Results
Expression Analysis of Suz12, Utp6 and Adap2 Genes in Mouse Reveals That Adap2 Is Expressed During Key Stages of Heart Development
In order to elucidate the expression pattern of Suz12, Utp6 and Adap2 genes in mouse, we performed in situ hybridisations using whole mounts at different stages of development, ranging from 7.5 to 11.5 dpc.
The gene which revealed the most interesting expression pattern was Adap2, since it was visible in heart between 9 and 10.5 dpc (figure 1) during fundamental phases of cardiac morphogenesis, namely heart looping (beginning at 8 dpc), endocardial cushion formation (10 dpc), and septation of the outflow tract, atria, and ventricles (10.5 dpc). In particular, the strongest Adap2 mRNA hybridisation signal was seen in the heart atria and ventricles at 9.5 dpc (figure 1E), but its expression in the heart was visible as of 9 dpc (figure 1C) and was still present in the atria and ventricles at 10.5 dpc (figure 1F). We also performed in situ hybridisations on cryosections of 15.5 dpc embryos in order to assess if Adap2 transcript is also present in the heart during the later stages of fetal cardiac development. Our experiments demonstrated that the expression of Adap2 in the heart continues to be maintained at least until 15.5 dpc, in the ventricles and atria (figure 1H).
(Enlarge Image)
Figure 1.
Expression of Adap2 in whole-mount mouse embryos and mouse cryosections. (A–G) Whole-mount in situ hybridisation on embryos from 8.25 dpc to 11.5 dpc with an Adap2 specific probe. (A) 8.25 dpc, expression at the midbrain/hindbrain boundary. (B) 8.5 dpc, expression in the gut tube. (C) 9 dpc, expression in forebrain, midbrain, hindbrain, heart (arrow), otic vesicles, gut tube. (D) 9.25 dpc, expression in forebrain, midbrain, hindbrain, otic vesicles, heart (arrow), posterior part of the gut tube. (E) 9.5 dpc, expression in forebrain, midbrain, hindbrain, otic vesicles, heart (arrow), gut tube. (F) 10.5 dpc, expression in forebrain, midbrain, hindbrain, otic vesicles, heart (arrow), gut tube. (G) 11.5 dpc expression in midbrain, inner ear, forelimbs, weakly in hindlimbs. (H) In situ hybridisation on cryosection of a 15.5 dpc embryo showing Adap2 expression in heart atrium (a) and ventricle (v).
Conversely, Suz12 evidenced a more spatially and temporally restricted expression in heart, with a clear hybridisation signal only at 10.5 dpc in the atrium, while Utp6 revealed no expression in heart at any analysed stages (see online supplementary figure S1 http://jmg.bmj.com/content/51/7/436/suppl/DC1).
Based on this evidence, we held ADAP2 the most interesting candidate gene for CVMs occurrence, and we used zebrafish as a model organism to investigate in vivo its role during vertebrate heart development.
adap2, the ADAP2 Zebrafish Ortholog, Is Required for Proper Cardiac Morphogenesis
In order to explore the spatio-temporal expression pattern of adap2, the ADAP2 zebrafish ortholog (Ensembl Gene ID: ENSDARG00000070565), we performed RT-PCR and whole-mount in situ hybridisation assays. adap2 transcript was detected by RT-PCR at all analysed stages, from cleavage up to 120 hpf (hours post-fertilisation), as well as in the oocytes, indicating that the gene is maternally and zygotically expressed. Furthermore, adap2 mRNA was present in all analysed adult tissues, including heart (see online supplementary figure S2 http://jmg.bmj.com/content/51/7/436/suppl/DC1). Whole-mount in situ hybridisation (WISH) revealed that adap2 transcript was present in the heart at 2 dpf (days post-fertilisation) and 3 dpf stages, in the region corresponding to bulbus arteriosus (see online supplementary figure S2 http://jmg.bmj.com/content/51/7/436/suppl/DC1).
In order to investigate the potential role of adap2 during zebrafish heart development in vivo, we performed loss-of-function experiments by injecting two independent translation-blocking morpholinos (adap2-MO and UTR-adap2-MO) which target the region surrounding adap2 translation start codon and the 5'-UTR region, respectively. The injection of a control morpholino (std-MO) with no targets in zebrafish was used as control of the microinjection. At 2 dpf, most of embryos injected with 0.3 pmol of adap2-MO (morphants), unlike std-MO injected embryos, displayed blood circulation defects and curved tail (figure 2). Lower doses caused no circulatory defects. For the analysis of injected embryos, we focused our attention on 2 dpf, stage at which the circulation is surely started and the cardiac looping occurred in control embryos. At this stage, 61% (n=94) of embryos injected with 0.3 pmol/embryo of adap2-MO showed one or more blood circulatory defects, such as the total loss of circulation (21%), accumulation of blood cells in the trunk and/or tail region (48%) and blood stases in the head (13%) (figure 2D–F,G). All these circulatory defects were noticed in both adap2 morphants which showed a body axis comparable with that of control embryos and morphants which displayed a bent tail phenotype (71%, n=94). The injection of the second translation-blocking MO, UTR-adap2-MO, caused in vivo qualitatively similar defects to the first injected MO, though with a different penetrance (see online supplementary figure S3 http://jmg.bmj.com/content/51/7/436/suppl/DC1).
(Enlarge Image)
Figure 2.
adap2 knockdown causes circulation defects in zebrafish. (A) Lateral view and (B) detailed image of the trunk-tail region of std-MO-injected embryos at 2 dpf. (C, E and F) Lateral view and (D) detailed image of the trunk-tail region of adap2-MO-injected embryos at 2 dpf. Anterior to the left. Black arrows: blood stases in the tail region; arrowhead: blood stasis in the head. (G) Percentage of circulation defects in adap2 morphants at 2 dpf (n=94): 21% of the adap2 morphants displayed no blood circulation, 48% blood stases in the trunk-tail region and 13% blood stases in the cephalic region.
To rule out that circulation defects could be caused by alterations of vascular development, we carried out adap2 loss-of-function experiments in the tg(flk1:EGFP) zebrafish transgenic line, where EGFP expression is controlled by the endothelial-specific flk1 promoter (see online supplementary figures S4 and S5 http://jmg.bmj.com/content/51/7/436/suppl/DC1). At 2 dpf, adap2 knocked-down embryos revealed no gross defects in vascular development, with correct development of main axial vessels, dorsal aorta (DA) and cardinal vein (CV), indicating a normal vasculogenesis. Weak defects in intersomitic vessels (Se) were observed only in those embryos with a marked curved tail, suggesting that these alterations were likely caused by structural defects of body axis rather than by angiogenesis abnormalities.
The evidence that two independent morpholinos gave the same in vivo phenotypes confirmed the specificity of the adap2 morpholinos. Consequently, we present here data obtained on embryos injected with the adap2-MO, which we indicate as adap2 morphants.
The evidence that circulatory defects in adap2 morphants were not caused by vascular defects suggested that they were most likely derived from an abnormal heart development and functionality. To test this hypothesis, we injected adap2-MO or std-MO in embryos belonging to the tg(gata1:dsRed);tg(flk1:EGFP) double transgenic line, in which erythrocytes are labelled in red and endothelial cells are labelled in green; we observed the injected embryos under a confocal microscope (figure 3). At 2 dpf, control embryos displayed a normal heart morphology (figure 3A), while adap2 morphants showed a reduction of atrioventricular (AV) canal bending, a partial lack of atrium and ventricle separation, as well as a reduced ventricle size (figure 3B,C). All analysed embryos displayed blood circulation.
(Enlarge Image)
Figure 3.
adap2 loss-of-function affects normal heart morphogenesis in zebrafish. The hearts of double transgenic tg(gata1:dsRed);tg(flk1:EGFP) embryos injected with std-MO or adap2-MO were examined in vivo by confocal microscopy at 2 dpf. Erythrocytes and endocardium are labelled in red and green, respectively. Confocal images of the heart in (A) std-MO-injected embryo, in (B) adap2 morphant displaying normal morphology and in (C) adap2-MO-injected embryo with bent tail. All analysed embryos presented blood circulation.
The in vivo analysis of adap2 phenotype in morphants prompted us to investigate their heart morphology by a molecular approach, through whole-mount in situ hybridisation assays with the cardiac-specific marker cmlc2 (cardiac myosin light chain 2) (figure 4, see online supplementary tables S1 and S2 http://jmg.bmj.com/content/51/7/436/suppl/DC1). At 26 hpf, std-MO-injected embryos showed the linear cardiac tube correctly positioned ventrally in the left region of the embryo (left jog) (figure 4A). On the contrary, only 39% (n=59) of adap2-MO-injected embryos displayed, at the same stage, the correct leftward cardiac jogging (figure 4B), while another 39% showed no jog, with the heart tube situated centrally along the midline of the embryo (figure 4C). Finally, the remaining 22% of adap2 morphants was characterised by an inverted cardiac jogging (right jog) (figure 4D). At 2 dpf, std-MO-injected embryos hybridised with the cmlc2-specific probe presented a normal S-shaped heart with the ventricle positioned on the right of the atrium, indicating a correct D-looping process (figure 4E). Differently, only 22% (n=49) of adap2 morphants showed a heart morphology comparable to control embryos (figure 4F). The remaining adap2-injected embryos displayed either an intermediate phenotype with reduced looping (18%), or absence of looping with a completely linear heart tube (47%), or a reversed heart looping with the ventricle on the left of the atrium (12%) (figure 4G–I). Moreover, whole-mount in situ hybridisation assays with the ventricle-specific marker vmhc (ventricular myosin heavy chain) evidenced, at 2 dpf, a marked reduction of ventricle size in 64% (n=39) of adap2 morphants, confirming the in vivo observations (see online supplementary figure S6 http://jmg.bmj.com/content/51/7/436/suppl/DC1). The reduction of ventricle size was observed regardless of the heart looping phenotype (D-loop, no loop or reversed loop). Notably, the percentage of embryos showing reduced ventricle size was similar in adap2 morphants with or without blood circulation, 65% (n=29) and 60% (n=10) respectively, suggesting no relation between this defect and circulatory complications.
(Enlarge Image)
Figure 4.
adap2 loss-of-function experiments perturb zebrafish heart jogging and heart looping. Analysis of cmlc2 expression by in situ hybridisation was performed on std-MO and adap2-MO-injected embryos at 26 hpf and 2 dpf. The heart position in injected embryos was scored as left jog (normal; A and B), no jog (C) and right (reversed) jog (D) at 26 hpf and as D-loop (normal; E and F), reduced loop (G), no loop (H) and reversed loop (I) at 2 dpf. V: ventricle; A: atrium. (A–D) Dorsal views through the head, anterior to the bottom; (E–I) frontal views, head to the top.
adap2 Loss-of-Function Affects AV Valve Development
In order to shed light on the effect of adap2 knockdown on cardiac functionality, we analysed AV valve formation in zebrafish by carrying out histological sections of AV valve in std-MO and adap2-MO-injected embryos at different developmental stages. At 3 dpf stage, control embryos displayed correctly formed endocardial cushions in the AV canal connecting the two cardiac chambers (figure 5A). The adap2 morphants morphologically more similar to std-MO-injected embryos still showed proper heart morphology with normal endocardial cushions, the only evident defect being a mild reduction of ventricle size, as already evidenced (figure 5B). In adap2-injected embryos which in vivo showed an intermediate phenotype (bent tail and presence of blood circulation), a visible alteration of the endocardial cushions was observed, with a marked disorganisation of the cellular elements that will be forming the mature AV valve (figure 5C). Embryos with severe phenotype, that is, curved tail and absent circulation, showed serious alterations in the heart morphology, making impossible any consideration on endocardial cushion formation (figure 5D). The histological analysis of std-MO-injected embryos at 5 dpf evidenced a properly developed mature valve, recognisable as two flap-like structures in correspondence to the AV canal (figure 5E). At this stage, adap2-MO-injected embryos showing an in vivo mild phenotype were already characterised by evident defects of mature AV valve, whose cells resulted disorganised and poorly compact (figure 5F). The morphology of mature valves in morphants with curved phenotype and with blood circulation appeared more compromised, structurally disorganised, without the typical valvular shape and with cells irregularly disposed (figure 5G). Finally, the most affected adap2 morphants showed severe cardiac malformations: the heart appeared essentially as a linear-shaped structure, without a clear separation between the two chambers, and consequently it was impossible to analyse mature cardiac valve conformation (figure 5H). Moreover, longitudinal histological sections of adap2 morphants at 5 dpf evidenced an endocardial detachment from the myocardial layer notably in the atrial chamber (figure 5F,G).
(Enlarge Image)
Figure 5.
adap2 knockdown impairs the normal endocardial cushions and mature valve formation. Histological sections of std-MO and adap2-MO-injected embryos at 3 dpf (transversal sections) and 5 dpf (longitudinal sections) stained with haematoxylin and eosin. (A and E) Heart sections of control embryos at 3 dpf (A) and 5 dpf, with magnification of the valve region (E). (B and F) Heart sections of adap2 morphants with blood circulation and morphology comparable to controls at 3 dpf (B) and 5 dpf, with magnification of the valve region (F). (C and G) Heart sections of adap2 morphants with blood circulation and bent tail at 3 dpf (C) and 5 dpf, with magnification of the valve region (G). (D and H) Heart sections of adap2 morphants with no blood circulation and curved tail at 3 dpf (D) and 5 dpf, with magnification of the valve region (H). Arrowheads: endocardial cushions; double arrows: extracellular matrix (cardiac jelly) located between myocardium and endocardium.
To characterise at molecular level the cardiac AV valve defects displayed by embryos as a consequence of adap2 functional inactivation, we analysed, by means of in situ hybridisation experiments, the expression pattern of two markers, bmp4 (bone morphogenetic protein 4) and notch1b (notch homolog 1b), which at 2 dpf are specifically expressed within the myocardial and endocardial component of AV canal, respectively (figure 6A,E). At 2 dpf stage, 91% (n=46) of control embryos showed a bmp4-specific hybridisation signal precisely marking the myocardial component of AV canal, as expected (figure 6B). Differently, 51% (n=41) of adap2-MO-injected embryos displayed a disorganised and ectopically expanded bmp4-specific expression domain, notably as the ventricular chamber is concerned (figure 6C,D). These defects were observed in all the phenotypic classes of heart development. Similar results were obtained from the analysis of notch1b marker at the same stage, with 49% (n=43) of adap2-MO-injected embryos displaying an expanded and disorganised notch1b expression pattern (figure 6G,H). All these data highlight adap2 function in fundamental processes of zebrafish cardiac morphogenesis, notably heart jogging, heart looping, determination of ventricular size and AV valve formation.
(Enlarge Image)
Figure 6.
The expression of atrio-ventricular boundary markers is affected in adap2 morphants. The analysis of bmp4 and notch1b expression by in situ hybridisation was performed on std-MO and adap2-MO injected embryos at 2 dpf. (A and E) Schematic representation of bmp4 and notch1b expression domain in zebrafish heart at 2 dpf. The myocardium and endocardium-specific territories of bmp4 and notch1b expression are depicted in magenta. (B and F) Embryos injected with std-MO displaying a normal hybridisation signal. (C, D and G, H) Embryos injected with adap2-MO displaying expanded and disorganised bmp4 and notch1b expression domains. Frontal views are shown.
Overall, our findings provide compelling evidence that ADAP2 is involved in heart development, pointing to it as the most plausible candidate gene for the occurrence of congenital CVMs in NF1 microdeletion syndrome and, more generally, for the occurrence of sporadic and familial congenital CVMs.