Knockdown of alpha myosin heavy chain disrupts the cytoskeleton and leads to multiple defects during chick cardiogenesis

Abstract

Atrial septal defects are a common congenital heart defect in humans. Although mutations in different genes are now frequently being described, little is known about the processes and mechanisms behind the early stages of atrial septal development. By utilizing morpholino-induced knockdown in the chick we have analysed the role of alpha myosin heavy chain during early cardiogenesis in a temporal manner. Upon knockdown of alpha myosin heavy chain, three different phenotypes of the atrial septum were observed: (1) the atrial septum failed to initiate, (2) the septum was initiated but was growth restricted, or (3) incorrect specification occurred resulting in multiple septa forming. In addition, at a lower frequency, decreased alpha myosin heavy chain was found to give rise to an abnormally looped heart or an enlarged heart. Staining of the actin cytoskeleton indicated that many of the myofibrils in the knockdown hearts were not as mature as those observed in the controls, suggesting a mechanism for the defects seen. Therefore, these data suggest a role for alpha myosin heavy chain in modelling of the early heart and the range of defects to the atrial septum suggest roles in its initiation, specification and growth during development.

Introduction

The heart is the first functional organ to form in vertebrate development, with the fused heart tube present at Hamburger and Hamilton (HH) stage 10 (33–38 h) in the chick (Hamburger & Hamilton, 1951; Sissman, 1970). Looping of this simple tube and beating of the cardiomyocytes commence shortly after fusion (Patten, 1925; Hamburger & Hamilton, 1951; Sissman, 1970). Atrial septation is initiated at HH14 (early day 3) and the septa meets and fuses with the endocardial cushions during day 5, ventricular septation is initiated at about HH23 (day 4) and fusion is complete by late HH29/30 (day 6–7) (Sissman, 1970; Hendrix & Morse, 1977; Morse, 1978; Ben-Shachar et al. 1985; Quiring, 2008). Normal atrial-ventricular septation requires the fusion of the superior and inferior endocardial cushions, mesenchymal cap and dorsal mesenchymal protrusion, with this protrusion important in closing the gap between the atrial septum and the cushions (Webb et al. 1998; Anderson et al. 1999; Wessels et al. 2000; Snarr et al. 2007; Goddeeris et al. 2008). Septation is a complex process that occurs rapidly and is susceptible to defects. In humans, abnormalities in atrial or ventricular septation are common, accounting for nearly 50% of congenital heart defects, with approximately 1% of newborns diagnosed with a congenital heart defect. Atrial septal defects (ASDs) allow abnormal left-to-right shunting of oxygenated blood between the atrial chambers and, if not treated, can lead to problems such as failure to thrive and pulmonary hypertension (Webb & Gatzoulis, 2006). In addition, significant ASDs have implications for morbidity and mortality rates (Campbell, 1970). The most common type of ASD is the secundum variety, usually due to maldevelopment of the septum primum or foramen ovale (Blom et al. 2005).

The molecular genetics of ASDs are starting to be elucidated, with mutations in certain genes found in some families with an ASD. Mutations have been reported in the transcription factors Nkx2.5 (Schott et al. 1998; McElhinney et al. 2003), Tbx5 (Basson et al. 1997; Li et al. 1997), Gata4 (Garg et al. 2003; Nemer et al. 2006), Tbx20 (Kirk et al. 2007) and, more recently, α cardiac actin (Matsson et al. 2008) and tolloid-like 1 (Stanczak et al. 2009). In addition, members of the myosin heavy chain (MHC) family of structural proteins also play a role. Members of a family carrying a mutation in MHC6 (αMHC) have an ASD (Ching et al. 2005). Although mutations in the closely related MHC7 (βMHC) gene are commonly associated with cardiomyopathy (Keren et al. 2008), a de-novomutation has recently also been associated with ASDs in four probands in a family with non-compaction of the ventricular myocardium (Budde et al. 2007). The mutations described in both MHC6 and MHC7 were missense mutations that led to a loss of function (Ching et al. 2005; Budde et al. 2007).

The MHC isoforms are traditionally acknowledged as major structural components of the heart muscle contractile apparatus, with the head domain of the MHC binding to the thin actin filaments and using ATP hydrolysis to generate contraction (Warrick & Spudich, 1987; Rayment et al. 1993). Within the heart, the two predominant MHC isoforms that are present throughout development and adult stages are β- and αMHC, with their expression patterns similar in both chick and human. βMHC is expressed throughout the early heart tube, becoming more abundant in the ventricles than in the atrial chambers as development proceeds, so in the adult only low levels of βMHC are present in the atrium (Wessels et al. 1991; Reiser et al. 2001; Somi et al. 2006). αMHC is expressed in the primitive atrial chamber prior to heart looping to the adult stage in a relatively restricted manner in both chick and humans (Wessels et al. 1991, 2000; Oana et al. 1998; Reiser et al. 2001; Somi et al. 2006). Although predominantly expressed in the atrium, low levels of αMHC have been observed in the ventricles, and mutations in MHC6 (human homologue of αMHC) have been associated with both ventricular hypertrophic and dilated cardiomyopathy (Niimura et al. 2002; Carniel et al. 2005).

Previously we have shown that a mutation in the MHC6 gene leads to secundum ASDs in humans, and that knockdown of the chick homologue (αMHC) leads to abnormal development of the atrial septum (Ching et al. 2005). The aim of this study was to provide a detailed analysis of the role that αMHC plays in the developing chick heart, and to provide insights into the effect that a reduction in αMHC would have on the actin cytoskeleton. As well as affecting the growth of the atrial septum, this study shows that αMHC plays a role in the initiation and specification of the septum, with the septum failing to initiate and develop appropriately upon knockdown in some embryos. In addition to the internal phenotypes observed, severe cardiac looping abnormalities were seen in some knockdown chicks. Analysis of the actin cytoskeleton showed that, upon knockdown of αMHC, the actin filaments often appeared immature compared with those observed in control animals. This study confirms the presence of αMHC in the chick atrial septum and shows its importance in septum initiation, specification and growth, and suggests a further role for the gene within cardiac looping.

Materials and methods

Morpholino design and optimization

Two experimental morpholinos (first experimental morpholino, 5′-TCCTTTTGGCTCGTTCCAGGTCCAT-3′; second experimental morpholino designed upstream, 5′-CGTATCTTTTTCTCCTGTTCCAGTG-3′) were designed against αMHC (AB004801, which is the full-length cDNA). In addition, a mismatch morpholino (5′-TCGTTTTCGCTCCT TCCAGCTCGAT-3′, designed to the same sequence as the first experimental morpholino but with 5 bp altered to prevent mRNA binding) was used as a negative control in addition to a GeneTools standard control morpholino (5′-CCTCTTACCTCAGTTACAATTTATA-3′, antisense to a mutated form of human β-globin). All morpholinos were supplied by GeneTools LLC, USA and were attached to a lissamine or fluorescein tag. The experimental morpholino sequences did not overlap and all sequences were analysed using bioinformatic procedures to ensure that only the intended gene was knocked down. A concentration of 250 µm for the morpholinos was utilized in this study, as preliminary data demonstrated that the morpholino precipitated out at 500 µm upon mixing with ice-cold pluronic gel, and a milder phenotype and lower morpholino uptake rate (just 30%) was observed when using the morpholino at 125 µm.

Application of F127 pluronic gel in ovo

White, fertile chicken eggs (Gallus gallus; Henry Stewart, Lincolnshire, UK) were incubated at 37 °C in a humidified atmosphere and underwent constant rotation prior to opening. All embryos were staged according to HH (Hamburger & Hamilton, 1951). At HH12 (approximately 50 h), HH14 (approximately 54 h) or HH16 (approximately 57 h), a small window was cut in the egg shell to expose the underlying embryos. Avoiding the chorioallantoic membrane, 3 mL of albumin was removed from beneath the embryo using a needle and syringe to facilitate embryonic manipulation and all extra-embryonic membranes (allantoic and vitelline) overlying the heart were carefully dissected. The morpholino/pluronic gel mixture [consisting of 7 µL of 30% F127 pluronic gel (BASF Corp., Germany) in Hank's Buffered Salt Solution (HBSS), mixed 1 : 1 with morpholino to give 250 µm final concentration] was applied directly to the exposed chick heart. The pluronic gel, morpholino and all pipette tips were stored on ice to prevent thermogelation during application. The shell window was then resealed using masking tape and the eggs were reincubated at 37 °C without rotation until harvested. This procedure was carried out for all αMHC, mismatch and standard control morpholinos. In addition, age-matched untreated control embryos were also subjected to the same experimental procedures, with the exception of morpholino/pluronic gel application. Although the second experimental αMHC morpholino was used for most of the studies, the first experimental αMHC morpholino was found to give the same phenotype with a similar penetrance. The numbers in Table 1 for the knockdown embryos are from both experimental morpholinos pooled. Control embryo values shown throughout this work refer to standard control, mismatch and untreated control chicks as, apart from one developmentally delayed chick (1/81), no phenotypic or statistical differences were observed between these groups. All phenotypic analysis was performed blind. All animal work was carried out according to national (UK Home Office) and institutional regulations and ethical policies.

Table 1.

Summary of phenotype observed after alpha myosin heavy chain (αMHC) knockdown

		External phenotype					Internal phenotype
		Heart size and shape					Atrial septation
Stage*	Embryo type†	Small	Enlarged	Looping abnormal	Normal	Total‡	Absent§	Small	Structure abnormal	Normal	Total‡
HH12/19	Control	–	–	–	12	12	–	–	–	12	12
	αMHC	–	–	1	8	9	5	4	–	–	9
HH14/19	Control	–	–	–	54	54	–	–	–	26	26
	αMHC	1	–	4	21	26	3	9	2	–	14
HH16/19	Control	1	–	–	14	15	–	1	–	14	15
	αMHC	–	1	–	9	10	3	5	–	1	9

^*Stage of development when knockdown/harvesting was performed.

^†Control embryos consist of standard control, mismatch and untreated embryos; αMHC denotes αMHC knockdown embryos.

^‡The total number of embryos for the external phenotype column is higher at some of the stages than for the internal phenotype. This is due to some embryos being analysed externally but, due to processing difficulties, not being analysed internally. In addition, embryos isolated for western analysis at Hamburger and Hamilton (HH) stage 14 were analysed externally and snap frozen, hence preventing internal analysis. The embryos stained with phalloidin and analysed by confocal analysis were analysed externally and internally and are included in the HH14/19 group.

^§In place of a normally-shaped septum, the dorso-cranial atrial wall appears to have an outgrowth and hence appears thickened.

Embryo isolation and processing

Except for the temporal analysis of morpholino uptake and the immunohistochemistry study on untreated control embryos, all embryos were harvested at HH19 (approximately 81 h development). Embryos that had morpholino applied were examined using an SV11 stereomicroscope (Zeiss, Germany). Morpholino uptake was confirmed in each embryo targeted (experimental, standard control and mismatch) by detection of the lissamine or fluorescein tag at the time of harvesting.

For the morpholino uptake study, standard control morpholino was applied to embryos at HH12, HH14 or HH16, and the embryos were reincubated until isolation at designated time intervals ranging from 15 min to 6 days subsequent to morpholino application. For the αMHC immunohistochemistry expression study, untreated control embryos were isolated at HH12, HH14, HH15, HH19 and HH24 and fixed in 4% paraformaldehyde (PFA) for between 45 min and 3 h.

For phenotypic and immunohistochemical analysis, embryos were fixed in 4% PFA for 1.5 h at room temperature (approximately 21 °C), washed in 1× phosphate-buffered saline, dehydrated in an ethanol series and paraffin embedded in a transverse orientation. Serial sections were taken at 8 µm intervals using a Leica DSC1 microtome, dewaxed, taken through an ethanol series to water and were either stained with Mayers haemalum (Raymond Lamb, UK) to allow visualization of the tissue for phenotypic analysis or underwent immunohistochemisty (see below) and were then dehydrated and mounted using Di-N-Butyle Phthalate in Xylene (DPX) mountant.

For western analysis, a total of three hearts per ‘sample’ were dissected free from the bodies of HH19 chicks, snap frozen in liquid nitrogen and stored at −80 °C until use. For phalloidin staining, morpholino-positive embryos were placed into 30% sucrose for 30 min, orientated in optimum cutting temperature embedding compound (OCT), frozen using liquid nitrogen-cooled isopentane and stored at −80 °C.

Morpholino uptake studies

Whole mounted chicks fixed in 4% PFA for 90 min (knocked down at HH12, HH14 or HH16 and harvested at HH19) were analysed for morpholino uptake using either an SP2 (Leica Microsystems, Germany) or an LSM 510 NLO and coherent chameleon with infrared pulse laser, excited at 800 nm, with emission collected at bandwidth filter 500–530 nm on a Zeiss Axio Vert 200 using either an EC Plan-neoflour 10×/0.3 lens or an EC Plan-neoflour 20×/0.5 lens (Zeiss, Germany) and a median sharpness filter was applied to images.

Immunohistochemical studies

After dewaxing and rehydration of the tissue sections, antigen unmasking was carried out (microwaving for 10 min in 10 mm sodium citrate, pH 6.0) and endogenous peroxidase activity was quenched (incubated in 1% hydrogen peroxidase for 10 min at room temperature). Sections were blocked with 5% normal goat serum (Santa Cruz, USA) for 1 h prior to incubation with a 1 : 50 dilution of primary antibody to αMHC, kindly provided by Professor Moorman (University of Amsterdam, The Netherlands) (de Groot et al. 1987). After washing, the avidin-biotin phosphatase amplification kit (StreptABComplex duet kit; Dako, Denmark) was used to visualize antibody binding. Slides were then counterstained in Mayers haemalum (Raymond Lamb, UK), dehydrated and mounted using DPX solution. Morphological analysis was performed using an Axioscope microscope (Zeiss, Germany) and Openlab software (Improvision, UK).

Western blotting analysis

Snap-frozen hearts from untreated controls, standard controls and αMHC knockdown embryos were collected following knockdown at HH14 and harvesting at HH19; each individual sample contained three hearts. Samples were separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis in gels [alongside precision plus protein standards (kDa); BioRad, UK], transferred onto nitrocellulose membranes (GE Healthcare), blocked with 5% marvel milk powder (Sigma, UK) Tris-buffered saline with 0.01% Tween-20, and then incubated overnight at 4 °C with S58 (αMHC detection; Developmental Studies Hybridoma Bank, Iowa, USA; used at a dilution of 1 : 100) or glyceraldehyde 3-phosphate dehydrogenase (GAPDH; loading control antibody at a dilution of 1 : 1000; AB8245, Abcam, UK). Following incubation for 1 h at room temperature in a mouse monoclonal secondary antibody (P0261; Dako, Denmark), the immunoreactive bands were visualized using Enhanced Chemiluminescence (ECL; GE Healthcare Ltd, UK) and densitometry was carried out using a GS710 imaging densitometer (BioRad, UK). Four different immunoblots containing untreated controls, standard control and αMHC knockdown heart samples were carried out, and densitometry results were utilized for statistical analysis (anovawith post-hoc testing; SPSS V11.0.1, SPSS Inc.).

Phalloidin staining and analysis

Serial sections of the embryos embedded in OCT were cut (20 µm), placed on microscope slides and fixed for 3 min in 4% PFA. Following fixation the sections were blocked for 30 min in 1% bovine serum albumin and stained using Alexa Fluor-conjugated phalloidin (Invitrogen, UK; 1 : 50 dilution; n = 5 control chicks and n = 4 αMHC knockdown chicks) for 20 min. The samples were then washed and counterstained with either 0.5 µg/mL propidium iodide (Dako, UK) or 1 µg/µL Hoechst (Invitrogen, UK) to indicate cell nuclei, washed in 1× phosphate-buffered saline and mounted using glycerol. Analysis was carried out using an LSM510uv META combi confocal microscope (Zeiss, Germany) and software (LSM Image Examiner, Carl Zeiss MicroImaging, Germany).

Results

Morpholino detection throughout the heart

Of the embryos that had 250 µmαMHC morpholino applied and were reincubated, on average a total of 59% were considered to have successful uptake of the morpholino in comparison to 57% observed in morpholino control animals. In addition, the total number of chicks alive at harvesting (following pluronic gel application) was monitored and untreated control embryos showed a survival rate of 91%, morpholino control survival rates were 81% and 88% of chicks that received 250 µmαMHC morpholino and morpholino were still alive at HH19. In order to determine the time period that a fluorescently-tagged morpholino persists in the developing chick embryo in ovo, a standard control morpholino was applied at either HH14 or HH16 and harvested at a range of time-points from 15 min up until embryonic day 6 (HH28/29). As shown in Fig. 1Aa and b, fluorescence throughout the embryo was detected at one of the earliest time-points studied (30 min), with similar levels recorded after 15 min (Supplementary Information Fig. S1). Consistently intense levels of fluorescence were found when embryos were left for a further 48 h, until about HH24 (Fig. 1Ac,d). Fluorescence was detected at later time-points but was found to be faint and less uniform. In addition, confocal analysis was performed using morpholino applied at HH14 and harvested at HH19, with fluorescence detected throughout the heart (Fig. 1B and Supplementary Information Fig. S2). Scans through the entire heart were performed at 5 µm intervals, and fluorescence was found throughout the heart, in both the atrial and ventricular chambers. Similar experiments were also performed following morpholino application at HH12 or HH16 and harvesting at HH19 (Supplementary Information Fig. S3A,B).

Fig. 1.

Morpholino knockdown and expression of alpha myosin heavy chain (αMHC). (A) Light (a and c) and dark (b and d) field microscopy of embryos that had 250 µm of fluorescein-tagged standard control morpholino (StC) applied. Upon application of the morpholino at Hamburger and Hamilton (HH) stage 16 with harvesting after 30 min (HH16/30 min; a and b), fluorescein uptake can clearly be seen throughout the embryo (b). When the morpholino was applied at HH14 and left until HH24 (HH14/24), the fluorescence was found to persist (d). Arrow denotes heart. Scale bars: a, 1000 µm (same in b); c, 500 µm (same in d). (B) Lissamine-tagged standard control was applied at HH14 and harvested at HH19 (HH14/19). The heart was then scanned by confocal microscopy with images captured at 15 µm intervals. Fluorescence can be seen to both the atrial (At) and ventricular (V) chambers. Scale bar: 300 µm. (C) Staining of αMHC was observed throughout the wall of the tubular heart (H) at HH12 (a). By HH15 (b) staining was found in the myocardial walls of the atrium (At) and in the small atrial septum that can be observed emerging from the roof of the atrial chamber. By HH19 (c), the atrial wall was clearly stained as was the septum as it extended towards the endocardial cushions (EC), which it fused with by HH24 (d). However, staining of the endocardial cushions and ventricular myocardium was not detected. Arrow, atrial septum; asterisk, trabeculae; L, liver; Untreated, untreated control embryo. Scale bars: a and b, 200 µm; c and d, 500 µm. (D) Graphical representation of densitometry data obtained from four immunoblots using S58 (specific αMHC) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH; loading control) antibodies. On average a 63% reduction in αMHC protein levels was observed in the αMHC morpholino knockdown (αMHC kd) hearts in comparison to age-matched controls. Western blot control tissue consisted of hearts taken from both untreated and standard control embryos. Densitometry analysis showed no statistical differences between the two control groups; therefore the results were combined and compared against knockdown heart data. anova was used to assess statistical significance between the control and αMHC knockdown hearts (P= 0.034). n = 4 separate blots, with each sample on each blot containing three hearts pooled. OD, optical density.

Alpha myosin heavy chain is expressed during all stages of early atrial septal development and knockdown leads to a reduction in protein

Previous studies have demonstrated that αMHC is expressed from HH9 specifically in the developing chick atrial chamber (Oana et al. 1998; Somi et al. 2006) and to the atrial chamber and septum in human embryonic tissue (Wessels et al. 1991, 2000). This study aimed to extend previous studies by verifying the immunoreactivity of αMHC not only in the developing atrial chamber but also in the emerging atrial septum in the chick. αMHC was found to be expressed in the HH12 heart prior to atrial septum initiation (Fig. 1Ca). Subsequently, a slight protuberance of the dorso-cranial wall could be observed at HH14, which was more clearly defined as a very small septum by HH15, with staining of αMHC throughout the atrial chamber and septum (Fig. 1Cb). This immunoreactivity to αMHC persisted at HH19 when the atrial septum was clearly defined and at HH24 when the septum was undergoing fusion with the endocardial cushions (Fig. 1Cc,d).

Western blot analysis on isolated hearts was performed in order to confirm that the αMHC morpholino was having a knockdown effect and leading to insufficiency of the protein. Knocking down at HH14 and harvesting at HH19 with 250 µm of experimental morpholino resulted in a 63% reduction of αMHC protein levels compared with control hearts (Fig. 1D; ‘control’ denotes both control types as the densitometry data for the untreated and standard control embryos were not statistically different from each other, their data was combined and compared against knockdown heart data). The reduction in αMHC protein levels in the knockdown embryos was shown to be statistically significant in comparison to control chicks (P = 0.034).

Knockdown of alpha myosin heavy chain can lead to an enlarged heart and aberrant looping

In the chick, cardiovascular looping is initiated at HH10 and consists of the heart rotating to the right hand side, with the bulboventricular region bulging ventrally and the atria and sinus venosus moving dorsally (Patten, 1925; Hamburger & Hamilton, 1951; Sissman, 1970). Upon knockdown at HH12 and harvesting at HH19, it was noted that the embryos appeared to have a normal phenotype externally except for one αMHC knockdown embryo (1/9; Table 1). In this abnormal embryo an over-rotation of the entire heart was observed (Fig. 2E shows an example of this phenotype in a chick knocked down at HH14). In all of the HH14 control hearts (n = 54; Fig. 2A,B) and the majority of αMHC knockdown embryos (22/26) observed in this study, cardiovascular looping appeared normal (Fig. 2C). In four of the HH14 knockdown chicks, however, abnormal external phenotypes were detected (Table 1;Fig. 2D,E). In three of the chicks an over-rotation of the entire heart was observed, which appeared to originate from the truncus arteriosus and sinus venosus regions (Fig. 2E). One of these embryos was also growth restricted in comparison to controls. The fourth chick observed appeared to have formed an additional groove in the outflow tract, which had undergone rotation but bulbo-ventricular and atrio-ventricular grooves had not fully formed (Fig. 2D). In all cases the hearts looped correctly to the right hand side. Later on in development, at HH16, one knockdown embryo appeared to have an abnormal external cardiac phenotype. This embryo had an enlarged heart, particularly to the ventricular chamber (1/10; Fig. 2F). These data are summarized in Table 1. All control embryos had a normal phenotype, except for one untreated control embryo in the HH16 control group, which had a small heart and upon internal analysis the atrial septum was classified as small (1/53; Table 1). This latter phenotype was similar to that observed in the αMHC knockdown embryos, although a small heart was only detected in one of these chicks. Due to the large number of controls analysed, however, it is assumed that this phenotype was due to natural variation in development.

Fig. 2.

Knockdown of alpha myosin heavy chain (αMHC) can lead to abnormal looping or an enlarged heart in the chick. In order to reduce αMHC expression, morpholino was applied at either Hamburger and Hamilton (HH) stage 14 (C–E) or HH16 (F) and appropriate untreated (A) and standard control morpholino (B) embryos were analysed using light microscopy. All embryos were harvested at HH19. In comparison to the controls (A and B) the αMHC knockdown (αMHC kd) embryos were either phenotypically normal (C), had an abnormally looped heart (D and E) or the heart was enlarged (F). The asterisk denotes the sharp bend between the outflow tract (OT) and ventricle (V) that was seen in some of the αMHC knockdown embryos (D–F) but not in the controls (A and B). H, head; StC, standard control; Scale bars: in A, 500 µm (same for all panels).

Knockdown of alpha myosin heavy chain leads to abnormal atrial septal development

In this study, the aim was to characterize the role of αMHC during early cardiogenesis in a temporal and stage-specific manner, with knockdown occurring before atrial septum initiation (HH12), as the septum is in the process of initiating (HH14) and once septation has been initiated (HH16). Upon knockdown of αMHC at all stages, the atrial septal phenotype was found to fall into two classes, of which representative examples along with controls are shown in Fig. 3A. Upon application of the standard control morpholino, the development of the septum was found to form normally and be indistinguishable from the untreated control phenotype (Fig. 3Aa–d shows representative examples of normal atrial septal development in control embryos at HH19). Upon knockdown of αMHC (and following blind analysis) the septum was sometimes observed to have formed but was growth restricted in comparison to controls (representative examples of this phenotype are shown in Fig. 3Ae,f). In contrast, in other embryos the septum often failed to form with a normal septal structure not encountered in any of the serially sectioned cardiac slices. It is interesting to note, however, that in these embryos the dorso-cranial wall appeared thickened with a slight outgrowth in the region where a septum would be expected (Fig. 3Ag,h). In addition, the frequency of the atrial septa phenotypes varied at the different stages analysed. Examination of atrial septation showed that, upon knockdown of αMHC at HH12 and harvesting at HH19, five of the embryos had absent septa, although a thickened dorso-cranial atrial wall was observed (5/9) and four embryos had a small septum (4/9) when compared with control embryos (Table 1). Following knockdown of αMHC at HH14, an absent septum, only a thickened wall with a slight outgrowth, was observed in three chicks (3/14) and nine embryos were observed with a small septum (9/14; Table 1). The final two embryos knocked down at HH14 were deemed to have ‘abnormally structured’ septa (see below). Of the embryos knocked down at HH16, three had no septum, although a thickened wall with a slight outgrowth was seen (3/9), and five had a reduced septum (5/9; Table 1; Fig. 3A). In contrast to HH12 and HH14, at HH16 the penetrance was not found to be 100%, with one embryo having a normal septum (1/9) in comparison to controls (Table 1). In addition to analysis of atrial septation in αMHC knockdown chicks, the gross morphology of other cardiac features was analysed. The endocardial cushions, mesenchymal cap, trabeculae, ventricles and outflow tract were investigated but all features were consistent with those observed in control hearts.

Fig. 3.

Abnormal atrial septal development upon knockdown of alpha myosin heavy chain (αMHC). (A) Untreated (a and b), standard control (StC) (c and d) and αMHC knockdown (αMHC kd) (e–h) embryos were sectioned, stained with the nuclear marker haemalum and phenotypically analysed using light microscopy. Morpholino was applied at Hamburger and Hamilton (HH) stage 16 and chicks were harvested at HH19 (HH16/19; c–h). Micrographs b, d, f and h are high magnifications of the atrial chambers seen in a, c, e and g, respectively. In the untreated and StC embryos, normal atrial septal development can be seen, with a large septum having grown from the dorso-cranial atrial wall (arrows in a–d denote atrial septum). In contrast, upon knockdown with αMHC an atrial septum was often observed but was reduced in comparison to control septa (arrows in e and f compared with controls a–d). Alternatively, an atrial septum was not detected but a thickening of the dorso-cranial wall was seen (arrows in g and h). At, atrium; OFT, outflow tract; V, ventricle. Scale bars: a, c, e and g, 500 µm; b, d, f and h, 100 µm. (B) A photomicrograph of an αMHC embryo that was knocked down at HH14 and harvested at HH19 (HH14/19), and stained with an antibody against αMHC (kindly provided by Professor Moorman) to specifically detect the atrial chamber and counterstained with haemalum. Micrographs b, d and f are high magnifications of the atrial septa seen in a, c and e, respectively. Analysis of serial sections taken throughout the atrial chamber (At) revealed abnormal atrial septation. Initially three septa were detected emerging from the dorso-cranial wall of the atrial chamber (a and b), of which two subsequently dominated (c and d). Further sectioning through the atrium showed these two septa fusing at their inferior end leaving a large opening between them (e and f). Arrows denote atrial septa. Scale bars: 100 µm.

Interestingly, in one embryo knocked down with αMHC morpholino at HH14 (1/14), severe anomalies in the architectural development of the septa itself were seen. Initially, instead of one normal septum emerging from the dorso-cranial wall, three septa were detected (Fig. 3Ba,b). Moving dorsally through the atrium, two of the septa dominated with both initially appearing with a normal structure (Fig. 3Bc,d). However, further sectioning through the atrial wall showed that these two septa fused at their most inferior end with a large cavity remaining between the septa (Fig. 3Be,f), before finally both septa terminated. In addition, one chick (1/14) that was used in the phalloidin study appeared to have a double septum with two small, but septal-like, structures emerging from the dorso-cranial wall of the atrium (Fig. 4F). None of the other knockdowns or control embryos had an abnormally structured septum at any of the stages analysed.

Fig. 4.

Reduction of embryonic myosin heavy chain leads to a disrupted actin cytoskeleton in the atrium. Untreated control embryos were harvested at Hamburger and Hamilton (HH) stage 19 (A–C), whereas HH14 embryos initially had alpha myosin heavy chain (αMHC) morpholino applied prior to harvesting at HH19 (HH14/19; D–F). Embryos were subsequently stained with phalloidin to label the actin filaments and counterstained with propidium iodide (PI) (A, B and D–F) or Hoechst (C) to label the nuclei. Upon confocal microscopy, a distinct septum (arrow in A and C) could be detected within the atrial chamber (At) of the untreated control embryos. In contrast, the αMHC knockdown (αMHC kd) embryos had a slight outgrowth from the atrial dorso-cranial wall (arrow in D), with a definitive septum not observed. Upon higher magnification, phalloidin staining can be clearly seen to label distinct actin fibrils in both the untreated control and the knockdown embryos (arrows in B and E). However, the fibrils in the αMHC kd atrial wall appear thinner and punctuate in comparison to the untreated fibrils. Boxed areas in A and D are the regions shown in B and E, respectively. A more intense staining was observed in four out of five chicks at the leading edge of the septum in control hearts (C) but this feature was not observed in the αMHC kd chicks (D). An αMHC knockdown embryo is also shown with two atrial septa emerging from the roof of the atrial chamber (F). Scale bars: A, 25 µm (same in C, D and F); B, 5 µm (same in E).

Reduced alpha myosin heavy chain leads to disrupted cytoskeleton in the heart

In order to determine the integrity of the actin cytoskeleton upon knockdown of αMHC, phalloidin staining of the actin filaments was performed on embryos knocked down at HH14 and harvested at HH19. The control embryos consisted of untreated controls (3/5) and standard controls (2/5). Upon phenotypic analysis of the knockdown embryos, the heart appeared to be normal externally in all embryos (these embryos are included in Table 1).

Upon internal phenotypic analysis of the developing atrial chamber stained with phalloidin, there was either an absent septum with a thickened wall or a small septum in all of the αMHC knockdown embryos (3/4) with the exception of the ‘double septa’ heart discussed above (1/4; Table 1; Fig. 4F). Interestingly, in both the knockdowns and controls, the phalloidin staining appeared to show actin filaments of good integrity, with the actin filaments being easily distinguishable in the atrium and often being seen surrounding cells in a cortical fashion (Fig. 4D,E in comparison to Fig. 4A,B). In contrast, the actin filaments in the knockdown atrial chambers appeared thinner and some had a discontinuous, punctate staining pattern when compared with control atria (Fig. 4E compared with Fig. 4B). In addition, in the control hearts the phalloidin staining was usually found to be slightly more intense at the atrial septum leading edge (4/5; Fig. 4C), a feature not seen in the αMHC knockdown hearts.

Discussion

The MHC6 (human homologue of chick αMHC) is known to be involved in normal atrial septal development, with two families with ASDs found to have mutations in MHC6 (Ching et al. 2005). In addition, mutations in MHC6 have also been reported in individuals with cardiomyopathy (Carniel et al. 2005). Mutations in MHC7 (human homologue of chick βMHC) are a major cause of cardiomyopathy (Keren et al. 2008), with some mutations associated with skeletal muscle myopathies (Tajsharghi et al. 2003, 2007; Meredith et al. 2004). In addition, four members of a family carrying a novel MHC7 mutation causing non-compaction of the ventricular myocardium (11 affected individuals in total) had an ASD (Budde et al. 2007). Interestingly, mutations in another structural protein, cardiac α actin, have also been associated in probands with ASDs (Matsson et al. 2008) or cardiomyopathy (Olson et al. 1998; Mogensen et al. 1999; Olson et al. 2000), with some individuals having both defects (Monserrat et al. 2007). Therefore, these studies suggest a link between structural proteins, cardiomyopathy and a common congenital heart defect (Wessels & Willems, 2008).

From the literature, mutations in MHC6 (human homologue of αMHC) can cause both dilated and hypertrophic cardiomyopathy, although at a much lower frequency in comparison to genes such as MHC7, troponin T and α tropomyosin (Keren et al. 2008). To date, a total of five different mutations in five probands have been described for MHC6, of which three had dilated and two hypertrophic cardiomyopathy (Niimura et al. 2002; Carniel et al. 2005). In contrast, cardiomyopathy was not observed in one family with eight members afflicted with an ASD and carrying a mutation in MHC6 (Ching et al. 2005). Interestingly, our data support a role for αMHC in cardiomyopathy at a low frequency, with only 1/45 embryos knocked down with αMHC having an enlarged heart. This may be explained by the very low expression of αMHC in the ventricles compared with the atrium during development and in the adult (Wessels et al. 1991; Oana et al. 1998; Reiser et al. 2001; Somi et al. 2006). It is possible that fewer sarcomeres would be affected in the ventricle with reduced αMHC in comparison to βMHC, which is the predominant MHC protein in the ventricular chambers.

This study has involved a more detailed analysis of the role that αMHC plays during cardiogenesis. αMHC was knocked down in a temporal manner at HH12, HH14 and HH16 in order to delineate the roles that αMHC plays in development and maturation of the atrial septum. A reduction of αMHC led to a failure of correct septal initiation with a thickening of the atrial roof observed in place of a definitive septum in 11/32 embryos. In a further 18/32 embryos, a definitive septum was present but it was reduced in size in comparison to controls. Both of the atrial septal phenotypes were observed at all stages of knockdown, and suggest that αMHC is important in initiation and in the normal growth of the atrial septum. In addition, in two embryos where septum initiation did occur, specification and growth were grossly abnormal, with defects to the myocardial architecture. In these two embryos, either two or three septa were seen emerging from the dorsocranial wall of the common atrial chamber and, in one of these hearts, two of these septal structures fused inferiorly with a large cavity remaining between. Therefore, these data suggest that, upon knockdown of αMHC, the embryos failed to correctly specify one septum to form and two or three septa were initiated instead. It is also of interest to note that there was a high penetrance of the phenotypes in the development of the septum, with 31/32 of the knockdown embryos demonstrating abnormal atrial septal development. There does seem to be variable penetrance of the phenotype, as both a failure of correct septal initiation and the presence of a small septum were seen in the same sub-groups. This could be due to normal variation, as septum development is an ongoing process at the stages of development in this study. Gene penetrance and expressivity are considered to be important in both normal and mutated genes and these factors have been implicated in humans with ASDs (Benson et al. 1998). Environment has also been widely accepted as a contributing factor towards phenotypic presence/severity (Gibson, 2008) and, although every care was taken to make environmental conditions as stable as possible throughout the experiments, it is impossible to control for the exact conditions within the egg. An alternative hypothesis is that the variable penetrance could be due to some embryos being knocked down more and some less than the average of 63% seen by western analysis, leading to some embryos having a more severe phenotype than others.

The endocardial cushions, mesenchymal cap and dorsal mesenchymal protrusion are important in normal atrial septation (Webb et al. 1998; Anderson et al. 1999; Wessels et al. 2000; Snarr et al. 2007; Goddeeris et al. 2008). It is interesting therefore that the endocardial cushions and mesenchymal cap appeared normal in the αMHC knockdown embryos analysed, with the dorsal mesenchymal protrusion difficult to distinguish as it was at an early stage of its formation. However, these structures are likely to be important in the later stages of atrial septation, rather than the early stages of septum initiation and early development. Upon disruption of sonic hedgehog signalling in a mouse model, the primary atrial septum was found to develop normally, although the dorsal mesenchymal protrusion did not form and maturation of the endocardial cushions was defective leading to a failure of fusion and atrioventricular septal defects (Goddeeris et al. 2008). Therefore, these data suggest that different mechanisms are important for early stages of atrial septation in comparison to later stages that involve fusion of the septum.

Upon knockdown at HH12 and HH14, when cardiac looping is already underway, in some embryos (1/9 and 4/26, respectively) the looping failed to progress normally, with the curvature between the outflow region and the ventricular chamber either more pronounced or with an extra groove when compared with controls. Similarly, abnormal looping was found upon knockdown of cardiac α actin (Matsson et al. 2008). Interestingly, a role for actin has been proposed in cardiac looping possibly by causing tension and therefore enabling dextro-looping (Cooper, 1987; Itasaki et al. 1991; Latacha et al. 2005). Furthermore, myosin and actin are known to be in close association in the sarcomere. Therefore, it may well be that disruption of either actin or myosin might disrupt this close relationship, hence disturbing the correct degree of tension that the heart requires for looping to progress normally.

Previous studies have demonstrated that, although the sarcomere components are present in the tubular heart (late HH9/early HH10), their localization in the immature is different to that of the mature myofibrils (Ehler et al. 1999; Du et al. 2008). Actin staining can be seen to be punctate and more diffuse in the immature filaments (Ehler et al. 1999). By the time that the heart starts beating in a regular fashion (late HH10/early HH11), the myofibrillar proteins of the sarcomere are fully assembled, with a striated staining pattern observed for actin. It is therefore possible that, as actin and myosin have an intimate relationship in the sarcomere, both may play a role in normal looping of the heart (this study and Matsson et al. 2008). Actin is also known to be important in cellular processes such as migration (Cramer et al. 1997; Cavey et al. 2008). By performing phalloidin staining on the knockdown hearts, we aimed to determine the integrity of the actin cytoskeleton. Clearly defined filaments were detected in all knockdown hearts (4/4); however, these filaments appear to be thinner in comparison to those in control chicks. In addition, many of the filaments in the knockdown, but not control, hearts had a discontinuous pattern, suggesting immaturity of the structures. These data therefore suggest that, upon knockdown of αMHC, many of the myofibrils appear immature, possibly because there are fewer αMHC proteins available to assemble into mature sarcomeres.

In this study we have observed that a reduction in the amount of αMHC protein can lead to a failure of a septum to emerge, for multiple septa rather than a single septum to be initiated and for the septum to be reduced or form abnormally. We propose that decreased αMHC protein leads to fewer mature actin/myosin bundles, which in turn leads to a failure of an atrial septum to develop normally, possibly by decreased cell migration, which is supported by the literature that suggests a role for actin in cell migration (Cramer et al. 1997; Cavey et al. 2008). The punctate staining pattern observed in some of the actin filaments in the αMHC knockdown hearts suggests that some of the actin/myosin functional units, the sarcomeres, were immature. In addition to abnormal looping, when α cardiac actin was knocked down in the chick, the septa were growth restricted (Matsson et al. 2008), similar to the αMHC knockdown phenotype seen in this study.

Therefore, appropriate interaction between actin and myosin, and normal formation and maturation of actin filaments in the sarcomeric cytoskeleton, may well be required for cardiac looping and atrial septation to progress normally. The data presented here suggest that insufficiency of αMHC in the chick, despite the variable phenotypes observed, is a good model for the traits observed upon loss of function due to a mutation of the homologous gene in human. In the chick we have shown that αMHC is critical for the initiation, specification and normal growth of the atrial septum, and plays an important role in modelling the early heart.

Supporting information

Additional Supporting Information may be found in the online version of this article:

Fig. S1 Morpholino uptake in Hamburger and Hamilton (HH) stage 16 chick embryos after 15 min. Dark (A) and light (B) field microscopy of embryos that had 250 μM of fluorescein-tagged standard control morpholino applied. Upon application of the morpholino at HH16 with harvesting after 15 min (HH16/15 min; A and B), fluorescein uptake can clearly be seen throughout the embryo (A). Arrow denotes heart. Scale bar, 500 μm (same in A).

Fig. S2 Morpholino uptake in Hamburger and Hamilton (HH) stage 14 chick heart. Lissamine-tagged standard control was applied at HH14 and harvested at HH19 (HH14/19). The heart was then scanned by confocal microscopy with images captured at 5 μm intervals. Fluorescence can be seen throughout the heart. At, atrium; V, ventricle.

Fig. S3 Morpholino uptake in Hamburger and Hamilton (HH) stage 12 and HH16 chick hearts. Fluorescein-tagged morpholino was applied at either HH12 (A) or HH16 (B) and chicks were harvested at HH19. The heart was scanned with infrared pulse laser and images captured through the heart at 42 μm intervals. Fluorescence can be observed throughout the heart. At, atrium; OFT, outflow tract; V, ventricle. Scale bars: 200 μm in A; 100 μm in B.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.