenesis the lateral plate mesoderm splits into two layers: the somatic and the splanchnic mesoderm forming the outer and the inner layer of the coelomic cavity (DeRuiter et al.,1992; Moreno-Rodriguez et al.,2006). The somatic mesoderm is involved in development of the body wall and extremities while myocardial precursors are restricted to the splanchnic mesoderm (Linask,1992; Linask et al.,1997). Subsequently, the left and right primary heart fields (cardiogenic plates) fuse at the ventral midline resulting in the primary linear heart tube, which starts looping at embryonic day (E) 8.5 (Stalsberg and DeHaan,1969; DeRuiter et al.,1992; Moreno-Rodriguez et al.,2006).

Previous studies of heart development (Stalsberg and DeHaan,1969; Viragh and Challice,1973; De la Cruz et al.,1977) have shown that further development of the heart tube is related to the addition of cells at both the arterial and venous pole of the heart, forming the outflow and the inflow tract myocardium, respectively. These early observations on the addition of the secondary myocardium have recently been supported by several studies describing the addition of myocardium at the arterial pole, being secondary (Waldo et al.,2001) and anterior heart fields (Kelly et al.,2001; Mjaatvedt et al.,2001; Meilhac et al.,2004; Kelly,2005) and the posterior heart field (PHF) at the venous pole (Cai et al.,2003; Meilhac et al.,2004; Christoffels et al.,2006; Gittenberger-de Groot et al.,2007) of the developing heart. The complete length of the splanchnic mesoderm contributing to the addition of myocardium at both poles of the heart was called second heart field (Cai et al.,2003) or second lineage (Cai et al.,2003; Meilhac et al.,2004; Kelly,2005).

Several genes and proteins, considered as early markers of the second heart field at the outflow and inflow tract, have been reported, such as MesP1 (Saga et al.,1999), fibroblast growth factor (Fgf) 8 and 10 (Kelly et al.,2001), BMP-2 and Nkx2.5 (Waldo et al.,2001; Christoffels et al.,2006; Gittenberger-de Groot et al.,2007), Isl1 (Cai et al.,2003), inhibitor of differentiation Id2 (Martinsen et al.,2004), GATA factors targeting Mef2c (Dodou et al.,2004; Verzi et al.,2005), Tbx1, Tbx3, and Tbx18 (Xu et al.,2004; Christoffels et al.,2006) and Shox2 (Blaschke et al.,2007). Recently, we have added podoplanin to this list as a novel gene in cardiac development.

As a coelomic and myocardial marker, podoplanin is specifically expressed in the mesenchyme and in the myocardium at the venous pole (Gittenberger-de Groot et al.,2007; Douglas et al.,2008; Mahtab et al.,2008). Studying the mesenchymal population, podoplanin expression was observed in the proepicardial organ and epicardium (Gittenberger-de Groot et al.,2007; Lie-Venema et al.,2007; Mahtab et al.,2008). In the cardiomyocyte population podoplanin staining was seen in major parts of the developing atrioventricular cardiac conduction system, in sinus venosus myocardium including the sinoatrial node, the venous valves, the dorsal mesocardium, the dorsal atrial wall and primary atrial septum. Also the myocardium surrounding the cardinal veins and the common pulmonary vein belongs to this population (Gittenberger-de Groot et al.,2007; Douglas et al.,2008; Mahtab et al.,2008). In earlier publications, podoplanin, a 43-kDa mucin type transmembrane glycoprotein, first named E11 antigen as a new marker for an osteoblastic cell line (Wetterwald et al.,1996), was also reported in the nervous system; the epithelia of lung, eye, esophagus, and intestine (Williams et al.,1996); the mesothelium of the visceral peritoneum (Wetterwald et al.,1996); the coelomic wall (pericardium) lining the pericardial cavity (Gittenberger-de Groot et al.,2007) and the epicardium (Mahtab et al.,2008); the podocytes of the kidney (Breiteneder-Geleff et al.,1997); and the lymphatic endothelium (Schacht et al.,2003).

To elucidate a possible functional role of podoplanin in cardiac development and more specifically in the development of the sinus venosus myocardium, we studied podoplanin knockout mouse embryos and used several immunohistochemical markers. To investigate the sinus venosus myocardium including the sinoatrial node we have used hyperpolarization-activated cyclic nucleotide-gated cation 4 (HCN4; Boyett et al.,2000; Liu et al.,2007; Mommersteeg et al.,2007b). In addition, atrial myosin light chain 2 (MLC-2a), NK2 transcription factor-related locus 5 (Nkx2.5) and connexin 43 (Cx43) were used. Furthermore, we studied E-cadherin, a cell to cell adhesion protein, and RhoA important for epithelial-to-mesenchymal transformation (EMT) of the coelomic epithelium, a process that allows epithelial cells to become mobile mesenchymal cells (Hay,2005). To visualize the epithelium and mesothelium of the coelomic cavity, the epicardium and sites of active EMT we have used Wilm's tumor suppressor protein (WT-1; Moore et al.,1999; Carmona et al.,2001; Perez-Pomares et al.,2002). It has been described that the PHF and resulting myocardium are derived from the epithelial lining of the coelomic cavity (splanchnic mesothelium) by EMT (Gittenberger-de Groot et al.,2007). During normal development, loss of E-cadherin is needed for proper EMT resulting in loss of epithelial features (Cano et al.,2000) and consecutive development into migratory mesenchymal cells. During abnormal development, an up-regulated state of E-cadherin, by for example, lack of podoplanin, presents an altered EMT (Martin-Villar et al.,2005). Podoplanin can, therefore, be considered as an inhibitor of E-cadherin stimulating EMT. In addition, RhoA activation by podoplanin and ezrin interaction have been described to lead into podoplanin-induced EMT (Martin-Villar et al.,2006). Similar to E-cadherin, lack of podoplanin might lead to down-regulated RhoA and altered EMT.

In the current study, we studied the role of podoplanin in the development of the sinus venosus myocardium derived from the PHF. We demonstrate that knocking out the podoplanin gene leads to myocardial abnormalities in sinus venosus myocardium at the venous pole of the developing mouse heart.

RESULTS

General Description

We studied the embryonic phenotype of the podoplanin knockout mice, which show an increased embryonic death of approximately 40% of the homozygote embryos between E10 and E16. Additionally, 50% of the neonatal homozygote knockout mice die within the first weeks of life, while heterozygous mutants reach sexual maturity. The cause of embryonic death has been correlated to the cardiac defects (Mahtab et al.,2008), while the cause of neonatal death is still unknown.

Several marked morphological cardiac abnormalities were observed in the knockout mouse embryos. In the younger stages, a hypoplastic proepicardial organ (E10.5) as well as ventricular myocardium were observed with discontinuous epicardium, a thin layer of the subepicardial mesenchyme and a diminished amount of epicardium-derived cells (EPDCs; Mahtab et al.,2008). These hearts also presented with outflow tract abnormalities such as severe dextroposition of the aorta, coronary artery abnormalities, spongious myocardium of the developing interventricular septum and impaired formation and fusion of the atrioventricular cushions. At the sinus venosus region, which is the area of our focus, myocardial hypoplasia and morphological abnormalities were observed.

Below, morphology and immunohistochemical expression patterns of the sinus venosus myocardium will be described in the knockout embryos and compared with wild-type embryos in subsequent stages of heart development.

Podoplanin Expression in the Heart

In wild-type mice the first expression of podoplanin can be recognized as early as E9.5 in the coelomic mesothelium and in the proepicardial organ (Gittenberger-de Groot et al.,2007; Mahtab et al.,2008). Moreover, podoplanin was observed in the myocardium of the medial wall of the left cardinal vein shortly before entering the sinus venosus. At E10.5, podoplanin staining in this region was more extensive and extended into the dorsal mesocardium. At the right side, the staining is evaluated for the first time at the level of the future sinoatrial node, in the medial wall of the right cardinal vein. The venous valves showed also podoplanin positivity. At E12.5, podoplanin was clearly observed in the pericardium, epicardium, sinoatrial node, venous valves, dorsal mesocardium, myocardium of the pulmonary vein, atrial septum and ventricular conduction system. Remarkably, podoplanin staining was also observed at the left side bordering the medial contour of the left cardinal vein. The pattern and intensity of this region was similar to the sinoatrial node region at the right side, although the left-sided region was smaller at this stage (Gittenberger-de Groot et al.,2007).

Sites of Epithelial-to-Mesenchymal Transformation

To demarcate the coelomic mesothelium and sites of active EMT, we used WT-1, E-cadherin, and RhoA staining. WT-1 was observed in both the coelomic mesothelium and proepicardial organ at E9.5 in the wild-type. In the mutants, compared with the wild-type embryos, these regions were smaller and E-cadherin was up-regulated. At E10.5 in wild-type embryos WT-1 positivity was found in the single layer of mesothelium of the coelomic cavity and epicardium (Fig. 1a,b). Marked staining was seen at the corners of the coelomic cavity at both sides adjacent to the cardinal veins, where the expression of WT-1 was more extensive and the cells of the coelomic mesothelium appeared to be cuboidal and well organized (Fig. 1b). In the podoplanin knockout embryos WT-1 was also present in these regions (Fig. 1e,f), however, the coelomic mesothelium was disorganized, the cells were irregular in shape and size and seemed to have maintained their epithelial confinement (Fig. 1, compare a and b with e and f). The defective spreading of the epicardium at several locations was shown after WT-1 staining that followed this pattern and did not normally cover the myocardium in the knockout mouse embryos (Fig. 1a,e). Increased E-cadherin staining has been observed clearly in both the epicardium and mesothelium lining the coelomic wall in the podoplanin knockout embryos (Fig. 1g) compared with the wild-type mouse embryos (Fig. 1c). Remarkably, in the mutants E-cadherin staining was not only up-regulated in the epithelium of the coelomic cavity but also in the underlying mesenchymal cells (Fig. 1c,g), supporting the observation of disturbed EMT. At E10.5 in wild-type embryos, major RhoA expression was seen in the mesenchyme and epithelium of the coelomic cavity and epicardium (Fig. 1d). In the mutants overall expression of RhoA was down-regulated (Fig. 1h).

Figure 1.

Transverse sections showing epithelial-to-mesenchymal transformation (EMT) sites of the coelomic cavity epithelium in the hearts of E10.5 wild-type (WT) and podoplanin knockout (podoplanin−/−) mouse embryos. ac,eg: To demarcate the coelomic mesothelium and sites of active EMT, we studied WT-1 staining (a,b,e,f) and E-cadherin staining (c,g). a,b,f: WT-1 positivity was found in the mesothelium (asterisk in b,f) and epithelium (CE) of the coelomic cavity and epicardium (EP) of the WT embryos (a,b). b: WT-1 staining was more extensive at the corners of the coelomic cavity (asterisk) below the left cardinal vein (LCV), where the CE appeared to be cuboidal and well organized. a,b,e,f: In the podoplanin−/− embryos, WT-1 was also present in these regions (e,f); however, the CE was disorganized and the cells were irregular in shape and size (compare a and b with e and f). c,g: E-cadherin staining appeared stronger in both the EP and CE of the podoplanin−/− embryos (g) as compared to the WT mouse embryos (c). Remarkably, in the podoplanin−/− E-cadherin staining was not only up-regulated in the CE but also in the underlying mesothelial cells (c,g). d: In WT embryos, RhoA expression was seen in the coelomic mesenchyme, CE, and epicardium. h: In the mutants overall expression of RhoA was down-regulated. LA, left atrium. Scale bars = 60 μm in a,e, 30 μm in b–d,f–h, 20 μm magnification box in c,d,g,h.

The Sinus Venosus Myocardium

The HCN4-positive region of the sinus venosus area (Figs. 2i,k, 3d,f,m,o) overlaps the Nkx2.5-negative (Figs. 2g,j, 3h–k) and MLC-2a–positive sinus venosus region (Figs. 2c–f, 3c,e,l,n). Podoplanin is also expressed in these regions of the sinus venosus myocardium, almost completely overlapping with the Nkx2.5-negative and HCN4-positive regions.

Figure 2.

ao: Craniodorsal view of three-dimensional reconstructions (a,b) and transverse sections (c–o) showing at embryonic day (E) 12.5 the hypoplasia of the sinoatrial nodal (SAN) region in the podoplanin knockout (podoplanin−/−) mouse embryos compared with the wild-type (WT) embryos. The intersection line c refers to sections c and d and the intersection line e refers to sections e and f. c–g,i–k: In the SAN (asterisks), the expression level of MLC-2a (c–f), Nkx2.5 (g,j), and HCN4 (i,k) was unaltered in the WT (c,d,g,i) and podoplanin−/− embryos (e,f,j,k). Section h shows podoplanin (Podo) expression in the SAN of the WT heart. The SAN in the podoplanin−/− is thin and hypoplastic compared with WT (compare d with f). c,e: In the mutants, the venous valves (VV), dorsal atrial wall (arrows in c and e) and primary atrial septum (AS) were also small and hypoplastic (compare c with e). Because of the deficient AS, the mutant hearts showed a large atrial septum defect (compare c with e) and the AS myocardium was not properly embedded in the atrioventricular cushion (AVC, arrow heads in c and e). The AVC was not fused properly with the ventricular septum resulting in a ventricular septum defect (compare c with e). l,m: In WT embryos, RhoA expression was seen in coelomic mesenchyme, CE as well as in SAN and epicardium (EP). n,o: In the mutants, RhoA expression was down-regulated in SAN as well as in CE and EP. Color codes: atrial myocardium, light brown; cardinal veins lumen, transparent blue; common pulmonary vein (PV) lumen, pink; sinus venosus myocardium, purple; ventricular myocardium (V), brown. LA, left atrium; LCV, left cardinal vein; RA, right atrium; RCV, right cardinal vein. Scale bars = 200 μm in c,e, 30 μm in d,f–l,n, 20 μm in m,o.

Figure 3.

ao: Craniodorsal view of three-dimensional reconstructions (a,b) and transverse sections (c–o) showing the myocardial hypoplasia of the common pulmonary vein (PV), atrial septum (asterisk), and the wall of the cardinal veins of embryonic day (E) 12.5 wild-type (WT) and podoplanin knockout (podoplanin−/−) embryonic mouse hearts. The intersection lines c,l,e and n refer to the sections c,l,e and n, respectively. c,e: In the MLC-2a sections (c,e) the hypoplasia of the myocardium around the PV in the podoplanin−/− is clearly visible. c–g: Moreover, parts of myocardium around the PV, which are also positive for HCN4 (d) and podoplanin (g), are missing in the podoplanin−/− (compare c,d with e,f). h–k: This lack of myocardium around the PV is also seen in the sections stained with Nkx2.5 (h–k) where the Nkx2.5 mosaic myocardium of the PV is missing in the podoplanin−/− (k,j) compared with the WT (h,i). c,e: The hypoplasia of the atrial septum (asterisk) is seen in the podoplanin−/− (e) compared with the WT (c). l–o: Sections l and n are stained with MLC-2a, and sections m and o show HCN4 expression of the same region demonstrating the hypoplasia and the perforations (arrows in n) of the myocardium of the cardinal veins in the podoplanin−/−. Color codes: atrial myocardium, light brown; cardinal veins lumen, transparent blue; PV lumen, pink; sinus venosus myocardium, purple; ventricular myocardium, brown. LA, left atrium; LCV, left cardinal vein; LV, left ventricle; RA, right atrium; RCV, right cardinal vein; RV, right ventricle. Scale bars = 60 μm in c,e,l,n, 30 μm in d,f, g,h,k, 20 μm in i,j, 200 μm in m,o.

Sinoatrial Node and Venous Valves

In the sinoatrial node the MLC-2a–positive (Fig. 2c–f) and Nkx2.5-negative (Fig. 2g,j) region were identical to the HCN4-postitive (Fig. 2i,k) and podoplanin-positive (Fig. 2h) areas. In contrast to the remaining sinus venosus myocardial structures, in the sinoatrial node HCN4 remained positive, while Nkx2.5-negative staining was maintained. In the knockout mice, the sinoatrial node was hypoplastic (Fig. 2a,b,d,f) and the venous valves were shorter and thinner (Fig. 2c,e); however, the expression pattern of MLC2a, Nkx2.5 HCN4 and Cx43 (not shown) was not changed compared with the wild-type (Fig. 2c–k). Comparable to E10.5 in wild-type embryos, RhoA expression was seen in the coelomic mesenchyme and epithelium as well as in the sinoatrial node and epicardium (Fig. 2l,m). In the mutants, RhoA expression was down-regulated in the sinoatrial node as well as in the coelomic epithelium and epicardium (Fig. 2n,o).

Primary Atrial Septum and Dorsal Atrial Wall

In both wild-type and podoplanin knockout embryos MLC-2a expression was present in the sinus venosus myocardium and the myocardium of the atrial and ventricular wall (Figs. 2a–f, 3a–c,e,l,n). Similar to the sinoatrial node and venous valves, the expression pattern of the mentioned markers was unchanged in the mutants compared with the wild-type embryos. In the mutants, myocardial hypoplasia was seen of the dorsal atrial wall (Fig. 2c–f). The atrial septum was thin (Fig. 3c–f) and deficient (Fig. 2c,e) with a large secondary foramen (Fig. 2c,e). The myocardialization process at the base of the atrial septum was almost absent (Fig. 2c,e). Additionally in the knockout embryos, the atrioventricular cushion was not fused properly to the top of the ventricular septum resulting into a persisting interventricular communication (Fig. 2c,e). There is marked dilatation of the atria in the mutant embryos compared with the wild-type (Fig. 2c,e).

Pulmonary and Cardinal Veins

In wild-type embryos the wall of the pulmonary and cardinal veins showed MLC-2a (Fig. 3c,l) and HCN4 (Fig. 3a,d,m) expression while the Nkx2.5 expression pattern was mosaic in the myocardium lining the pulmonary vein (Fig. 3h,i) and negative in the wall of the cardinal veins (Fig. 3h). The staining for HCN4 was diminished at E15.5 compared with earlier stages, whereas the initially Nkx2.5-negative myocardium of the cardinal veins and Nkx2.5 mosaic myocardium of the pulmonary vein wall became positive.

In the knockout embryos, myocardium of the pulmonary vein wall was hypoplastic and almost absent compared with the wild-type (Fig. 3c,e). The extent of hypoplasia and lack of myocardium of the pulmonary vein corresponded with the regions normally expressing HCN4 (Fig. 3d,f) and podoplanin (Fig. 3g). Comparable to the MLC-2a stained myocardium around the pulmonary vein as described above, the Nkx2.5 mosaic area was hypoplastic and almost absent in the knockout mice (compare Fig. 3h,i with j,k).

The myocardium of the cardinal veins was also hypoplastic and showed several fenestrations (Fig. 3l–o). Both the atrial lumen and the lumen of the cardinal veins were dilated. The myocardial volume of the sinus venosus (Fig. 4a) and separately the sinoatrial node (Fig. 4b), which was estimated by myocardial morphometry, showed a significant (P < 0.05) decrease of myocardial volume in the knockout embryos compared with the wild-type embryos.

Figure 4.

a,b: Sinus venosus (SV, a) and sinoatrial node (SAN, b) myocardial volume estimation of 12 wild-type (WT) mouse hearts of E11.5 (n = 3), E12.5 (n = 3), E13.5 (n = 3), and E14.5 (n = 3) and 12 podoplanin knockout (KO) mouse hearts embryonic day (E) 11.5 (n = 3), E12.5 (n = 3), E13.5 (n = 3), and E14.5 (n = 3). Podoplanin knockout embryos have a significant smaller SV and SAN myocardial volume (*P < 0.05) compared with WT embryos.

DISCUSSION

This study was conducted to elucidate the role of podoplanin in the development of the sinus venosus myocardium derived from the specific area of the second heart field which we have named PHF (Gittenberger-de Groot et al.,2007; Mahtab et al.,2008). Our observations are based on the study of the podoplanin gene and its protein expression in different embryonic stages during cardiac development. We have generated podoplanin knockout mouse embryos and used several immunohistochemical markers to study the mutants and wild-types. These findings have consequences for the development and contribution of the PHF to the sinus venosus myocardium.

Sinoatrial Node and Venous Valves

The sinoatrial node is a complex structure that plays a fundamental role in cardiac pacemaker activity (Boyett et al.,2000). Despite its essential role in cardiac conduction the origin of the sinoatrial node is still not well understood. Recently, the formation and differentiation of the sinoatrial node at the venous pole of the heart has been described from the novel Nkx2.5-negative and Tbx18-positive precursor cells (Christoffels et al.,2006), which were positive for podoplanin (Gittenberger-de Groot et al.,2007). In the current study, we report HCN4 expression in the podoplanin-positive and Nkx2.5-negative sinoatrial node in accordance with observations in other studies (Garcia-Frigola et al.,2003; Christoffels et al.,2006; Mommersteeg et al.,2007b). The specific combination of Nkx2.5-negative and podoplanin- and HCN4-positive expression in the sinoatrial node during early heart development is in contrast with the expression of these markers in the primary atrial myocardium. The latter suggests a different precursor for the sinoatrial node and the primary atrial myocardium. In contrast to the primary atrial myocardium, which derives from the primary heart field, the sinus venosus myocardium including the sinoatrial node originates from the second heart field, as concluded from Isl1, Tbx3 and Tbx18 expression at the venous pole of the heart (Cai et al.,2003; Christoffels et al.,2006; Hoogaars et al.,2007; Mommersteeg et al.,2007b). In podoplanin mutants we found a hypoplastic sinus venosus myocardium. Another interesting gene involved in the development of the sinus venosus myocardium is Shox2 (Blaschke et al.,2007). Shox2 mutants showed severe hypoplasia of the sinus venosus myocardium comparable to our observations in the podoplanin null mice. Moreover, the hypoplastic sinoatrial node in Shox2 mutants showed aberrant expression of Cx43 combined with abnormal Nkx2.5 positivity. In contrast to the Shox2 mutants, Nkx2.5 and Cx43 expression patterns remained unchanged in the podoplanin knockout mouse hearts suggesting a different role for podoplanin in the sinoatrial node pacemaking development than Shox2. Electrophysiological experiments will be carried out to investigate possible arrhythmias in these mutants to solve the mentioned neonatal death.

Primary Atrial Septum and Dorsal Atrial Wall

Lineage tracing experiments studying Isl1 (Cai et al.,2003), Fgf 10 (Kelly,2005), and Tbx5 (Bruneau et al.,1999; Liberatore et al.,2000), 18 (Christoffels et al.,2006), and Shox2 (Blaschke et al.,2007) have demonstrated the contribution of the second heart field to the formation of the atrial myocardium which has a distinct molecular composition compared with the heart tube derived from the primary heart field (Cai et al.,2003; Kelly,2005; Anderson et al.,2006).

Podoplanin is expressed in PHF as well and not only stains the proepicardial organ derivatives but also the Nkx2.5-negative myocardium in the dorsal mesocardium (Gittenberger-de Groot et al.,2007). This myocardium is supposed to form part of the dorsal atrial wall as well as the atrial septum. In the podoplanin mutant mouse, this myocardium is hypoplastic, probably due to diminished PHF-derived myocardial contribution. Another option, is that abnormal epicardial–myocardial interaction plays a role in development of deficient myocardium as is seen in SP3 mutant mouse (Van Loo et al.,2007). We already described the deficient EPDC contribution in the podoplanin mutant (Lie-Venema et al.,2007; Mahtab et al.,2008). Therefore, both the hypoplasia of the atrial septum as well as the dorsal atrial wall observed in the current study, are related to the altered contribution of myocardial and epicardial cells from the PHF.

Pulmonary and Cardinal Veins

A controversy regarding the development of the venous pole concerns the origin of the pulmonary vein. The pulmonary pit develops either in the dorsal mesocardium at the midline (Webb et al.,1998; Wessels et al.,2000; Jongbloed et al.,2004), at the left (Tasaka et al.,1996) or the right (DeRuiter et al.,1995; Tasaka et al.,1996; Blom et al.,2001) side of the embryo as a solitary unpaired structure that arises from the sinus venosus (DeRuiter et al.,1995; Tasaka et al.,1996; Blom et al.,2001) or primitive atrium (Webb et al.,1998; Soufan et al.,2004; Anderson et al.,2006). Recently, Manner and Merkel (Manner and Merkel,2007) have described the pulmonary pit as a bilaterally paired structure. We have described that the early common pulmonary vein is surrounded by Nkx2.5 mosaic cells which are positive for MLC-2a and podoplanin (Gittenberger-de Groot et al.,2007; Douglas et al.,2008). Part of this myocardium has also been reported as “mediastinal myocardium” (Soufan et al.,2004). The Nkx2.5 mosaic area forms a myocardial sleeve around the pulmonary vein extending to the atrial septum and Nkx2.5-negative cardinal veins in contrast to the primary atrial myocardium which is completely Nkx2.5-positive. These data support the formation of the wall of the pulmonary vein to be from the surrounding mesodermal precursor cells at the dorsal mesocardium postulated to be derived from the PHF (Gittenberger-de Groot et al.,2007; Douglas et al.,2008). Concomitant with the higher proliferation rate of the pulmonary vein myocardium (Mommersteeg et al.,2007a), the Nkx2.5 mosaic area became Nkx2.5-positive in contrast to the sinoatrial node and cardinal vein myocardium that remained Nkx2.5-negative but HCN4-positive. This suggests a distinct differentiation rate of the pulmonary vein myocardium compared with the sinoatrial node and cardinal vein myocardium. At E15.5, the cardinal vein myocardium became gradually positive for Nkx2.5, whereas the HCN4 expression was diminished suggesting the gradual completion of the differentiation process at the venous pole.

In the mutant embryos, the diminished myocardial contribution to the wall of the pulmonary vein and cardinal veins is evident. It is not clear whether MLC-2a and Nkx2.5 are directly regulated by podoplanin or whether this is due to altered addition of secondary myocardium from the PHF region by lack of podoplanin.

Role of podoplanin in EMT of the Coelomic Epithelium

An important feature of EMT is down-regulation of E-cadherin (Batlle et al.,2000; Cano et al.,2000; Carmona et al.,2000) which is correlated with podoplanin (Martin-Villar et al.,2005). We have observed up-regulation of E-cadherin in the coelomic cavity epithelium of the podoplanin knockout embryos. Regarding the addition of myocardium from the second heart field, the observed hypoplasia at the sinus venosus region might be caused by up-regulated E-cadherin in the podoplanin knockout embryos, which causes abnormal EMT of the coelomic epithelium at specific sites of the sinus venosus myocardium (Mahtab et al.,2008).

Moreover, podoplanin is involved in motility of cells where it colocalizes with the ezrin, radixin and moesin (ERM) protein family (Scholl et al.,1999; Martin-Villar et al.,2005). ERM proteins bind to the podoplanin ERM-binding site to activate RhoA, a member of the Rho GTPase protein family controlling a wide variety of cellular processes including proliferation, differentiation, cell morphology and motility (Hall,1994; Van and Souza-Schorey,1997). The increased RhoA activity leads to “podoplanin-induced” EMT (Martin-Villar et al.,2006). With regard to this mechanism, lack of podoplanin has resulted in diminished expression of RhoA protein which may prevent the “podoplanin-induced” EMT with subsequent myocardial abnormalities of the venous pole.

Taken together, we show severe hypoplasia and myocardial abnormalities of the sinus venosus myocardium by lack of podoplanin. In addition, we postulate not only a common origin of the sinoatrial node, dorsal atrial wall, atrial septum, pulmonary and cardinal veins deriving from the PHF, but also provide a link between the pulmonary vein and sinus venosus myocardium rather than the pulmonary vein and primary atrial myocardium.

Clinical Implications

Parts of the cardiac conduction system derive from the secondary heart field, which may suggest a role in the etiology of clinical syndromes. Several transgenic mice present with sick sinus syndrome, occurring in a familial form, are due to the lack of Ca2+ and other ion channels as well as gap junctions (Dobrzynski et al.,2007). The dysfunctions include bradycardia, sinus dysrhythmia, and sinus node exit block (Dobrzynski et al.,2007). In our mutant embryos, we have observed a hypoplastic sinoatrial node, while podoplanin is involved in water transport (Williams et al.,1996), cationic, anionic, and amino acid transport (Boucherot et al.,2002) and Ca2+-dependent cell adhesiveness (Martin-Villar et al.,2005). It is relevant to perform functional studies in these hearts in future to show dysfunctions such as bradycardia (sick sinus syndrome) comparable to Shox2 mutants (Blaschke et al.,2007).

Several studies suggested an embryonic background of atrial fibrillation originating from the pulmonary and caval veins and based on expression patterns of molecular and immunohistochemical markers (Blom et al.,1999; Jongbloed et al.,2004; Melnyk et al.,2005). In the current study, we observed HCN4 expression in the sinoatrial node and in the myocardium of the wall of the cardinal veins and the pulmonary vein. In the mutant mice, this population of HCN4-positive cells is diminished in the sins venosus myocardium, which may provide a developmental background of arrhythmias originating from this area.

Next to disturbances in cardiac conduction the observed deficient myocardial as well as epicardial contribution results in atrial and ventricular septal defects in addition to the already observed myocardial and coronary vascular abnormalities (Mahtab et al.,2008).

EXPERIMENTAL PROCEDURES

Generation of podoplanin−/− Mice and Harvesting of Embryos

The podoplanin wild-type as well as knockout mouse embryos were kindly provided by Prof. Dr. B.R. Binder from Medical University of Vienna, Austria. The knockout mice were generated by homologous recombination in embryonic stem cells from the 129S/v mouse line by inserting a neomycin phosphotransferase cassette in a 7.7 kb genomic fragment encompassing exons II to V. The complete description of this model was reported previously (Mahtab et al.,2008). Briefly, the heterozygous embryonic stem (ES) cell clones were test-bred for germline transmission with Swiss mice to generate podoplanin+/− offspring (50% 129S/v: 50% Swiss genetic background) using standard procedures. These mice were crossed to obtain podoplanin−/− embryos and podoplanin+/+ (wild-type) littermates. The morning of the vaginal plug was stated embryonic day (E) 0.5. Pregnant females were killed and embryos were harvested.

General Description

We investigated the lining of the coelomic cavity and the morphology of the sinus venosus myocardium of the heart in 27 wild-type mouse embryos of E9.5 (n = 4), E10.5 (n = 4), E11.5 (n = 3), E12.5 (n = 4), E13.5 (n = 5), E14.5 (n = 4), and E15.5 (n = 3) and compared these with 37 podoplanin knockout mouse embryos of E9.5 (n = 4), E10.5 (n = 4), E11.5 (n = 6), E12.5 (n = 8), E13.5 (n = 6), E14.5 (n = 5), and E15.5 (n = 4). All embryos were fixed in 4% paraformaldehyde (PFA) and routinely processed for paraffin immunohistochemical investigation.

Immunohistochemistry

Immunohistochemistry was performed with antibodies against MLC-2a (1/6,000, kindly provided by S.W. Kubalak, Charleston, SC), Nkx2.5 (1/4,000, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, SC-8697), podoplanin (clone 8.1.1., 1/500, Hybridoma bank, Iowa City, Iowa), WT-1 (1/1000, Santa Cruz Biotechnology, Inc., sc-192), E-cadherin (1/150, Santa Cruz Biotechnology, Inc., SC-7870), HCN4 (1/1000, Alomone labs, The Netherlands, APC-052), RhoA (1/2000, Santa Cruz Biotechnology, Inc., SC-418), and Cx43 (1/200, Sigma-Aldrich, St. Louis, MO, C6219). The primary antibodies were dissolved in phosphate buffered saline (PBS) –Tween-20 with 1% bovine serum albumin (BSA, Sigma Aldrich). Between subsequent incubation steps, all slides were rinsed in PBS (2×) and PBS-Tween-20 (1×). The slides were incubated with the secondary antibody for 45 min for MLC-2a, WT-1, E-cadherin, HCN4 and Cx43 with 1/200 goat anti-rabbit biotin (Vector Laboratories, Burlingame, CA, BA-1000) and 1/66 goat serum (Vector Laboratories, S1000) in PBS-Tween-20; for Nkx2.5 with 1/200 horse anti-goat biotin (Vector Laboratories, BA-9500) and 1/66 horse serum (Brunschwig Chemie, Switzerland, S-2000) in PBS-Tween-20; for podoplanin with 1/200 goat anti–Syrian hamster biotin (Jackson Imunno Research, West Grove, PA, 107-065-142) and 1/66 goat serum in PBS-Tween-20 and for RhoA with 1/200 horse anti-mouse biotin (Santa Cruz Biotechnology, Inc., SC-9996-FITC) and 1/66 horse serum in PBS-Tween-20. The slides were incubated with ABC-reagent (Vector Laboratories, PK 6100) for 45 min. For visualization, the slides were incubated with 400 μg/ml 3-3′di-aminobenzidin tetrahydrochloride (DAB, Sigma-Aldrich, D5637) dissolved in Tris-maleate buffer pH 7.6 to which 20 μl H2O2 was added: MLC-2a, and E-cadherin 5 min; Nkx2.5, HCN4, Cx43, WT-1 and podoplanin 10 min. Counterstaining was performed with 0.1% hematoxylin (Merck, Darmstadt, Germany) for 5 sec, followed by rinsing with tap water for 10 min. All slides were dehydrated and mounted with Entellan (Merck).

Three-Dimensional Reconstructions

We made three-dimensional reconstructions of the sinus venosus myocardium based on MLC-2a, Nkx2.5, and HCN4 stained sections of wild-type as well as podoplanin knockout embryos of E12.5 in which the morphological differences were shown. The reconstructions were made as previously described (Jongbloed et al.,2005) using the AMIRA software package (Template Graphics Software, San Diego, CA).

Morphometry of the Myocardium

Based on HCN4-, MLC-2a-, and Nkx2.5-stained sections, sinus venosus and separately sinoatrial node myocardial volume estimation was performed of 12 wild-type mouse hearts of E11.5 (n = 3), E12.5 (n = 3), E13.5 (n = 3), E14.5 (n = 3) and 12 podoplanin knockout mouse hearts of E11.5 (n = 3), E12.5 (n = 3), E13.5 (n = 3), and E14.5 (n = 3) based on Cavalieri's principle as described previously (Gundersen and Jensen,1987). Statistical analysis was performed with an independent sample-t-test (P < 0.05) using the SPSS 11.0 software program (SPSS Inc, Chicago, IL). In summary, regularly spaced points (100 mm2 grid for sinus venosus myocardium and 49 mm2 grid for sinoatrial node myocardium) were randomly positioned on the HCN4 stained myocardium. The distance between the subsequent sections of the slides was 0.075 mm for sinus venosus myocardium and 0.025 mm for sinoatrial node. The volume measurement was done using the HB2 Olympus microscope with a ×100 “final magnification” objective for sinus venosus myocardium and ×200 “final magnification” for sinoatrial node.

Acknowledgements

We thank Jan Lens for expert technical assistance with the figures.

Ancillary