1. Developmental Biology
Download icon

Mechanically activated piezo channels modulate outflow tract valve development through the Yap1 and Klf2-Notch signaling axis

  1. Anne-Laure Duchemin
  2. Hélène Vignes
  3. Julien Vermot  Is a corresponding author
  1. Institut de Génétique et de Biologie Moléculaire et Cellulaire, France
  2. Centre National de la Recherche Scientifique, France
  3. Institut National de la Santé et de la Recherche Médicale, France
  4. Université de Strasbourg, France
Research Article
  • Cited 0
  • Views 981
  • Annotations
Cite this article as: eLife 2019;8:e44706 doi: 10.7554/eLife.44706

Abstract

Mechanical forces are well known for modulating heart valve developmental programs. Yet, it is still unclear how genetic programs and mechanosensation interact during heart valve development. Here, we assessed the mechanosensitive pathways involved during zebrafish outflow tract (OFT) valve development in vivo. Our results show that the hippo effector Yap1, Klf2, and the Notch signaling pathway are all essential for OFT valve morphogenesis in response to mechanical forces, albeit active in different cell layers. Furthermore, we show that Piezo and TRP mechanosensitive channels are important factors modulating these pathways. In addition, live reporters reveal that Piezo controls Klf2 and Notch activity in the endothelium and Yap1 localization in the smooth muscle progenitors to coordinate OFT valve morphogenesis. Together, this work identifies a unique morphogenetic program during OFT valve formation and places Piezo as a central modulator of the cell response to forces in this process.

https://doi.org/10.7554/eLife.44706.001

Introduction

Heart pumping and shaping take place concomitantly during embryonic development. These two processes require a tight and dynamic coordination between mechanical forces and tissue morphogenesis. Heart valve development is a great model for studying these interactions when considering that heart valve defects are common congenital cardiac malformation in human (Hoffman and Kaplan, 2002; Øyen et al., 2009). The four-chambered heart contains two different sets of valves (Lin et al., 2012):

- Tricuspide valves including the two semilunar (SL) valves, the aortic valve and the pulmonary valve, as well as the tricuspide valve located between the right atrium and the right ventricle.

- A bicuspide valve called the mitral valve separating the left atrium from the left ventricle. Abnormalities of the arterial valve leaflets are the most-common congenital malformations, in particular bicuspid aortic valve (Hoffman and Kaplan, 2002). Aortic valves mainly derive from endocardial cushion progenitors with a potential contribution from other cellular sources (epicardial cells and neural crest) (Wu et al., 2017). Generally, valve formation depends on two main events: cell proliferation, which is mainly mediated by the vegf-nfat1 pathway, matrix deposition and an endothelial-to-mesenchymal transformation (endMT) under the control of Gata transcription factors (Laforest et al., 2011; Stefanovic et al., 2014), Notch signaling (Luxán et al., 2016), Smad/tgf-beta/Bmp, and Wnt-beta catenin signals (Combs and Yutzey, 2009). Nevertheless, recent evidence suggests that arterial valves develop differently from atrioventricular valves by differentiating directly from progenitors in the outflow wall independently from endMT in mouse (Eley et al., 2018).

Congenital valve defects may originate from developmental origins and/or abnormal haemodynamic forces between these two sets of valves, and it remains unclear how general these developmental programs are. Aortic valves are located in areas of high flow velocity and mechanical forces have a great impact on valve morphogenesis (Butcher et al., 2008). Abnormal blood circulation is widely recognized as a cardiovascular risk factor and abnormal mechanotransduction has been associated with valvulopathies (Bäck et al., 2013). Congenital heart valve malformations are usually associated with genetic mutations in genes essential for heart valve development, such as signaling factors (Notch1, TGFβ) for the aortic valves (Bäck et al., 2013), and actin-binding proteins (FilaminA) for the mitral valves (Sauls et al., 2012). The reoccurring discovery of genetic mutations linking valve defects with genes involved in controlling developmental programs (e.g., in NOTCH1, TBX5, GATA4, TBX20, LMCD1, TNS1, and DCHS1) (PROMESA investigators et al., 2015; Durst et al., 2015; Garg et al., 2005; Richards and Garg, 2010), has spurred interest in valve morphogenesis. A key issue is to further define the genetic or environmental causes of valve malformation.

The zebrafish constitutes a powerful model to study cardiac valve development and the role of mechanical forces at the cellular scale. Zebrafish heart is two chambered and contains three sets of valves (the outflow tract (OFT), atrioventricular (AVC) and the inflow tract (IFT) valve [Figure 1A]) that are all bicuspid (Beis et al., 2005; Hsu et al., 2019; Tessadori et al., 2012). While the developmental programs driving mitral valve development in response to mechanical forces start to be well established in zebrafish, less is known about OFT and IFT valves (Paolini and Abdelilah-Seyfried, 2018; Steed et al., 2016a). The cellular processes leading to valve formation are dynamic and are particularly challenging to address in vivo. Zebrafish heart valves originate from progenitors located in the ventricle and atrium that generate the valve leaflets through a coordinated set of endocardial tissue movements (Boselli et al., 2017; Pestel et al., 2016; Steed et al., 2016a; Steed et al., 2016b; Vermot et al., 2009). The sequence of cellular events leading to AVC valve formation in zebrafish embryonic hearts is initiated through cell shape changes that lead to EC convergence towards the AVC (Boselli et al., 2017) and cellular rearrangements that will form a multilayered tissue (Beis et al., 2005; Pestel et al., 2016; Scherz et al., 2008; Steed et al., 2016b). In the zebrafish AVC, blood flow and Klf2a control notch1b and bmp4 expression, both of which are necessary for valve formation (Vermot et al., 2009). Klf2a regulates the deposition of matrix protein (in particular Fibronectin1) in the valve forming area (Steed et al., 2016b), as well as Wnt signaling by controlling wnt9b expression (Goddard et al., 2017). The latter is consistent with the fact that canonical Wnt signals arise specifically in sub-endocardial, abluminal cells and that these Wnt signals are dependent upon hemodynamic forces in zebrafish (Pestel et al., 2016). In addition, Notch signaling is essential for aortic valve formation (Garg, 2016) and OFT development (MacGrogan et al., 2016; Wang et al., 2017). The role of mechanical forces during OFT valve development at the cellular and molecular scale, however, is largely unknown.

The OFT develops from 56 hpf to form functional valves at 144 hpf.

(A) Top: Z-section of the double transgenic line Tg(kdrl:nls-mCherry; myl7:GFP) showing the overall structure of the heart. Bottom: Scheme of the zebrafish heart with the endocardium, myocardium and smooth muscles and zoom on the OFT structure. The OFT includes the CA and the BA. The CA is the zone of the myocardial connection of the ventricle to the BA. The BA is after the ventricle and is surrounded by smooth muscles. Scheme adapted from Felker et al. (2018). OFT: outflow tract, IFT: inflow tract, AVC: atrioventricular canal, At: atrium, V: ventricle, BA: bulbus arteriosus, CA: conus arteriosus. Scale bar: 20 µm. (B) Z-sections of the double transgenic line Tg(fli:lifeact-EGFP; kdrl:nls-mCherry) at different time-points showing the endocardial OFT structure. Scale bar: 25 µm. (C) Schematic representation summarizing the formation of the valve leaflets over time and flow profile in the OFT during development (from 56 hpf to 144 hpf) showing the forward flow (black), retrograde flow (white) and no flow (grey) fractions with the velocity of the red blood cells (in mm/s) using the double transgenic line Tg(gata1:ds-red; kdrl:EGFP). V: ventricle.

https://doi.org/10.7554/eLife.44706.002

Fluid shear stress is an important environmental cue that governs vascular physiology and pathology (Baeyens et al., 2016), but the molecular mechanisms that mediate endocardial responses to flow are only partially understood. In zebrafish, the mechanosensitive channels Trpv4 and Trpp2 modulate endocardial calcium signaling and klf2a expression is necessary for valve morphogenesis and downstream pathway activation (Heckel et al., 2015; Steed et al., 2016b). Notch signaling is tightly involved in cellular mechanosensitive responses in human aortic valves (Godby et al., 2014) and Notch1 is a potent mechanosensor in adult arteries (Mack et al., 2017). More recently, it has been shown that stretch-sensitive channels from the Piezo family (Murthy et al., 2017) are important for vascular development (Li et al., 2014; Ranade et al., 2014) and lymphatic valve formation (Nonomura et al., 2018). In the embryo, Piezo channels exert essential roles during cell differentiation (He et al., 2018) and can affect lineage choice by modulating the nuclear localization of the mechanoreactive transcription coactivator Yap (Pathak et al., 2014). Nevertheless, the role of Piezo-mediated mechanotransduction during cardiac development and its potential targets remain unclear.

In the developing cardiovascular system, biomechanics is key for modulating flow propagation (Anton et al., 2013). In the teleost heart, the OFT constitutes a specialized organ comprising the conus arteriosus (CA) and the bulbus arteriosus (BA) (Figure 1A). The BA dampens the pressure wave down the arterial tree (Braun et al., 2003b). To perform its function, the BA expresses elastic fiber genes that are thought to provide the mechanical properties necessary for its physiological function (Braun et al., 2003a; Braun et al., 2003b; Keith et al., 1977). The BA is separated from the ventricle by the OFT valve and is composed of smooth muscle. The extracellular matrix (ECM) gene, elastin b, contributes to the development of the BA by regulating cell fate determination of cardiac precursor cells into smooth muscle via a process that involves the mechanotransducer Yap1 (Moriyama et al., 2016). How these factors contribute to OFT valve development and interact with other mechanosensitive pathways remains unclear.

In this study, we investigated the signaling events taking place during OFT valve formation and addressed their regulation by the mechanosensitive channels Piezo and transient potential channels (Trp) as well as the flow-responsive transcription factor Klf2a. We show that OFT valve formation proceeds via an initial stage of endothelial cell folding, which is associated with the generation of a cluster of smooth muscle cell progenitors surrounding the endothelial layer. Subsequent global tissue remodeling events result in the appearance of functional leaflets, which defines a unique process of valvulogenesis. Using live reporters to highlight the signaling changes accompanying these temporally coordinated cell-movement events and genetics, we identified Notch and Klf2 as key flow-dependent factors as well as Yap1 as necessary factors for the correct coordination of OFT valvulogenesis. We show that Piezo and Trp channels are key regulators of klf2 activity in the endothelium and that piezo modulates Yap1 localization in the smooth muscle cells, providing a molecular link between mechanosensitivity and cell signaling in the multilayered valve structure. These data describe the cell responses that are coordinated by the mechanical environment and mechanotransduction via mechanosensitive channels in the endothelium.

Results

Outflow tract valve morphogenesis is unique

In order to better understand the roles played by blood flow during outflow tract (OFT) valve development, we have developed imaging techniques to capture cardiac motion and analyze blood flow in the OFT. Live imaging of the double transgenic line Tg(gata1:ds-red; kdrl:EGFP) to follow red blood cells and endothelial cell wall movements reveal dramatic changes in intracardiac blood flow patterns during OFT valve development: as the heart matures, blood flow in the OFT is bidirectional until functional valve leaflets emerge in the OFT at 144 hpf (Figure 1A–C). Throughout development, the periods within the cardiac cycle in which reversing flow can be observed decrease in length until 144 hpf, the stage at which we could not observe reversing flow anymore (Figure 1C). Using the Tg(fli:lifeact-EGFP; kdrl:nls-mCherry) line, which labels endothelial cells, we found that these flow profile modifications are linked to changes in OFT tissue geometry and the state of OFT maturation (Figure 1B). At 56 hpf, the endothelium resembles a tube (Figure 1B,C), maturing into cushions by 72 hpf, into premature leaflets by 96 hpf and finally into elongated, thin leaflets by 144 hpf (Figure 1B,C).

To better characterize how the endothelium changes shape over time and how cells reorganize to form OFT valves, we performed photoconversion experiments using the Tg (fli:gal4FF;UAS:kaede) (Figure 2A). We photoconverted Kaede from green to red in the cells located in the anterior, middle or posterior part of the valve at 72 hpf and assessed their position at 96 hpf and 120 hpf (Figure 2A,B). The results suggest that the endothelium folds to form the valve without multilayering (Figure 2B,C and Figure 2—figure supplement 1A). Indeed, we could observe that the photoconverted cells remain attached to each other and do not show signs of delamination as observed in the AVC (Figure 2B,C and Figure 2—figure supplement 1B).

Figure 2 with 1 supplement see all
The endothelium contribution to emerging OFT valve leaflets.

(A) Experimental set-up for the photoconversion studies. Heart was stopped at 72 hpf, the region of interest exposed to 405 nm light to convert kaede from green to red (shown in magenta) fluorescent form and heart contraction was resumed until 120 hpf. Beating hearts were imaged at 96 hpf and 120 hpf by spinning disk microscopy. A. Anterior, P. Posterior. (B) Z-sections of the Tg(fli:gal4FF; UAS:Kaede) line just after photoconversion (72 hpf), and at 96 hpf and 120 hpf. The star highlights the photoconverted cell in the top (n = 6), middle (n = 5) and bottom (n = 4) part of the OFT valve. The other photoconverted cell anteriorly goes out of the frame. Scale bar: 20 µm. Results obtained from three independent experiments. (C) Schematic representation of the results of the photoconversion studies showing the folding of the endothelium in the OFT.

https://doi.org/10.7554/eLife.44706.003

Together, these results suggest that OFT valves form by a folding process that might involve the adjacent tissue.

The cellular contribution of the OFT valve led us to the hypothesis that the surrounding tissue could contribute to valve morphogenesis. We analyzed Fibronectin1 (Fn1) expression in the OFT by counterstaining the Tg(kdrl:EGFP) with the Fn1 antibody (Figure 3A–D). We found that Fn1 deposition in the OFT is different from that in the AVC (Steed et al., 2016a). At 72 hpf, cushions appear and Fn1 is observed in the cells around the endothelial cells, that are themselves surrounded at their base by myocardium in the posterior part (Figure 3A,B,C,D,E). Fn1 expression level increases at 96 hpf and delineates a group of cells surrounding the OFT as well as in the basal side of a few endothelial cells that form the cushions (Figure 3C, arrows, E). In the developed leaflets at 120 hpf, Fn1 expression is maintained within the leaflets (Figure 3D,E). We found that most of the cells surrounding the endothelium expressing Fn1 also express Elastin b (Elnb, Eln2 or Tropoelastin), a marker of smooth muscle cells (Grimes et al., 2006; Miao et al., 2007; Paffett-Lugassy et al., 2017) (Figure 3B’, C’, D’). We could confirm that the cells surrounding the Fn1 cells are not myocardial cells as the myocardium stops just before the BA region (Figure 3A,E). To better characterize the smooth muscle identity and their activity, we performed a DAF-FMDA assay and counterstained with Fn1 at 72 hpf, 96 hpf and 120 hpf (Figure 3—figure supplement 1A). These results suggest that all cells expressing Fn1 are also DAF-FMDA-positive (Figure 3—figure supplement 1A). Some of these cells also express Tg(wt1a:GFP) (asterisks in Figure 3—figure supplement 1B) a marker of epicardial cells (Peralta et al., 2013). One hypothesis could therefore be that they might have originated from epicardial precursors. Together, these results suggest a developmental sequence of OFT morphogenesis in vivo where endothelial cell reorganization is associated with changes in gene expression in the surrounding smooth muscle cell progenitors. This indicates that the OFT morphogenesis involves remodeling of not just the endothelium, but also of a group of smooth muscle cells that express Fn1 and Elnb, and are functionally active.

Figure 3 with 1 supplement see all
The OFT endothelium is surrounded by smooth muscle cell progenitors expressing fibronectin and elastin.

(A) Staining of Fibronectin1 (magenta) on Tg(myl7:GFP; fli1a:nls-mCherry), highlighting the myocardium (white) and the endothelium (green) at 72 hpf. Scale bar: 20 μm. V: ventricle. Fibronectin1 (anti-Fn1, magenta) counterstaining on Tg(kdrl:GFP) and Elastinb (anti-Elnb, magenta) counterstaining on Tg(fli:lifeact-EGFP) showing their expressions in the OFT at 72 hpf (B), (B’) respectively) at 96 hpf (C), (C’) respectively) and at 120 hpf (D, D’ respectively). Scale bar: 20 µm. Arrows show the Fn1 localisation within the valve leaflets. Results obtained from three independent experiments. (E) Scheme of the three layers shown in A’, B, C and D (magenta, smooth muscles; green, endothelium; grey, myocardium; Fibronectin1, magenta lines) at 72hpf, 96hpf and 120hpf.

https://doi.org/10.7554/eLife.44706.005

We conclude that the tissue remodeling occurring during OFT valve development is significantly different from AVC valve development where the endocardium is the main remodeling tissue (Beis et al., 2005; Pestel et al., 2016; Steed et al., 2016b).

Klf2, Notch signaling, and Hippo pathways are active in the OFT in different cell layers and are all necessary for proper OFT valve development

To elucidate how these early events are regulated, we sought to determine the mechanosensitive signaling pathways activated at these early stages of OFT valve formation. We first assessed the activity of a Klf2a reporter line (Tg(klf2a:H2B-GFP)) (Figure 4A), which is a well described flow responsive reporter (Heckel et al., 2015; Steed et al., 2016b) and the Notch reporter Tg(tp1:dGFP) (Figure 4B) which is well active in the progenitors of the AVC cardiac valves (Pestel et al., 2016). We could observe a specific activation of the Klf2a reporter in the OFT endothelium (Figure 4A,C and Figure 4—figure supplement 1), in particular in the ventricular part of the valve (posterior) from 72 hpf to 120 hpf (Figure 4C,D and Figure 4—figure supplement 1). Similarly, the Notch reporter Tg(tp1:dGFP) is specifically expressed in the OFT endothelium (Figure 4B,C), in the ventricular part of the valve from 72 hpf to 120 hpf (Figure 4C,D). Interestingly, the spatial activation of the reporter varies within the valve forming area - the transgene activation is stronger in the posterior part of the valve than the anterior part of the valve throughout the process of valve maturation (Figure 4C,D). We made similar observations for the Notch reporter (Figure 4C,D). These results suggest that Klf2a and Notch signaling are activated specifically in the part of the valve corresponding to where the OFT has the smallest diameter and where shear stress is expected to be the highest, consistent with the hypothesis that Klf2a and Notch signaling are flow-responsive.

Figure 4 with 1 supplement see all
Klf2a and Notch reporters are activated in the OFT endothelium Confocal z-section of the Tg(klf2a:H2B-GFP; fli:nls-mCherry).

(A) and Tg(tp1:dGFP) (B) at 72hpf. OFT: outflow tract, At: atrium, V: ventricle. Scale bar: 20 µm. (C) Confocal z-section of OFT valves expressing the Klf2a reporter and Notch reporter at 72 hpf, 96 hpf, and 120 hpf. Schemes explaining the considered anterior (grey) and posterior (green) parts of the valve at 72 hpf, 96 hpf, and 120 hpf. A: anterior, P: posterior. V: ventricle. Scale bar: 10 µm. (D) Quantification of the fluorescent intensity of the Klf2a (GFP over mCherry) and Notch (GFP over background) reporters in the anterior versus posterior part of the valves at 72 hpf (n = 6 embryos, p=0001 and n = 5 embryos, p=0,02 respectively), 96 hpf (n = 5 embryos, p=0005 and n = 6 embryos, p=0,0007 respectively) and 120 hpf (n = 6 embryos, p=0,01 and n = 5 embryos, p=0,01) in wild-type embryos using the student’s t-test. Boxplots: Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots. Results obtained from three independent experiments.

https://doi.org/10.7554/eLife.44706.007

We next investigated the expression of the Hippo effector Yap1 and a reporter of the Hippo pathway in vivo. We found that Yap1 is localized in the heart at 72 hpf, in particular in the OFT smooth muscle cells (Figure 5—figure supplement 1A,B). To better characterize its tissue-specific expression, we used the double transgenic line Tg(fli:lifeact-EGFP; kdrl:nls-mCherry) and could show that the smooth muscle cells surrounding the OFT, but also some OFT endothelial cells express Yap1 (asterisks in Figure 5—figure supplement 1B). To assess the activity of the Hippo pathway in the heart at 72 hpf, we made use of the Tg(4xGTIIC:d2GFP) and could see that the Yap/Wwtr1-Tead reporter was activated in the OFT endothelial cells (labelled using the kdrl:membrane-mCherry line) as well as the smooth muscle cells surrounding the OFT (labelled by the Elnb staining) (Figure 5A). To assess the implication of Yap1 during OFT valve development, we followed the same embryos over time and looked at the valve phenotype in yap1 mutant embryos, yap embryos and yap1 control embryos (Agarwala et al., 2015) (Figure 5B). When analyzed from 72 hpf until 120 hpf, a significant fraction of yap1-/- embryos displayed abnormal valves (17% at 72 hpf, 8% at 96 hpf, 17% at 120 hpf, n = 12) and an increasing fraction of yap1 embryos did not have recognizable OFT valves (58% at 120hpf, n = 12). Thus, these results suggest that yap1 is involved during OFT valve morphogenesis (Figure 5B) and that smooth muscle cell progenitors are likely to play a role in the process.

Figure 5 with 1 supplement see all
Hippo pathway effector Yap1 is active and Yap1 is essential for valve formation in the OFT.

(A) Confocal z-sections of the double transgenic line 4xGTIIC:d2GFP; kdrl:membrane-mCherry counterstained with the Elnb antibody and focused on the OFT. Scale bar: 20 µm. (B) Example of the valve phenotypes (normal, thick, delayed and no valve) and quantification of the phenotypes in the yap1+/+ controls embryos, yap1+/- and in yap1-/- mutant embryos. Chi-square test. n.s.: non significant, ***: p<10−3. Scale bar: 10 µm. Results obtained from two independent experiments.

https://doi.org/10.7554/eLife.44706.010

Together, these results show that Klf2a, Notch signaling, and Hippo pathways are active in the OFT during valve morphogenesis.

Klf2a and notch reporter activity is flow-dependent in the OFT

As blood flow is an important regulator of klf2a expression and cardiac valve formation, we next wanted to assess whether changes in flow properties impact Klf2a and the Notch reporter activity in the OFT.

We analyzed the reporters activity following injection of a morpholino specific for troponin T2a (tnnt2a), which is necessary for heart contraction and reliably mimics the sih mutants (Sehnert et al., 2002), to determine whether these signaling pathways were impacted when heart contraction is abnormal and/or are activated upon shear stress forces. As the absence of heart contraction can impact heart morphogenesis, we injected highly diluted tnnt2a morpholino (hypomorphic condition) into these two reporter lines (Figure 6A,B). Such treatment allows us to decrease heart function and flow without dramatically altering heart shape (Figure 6A,B, Videos 1 and 2). Depending on the knockdown efficiency in single embryos, this treatment leads to the generation of two groups of embryos: group1 where the heart beats ‘normally' (normal heart rate at 2–3 Hz and function) and group2 where the heartbeat is slower. In the group of ‘beating heart’ embryos (group1, Video 1), we still observe stronger GFP expression in the posterior part of the valve for both reporters at every stage analyzed (Figure 6A). In the group of embryos where the heart is still beating but at an abnormal slow rate (less than 2 Hz, group2, Video 2), we observe no difference between the anterior and posterior part of the valve (p=0.99) for the Klf2a reporter at 72 hpf (Figure 6B). For the Notch reporter, we observed no difference in fluorescence intensity at 72 hpf (p=0.1) (Figure 6B). In addition, we analyzed the localization of Fn1, Elnb and Yap1 by immunostaining in both ‘beating heart’ (n = 4, n = 7 and n = 7, respectively) and ‘slow beating heart’ (n = 6, n = 7 and n = 7 respectively) groups and could observe that Fn1, Elnb, and Yap1 are down-regulated in the smooth muscle cells of the ‘slow beating heart’ embryos (Figure 6D,F). Moreover, we assessed the BA diameter and the activity of the smooth muscle using the DAF-FMDA assay (Figure 6E). The results suggest that the ‘slow beating heart’ group has a smaller BA (p<0,001) and the smooth muscle are much less active (p=0,05) (Figure 6E,F). We next assessed valve morphology in which blood flow is altered due to slow heart contraction and selected the fish with almost no contraction (group2). All the tnnt2aMO-injected embryos have no valve (n = 9/9, p<10−6) (Figure 6C). To confirm the role of flow in the process, we analyzed the effect of altered blood viscosity and shear stress by lowering red blood cell content in the gata1 mutants (Vlad Tepes) as previously described (Steed et al., 2016b; Vermot et al., 2009). In vlad tepes mutant embryos (n = 21, p<10−2), more than 80% of the mutants displayed abnormal OFT valves (Figure 6C).

Klf2a and notch response, as well as the smooth muscle cell identity, are flow-dependent.

Quantification of the Klf2a and Notch reporter expressions in tnnt2a-morpholino injected embryos showing a ‘beating heart’ (p=0,27 and p=0,01 respectively). (A) and a ‘slow beating heart’ (p=0,7 and p=0,1 respectively) (B) at 72 hpf. N = 2 independent experiments. (C) Quantification of the phenotypes in the control (n = 24), vlad tepes mutant (n = 21 embryos from two independent experiments), tnnt2aMO-injected embryos (n = 9 embryos from two independent experiments). Chi-square test. **: p<10−2, ******: p<10−6. (D) Z-sections of the Tg(fli:lifeact-eGFP) counterstained with either Fibronectin1 (Fn1), elastin (Elnb) or Yap1 in tnnt2a-morpholino injected embryos (slow beating and beating heart). Scale bar: 20 µm. N = 2 independent experiments. (E) Z-section and quantification of the BA diameter and DAF-FMDA intensity in tnnt2a-morpholino injected embryos (p=0,0005 and p=0,05 respectively). Student’s t-test. Boxplots: Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots. (F) Scheme summarizing the down-regulation of the smooth muscle markers in ‘slow beating heart’ embryos compared to ‘beating heart”embryos.

https://doi.org/10.7554/eLife.44706.013
Video 1
Bright field videomicroscopy of the typical heart function of a tnnt2a MO group1 embryo where the heart is still beating normally.
https://doi.org/10.7554/eLife.44706.015
Video 2
Bright field videomicroscopy of the typical heart function of a tnnt2a MO group2 embryo where the heart is still beating but at an abnormal slow rate (less than 2 Hz).
https://doi.org/10.7554/eLife.44706.016

Together, these results suggest that the expression of both Klf2a and Notch reporters is flow-dependent and that mechanical forces associated with heart activity are necessary for valve development and smooth muscle cell identity in the OFT.

Klf2a regulates OFT valve morphogenesis via notch signaling activation and is necessary for the smooth muscle cells differentiation

As notch1b and klf2a are expressed in the OFT endothelium in response to flow forces, we hypothesized that they have a role during OFT valve morphogenesis. Therefore, we imaged the Tg(fli:gal4/UAS:kaede) and Tg(flli:lifeact-EGFP; kdrl:nls-mCherry) transgenic lines (Figure 7A,B). We looked at the phenotype of OFT valve endothelium in the klf2a-/- and notch1b-/- embryos at 72 hpf, 96 hpf and 120 hpf (Figure 7A,B). We found that most of the klf2a-/- embryos display proper valves at 72 hpf (n = 7/12), 96 hpf (n = 6/12) and 120 hpf (n = 6/12). However, 33% of the klf2a-/- embryos have abnormal ‘delayed phenotype’ at 72 hpf (n = 4/12), at 96 hpf (n = 4/12) and 120 hpf (n = 4/12) (Figure 7A). In this case, the OFT valves are larger, leading to big cushions instead of thin valve leaflets (Figure 7A). In addition, notch1b is also necessary for proper valve formation since 30% (n = 6/20) of the notch1b-/- embryos have a ‘delayed phenotype’ at 72 hpf, 35% (n = 7/20) at 96 hpf and 45% (n = 9/20) at 120 hpf (Figure 7B) while almost all control embryos have proper valves at 72 hpf (n = 6/7), 96 hpf (n = 7/7) and 120 hpf (n = 7/7) (Figure 7B). Next, we wondered whether a cross-regulation between notch1b and klf2a exists since they are both necessary for proper valve formation. First, we compared notch1b expression between klf2a-/- embryos and klf2a+/+ control embryos (Figure 7—figure supplement 1A). In controls, 85% and 91% of the embryos have notch1b expression in the OFT at 48 hpf and 72 hpf, respectively (Figure 7—figure supplement 1A–C). However, in klf2a-/- embryos, notch1b expression is altered with no clear expression defining the OFT region at 48 hpf (64%) and at 72 hpf (78%) (Figure 7—figure supplement 1A,B). However, in the reverse experiment, most of the embryos show proper klf2a expression in notch1b -/- embryos (61% at 48 hpf and 51% at 72 hpf) (Figure 7—figure supplement 1C).

Figure 7 with 1 supplement see all
Klf2a and notch are necessary for valve formation.

Quantification of the valve phenotypes at 72 hpf, 96 hpf and 120 hpf (normal, thick, delayed) in klf2a+/+(n = 11), and klf2a-/- (n = 12) using the Tg(fli:gal4FF/UAS:Kaede). (A) and notch1b+/+ (n = 7) and notch1b-/- (n = 20) using Tg(fli:lifeact-EGFP; kdrl:nls-mCherry) embryos. (B) Scale bar: 10 µm. N = 3 independent experiments. (C) Confocal z-sections of the Tg(tp1:dGFP) in klf2a+/+ and klf2a-/- embryos at 72 hpf, 96 hpf and 120 hpf. V: ventricle. Scale bar: 10 µm. (C’) Quantification of the fluorescent intensity of the Notch reporter (GFP over background) in the anterior versus posterior parts of the valves in in klf2a+/+ (n = 5) and klf2a-/- (n = 4) embryos. Statistical test were performed to compare the posterior intensities in klf2a+/+versus klf2a-/- at 72 hpf (p=0,05), 96 hpf (p=0,03) and 120 hpf (p=0,04). Student’s t-test. Boxplot: Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots. Results obtained from three independent experiments.

https://doi.org/10.7554/eLife.44706.017

In order to assess the regulation of the Notch pathway activity by Klf2a in the endothelium, we analyzed the Notch reporter activity in the posterior and anterior parts of the valve in klf2a+/+ (n = 5) versus klf2a-/- (n = 4) embryos. As for Notch reporter expression in wild-type (Figure 4C,D), the Notch reporter is significantly more expressed in the posterior part compared to the anterior part of the valves at 72 hpf (p<10−2), 96 hpf (p<10−2) and 120 hpf (p<10−1) (Figure 7C,C’) in the klf2a+/+ embryos. Interestingly, the stronger posterior expression is lost in the klf2a-/- embryos and the fluorescent intensity in the posterior part of the valve is significantly reduced in the klf2a-/- compared to klf2a+/+ at 72 hpf (p<10–1), 96 hpf (p<10−1) and 120 hpf (p<10−1) (Figure 7C,C’). In comparison, the expression in the posterior part of the valve in the notch1b-/- embryos compared to the notch1b+/+ control embryos is not significantly different at any time point analyzed (Figure 7—figure supplement 1D). To further assess whether Klf2 has an effect on the OFT formation, we performed Fn1, Elnb and Yap1 stainings on klf2a-/- and controls at 72 hpf. We could observe that Fn1, Elnb, and Yap1 are properly localized in controls (n = 13/13, n = 7/7 and n = 4/4, respectively), klf2a-/- (n = 5/6, n = 3/4, and n = 3/3) (Figure 7—figure supplement 1E).

These results suggest that klf2a and notch1b are involved for proper valve morphogenesis and that Klf2a modulates notch expression in the process. Moreover, klf2a does not seem necessary for the smooth muscle cell differentiation surrounding the endothelium.

Piezo channels regulate both the endothelial and smooth muscle cell markers expression

In order to decipher whether flow and mechanosensitive channels could be involved in the regulation the OFT valve formation, we proceeded to analyze the phenotype of different potent mechanosensitive channel mutants. We focused on the non-selective ion channels Trpp2, Trpv4, Piezo1 and Piezo2a (Figure 8—figure supplement 1A,B,C) that are known for their mechanosensitive properties (Coste et al., 2010; Köttgen et al., 2008; Li et al., 2014; Sharif-Naeini et al., 2009; Thodeti et al., 2009). First, we evaluated the relative fractional shortening (RFS) at 72hpf in the atrium and in the ventricle of each mutant and their corresponding controls (Figure 8—figure supplement 1A). We did not observe any significant difference in the RFS between mutants and their respective controls neither in the atrium nor in the ventricle, suggesting that heart function in these mutants is normal. As another readout of flow forces and heart function, we quantified the retrograde flow fraction (RFF) at 72hpf and 120hpf in these mutants and the time windows for forward, reverse or no flow. We found that they are equivalent in all mutants when compared to their respective controls (Figure 8—figure supplement 1B), confirming that heart function is not different in these mutants. We looked at valve morphology at 120 hpf when leaflets are normally fully formed. In control embryos, the valves are extending into the lumen (Figure 8—figure supplement 1C). We found that 33% of the trpv4 mutant embryos (n = 7) have normal valves and mainly display thick valves (67%). In the trpp2 mutant embryos (n = 20), a stronger phenotype is observed with only 10% of normal valves. We observed that some embryos are delayed with respect to valve phenotype, meaning that the valve forming area still displays cushions at 120 hpf instead of having leaflets (35%). The trpv4-/-; trpp2+/- (n = 8) has an intermediate phenotype with 25% of the embryos having normal valves. Finally, the trpp2-/-; trpv4-/- embryos (n = 11) have a stronger phenotype, with none of the embryos showing proper OFT valve development and displaying mainly a delayed valve phenotype (45%). Interestingly, the piezo1 mutant embryos (n = 14) show mainly a delayed valve formation (50%), similarly to the trpv4-/-; trpp2-/- embryos. The piezo2a-/- (n = 9) has a less stringent phenotype with mainly normal valves (44%) but nevertheless 11% of piezo2a-/- fish do not have valves at all. This phenotype is even more prevalent in piezo1-/-; piezo2a-/- embryos (n = 11), where none of the fish display proper valve development and most of them do not form any valves (36%).

To better characterize the valve phenotype in trpp2-/- and piezo1-/- embryos, we made use of the Tg(fli:lifeact-EGFP; kdrl:nls-mCherry) transgenic line and assessed the shape of the endothelium at 72 hpf (Figure 8A and Figure 8—figure supplement 2A), 96 hpf (Figure 8B and Figure 8—figure supplement 2A) and 120 hpf (Figure 8C and Figure 8—figure supplement 2A). trpp2-/- embryos display mainly thick valves at all time points (n = 4/8 at 72 hpf, n = 8/13 at 96 hpf and n = 5/11 at 120 hpf) (Figure 8A–C and Figure 8—figure supplement 2A). Indeed, the endothelial layer is not a single cell layer anymore (Figure 8B) which leads to a thicker leaflet at 120 hpf (Figure 8C). Although piezo1-/- embryos have mainly normal valves at 72 hpf (Figure 8A and Figure 8—figure supplement 2A) (n = 6/10), they display a delayed phenotype (still cushions) at 96 hpf (Figure 8B and Figure 8—figure supplement 2) (n = 4/10) and thick or delayed phenotype at 120 hpf (Figure 8C and Figure 8—figure supplement 2A) (n = 3/10 and n = 3/10 respectively). To assess the redundancy between trpp2 and piezo1, we performed the same experiment in piezo1-/-; trpp2-morpholino injected embryos at 72 hpf (Figure 8A), 96hpf (Figure 8B) and 120 hpf (Figure 8C). Although the trpp2-MOrpholino injected embryos do not show a phenotype as strong as the trpp2-/- mutant embryos at 96 hpf and 120 hpf (possibly due to less effective morpholinos at these stages), the double piezo1-/-; trpp2-MO do not clearly show a stronger phenotype than the single trpp2-/- or piezo1-/- embryos (normal valves in n = 6/7, n = 2/7 n = 4/7 and at 72 hpf, 96 hpf and 120 hpf respectively). These results suggest that Trp and Piezo channels are necessary for the proper folding of the endothelium from 72 hpf to 120 hpf.

Figure 8 with 3 supplements see all
Flow and mechanosensitive channels are necessary for proper OFT valve formation.

(A) Z-sections and quantifications of the valves phenotypes (normal, thick, delayed) of the Tg(fli:lifeaAct-EGFP; kdrl:nls-mCherry) at 72 hpf (B), 96 hpf (C) and 120 hpf (D) in trpp2+/+(n = 11, n = 10, n = 10), trpp2-/- (n = 8, n = 13, n = 13 from three independent experiments), piezo1+/+ (n = 10, n = 10, n = 9), piezo1-/-(n = 10, n = 10, n = 9 from two independent experiments), piezo1+/+; trpp2+/+ (n = 7), trpp2-morpholino injected embryos (n = 9) and piezo1-/-; trpp2-morpholino injected embryos (n = 7). Scale bar: 10 µm. (D) Z-section of the OFT stained with DAF-FMDA in piezo1+/+ and piezo1-/-. Scale bar: 20 µm. (E) Fibronectin1 (Fn1), elastin (Elnb) and Yap1 staining on Tg(fli:lifeact-eGFP) in piezo1+/+ (n = 12, n = 4 and n = 10 respectively) and piezo1-/- (n = 12, n = 4 and n = 10 respectively from three independent experiments). Scale bar: 20 µm. Z-sections (F) and quantification (F’) of the klf2a reporter (GFP over background) in the anterior and posterior parts of the valves in piezo1+/+ (n = 6) and piezo1-/- (n = 6) obtained from two independent experiments). Scale bar: 20 µm. Z-sections (G) and quantification (G’) of the klf2a reporter (GFP over background) in the anterior and posterior parts of the valves in trpp2+/+ (n = 6) and trpp2-/- (n = 6) (obtained from two independent experiments). Scale bar: 20 µm. Student’s t-test. Boxplot: Center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots.

https://doi.org/10.7554/eLife.44706.020

To better characterize the cell layer affected by piezo1 loss of function, we performed immunohistochemistry against Yap1 and two smooth muscle identity markers (Fn1 and Elnb) as well as their active functionality in piezo1+/+ controls and piezo1-/- mutants at 72 hpf (Figure 8D and Figure 8—figure supplement 2B). A bit more than two thirds of piezo1-/- embryos (n = 7/10) display reduced Yap1 expression as well as a down-regulation of Elnb (n = 9/12) and Fn1 (n = 3/4) (Figure 8E). However, the BA diameter as well as the DAF-FMDA fluorescence intensity are not affected in the piezo1-/- compared to piezo1+/+ (Figure 8D and Figure 8—figure supplement 2B) demonstrating that smooth muscle cells are still present in the piezo mutants. These results suggest a selective role of piezo1 in the regulation of the smooth muscle cell maturation and proper Yap1 localization in the OFT. We next assessed Klf2a reporter expression in piezo1-/- and trpp2-/-. We found that Klf2a expression was misregulated in piezo1-/- mutant embryos with stronger Klf2a reporter activation in the anterior and posterior part of the valve endothelium when compared to controls (p<10−3 for the anterior part and p<10−2 for the posterior part of the valves) (Figure 8F,F'). By contrast, we found that the posterior part of the valve has a decreased expression of GFP (p<10−2), highlighting a down-regulation of klf2a in the trpp2-/- embryos (Figure 8G,G’). These results suggest that Piezo1 inhibits klf2a overall the valve endothelium (Figure 8F,F’) while Trpp2 is required for klf2a activation in the posterior part of the OFT valve endothelium (Figure 8G,G’). To assess if the localization of piezo1 and trpp2 mRNA could explain the differential function of these channels, we performed RNAscope assay at 72hpf (Figure 8—figure supplement 3A). We found that trpp2 is ubiquitously expressed in the embryo, including in the different layers composing the OFT. Similarly, we found that piezo1 is expressed in both endothelium and smooth muscles, albeit at a lower level than trpp2 (Figure 8—figure supplement 3A). To confirm these results, we generated a transgenic reporter line with 3 kb of the piezo1 promoter upstream of the start codon (piezo1:nls-Venus). We observed the expression of the reporter line mostly in smooth muscles at 72hpf (Figure 8—figure supplement 3B) and cells of the endothelium of the OFT valve. Trpp2 immunohistochemistry showed that Trpp2 is expressed in the endothelium and the smooth muscles confirming that trpp2 is ubiquitously expressed in the OFT (arrow in Figure 8—figure supplement 3B). We conclude that Piezo1 plays a dual role in the OFT: it modulates klf2a expression in the endothelium and it is necessary for proper smooth muscle cell maturation around the OFT endothelium.

Discussion

Using cardiac live imaging and functional studies combined with in vivo reporter analysis, we uncover key mechanosensitive signaling pathways involved in OFT valve morphogenesis (Figure 9). We identify two tissue layers sensitive to mechanical forces in the OFT: (1) The endothelial cells where klf2a expression is modulated both by Piezo and Trp channels (2) The smooth muscles surrounding the endothelium, where Piezo channels regulate Yap1 localization and smooth muscle cell specific marker expression (Figure 9). These observations enable us to confirm the universal role of mechanical forces in cardiac valve morphogenesis and suggest a specific mechanism for OFT valve morphogenesis in which the origins of the valve progenitors, the implication of particular groups of cells, the mechanosensors involved and the impact of the mechanotransduction cascade are identified.

Working model summarizing OFT valve morphogenesis in response to mechanical forces.

Two cell layers forming the OFT respond to piezo1 activity: the endothelium (green) and smooth muscle cells (magenta). Endothelial klf2a expression is repressed by Piezo1. In the smooth muscle cells, the expression of elastin (Elnb), fibronectin (Fn1) and Yap1 is modulated by Piezo1. Fibronectin is localized in the smooth muscle cell layer and within the forming valve (magenta lines). Future work will help to decipher how the two cell layers interact with each other to modulate OFT valve formation. V, ventricle.

https://doi.org/10.7554/eLife.44706.025

Klf2a modulates notch signaling specifically in the OFT endothelium

Valve morphogenesis occurs in complex mechanical environments. In the AVC, endocardial cells experience both shearing forces and mechanical deformation due to the contraction of the heart and its associated blood flow. The situation is slightly different in the OFT because endothelial valvular progenitors are not surrounded by contractile cardiomyocytes but passive smooth muscle cells that can provide a counter force to flow and pressure. Here, we show that the main mechanosensitive pathways involved in AVC valve development are required for OFT valve development, even though the mechanical stimuli vary greatly between these two sets of valves. Previous studies have proposed that klf2a and notch1b are important during the formation of functional cardiac valve leaflets in zebrafish. Both are specifically expressed in the AVC (Beis et al., 2005; Pestel et al., 2016; Steed et al., 2016b; Vermot et al., 2009) and are transcriptionally misregulated in models of Cerebral cavernous malformation (CCM) where cardiac valve development is altered (Donat et al., 2018). Here, we show that both Notch signaling and klf2a are active in the endothelial cells of the OFT between 48 hpf and 120 hpf. During the process in which zebrafish cardiac AVC cushion remodel into valve leaflets, endocardial Klf2a expression and Notch activity are high on the luminal side of the developing valve leaflet, which is exposed to blood-flow, whereas their expression is lower on the abluminal side of the leaflet (Pestel et al., 2016; Steed et al., 2016b). In the OFT, Klf2a expression and Notch activation follow a different pattern because we could not clearly identify abluminal cells in the OFT. This might reflect a complete lack of endothelial to mesenchymal transition in the OFT by comparison to the AVC. Nevertheless, the impact of both pathways on valve morphogenesis remains the same as we found that klf2a-/- and notch1b-/- embryos show similar valve phenotypes and Notch signaling pathway is significantly decreased in klf2a-/- mutants. Even though the downstream players of the notch pathway remain to be established in the OFT, these results validate previous observations, suggesting that klf2a acts upstream of Notch signaling in the endocardium (Donat et al., 2018; Samsa et al., 2015; Vermot et al., 2009). Considering that Notch has been proposed to be mechanosensitive in blood vessels (Mack et al., 2017), an attractive hypothesis is that the Klf2-Notch axis could be a general mechanosensitive cascade in endothelial cells. In that case, it would be interesting to address notch activity in a flow related context, similarly to what is currently being done with Nf-kb, Klf2a, and other flow responsive pathways (Feaver et al., 2013). This assumption is particularly interesting in the context of sprouting angiogenesis and other aspects of angiogenesis where Notch is broadly required (Choi et al., 2017; Hasan et al., 2017; Pitulescu et al., 2017; Tammela et al., 2011). We further show that the expression of Fibronectin1 and Elastinb in the smooth muscle cells surrounding the OFT is not altered in the klf2a-/-embryos. Considering that Piezo is emerging as an important contributor to human diseases related to blood cells (Ma et al., 2018; Zarychanski et al., 2012) and lymphatic diseases in humans (Lukacs et al., 2015), our work should motivate the search of potential involvement of Piezo in cardiac pathologies such as valvulopathies.

Piezo is necessary for proper OFT smooth muscle cell identity and endothelial klf2a expression in the OFT

Cardiac valve development is highly dependent on endothelial/endocardial cell mechanosentivity. Previous work identified the membrane-bound mechanosensitive channels (Trpp2 and Trpv4) and a calcium-activated intracellular signaling cascade leading to Klf2a expression and valve morphogenesis as key elements of the endocardial mechanodetection-signaling pathway (Heckel et al., 2015 ) Consistently, we found that these Trp channels are also required for endothelial Klf2a expression and OFT valve formation. In addition, we identify another type of mechanosensitive channel belonging to the Piezo family. Piezo channel mutants show valve dysgenesis phenotype, suggesting their requirement during OFT valve morphogenesis. Interestingly, Piezo channels are important both for modulating Klf2a expression in the endothelium and for smooth muscle cell-specific expression of Elnb and Fn1. Piezo function in these two cell layers is consistent with the fact that Piezo1 is expressed in smooth muscle cells and is important for tissue remodeling in mouse arteries (Retailleau et al., 2015) as well as in endothelial cells in mouse vasculature (Li et al., 2014; Ranade et al., 2014) and lymphatic valves (Nonomura et al., 2018). These studies along with our results suggest that Piezo1 might be necessary for both layers to regulate distinguishable functions: activation of signaling cascade upon shear stress in the OFT endothelium and proper cell identity acquisition in smooth muscles. Both tissue layers might be sensing different stimuli: shear plus strain for the endothelial cells and strain for the smooth muscle cells. This hypothesis is consistent with the fact that Piezo is sensitive to stretch (Ranade et al., 2015), compression (Lee et al., 2014; Qi et al., 2015), and rhythmic mechanical stimuli (Lewis et al., 2017). Importantly, Piezo-dependent mechanisms can transduce forces at the cell-cell or cell-matrix interface (Eisenhoffer et al., 2012; Poole et al., 2014). Thus, Piezo can have different mechanosensitive roles both in smooth muscle cells and endothelium to coordinate the expression of Klf2 and ECM proteins. An interesting hypothesis is that the smooth muscle cell layer participates in the shaping up of the OFT valve and that both tissue layers establish paracrine interactions to fine tune the morphogenetic process. Further work will be needed to identify if this is the case and how it affects OFT valve morphogenesis at the cellular scale.

Yap1 and hippo pathways are regulators of the OFT valve formation in the endothelium and smooth muscles

Our work shows that Yap1 is specifically localized in the OFT in both endothelial cells and smooth muscle cells and is required for valvulogenesis. Accordingly, the Hippo pathway is active in smooth muscles and in the endothelium. Our work suggests that Piezo constitutes a plausible regulator of Yap1 localization as Piezo1 seems important for its localization in smooth muscle cells. Interestingly, Yap1 has been shown to translocate less in the nucleus in Piezo1 mutant mice neural stem cells (Pathak et al., 2014) and to induce proliferation in smooth muscles during cardiovascular development in mouse (Wang et al., 2014). Even though we were not able to assess Yap1 subcellular localization in vivo, the regulation of Yap1 activity and ECM assembly might be a general feature of Piezo function. The connection between Piezo and Yap1 is particularly interesting in the context of OFT development and function. In teleost, the OFT has an important biomechanical role for the control of flow propagation within the vascular networks by contributing to the dampening of the pressure wave down the arterial tree (Braun et al., 2003b). In zebrafish, Yap1 is involved in the determination of cardiac precursor cells into smooth muscle cell fate via a process that involves the regulation of Elastinb expression (Moriyama et al., 2016). Besides the control of cell identity, ECM contributes to biomechanical properties of tissues (Dzamba and DeSimone, 2018). It is thus possible that Piezo acts as a regulator of tissue mechanical properties by regulating Yap1 expression both in the OFT and in the vascular system where Yap1 expression is flow inducible (Nakajima et al., 2017).

In summary, this study reveals a novel function for mechanosensitive Piezo and Trp channels in modulating OFT valve development as well OFT smooth muscle cell maturation. It will be important to further investigate endothelial-smooth muscle cells interactions during OFT cardiac wall maturation.

Materials and methods

Zebrafish strains, husbandry, embryo treatments, and morpholinos

Request a detailed protocol

Animal experiments were approved by the Animal Experimentation Committee of the Institutional Review Board of the IGBMC (reference numbers MIN APAFIS#4669–2016032411093030 v4 and MIN 4669–2016032411093030 v4-detail of entry 1). Zebrafish lines used in this study were Tg(fli1a:lifeact-EGFP) (Phng et al., 2013), Tg(kdrl:nls-mCherry) (Nicenboim et al., 2015), Tg(kdrl:EGFP) (Jin et al., 2005), Tg(fli1a:nls-mCherry) (Heckel et al., 2015), Tg(myl7:egfp) (Huang et al., 2003), Tg(−26.5Hsa.WT1-gata2:eGFP)cn12 (Sánchez-Iranzo et al., 2018), Tg(klf2a(6 kb):H2B-eGFP) (Heckel et al., 2015), Tg(tp1:dGFP) (Ninov et al., 2012), Tg(4xGTIIC:d2GFP) (Miesfeld and Link, 2014), vlad tepesm651 (Lyons et al., 2002), cuptc321 (Schottenfeld et al., 2007), piezo1sa12608 (EZRC), piezo2asa12414 (ZIRC), trpv4sa1671 (ZIRC), notch1bsa11236 (ZIRC), klf2aig4 (Steed et al., 2016b), yap1fu48 (Agarwala et al., 2015) and wild-type AB. Cup mutant embryos were phenotyped based on the curled tail phenotype. The Tg(piezo1:nls-Venus) was generated by injection of the piezo1:nls-Venus plasmid and the mRNA of the Tol2 transposase. The plasmid was generated by cloning of 3 kb of the zebrafish piezo1 promoter upstream of the ATG start site and 3xnls-Venus into a pTol2-GAGGS vector. All animals were incubated at 28.5°C for 5 hr before treatment with 1-phenyl- 2-thiourea (PTU) (Sigma Aldrich) to prevent pigment formation. Morpholino specific for tnnt2a (Sehnert et al., 2002) (5’-CATGTTTGCTCTGATCTGACACGCA-3’) were obtained from GeneTools. It was injected into the yolk at the one-cell stage at a concentration of 5,8 ng to stop the heart. It was diluted 40 times in order to get fish with either a decreased heartbeat (‘slow beating heart’ group in this study) or close to the non-injected fish heartbeat (‘beating heart’ group in this study).

Immunofluorescence

Request a detailed protocol

Embryos were fixed at the desired stage in 4% paraformaldehyde overnight at 4°C. After washing in 1X PBST (PBS-0.1% Tween-20), embryos were permeabilized in 1X PBST containing 1% Triton-X 100 for 30 min at room temperature or overnight at 4°C. For Fibronectin1 staining, the pericardial cavity was then carefully pierced with the tip of a needle to facilitate antibody penetration before blocking in permeabilization buffer supplemented with 5% BSA. Embryos were incubated in blocking solution containing 5% BSA and 15% NGS (α-Fn1), 1% BSA, 2% NGS (anti-Elnb) and 2% BSA, 2% MgCl2 (1M), 5% NGS supplemented by 1,5% Tween-20 (anti-Yap1) for 2 hr at room temperature or overnight at 4°C. Primary antibodies were added to the relevant blocking solution and incubated 2 overnights at 4°C. Secondary antibodies were added in blocking solution after thorough washing in PBST and incubated for 2 days at 4°C. Embryos were thoroughly washed in PBST and mounted for imaging on a Leica SP8 confocal. Antibodies were used as follows: rabbit α-Fn1 (F3648, Sigma) 1:100, rabbit anti-Elnb (Miao et al., 2007; kind gift from Burns lab; Paffett-Lugassy et al., 2017) 1:1000, rabbit anti-Yap1 (generated by the Lecaudey lab; Agarwala et al., 2015) 1:300, rabbit anti-Trpp2 (kind gift from Drummond lab) 1:100 and Alexa Fluor 647 goat anti-rabbit IgG secondary antibody (A21245, Life Technologies) were used at 1:500.

In situ hybridization

Request a detailed protocol

In situ hybridization was performed as in Thisse and Thisse (2008). Anti-sense probes for notch1b and klf2a were generated from a plasmid containing the cDNA of zebrafish notch1b in pCR-script SK+ (provided by the Bakkers lab, The Netherlands) and zebrafish klf2a in IRBOp991B0734D (provided by RPDZ, Berlin; Vermot et al., 2009) and subsequently transcribed using the T3 polymerase and T7 polymerase, respectively. After ISH, embryos were incubated subsequently in 45% and 90% D-fructose (Sigma F0127) containing 0.5% of 1-Thioglycerol (Sigma M6145) for 20 min. Imaging of ISH was done using a Leica M165 macroscope with a TrueChrome Metrics (Tucsen) with a Leica 1.0X objective (10450028).

RNAscope

Request a detailed protocol

72 hpf wildtype zebrafish embryos were fixed in 4% PFA overnight at 4°C. The fixed embryos were dehydrated to 100% ethanol gradually. Embryos were stained using the RNAscope Fluorescent Multiplex kit (Advanced Cell Diagnostics).

Confocal imaging

Request a detailed protocol

For live imaging, zebrafish embryos were staged, anaesthetized with 0.02% tricaine solution or 50 mM BDM, to stop the heart when necessary, and mounted in 0.7% low melting-point agarose (Sigma Aldrich). Confocal imaging was performed either on a Leica SP8 confocal microscope (experiments with BODIPY-ceramide or fixed samples) or a Leica spinning disk (valve structure, flow profile, reporter experiments). Fast confocal imaging to image valve leaflets from 72 hpf to 120 hpf stained with BODIPY-ceramide was performed using the resonant scanner mode of the Sp8 microscope. Images were acquired with a low-magnification water immersion objective (Leica HCX IRAPO L, 25X, N.A. 0.95). The optical plane was moved 2 µm between the z-sections until the whole OFT was acquired. 2-colored fast confocal imaging was used to image valve structure, red blood cells, and reporter activities from 56 hpf to 144 hpf was performed using a Leica DMi8 combined with a CSU-X1 (Yokogawa) spinning at 10 000 rpm, two simultaneous cameras (TuCam Flash4.0, Hamamatsu) and a water immersion objective (Leica 20X, N.A. 0.75 or Leica 40X, N.A. 1.1). 1 ms exposure was used for red blood cells imaging and 20 ms exposure for valve structure and reporter activity experiments. 50% of 488 laser power and 40% of 561 laser power were used for reporter activity experiments.

BODIPY-ceramide imaging

Request a detailed protocol

Embryos were incubated with 4 mM BODIPY-ceramide (Molecular Probes) overnight and then processed as in Heckel et al. (2015) and Vermot et al. (2009) to visualize the valve shape.

DAF-FMDA labelling

Request a detailed protocol

To reveal the presence of NO, embryos were incubated in zebrafish medium containing 5 µM DAF-FM DA (Life Technologies, D23842) for 30 min in the dark at 28°C. Fluorescence intensities of the smooth muscles were measured using ImageJ software.

Photoconversion

Request a detailed protocol

Photoconversion was performed using the FRAP module on a SP8 confocal microscope and a Leica HCX IRAPO L, 25X, NA0.95 water immersion objective. Tg(fli1a:Gal4FF; UAS:Kaede) embryos were mounted in 0.7% low melting-point agarose supplemented with 50 mM BDM to inhibit heart contraction for the duration of the procedure. A region of interest corresponding to the anterior, middle or posterior part of the valve was selected and exposed to 405 nm light (25% laser power). One pre-bleach frame was acquired, followed by 3–6 bleach pulses (3–5 ms each) without acquisition to achieve conversion of the kaede protein to its red form. A z-stack of the photoconverted heart was then acquired in the standard confocal mode to record the starting point of each experiment. Embryos were then carefully dissected from the agarose, placed in fish water for 5–10 min until heart contraction resumed and then put at 28.5°C to develop individually under standard conditions until the next time point of interest.

Fractional shortening

Request a detailed protocol

Imaging was performed on a Leica DMIRBE inverted microscope using a Photron SA3 high-speed CMOS camera (Photron, San Diego, CA) and water immersion objective (Leica 20X, NA 0.7). Image sequences were acquired at a frame rate of 1000 frames per second. FS% = (Dm diastole- Dm systole)/ (Dm diastole) and where Dm is the diameter of the chamber of interest (atrium or ventricle).

Flow analysis

Request a detailed protocol

Red blood cells were manually tracked through the OFT and their velocity calculated from image sequences of the Tg(gata1:ds-red; kdrl:EGFP) beating heart, acquired at 1000 frames per second as described previously. Red blood cells transiting through the OFT were tracked manually on Imaris and their velocity calculated. The tracks of multiple cells in at least four embryos per stage were assembled to obtain an estimate of the flow velocity over multiple cardiac cycles (typically 3). Velocities estimated at the same time point by tracking different cells were averaged.

Image analysis

Request a detailed protocol

For fluorescence intensity analysis of the Klf2a reporter, the Tg(klf2a:H2B-EGFP) reporter line was crossed with the Tg(fli1a:nls-mCherry) line and the mCherry signal was used for normalization. The maximum intensity of each channel on a single Z-section through the valves was quantified and the EGFP over mCherry ratio generated. For fluorescence intensity analysis of the Notch reporter, the ratio of the maximum intensity of the GFP signal from the Tg(tp1:dGFP) reporter line on a single Z-section through the valves over the maximum intensity of the background was generated.

These ratios were then averaged for three cells in the anterior part and three cells in the posterior part of both valves in the OFT of individual embryos. Finally, the averages of the anterior and posterior parts were compared.

Statistical analyses

Request a detailed protocol

We did not compute or predict the number of samples necessary for statistical differences because the standard deviation of our study’s population was not known before starting our analysis. Biological replicate corresponds to the analysis of different embryos of the same stage. Technical replicate corresponds to the analysis of the same embryo imaged the same way. The sample size (biological replicate and number) to use was as defined by our ability to generate our datasets. For analyses between two groups of embryos, differences were considered statistically significant when the p-value<0.05, as determined using a two-tailed and paired Student’s t-test (klf2a and notch reporter expression). For boxplots, center lines show the medians, crosses show the means, box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, data points are represented as circles.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
    Mechanobiology of the aortic heart valve
    1. JT Butcher
    2. CA Simmons
    3. JN Warnock
    (2008)
    The Journal of Heart Valve Disease 17:62–73.
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
    Notch Signaling in Aortic Valve Development and Disease
    1. V Garg
    (2016)
    In: T Nakanishi, R. R Markwald, H. S Baldwin, B. B Keller, D Srivastava, H Yamagishi, editors. Etiology and Morphogenesis of Congenital Heart Disease: From Gene Function and Cellular Interaction to Morphology. Springer. pp. 371–376.
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66
  67. 67
  68. 68
  69. 69
  70. 70
  71. 71
  72. 72
  73. 73
  74. 74
  75. 75
    Hemodynamics driven cardiac valve morphogenesis
    1. E Steed
    2. F Boselli
    3. J Vermot
    (2016a)
    Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 1863:1760–1766.
    https://doi.org/10.1016/j.bbamcr.2015.11.014
  76. 76
  77. 77
  78. 78
  79. 79
  80. 80
  81. 81
  82. 82
  83. 83
  84. 84
  85. 85

Decision letter

  1. Holger Gerhardt
    Reviewing Editor; Max Delbrück Center for Molecular Medicine, Germany
  2. Didier Y Stainier
    Senior Editor; Max Planck Institute for Heart and Lung Research, Germany
  3. Holger Gerhardt
    Reviewer; Max Delbrück Center for Molecular Medicine, Germany

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for sending your article entitled "Mechanically activated Piezo channels control outflow tract valve development through Yap1 and Klf2-Notch signaling axis" for peer review at eLife. Your article is being evaluated by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation is being overseen by Didier Stainier as the Senior Editor.

Given the list of essential revisions, including new experiments, the editors and reviewers invite you to respond within the next two weeks with an action plan and timetable for the completion of the additional work. We plan to share your responses with the reviewers and then issue a binding recommendation.

Whilst all reviewers find merit in your work, two of the reviewers unanimously raise concerns regarding the low number of replicates in many experiments, both in those that use signalling reporters for klf2 and notch that are deemed inherently highly variable, and those that use mutants to assess phenotypes. Please see the details below.

Furthermore, the reviewers are concerned that the overall model is not sufficiently supported by the data, and that the phenotypic descriptions appear somewhat overstated. There is also a perceived disconnect between the described effects on the smooth muscle cells and the endothelial cells, with insufficient insight into how the mechanosensitive ion channels regulate the phenotypes of either of the two, and no information of potential collective effects or interaction in the morphogenic process.

The reviewers find that based on these weaknesses, the overall level of mechanistic insight is unsatisfactory.

The reviewers and the BRE have discussed extensively whether the necessary revisions will be feasible within the normal revision period. There is a consensus in that there could be benefits in delineating more carefully either the klf2a Notch axis or the mechanisms of the ion channels to provide focus and the level of detail required to substantiate the model.

For more information on the details, we have appended the full comments of the reviewers below.

Reviewer #1:

The manuscript by Duchemin and colleagues analyses the morphogenic events of outflow tract development during embryogenesis in the zebrafish. The authors are experts in this area and have a strong track record in studying flow dependent signalling and morphogenic regulation in the vasculature and the heart in fish. They exploit their excellent imaging skills to describe the changing shape of the outflow tract, tracking endothelial cells by photoconversion and imaging between 72 and 144hpf. The authors report that the endothelium shows a prominent tissue folding process without signs of delamination, likely driven by morphogenic events between both the endothelium and the surrounding smooth muscle cells, but unlikely involving the myocardium. Matrix labelling highlights prominent expression of fibronectin by the smooth muscle cells during the process.

The authors use klf2a and tp1 reporter lines to address when and where Klf2a and Notch signalling is activated, identifying a strong upregulation in the posterior part of the valve formation. Slowing heart function with hypomorphic Tnnt2a MO demonstrates that both signals are flow dependent, whilst mutants identify that klf2a is important for Notch activity, but not vice versa. The authors conclude that Klf2a regulates valve morphogenesis via Notch and show that Notch activity loses some of its polarity in klf2a mutants. Using Yap and tead reporter fish, the authors further show that hippo signalling is active in the smooth muscle cells and also some endothelial cells. Here the resolution is less clear in the Yap1-reporter, but the data do suggest that the smooth muscle cells show the strongest signal. Yap signal is also flow dependent, and YAP mutants like Klf2a mutants show delayed or defective valve formation at varying frequency. Searching for upstream activators of these mechanosensitive pathways in endothelial cells and smooth muscle, the authors finally investigate mutants for 4 mechanosensitive ion channels, Piezo1, Piezo2a, Trpp2 and TrpV4.

The results suggest some level of redundancy or cooperativity, with Piezo1 and Trpp2 important for endothelial folding, and Piezo1 involved in smooth muscle cell activation of Yap1 to modulate outflow tract development.

The work is overall very carefully executed, includes all relevant controls, beautifully illustrated and very comprehensively presented throughout. The details of endothelial morphogenesis, gene regulation, and the mechanosensitive signalling cascade provide for a much deeper understanding of this developmental process. The schematic illustrations accompanying the original imaging data is extremely helpful. Overall, this is an outstanding piece of work that is deemed of interest for not only the cardiovascular development community, but also for those interested in translational aspects given the importance of inherited cardiac defects associated with various mutations in pathways addressed in this work.

If I was to point out a weakness of the work, it would be the lack of commenting and possibly analysing the crosstalk between the smooth muscle cells and the endothelium in the process. The final concluding schematic places Piezo1 upstream of different components in both cell compartments, but given that the downstream effectors of Yap1 include modulators of matrix, Bmp and tgfb signalling etc., which will impact on mechanical properties of both endothelium and the smooth muscle cells, and that Notch activity also regulates similar targets, it would seem justified to comment on potential paracrine interactions between the two. I realize however that a full investigation here would go beyond the scope of the current work.

A second aspect that deserves a comment would be how Piezo1 inhibits klf2a whilst Trpp2 is deemed a positive regulator. How is this dichotomy achieved?

Finally, the kdrl:nls-mcherry line appears to label only very few endothelial cells in the posterior OFT and the images here do not really match the schematic. I assume the authors use the fli:lifeact GFP signal to delineate those kdrl negative cells, but this would seem important to comment for technical reasons. On a mechanistic level, it would be important to comment on this cellular heterogeneity in the posterior OFT?

Reviewer #2:

In the submitted manuscript, Duchemin and colleagues tackle the formation of the OFT valve in the zebrafish model and its dependency on mechanic stimuli. Their work indicates a first rudimentary pathway of mechanosensation-implicated players in the formation of the OFT valve, and describes the different morphogenesis of the OFT valve compared to the classically studied AV canal valve in the zebrafish heart.

Overall, the work raises general questions about the overall mechanism of flow-dependent morphogenetic processes. The study covers several valuable points, including the first analysis of piezo mutants in zebrafish. The figures include data points and schematics that will be valuable for future discussions and work in the field.

Nonetheless, at times, the manuscript makes strong conclusions based on small sample sizes (i.e. total embryos analyzed, see below) and the expression of isolated marker genes. Several mutant phenotypes have highly variable penetrance and expressivity (i.e. yap1 mutants, subsection “Klf2, Notch signaling, and Hippo pathways are active in the OFT in different cell layers and are all necessary for proper OFT valve development”), the analysis of which is again hampered by low sample sizes and ambiguous description as to how mutants are identified (assumed Mendelian ratio? Genotyping of individual embryos? etc.). Several statements of necessity and "dramatic" phenotypes need reassessment, as the phenotypes rather indicate the studied genes contribute to robust OFT valve formation rather than being necessary (i.e. notch1b, klf2; see also below). Further, OFT valve formation is a somewhat specialized developmental process, yet the authors seem tempted to draw broad-stroke generalizations about mechanosensing from their findings in the Discussion.

* The authors analyze several mutants and state necessity and importance for valve formation. Overall, however, the phenotype analysis is plagued by several issues as presented, and rather suggests the studied factors might be at times dispensable yet contribute to robust valve formation.

In all figures with box diagrams, at times vast variability is captured (i.e. Figure 8G'), yet individual data points not shown. Individual data points should be added. Further, the overall sample size in the majority of the mutant analyses seem low and spotty in the reported details, is shown in percent that are misleading in small sample sizes, and affects the phenotype interpretation. Percentage is used with exceedingly small sample sizes and becomes misleading in significance. Similarly for p-values, that are not sufficient to cover vastly incomplete penetrance of phenotypes. As examples:

a) The yap1 mutants have a seemingly increasing penetrance and expressivity of OFT valve phenotypes – yet the phenotype is reported for only 12 embryos (also, are these always the same 12 embryos?), resulting in 58% (6 or 7) of embryos with no valves at 120 hpf; that hardly seems to justify an "important role" but rather a contribution to robust valve formation.

b) Mutant notch1b embryos have a 30% incidence for valve issues, also not warranting the moniker "necessary" but rather contributing. Similarly, the authors state "most" notch1b embryos show proper klf2 expression, yet it's about half at only n=4 (Figure 7).

c) The authors state that a "large fraction" of piezo1 embryos have less Yap1 at n=7/10, again seeming like an overstatement (can also be called "a bit more than half"?).

Such examples go throughout all mutant analysis (also trpp2, etc.) and should be re-phrased and possibly remedied by increasing the sample size to achieve better insights into variability of the phenotypes.

Also, how do yap1, piezo, trpp2, etc. mutants look like overall? A description of other phenotypes (i.e. cardiac edema, endothelial problems, viability, etc.) would greatly help to gauge the non-autonomous impact of the described defects (or their dependence thereof).

* The authors conclude that loss of piezo function impacts smooth muscle "identity" of the BA based on downregulated Fn1 and Elnb staining. While interesting, Fn1 and Elnb are functional/differentiation markers and not necessarily determinants of fate/identity (i.e. Moriyama et al., 2016), yet these observations might rather indicate that flow has a functional contribution to BA maturation. The authors are encouraged to perform a functional BA assay, i.e. active NO metabolism using DAF-2DA (Grimes et al., 2006) to underline the functional impact on BA formation. Further, the claim of Fn1 expression in smooth muscle is not well-funded based on the provided images in Figure 3 and should be shown more clearly.

Reviewer #3:

This manuscript addresses how mechanical forces and genetic programs interact to control valve morphogenesis. The authors assess in vivo morphogenesis of the OFT valve and how mechanosensitive pathways: Yap, Klf2, and Notch1b contribute to this process. Finally they test the role of mechanosensors from the Trp and Piezo family.

While the topic of the manuscript is very interesting and timely, the conclusions drawn from the data are not based on enough evidence.

Throughout, the number of independent experiments should be clearly stated; the n in most of the experiments is very low. This makes any rigorous assessment futile and greatly undermines the quality of the manuscript.

All the quantification/bar graphs should show individual data points wherever possible. The statistical tests should be clearly stated in the figure legends. The manuscript text would benefit from revision; e.g. the last sentence of the Introduction is not clear, or adjectives as reversing flow or tissular remodeling are used and are not correct English.

1) In the Introduction: "The multichambered heart contains two different set of valves: arterial valves that are semilunar and mitral (atrioventricular) valves that are tricuspid." Where is the reference to that? Mitral valve is also called the left atrioventricular valve. Tricuspid valve is also called the right atrioventricular valve. Only mitral valve is bicuspid, all other valves are tricuspid in 4 chambered heart. In zebrafish all valves are bicuspid, they are at IFT, AVJ, and OFT. The authors do not mention IFT valve for some reason at all. The valve anatomy in 4 vs 2 chambered heart should be corrected and properly referenced; a simple schematic would be helpful for clarity.

2) "… the developmental programs driving mitral valve…,less is known about arterial valves" The reference Wu et al. cites more than 20 genes involved in arterial valve development. The rest of the Introduction mentions several signaling pathways involved in aortic valve formation. So meanwhile it is not entirely true that less is known about arterial valves. The Introduction would benefit if it is shortened and refocused on the current knowledge of the effect of mechanical forces on valvulogenesis.

3) Figure 1A-C: in the text a different transgenic is referred to then shown in the panels. What is the time point in Figure 1A?

4) Figure 2B: in the third panel, there are photoconverted cells anteriorly and posteriorly, where is the anterior cell located at 120 hpf? The photoconverted cells should be highlighted by a star in all panels, not just in one.

Consecutive panels for at least 3 embryos should be shown in the supplement.

5) The claim that AVC valve is delaminating should be referenced. Scherz et al., 2008, showed that in zebrafish AVC valve is forming by invagination and not via endocardial cushion formation, similarly to the process of OFT formation that the authors observe here. The conclusion that the OFT valve morphogenesis is unique is thus overstated. It is not clear what the authors mean when they say "unique".

6) The rationale that invagination of the OFT endothelium might be aided by the adjacent tissue should be introduced more clearly. In Figure 1 and 2 only endothelial contribution is observed and shown, yet the authors claim that the cellular contribution of the valve lead them to the hypothesis of adjacent tissue contribution. What is the model?

In Figure 3, besides merged images, individual channels for Fibronectin, Elastin b, and endothelial marker should be shown. Counterstaining for Fibronectin and Elastin b should be performed together; from the images presented, not all fibronectin positive cells seem to be Elastin b positive, and the schematic is therefore misleading. If panel A' (Figure 3) is a zoom of panel A, and both are scale bar of 10 microns, why is the scale bar in panel A longer than in panel A'? it does not look like much of a zoom either. This should be revised.

The vascular smooth muscle cells of BA of the OFT are accrued to the ventricle after 48 hpf. This process was recently described e.g. by Felker et al., 2018. Even though the OFT valve formation is occurring concomitantly with the formation of BA, these two events should not be equated, as stated in the conclusions describing Figure 3. If the authors wished to compare OFT valve formation to AVC valve formation, they should provide additional panels of AVC valve for comparison or at least briefly recap the events of AVC valve formation.

7) What is the stage in Figure 4A and B? The reporters should be visualized with the markers for endothelium and elastin b. The channel for fli:nls-mCherry should be shown especially if the GFP/mCherry ratio is used for the quantification in 4D. How is anterior/posterior boundary defined at different stages for the purpose of the quantification in 4D if the endothelium is invaginating in the direction from posterior to anterior as the authors show in Figure 2? This seems to be arbitrary and misleading due to the dynamic nature of the forming valve. The division medial/lateral or luminal might be more appropriate. What are the individual data points in D? the n should be increased for 120hpf for klf2 reporter; n=2 is not enough! The description of how exactly the quantification was performed should be included, as this read-out is used in the subsequent experiments; is the intensity for both leaflets per embryo averaged? From a single section or z-stack?

Statistical test should be named in the figure legend.

8) In Figure 5, the GTIIC:d2GFP Yap reporter should be used together with the endothelial and nuclear marker and elastin b; from the images presented it is not clear how the dashed line was drawn in 5B. Furthermore, Yap reporter should be used together with the nuclear marker, and nuclear/cytoplasmic ratio should be quantified. The antibody staining for Yap appears to be cytoplasmic, which does not confer anything about its signaling activity.

The description of 5C should be rewritten; "…a significant fraction of yap1 mutants displayed abnormal valves (17% at 72 hpf…", – 17% show normal valves, the text is misleading. The quantification of the phenotypes should be performed from at least 3 independent experiments, so that appropriate statistical test can be performed.

Yap1 mutants should be properly referenced.

If the Yap reporter is active in both endothelial cells and smooth muscle cells of BA, the statement that smooth muscle progenitors are likely to play a role in valve morphogenesis is a speculation and should be corrected. That Yap is essential for OFT valve development is an overstatement when 40% of the mutant embryos still form some sort of a valve structure.

Is the 5A' zoom of A? the ROI in A does not correspond to A'?

9) The myl7:GFP reporter line should be used to determine the heart morphology upon different tnnt2 MO doses. From the images presented nothing can be concluded about the MO effect, except perhaps a mild edema in B. The MO doses should be indicated either in the figure directly or figure legend. The BA size should be measured under these MO conditions; it is known that cardiac function may affect the BA size. What is the cut off for the heart rate in the "slow-beating" group? Also the authors should be consistent with the naming, use group 1/2 or beating/slow beating, but not both. The sample number of klf2 and notch reporter has to be increased; the reporters, especially the ones using destabilized GFP are very variable, n of 3 with the usual clutches of at least 100 zebrafish embryos under normal breeding conditions is just not acceptable. The activity of klf2 and notch reporter should be assessed at later time point – 120 hpf when the flow effects are more pronounced, assuming that the data presented are at 72 hpf; this information is missing.

In the experiments assessing the Fb1, Eln B and Yap abundance, both the smooth muscle cells of BA and endothelial cells have to counterstained, e.g. NO fluorescent indicator DAR4-AM is a reliable marker of the BA smooth muscle cells. It is not clear if the smooth muscles cells are present at all with the high dose of tnnt2 MO, and as such the "missing" staining can be simply a result of missing cells. In Figure 6D, at what time point the valve presence was quantified? n=9 for tnnt2 MO is very low; are the authors really injecting just 9 embryos?

The controls are completely missing in the experiments presented in Figure 6, with the exception of Figure 6D.

While the conclusion that klf2 and notch activity are flow-dependent is somewhat acceptable, as it also corroborates previously published data, the statement that BA smooth muscle cell identity depends on the mechanical forces is a speculation and not supported by any evidence. This should be corrected.

10) The penetrance of the phenotype of both klf2a and notch1b mutants is very low, so while these factors certainly contribute to OFT valve morphogenesis, the effect is mild. This fact should not be overlooked. What is the effect of the combined loss of klf2a and notch1b? n is very low, at least 3 independent experiments should be performed. The font of the graphs is too small and unreadable.

What do the numbers 64% and 56% represent at 48hpf regarding the notch1b expression in the manuscript text? From the graph it looks like about 70% of embryos do not show any notch1b expression in klf2a mutants at 48hpf. Similar issue is with 72hpf description of the effect. How do the authors define the difference between altered and absent expression? In the examples shown as altered no gene expression can be detected; the exemplary images for all groups should be shown. While loss of klf2a reduces notch1b expression it is not completely required.

In the absence of notch1b, the expression of klf2 is randomized and not mostly proper as concluded.

The n is too low in the experiments where notch signaling activity is assessed in klf2 mutants. Again at least 3 independent experiments should be performed. Individual data points should be presented. P values for Figures 7C, C' are wrongly written; (10-1 with 1 in superscript). What is the statistical test used?

The images of klf2a reporter expression in notch1b+/- and -/- should be shown.

The previous characterization of the mutants should be properly referenced.

The notch signaling activity seems to be reduced but not gone in the absence of klf2a.

The conclusion that klf2a regulates notch signaling in OFT valve endothelium is an overstatement, and should be corroborated. The data presented are not sufficient to claim this, and are based on assessment of 4 embryos of the reporter that is notoriously variable.

11) How many embryos were used to assess fractional shortening and RFF in the listed mutants? What is the time point for RFS? Individual data points should be plotted. Is the heart rate also unaffected? RFF is a measure for valve function not heart function, this should be corrected or explained better. RFF is assessed at 72 hpf when the valve only starts forming, what is the RFF at 120hpf when the valve leaflets formed?

The assessment of OFT valve morphology is performed in a low number of mutant embryos. Even in controls, 30% of embryos have still thickened valves, and thus the results should not be overstated; in a considerable percentage of the mutants or their combinations, the valves are still formed. The great variability of the phenotypes is especially apparent when the results of A and B are compared.

Therefore it is not evident that Trp and Piezo channels are necessary for the folding of the endothelium; they contribute to some extent to the process.

The data do suggest a redundancy between different channels though. What is the combined effect of trp and piezo channels? Especially the combination of trpp2 and piezo1 would be important to assess.

Where are these channels expressed, in the endothelium only or also in the vascular smooth muscle of the BA?

For the analysis of the Fn1, Elnb, and Yap markers again not sufficient number of embryos is assessed. Fn1 is also expressed in the endothelial cells as the authors show themselves in the Figure 3, and Elnb is a differentiation marker, thus using those two to conclude anything about the identity of the smooth muscle cells of the BA is not appropriate. The concerns here are similar as in the data presented in the Figure 6.

How do these markers look in trpp2 mutants?

The results of the experiments using klf2a reporter in piezo1 and trpp2 mutants are intriguing as they point out to the opposing effects of mechanosensors in regulating klf2a transcriptional activity.

There is not enough evidence that Piezo1 regulates the identity of smooth muscle cells of BA. What is the effect of loss of piezo1 on Yap reporter. Mere presence or absence of Yap in the cells is not sufficient to claim Piezo's role in Yap signaling.

12) The model in Figure 9 is speculative, and is not supported by enough corroborating evidence. The origin of the valve leaflets is tracked only by the optogenetic photoconversion of Kaede to endothelium, the authors do not trace any other cell type within the valve leaflets. While the presence of the vascular smooth muscle cells might affect the OFT valve morphogenesis, the authors do not test any active role of these cells in the valve formation. The role of mechanosensors of Trp an Piezo family in the process of OFT formation is overstated. The evidence points to the modulatory role; they certainly contribute but are not completely necessary for the process of OFT valve formation. There is not enough evidence presented that the mechanosensors play any role in the modulating cell fate or identity of the vascular smooth muscle cells of the BA. The Klf2a regulation of notch1b expression and its signaling activity in the valve endothelium is not supported by enough evidence and should be corroborated by functional assays; as presented now it is only a feasible hypothesis.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Mechanically activated Piezo channel modulate outflow tract valve development through Yap1 and Klf2-Notch signaling axis" for further consideration at eLife. Your revised article has been favorably evaluated by Didier Stainier as the Senior Editor, a Reviewing Editor, and two reviewers.

They praise your efforts and find the revised manuscript substantially improved. The critical issues have been adequately addressed and the reviewers believe the work will receive considerable attention in the field. They however identified a few remaining points, as outlined below:

The introductory text on the architecture of the valves, regarding tricuspid or bicuspid nature, needs to be corrected as detailed in the comments. Further referring to the initial question of robustness and sample size, please add this information to either the main text or figure legends throughout the manuscript.

Reviewer #2:

In the revised version of the manuscript, Duchemin et al. have addressed a majority of the criticisms from all reviewers. In particular, the increase in sample sizes for individual experiments is laudable. The finding that Fn1 and ElnB expression should be coupled with functional analysis should receive attention in the field as well.

Remaining points:

* Sample size: while the authors' issues with increasing the sample numbers due to embryo identification, the involved genotyping, etc. are understandable, stating that other papers did use similar numbers is not really a valid argument. Especially considering eLife's emphasis on transparency and reproducibility, two repetitions of an experiment seem less than ideal. The authors should clearly state the full number of repetitions, etc. throughout the manuscript so the reader can make up her/his own mind about the numbers.

* Please check the issue with bicuspid/tricuspid valve architecture, as also raised by reviewer 3.

* The added schematic in Figure 1A is highly helpful. However, it has a remarkable resemblance (down to individual trabeculae shapes, color scheme, and labeling) to the schematic provided in Felker et al., 2018. This work should be at least referenced in the manuscript.

Reviewer #3:

The authors addressed most of the points raised by all reviewers satisfactorily.

There are few remaining issues that need to be revised.

https://doi.org/10.7554/eLife.44706.028

Author response

Reviewer #1:

[…] If I was to point out a weakness of the work, it would be the lack of commenting and possibly analysing the crosstalk between the smooth muscle cells and the endothelium in the process. The final concluding schematic places Piezo1 upstream of different components in both cell compartments, but given that the downstream effectors of Yap1 include modulators of matrix, Bmp and tgfb signalling etc., which will impact on mechanical properties of both endothelium and the smooth muscle cells, and that Notch activity also regulates similar targets, it would seem justified to comment on potential paracrine interactions between the two. I realize however that a full investigation here would go beyond the scope of the current work.

We agree with the reviewer that this point is interesting and deserves further investigation. It is our plan to study this further in the near future, in particular, the interplays between the mechanical properties of both the endothelium and the smooth muscle cells. We agree that this would go beyond the scope of the current work but we now comment on the potential paracrine interactions between the two tissues in the Discussion.

A second aspect that deserves a comment would be how Piezo1 inhibits klf2a whilst Trpp2 is deemed a positive regulator. How is this dichotomy achieved?

This is another interesting observation of our study. While we can just speculate at this point, this can be attributed to the fact that piezo and trpp2 may act in different tissues and/or activate the release of different paracrine factors. To assess if the localization of piezo1 and trpp2 mRNA could explain the differential function of these channels, we performed RNAscope assay at 72hpf (Figure 8—figure supplement 3A). We found that trpp2 is ubiquitously expressed in the embryo, including in the different layers composing the OFT. Similarly, we found that piezo1 is expressed in both endothelium and smooth muscles, albeit at a lower level than trpp2. To confirm these results, we generated a transgenic reporter line with 3kb of the piezo1 promoter upstream of the start codon (piezo1:nls-Venus). We observed the expression of the reporter line mostly in smooth muscles at 72hpf (Figure 8—figure supplement 3B) and cells of the endothelium of the OFT valve. Trpp2 immunohistochemistry showed that Trpp2 is expressed in the endothelium and the smooth muscles confirming that trpp2 is ubiquitously expressed in the OFT (arrow in Figure 8—figure supplement 3B). We agree that this point deserves more work to be clear. For clarity, we simplified the working model shown in Figure 9 since trpp2 and piezo1 may be active in both cell layers.

Finally, the kdrl:nls-mcherry line appears to label only very few endothelial cells in the posterior OFT and the images here do not really match the schematic. I assume the authors use the fli:lifeact GFP signal to delineate those kdrl negative cells, but this would seem important to comment for technical reasons. On a mechanistic level, it would be important to comment on this cellular heterogeneity in the posterior OFT?

We used the double transgenic line fli1a:lifect-EGFP; kdrl:nls-mCherry at 72hpf and 96hpf to show that the kdrl:nls-mCherry transgenic line indeed highlights all nuclei of the endothelium, albeit not all nuclei with the same fluorescence intensity. You can see in Author response image 1 the labeling of the OFT endothelium by the line fli1a:lifect-EGFP; kdrl:nls-mCherry. Arrowheads in A show cells that express less kdrl transgene. To assess if the heterogeneity in expression is also visible at the mRNA level, we performed RNAscope assay at 72hpf using a kdrl probe (B). We found that kdrl expression might also be heterogenous at the mRNA level. This point deserves more in-depth and proper quantitative analysis to make strong conclusions about the heterogeneity of kdrl expression and is beyond the scope of this study. Nevertheless, even though we are not sure if this kdrl:nlsmCherry line may display different expression levels because of the transgene insertion, it is possible that kdrl expression is heterogeneous in endothelial cells of the OFT.

Author response image 1
(A) Z-section of the kdrl:nls-mCherry and fli1a:lifeact-eGFP in the OFT (B) Z-projection of the RNAscope of kdrl in the head and in the OFT (n=6 embryos).

Reviewer #2:

[…] At times, the manuscript makes strong conclusions based on small sample sizes (i.e. total embryos analyzed, see below) and the expression of isolated marker genes. Several mutant phenotypes have highly variable penetrance and expressivity (i.e. yap1 mutants, subsection “Klf2, Notch signaling, and Hippo pathways are active in the OFT in different cell layers and are all necessary for proper OFT valve development”), the analysis of which is again hampered by low sample sizes and ambiguous description as to how mutants are identified (assumed Mendelian ratio? Genotyping of individual embryos? etc.). Several statements of necessity and "dramatic" phenotypes need reassessment, as the phenotypes rather indicate the studied genes contribute to robust OFT valve formation rather than being necessary (i.e. notch1b, klf2; see also below). Further, OFT valve formation is a somewhat specialized developmental process, yet the authors seem tempted to draw broad-stroke generalizations about mechanosensing from their findings in the Discussion.

* The authors analyze several mutants and state necessity and importance for valve formation. Overall, however, the phenotype analysis is plagued by several issues as presented, and rather suggests the studied factors might be at times dispensable yet contribute to robust valve formation.

In all figures with box diagrams, at times vast variability is captured (i.e. Figure 8G'), yet individual data points not shown. Individual data points should be added.

We apologize for this oversight. We have now added the individual data points on the graphs for all figures with data points missing: Figure 7C’, Figure 8F’ and Figure 8G’.

Further, the overall sample size in the majority of the mutant analyses seem low and spotty in the reported details, is shown in percent that are misleading in small sample sizes, and affects the phenotype interpretation. Percentage is used with exceedingly small sample sizes and becomes misleading in significance. Similarly for p-values, that are not sufficient to cover vastly incomplete penetrance of phenotypes. As examples:

a) The yap1 mutants have a seemingly increasing penetrance and expressivity of OFT valve phenotypes – yet the phenotype is reported for only 12 embryos (also, are these always the same 12 embryos?), resulting in 58% (6 or 7) of embryos with no valves at 120 hpf; that hardly seems to justify an "important role" but rather a contribution to robust valve formation.

N=12 corresponds to 2 independent experiments. These embryos were analysed and followed over time (i.e. at 72hpf, 96 hpf, 120hpf), so it is n=12 for each time point. These experiments are performed in blind and require that each embryo is genotyped retrospectively, once the imaging has been completed. These embryo numbers are close to what has been published in the recent literature in other zebrafish cardiac studies recently published in eLife: See Guerra et al., 2018: Figure 5, n=17, n=12, n=7 or see Nguyen-Chi et al., 2015: Figure 1, n=12, n=3, n=11 or see Semmelhack et al., 2014: Figure 2, n=9).

We changed the text explanation and modified the sentence from “important” to “involved”.

To better quantify the penetrance of the yap1 mutation, we now show both yap1+/+ and yap1+/- data in Figure 5B. Indeed, this shows that the yap1+/-have an intermediate phenotype compared to yap1+/+and yap1-/-.

b) Mutant notch1b embryos have a 30% incidence for valve issues, also not warranting the moniker "necessary" but rather contributing. Similarly, the authors state "most" notch1b embryos show proper klf2 expression, yet it's about half at only n=4 (Figure 7).

In order to confirm that klf2 expression is not affected by notch1b, we performed additional experiments using notch1b+/+ and notch1b-/- embryos in the klf2a:H2B-GFP reporter line background. As previously done, we quantified the klf2a reporter expression in notch1b+/+ and notch1b-/-embryos (n=8 and n=6 respectively now) at 72hpf, 96hpf and 120hpf (Figure 7—figure supplement 1). This did not change the final conclusion that we previously stated: notch1b mutation does not affect klf2a expression.

In addition, we added the confocal images for the klf2a reporter and the kdrl:nls-mCherry in notch1b+/+ and notch1b-/- at 72hpf, 96hpf, and 120hpf to display individual data points (Figure 7—figure supplement 1).

c) The authors state that a "large fraction" of piezo1 embryos have less Yap1 at n=7/10, again seeming like an overstatement (can also be called "a bit more than half"?).

We changed a ‘large fraction’ to ‘a bit more than two thirds’ in the text to be more precise.

Such examples go throughout all mutant analysis (also trpp2, etc.) and should be re-phrased and possibly remedied by increasing the sample size to achieve better insights into variability of the phenotypes.

We performed additional immunostaining of Elnb, Fn1, Yap1 in piezo1+/+, piezo1-/-, tnnt2aMO normal beating, and tnnt2aMO slow beatingembryos to increase the n and reach 3 independent experiments. These stainings correspond to data on Figure 6D and Figure 8D. Please find in Author response table 1 a recapitulative table showing the numbers for the initial submission and the revised version.

ElnbFn1Yap1
Initial
submission
Revised versionInitial
submission
Revised versionInitial
submission
Revised version
piezo1+/+4/411/1411/1216/2010/1010/10
piezo1-/-3/411/159/1220/247/1010/14
tnnt2aMO normal7/719/224/47/117/711/14
tnnt2aMO slow5/719/266/614/144/710/16

Author response table 1: table summarizing the additional immunostainings performed for each mutant in comparison to the previous submission.

Also, how do yap1, piezo, trpp2, etc. mutants look like overall? A description of other phenotypes (i.e. cardiac edema, endothelial problems, viability, etc.) would greatly help to gauge the non-autonomous impact of the described defects (or their dependence thereof).

Some overall phenotypes were previously described: yap1 in Agarwala et al., 2015; trpp2 in Schottenfeld J et al., 2007 and trpv4 in Heckel et al. 2014. In addition, we imaged control and mutant embryos (see Author response table 2) and evaluated their overall shape, heart structure, heartbeat and endothelial phenotype (blood flow in the ISV). We also provided pictures of the overall embryo shape and heart structure.

Overall shapeHeart shapeBeatingEndothelial problems
normalabnormalnormalSitus inversusedemaNormal or abnormalBlood flow in ISV
piezo1+/+10/100/1010/100/100/107/7 normalNormal vasculature
See Shmukler et al., 2015
See Shmukler et al., 2016
See Kok et al., 2016
piezo1-/-11/143/14
(tail bent up)
13/140/141/1412/12 normal
trpp2+/+, +/- 18/180/1816/182/180/1818/18 normalNormal vasculature
See Goetz et al.,
2014
Schottenfeld J et al., 2007
trpp2-/-0/1616/16
(tail bent up)
10/166/160/1610/10 normal
trpv4+/+4/62/6
(tail bent up)
7/70/70/77/7 normalNormal vasculature (4/4)
trpv4-/-4/84/8
(tail bent up)
6/82/80/84/7 normal
3/7 slow
Normal vasculature no flow (2/4)
notch1b+/+10/100/109/100/101/1010/10 normalNormal vasculature (21/22)
notch1b-/-10/100/1025/261/260/2610/10 normalNormal vasculature (5/6)
klf2a+/+50/500/5010/100/100/1010/10 normalNormal vasculature (5/5)
klf2a-/-10/100/1010/211/2111/219/10 normal 1/10 slowNormal vasculature
(4/5)
1/5 no flow (1/5)
tnnt2aMO normal17/170/1715/172/170/1710/10 normalNormal vasculature
(5/5)
tnnt2aMO slow7/158/15
(tail bent up)
0/150/1515/156/6 slowno flow (5/5)
yap1+/+ 17/170/1717/170/170/1717/17 normalSubtle vascular
defects
See Nakajima et al.,
2017
yap1-/- 27/281/28
(tail bent up)
25/280/283/2826/28 normal, 2/28 no beating

Normal beating: 2-3 Hz

Slow beating: <2 Hz

No beating: 0 Hz

Author response table 2: table summarizing the overall phenotype observed for each mutant.

Author response image 2
(A) Pictures of the different mutants and their controls showing overall shape and heart shape.

* The authors conclude that loss of piezo function impacts smooth muscle "identity" of the BA based on downregulated Fn1 and Elnb staining. While interesting, Fn1 and Elnb are functional/differentiation markers and not necessarily determinants of fate/identity (i.e. Moriyama et al., 2016), yet these observations might rather indicate that flow has a functional contribution to BA maturation. The authors are encouraged to perform a functional BA assay, i.e. active NO metabolism using DAF-2DA (Grimes et al., 2006) to underline the functional impact on BA formation. Further, the claim of Fn1 expression in smooth muscle is not well-funded based on the provided images in Figure 3 and should be shown more clearly.

To clarify the expression domain of Fn1, we used DAF-FMDA assay (similar to the DAF-2DA assay, it reveals active NO metabolism) on wild-type embryos counterstained with Fn1 antibody. These co-stainings revealed the presence of DAF-FMDA activity and Fn1 in the same cells of the OFT at 72hpf, 96hpf and 120hpf, except for the Fn1 staining within the valves. These data are now provided in Figure 3—figure supplement 1. We conclude that Fn1 is localized in smooth muscle cells of the OFT.

For clarity, we also changed the figure panel to provide inverted LUT images of Fn1 and Elnb to make this clearer (Figure 3B, C, D, B’, C’, D’).

To assess the functional impact of mechanical forces on BA formation, we performed the DAFFMDA assay on the tnnt2a-morpholino injected embryos and piezo1 mutants. We also used this assay to measure the BA diameter. In tnnt2a-morpholino injected embryos with slow beating (n=6), we observed a smaller BA (diameter=36,1µm ± 4,5, p=10-3) and lower fluorescence intensity (intensity=1,4 ± 0,4, p<0,05) than in the tnnt2a-morpholino injected embryos displaying normal heartbeat (n=7, BA diameter=47,4µm ± 3,7 and intensity=3,0 ± 1,6). By contrast, the BA size as well as the fluorescence intensity in the piezo1-/- embryos (n=8, BA diameter=50,5µm ± 4,1, intensity=2,7 ± 1,5) were not significantly different from the piezo1+/+ (n=7, BA diameter=51,3µm ± 3,9, intensity=2,7 ± 1,4) measurements. We conclude that heart activity is required for BA maturation but not piezo1, even though piezo mutant display abnormal Fn1 and Elnb expression. We added these data to Figure 6E and Figure 8D and Figure 8—figure supplement 2B. As suggested by the reviewer, the results obtained in piezo mutants suggest that Fn1 and Elnb may indeed not be functional/differentiation markers and are not necessarily determinants of smooth muscle cell identity. We changed the text of our manuscript accordingly.

Reviewer #3:

[…] While the topic of the manuscript is very interesting and timely, the conclusions drawn from the data are not based on enough evidence.

Throughout, the number of independent experiments should be clearly stated; the n in most of the experiments is very low. This makes any rigorous assessment futile and greatly undermines the quality of the manuscript.

All the quantification/bar graphs should show individual data points wherever possible. The statistical tests should be clearly stated in the figure legends. The manuscript text would benefit from revision; e.g. the last sentence of the Introduction is not clear, or adjectives as reversing flow or tissular remodeling are used and are not correct English.

1) In the Introduction: "The multichambered heart contains two different set of valves: arterial valves that are semilunar and mitral (atrioventricular) valves that are tricuspid." Where is the reference to that? Mitral valve is also called the left atrioventricular valve. Tricuspid valve is also called the right atrioventricular valve. Only mitral valve is bicuspid, all other valves are tricuspid in 4 chambered heart. In zebrafish all valves are bicuspid, they are at IFT, AVJ, and OFT. The authors do not mention IFT valve for some reason at all. The valve anatomy in 4 vs 2 chambered heart should be corrected and properly referenced; a simple schematic would be helpful for clarity.

We agree with the reviewer and clarified these issues in the text. In addition, we generated a drawing of the heart and OFT structures in Figure 1.

2) "… the developmental programs driving mitral valve…,less is known about arterial valves" The reference Wu et al. cites more than 20 genes involved in arterial valve development. The rest of the Introduction mentions several signaling pathways involved in aortic valve formation. So meanwhile it is not entirely true that less is known about arterial valves. The Introduction would benefit if it is shortened and refocused on the current knowledge of the effect of mechanical forces on valvulogenesis.

We shortened the Introduction to refocus it on mechanical forces during valvulogenesis.

3) Figure 1A-C: in the text a different transgenic is referred to then shown in the panels. What is the time point in Figure 1A?

We apologize for the misunderstanding and clarified the figure panels. Figure 1A now shows the overall heart structure using the kdrl:nls-mCherry (highlights the endocardium) and myl7:GFP (highlights the myocardium). Figure 1B is based on the combination of the fli:lifeacteGFP and kdrl:nls-mCherry lines to highlight the actin and nuclei within the OFT, while Figure 1C uses the combination of the kdrl:eGFP and gata1:dsRed line for assessing the flow profile. The time point in A is 96hpf and was added on the figure.

4) Figure 2B: in the third panel, there are photoconverted cells anteriorly and posteriorly, where is the anterior cell located at 120 hpf? The photoconverted cells should be highlighted by a star in all panels, not just in one.

Consecutive panels for at least 3 embryos should be shown in the supplement.

The cell located anteriorly goes out of frame. This is now indicated in the Legend. Asterisks are included on each panel. We added a Figure 2—figure supplement 1A showing 3 photoconverted embryos for each position (top, middle and bottom).

5) The claim that AVC valve is delaminating should be referenced. Scherz et al., 2008, showed that in zebrafish AVC valve is forming by invagination and not via endocardial cushion formation, similarly to the process of OFT formation that the authors observe here. The conclusion that the OFT valve morphogenesis is unique is thus overstated. It is not clear what the authors mean when they say "unique".

Thanks for pointing that out. Pestel et al., 2016; Grunewald et al., 2018 and Steed et al., 2016 have recently described AVC valve morphogenesis at cellular resolution and showed that the AVC valve forms via endocardial cushion formation. This has been extensively discussed in different research and review articles (Paolini et al., 2018; Donat et al., 2018, Steed, Boselli et al., 2016) and it is now well accepted in the field that zebrafish AVC valves do not form solely via invagination. This is why we consider OFT valve formation to be different from AVC valve formation. This is now clarified in the text. In addition, we have added a Figure 2—figure supplement 1B where we compare the valve formation in the AVC and OFT.

6) The rationale that invagination of the OFT endothelium might be aided by the adjacent tissue should be introduced more clearly. In Figure 1 and 2 only endothelial contribution is observed and shown, yet the authors claim that the cellular contribution of the valve lead them to the hypothesis of adjacent tissue contribution. What is the model?

We agree that this is an exciting aspect of our findings. At this point we can only speculate about the contribution of the smooth muscle cells and endothelial cells in the morphogenetic process. We are currently establishing protocols that allow us to follow OFT valve morphogenesis though time-lapse analysis. We hope that once established, live imaging will provide us with a better understanding of the contribution of these cell layers during OFT valve morphogenesis.

In Figure 3, besides merged images, individual channels for Fibronectin, Elastin b, and endothelial marker should be shown. Counterstaining for Fibronectin and Elastin b should be performed together; from the images presented, not all fibronectin positive cells seem to be Elastin b positive, and the schematic is therefore misleading. If panel A' (Figure 3) is a zoom of panel A, and both are scale bar of 10 microns, why is the scale bar in panel A longer than in panel A'? it does not look like much of a zoom either. This should be revised.

We now show the individual channels. Fibronectin and Elastinb antibodies have been raised in the same organisms so it is not possible to perform co-labeling but we agree that it is possible that not all fibronectin positive cells are elastin b positive. We revised the figures and text accordingly.

We added the single channels for Fn1 and Elnb in Figure 3 as well as revised the scale bars from panels A, A’.

The vascular smooth muscle cells of BA of the OFT are accrued to the ventricle after 48 hpf. This process was recently described e.g. by Felker et al., 2018. Even though the OFT valve formation is occurring concomitantly with the formation of BA, these two events should not be equated, as stated in the conclusions describing Figure 3. If the authors wished to compare OFT valve formation to AVC valve formation, they should provide additional panels of AVC valve for comparison or at least briefly recap the events of AVC valve formation.

We agree with the reviewer that a comparison with AVC valve formation requires clarification and we have now added Figure 2—figure supplement 1B comparing the formation of AVC and OFT valves.

7) What is the stage in Figure 4A and B? The reporters should be visualized with the markers for endothelium and elastin b. The channel for fli:nls-mCherry should be shown especially if the GFP/mCherry ratio is used for the quantification in 4D. How is anterior/posterior boundary defined at different stages for the purpose of the quantification in 4D if the endothelium is invaginating in the direction from posterior to anterior as the authors show in Figure 2? This seems to be arbitrary and misleading due to the dynamic nature of the forming valve. The division medial/lateral or luminal might be more appropriate. What are the individual data points in D? the n should be increased for 120hpf for klf2 reporter; n=2 is not enough! The description of how exactly the quantification was performed should be included, as this read-out is used in the subsequent experiments; is the intensity for both leaflets per embryo averaged? From a single section or z-stack?

We corrected the figure legend to include the embryonic stages and included the kdrl:nlsmcherry channel in Figure 4—figure supplement 1. We revised the Materials and methods section to explain better how the boundaries are defined and which statistical tests were performed. We also revised the figure legend to show individual data points.

In addition, we performed an additional experiment to analyze the klf2a reporter expression in wild-type at 120hpf to reach n=6 embryos (Figure 4D).

Statistical test should be named in the figure legend.

This has now been done.

8) In Figure 5, the GTIIC:d2GFP Yap reporter should be used together with the endothelial and nuclear marker and elastin b; from the images presented it is not clear how the dashed line was drawn in 5B. Furthermore, Yap reporter should be used together with the nuclear marker, and nuclear/cytoplasmic ratio should be quantified. The antibody staining for Yap appears to be cytoplasmic, which does not confer anything about its signaling activity.

The description of 5C should be rewritten; "…a significant fraction of yap1 mutants displayed abnormal valves (17% at 72 hpf…", – 17% show normal valves, the text is misleading. The quantification of the phenotypes should be performed from at least 3 independent experiments, so that appropriate statistical test can be performed.

This has now been done.

Yap1 mutants should be properly referenced.

This has now been done.

If the Yap reporter is active in both endothelial cells and smooth muscle cells of BA, the statement that smooth muscle progenitors are likely to play a role in valve morphogenesis is a speculation and should be corrected. That Yap is essential for OFT valve development is an overstatement when 40% of the mutant embryos still form some sort of a valve structure. Is the 5A' zoom of A? the ROI in A does not correspond to A'?

For the reporter quantification, we are sorry for the misunderstanding: this reporter is cytoplasmic as it reveals the transcriptional activation of tead/yap and it is not a localization reporter. The reporter highlights yap activity and we agree that this is not the case for the antibody.

We provided an Elnb staining on kdrl:membrane-mCherry x 4xGTTIIC:d2GFP in Figure 5A. The reporter is expressed in some endothelial cells as well as in the smooth muscles.

We rewrote the text so that we are careful about our conclusions on the role of yap1 during OFT valve formation.

9) The myl7:GFP reporter line should be used to determine the heart morphology upon different tnnt2 MO doses. From the images presented nothing can be concluded about the MO effect, except perhaps a mild edema in B. The MO doses should be indicated either in the figure directly or figure legend. The BA size should be measured under these MO conditions; it is known that cardiac function may affect the BA size. What is the cut off for the heart rate in the "slow-beating" group? Also the authors should be consistent with the naming, use group 1/2 or beating/slow beating, but not both. The sample number of klf2 and notch reporter has to be increased; the reporters, especially the ones using destabilized GFP are very variable, n of 3 with the usual clutches of at least 100 zebrafish embryos under normal breeding conditions is just not acceptable. The activity of klf2 and notch reporter should be assessed at later time point – 120 hpf when the flow effects are more pronounced, assuming that the data presented are at 72 hpf; this information is missing.

We agree with the reviewer that this part should have been clearer: we selected by eye the embryos showing heart looping, but either with normal heartbeat (called “normal heartbeat”, 2-3Hz) or slow heartbeat (less than 2Hz). We now provide videos of hearts from these two categories for the reader to visualize the different phenotypes. Unfortunately, it is not possible to study 120hpf stages as embryos are not viable at this stage.

As suggested by the reviewer, we used the DAF-FMDA staining for measuring BA width in tnnt2aMO injected embryos. We found that it is smaller in the low flow embryos. We added this data to Figure 6E. We conclude from these results that mechanical forces are important for the functional maturation of the OFT.

We performed additional experiments imaging the Notch reporter in klf2a+/+ and klf2a-/- to reach n=8 and n=8 respectively, 3 independent experiments. Moreover, we added the individual data point on the graph in Figure 7C’.

We also performed additional imaging of the klf2a reporter in notch1b+/+ and notch1b-/- to reach n=8 and n=6 respectively, 3 independent experiments. We now show in addition the kdrl:nlsmCherry channel that was used for the normalization of the klf2a reporter intensity. These data are in Figure 7—figure supplement 1.

In the experiments assessing the Fb1, Eln B and Yap abundance, both the smooth muscle cells of BA and endothelial cells have to counterstained, e.g. NO fluorescent indicator DAR4-AM is a reliable marker of the BA smooth muscle cells. It is not clear if the smooth muscles cells are present at all with the high dose of tnnt2 MO, and as such the "missing" staining can be simply a result of missing cells. In Figure 6D, at what time point the valve presence was quantified? n=9 for tnnt2 MO is very low; are the authors really injecting just 9 embryos?

The controls are completely missing in the experiments presented in Figure 6, with the exception of Figure 6D.

We consider the normal beating heart category as our internal controls for tnnt2aMO-injected embryos with slow beating heart from the same clutch. Indeed, the injection of the diluted morpholino leads to various heartbeat phenotypes and we selected embryos with a heartrate of about 2-3 Hz for controls, and the ones with slower heartrate as our experimental condition.

While the conclusion that klf2 and notch activity are flow-dependent is somewhat acceptable, as it also corroborates previously published data, the statement that BA smooth muscle cell identity depends on the mechanical forces is a speculation and not supported by any evidence. This should be corrected.

To address this point, we performed new experiments with DAF-FMDA in tnnt2aMO-injected embryos with normal beating heart and slow beating heart to confirm the presence of the smooth muscles. However, they are not as functional as in the controls since DAF-FMDA expression is lower than in “normal beating heart” embryos. We thus conclude that BA smooth muscle cells do not mature properly when mechanical forces are abnormal. We added these data to Figure 6E.

10) The penetrance of the phenotype of both klf2a and notch1b mutants is very low, so while these factors certainly contribute to OFT valve morphogenesis, the effect is mild. This fact should not be overlooked. What is the effect of the combined loss of klf2a and notch1b? n is very low, at least 3 independent experiments should be performed. The font of the graphs is too small and unreadable.

We performed additional experiments imaging the Notch reporter in klf2a+/+ and klf2a-/- to reach n=8 and n=8 respectively, 3 independent experiments. Moreover, we added the individual data point on the graph in Figure 7C’. We also performed additional imaging of the klf2a reporter in notch1b+/+and notch1b-/-to reach n=8 and n=6 respectively, 3 independent experiments. We additionally show the kdrl:nls-mCherry channel that was used for the normalization of the klf2a reporter intensity. These data are in Figure 7—figure supplement 1.

We are also interested to know the effect of the combined klf2a; notch1b mutants but unfortunately, we do not have these combined mutants growing in our facility. Generating the compound mutants would require a much longer period of time than that given to us to perform the revisions so we did not perform this experiment. However, we plan to study these mutants in the near future.

What do the numbers 64% and 56% represent at 48hpf regarding the notch1b expression in the manuscript text?

64% corresponds to the% of embryos with altered expression of notch1b in the klf2a-/-. Thanks to the reviewer, we realized that the 56% and the “50%” are not referring to data shown in this figure. We revised the text accordingly.

From the graph it looks like about 70% of embryos do not show any notch1b expression in klf2a mutants at 48hpf. Similar issue is with 72hpf description of the effect. How do the authors define the difference between altered and absent expression? In the examples shown as altered no gene expression can be detected; the exemplary images for all groups should be shown. While loss of klf2a reduces notch1b expression it is not completely required.

‘Altered’ expression means that there is still expression in the heart but the expression pattern is abnormal, mislocalized. ‘Absent’ means that no ISH staining was observed in these embryos. We agree that klf2a-/- display a reduction in notch1b expression and it is not completely required. We provide representative images in Figure 7—figure supplement 1.

In the absence of notch1b, the expression of klf2 is randomized and not mostly proper as concluded.

The n is too low in the experiments where notch signaling activity is assessed in klf2 mutants. Again at least 3 independent experiments should be performed. Individual data points should be presented. P values for Figures 7C, C' are wrongly written; (10-1 with 1 in superscript). What is the statistical test used?

We performed additional experiments imaging the Notch reporter in klf2a+/+ and klf2a-/- to reach n=8 and n=8 respectively, 3 independent experiments. Moreover, we added the individual data point on the graph in Figure 7C’. We also performed additional imaging of the klf2a reporter in notch1b+/+ and notch1b-/- to reach n=8 and n=6 respectively, 3 independent experiments. We now show in addition the kdrl:nls-mCherry channel that was used for the normalization of the klf2a reporter intensity. These data are in Figure 7—figure supplement 1. The statistical test used is the Student t-test. This is also included in the Materials and methods section.

The images of klf2a reporter expression in notch1b +/- and -/- should be shown.

We have not considered the notch1b+/-for that experiment.

The previous characterization of the mutants should be properly referenced.

All the published mutants are now referenced in our revised version.

The notch signaling activity seems to be reduced but not gone in the absence of klf2a.

The conclusion that klf2a regulates notch signaling in OFT valve endothelium is an overstatement, and should be corroborated. The data presented are not sufficient to claim this, and are based on assessment of 4 embryos of the reporter that is notoriously variable.

We performed additional experiments imaging the Notch reporter in klf2a+/+ and klf2a-/-to reach n=8 and n=8 respectively, 3 independent experiments. Moreover, we added the individual data point on the graph in Figure 7C’.

11) How many embryos were used to assess fractional shortening and RFF in the listed mutants? What is the time point for RFS? Individual data points should be plotted. Is the heart rate also unaffected? RFF is a measure for valve function not heart function, this should be corrected or explained better. RFF is assessed at 72 hpf when the valve only starts forming, what is the RFF at 120hpf when the valve leaflets formed?

At least 3 embryos were used for each category for RFF and at least 3 other embryos for the fractional shortening. The embryos were analysed at 72 hpf.

To answer this point, we assessed the RFF at 120hpf. A table for the RFF at 120hpf is shown in Figure 8—figure supplement 1B. We also looked at the heartbeat in all mutants (see Author response table 3). RFFs were not dramatically altered in these mutants suggesting that the valves are still operating in the mutants even though the morphogenetic program is affected.

Overall shapeHeart shapeBeatingEndothelial problems
normalabnormalnormalSitus inversusedemaNormal or abnormalBlood flow in ISV
piezo1+/+10/100/1010/100/100/107/7 normalNormal vasculature
See Shmukler et al.,
2015
See Shmukler et al.,
2016
See Kok et al., 2016
piezo1-/-11/143/14
(tail bent up)
13/140/141/1412/12 normal
trpp2+/+, +/- 18/180/1816/182/180/1818/18 normalNormal vasculature
See Goetz et al., 2014
Schottenfeld J et al., 2007
trpp2-/-0/1616/16
(tail bent up)
10/166/160/1610/10 normal
trpv4+/+4/62/6
(tail bent up)
7/70/70/77/7 normalNormal vasculature (4/4)
trpv4-/-4/84/8
(tail bent up)
6/82/80/84/7 normal
3/7 slow
Normal vasculature no flow (2/4)
notch1b+/+10/100/109/100/101/1010/10 normalNormal vasculature (21/22)
notch1b-/-10/100/1025/261/260/2610/10 normalNormal vasculature
(5/6)
klf2a+/+50/500/5010/100/100/1010/10 normalNormal vasculature (5/5)
klf2a-/-10/100/1010/211/2111/219/10 normal
1/10 slow
Normal vasculature
(4/5)
1/5 no flow (1/5)
tnnt2aMO normal17/170/1715/172/170/1710/10 normalNormal vasculature
(5/5)
tnnt2aMO slow7/158/15
(tail bent up)
0/150/1515/156/6 slowno flow (5/5)
yap1+/+ 17/170/1717/170/170/1717/17 normalSubtle vascular
defects
See Nakajima et al.,
2017
yap1-/- 27/281/28
(tail bent up)
25/280/283/2826/28 normal, 2/28 no beating

Normal beating: 2-3 Hz

Slow beating: <2 Hz

No beating: 0 Hz

Author response table 3: table summarizing the overall phenotype observed for each mutant.

The assessment of OFT valve morphology is performed in a low number of mutant embryos. Even in controls, 30% of embryos have still thickened valves, and thus the results should not be overstated; in a considerable percentage of the mutants or their combinations, the valves are still formed. The great variability of the phenotypes is especially apparent when the results of A and B are compared.

Numbers for OFT valve morphology in trpp2 and piezo1 mutants are: in trpp2+/+ (n=11, n=10, n=10), trpp2-/- (n=8, n=13, n=13), piezo1+/+ (n=10, n=10, n=9) and piezo1-/- (n=10, n=10, n=9) at 72hpf, 96hpf and 120 hpf respectively, which are standards for live-imaging. Moreover, these embryos are followed over time.

Therefore it is not evident that Trp and Piezo channels are necessary for the folding of the endothelium; they contribute to some extent to the process.

The data do suggest a redundancy between different channels though. What is the combined effect of trp and piezo channels? Especially the combination of trpp2 and piezo1 would be important to assess.

In addition, we analyzed the valve phenotype of piezo1-/-; trpp2 MO to address the redundancy between trpp2 and piezo1 channels (Figure 8A, B, C). The trpp2 MO recapitulates well the trpp2 mutant phenotype and is well accepted by the community.

Where are these channels expressed, in the endothelium only or also in the vascular smooth muscle of the BA?

Trpp2 is believed to be ubiquitously expressed. For piezo1, we generated a transgenic line expressing venus FP under the 3kb piezo promoter. The results suggest that the expression of the piezo1 reporter is mainly in the smooth muscles.

In addition, we performed Trpp2 antibody staining on our piezo1 reporter line. Trpp2 seems to be expressed in smooth muscles as well as endothelium whereas piezo1 seems to be expressed mainly in smooth muscles (Figure 8—figure supplement 3B).

Finally, we performed RNAscope using the trpp2 and piezo1 fluorescent probes to better assess the localization of these channels. The results suggest that both channels are expressed in the endothelium and smooth muscles. However, trpp2 is expressed in many cells of the OFT while piezo1 is expressed in only a subset of cells in the OFT.

For the analysis of the Fn1, Elnb, and Yap markers again not sufficient number of embryos is assessed. Fn1 is also expressed in the endothelial cells as the authors show themselves in the Figure 3, and Elnb is a differentiation marker, thus using those two to conclude anything about the identity of the smooth muscle cells of the BA is not appropriate. The concerns here are similar as in the data presented in the Figure 6.

We performed additional immunostaining of Elnb, Fn1, Yap1 in piezo1+/+, piezo1-/-, tnnt2aMO normal beating, and tnnt2aMO slow beatingembryos to increase the n and reach 3 independent experiments. These stainings correspond to data on Figure 6D and Figure 8D. Author response table 4 is a recapitulative table with the numbers for the initial submission and the revised version.

ElnbFn1Yap1
Initial
submission
Revised versionInitial
submission
Revised versionInitial
submission
Revised version
piezo1+/+4/411/1411/1216/2010/1010/10
piezo1-/-3/411/159/1220/247/1010/14
tnnt2aMO normal7/719/224/47/117/711/14
tnnt2aMO slow5/719/266/614/144/710/16

Author response table 4: table summarizing the additional immunostainings performed for each mutant by comparison to previous submission.

How do these markers look in trpp2 mutants?

These markers are not affected in the trpp2 mutants.

The results of the experiments using klf2a reporter in piezo1 and trpp2 mutants are intriguing as they point out to the opposing effects of mechanosensors in regulating klf2a transcriptional activity.

There is not enough evidence that Piezo1 regulates the identity of smooth muscle cells of BA. What is the effect of loss of piezo1 on Yap reporter. Mere presence or absence of Yap in the cells is not sufficient to claim Piezo's role in Yap signaling.

We do not know the effect of piezo in YAP reporter. Here we present Yap1 expression analysis using the antibody. However, we did not intend to claim that piezo modulates yap signaling. To clarify the BA phenotype, we provide analysis of the DAF-2DA smooth muscle marker in the piezo1+/+ and piezo1-/-mutants. The BA size as well as the fluorescence intensity in the piezo1-/- embryos (n=8, BA diameter=50,5µm ± 4,1, intensity=2,7 ± 1,5) were not significantly different from the piezo1+/+ (n=7, BA diameter=51,3µm ± 3,9, intensity=2,7 ± 1,4) measurements. We added these data to Figure 8.

12) The model in Figure 9 is speculative, and is not supported by enough corroborating evidence. The origin of the valve leaflets is tracked only by the optogenetic photoconversion of Kaede to endothelium, the authors do not trace any other cell type within the valve leaflets. While the presence of the vascular smooth muscle cells might affect the OFT valve morphogenesis, the authors do not test any active role of these cells in the valve formation. The role of mechanosensors of Trp an Piezo family in the process of OFT formation is overstated. The evidence points to the modulatory role; they certainly contribute but are not completely necessary for the process of OFT valve formation. There is not enough evidence presented that the mechanosensors play any role in the modulating cell fate or identity of the vascular smooth muscle cells of the BA. The Klf2a regulation of notch1b expression and its signaling activity in the valve endothelium is not supported by enough evidence and should be corroborated by functional assays; as presented now it is only a feasible hypothesis.

We agree with the reviewer and apologize for the misunderstanding. The model is indeed speculative and is an attempt to provide discussion material. We carefully reassessed the text and discussion to make sure that the reader understands that this is a working model. We agree that mechanosensors have a modulatory role in the process of OFT morphogenesis and changed the title of our manuscript accordingly.

The regulation of notch1b by klf2a will be reinforced by the experiments performed. We have now reached n=8 and n=8, respectively, and 3 independent experiments.

We also performed additional imaging of the klf2a reporter in notch1b+/+ and notch1b-/- to reach n=8 and n=6 respectively, 3 independent experiments. We now show the kdrl:nls-mCherry channel that was used for the normalization of the klf2a reporter intensity. These data are in Figure 7—figure supplement 1.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Reviewer #2:

[…] Remaining points:

* Sample size: while the authors' issues with increasing the sample numbers due to embryo identification, the involved genotyping, etc. are understandable, stating that other papers did use similar numbers is not really a valid argument. Especially considering eLife's emphasis on transparency and reproducibility, two repetitions of an experiment seem less than ideal. The authors should clearly state the full number of repetitions, etc. throughout the manuscript so the reader can make up her/his own mind about the numbers.

This is now corrected.

* Please check the issue with bicuspid/tricuspid valve architecture, as also raised by reviewer 3.

Sincere apologies for this oversight, this is now corrected.

* The added schematic in Figure 1A is highly helpful. However, it has a remarkable resemblance (down to individual trabeculae shapes, color scheme, and labeling) to the schematic provided in Felker et al., 2018. This work should be at least referenced in the manuscript.

The reviewer is right thanks for pointing this out. This is now corrected.

https://doi.org/10.7554/eLife.44706.029

Article and author information

Author details

  1. Anne-Laure Duchemin

    1. Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France
    2. Centre National de la Recherche Scientifique, Illkirch, France
    3. Institut National de la Santé et de la Recherche Médicale, Illkirch, France
    4. Université de Strasbourg, Illkirch, France
    Contribution
    Conceptualization, Data curation, Formal analysis, Methodology, Writing—original draft
    Competing interests
    No competing interests declared
  2. Hélène Vignes

    1. Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France
    2. Centre National de la Recherche Scientifique, Illkirch, France
    3. Institut National de la Santé et de la Recherche Médicale, Illkirch, France
    4. Université de Strasbourg, Illkirch, France
    Contribution
    Data curation
    Competing interests
    No competing interests declared
  3. Julien Vermot

    1. Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France
    2. Centre National de la Recherche Scientifique, Illkirch, France
    3. Institut National de la Santé et de la Recherche Médicale, Illkirch, France
    4. Université de Strasbourg, Illkirch, France
    Contribution
    Conceptualization, Supervision, Funding acquisition, Writing—original draft, Project administration
    For correspondence
    julien@igbmc.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8924-732X

Funding

H2020 European Research Council (682938 - EVALVE)

  • Julien Vermot

Fondation pour la Recherche Médicale (DEQ29553)

  • Julien Vermot

Agence Nationale de la Recherche (ANR-15-CE13-0015-01)

  • Anne-Laure Duchemin
  • Hélène Vignes
  • Julien Vermot

European Molecular Biology Organization (Young Investigator Program)

  • Julien Vermot

Fondation Lefoulon Delalande

  • Anne-Laure Duchemin

Agence Nationale de la Recherche (ANR-10-IDEX-0002-02)

  • Anne-Laure Duchemin
  • Hélène Vignes
  • Julien Vermot

Agence Nationale de la Recherche (ANR-12-ISV2-0001-01)

  • Anne-Laure Duchemin
  • Hélène Vignes
  • Julien Vermot

Agence Nationale de la Recherche (ANR-10-LABX-0030-INRT)

  • Anne-Laure Duchemin
  • Hélène Vignes
  • Julien Vermot

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank the Vermot laboratory for discussion and H Fukui and R Chow for thoughtful comments on the manuscript. We thank C Burns, V Lecaudey and I Drummond for providing antibodies, mutants and protocols for immunohistochemistry. We thank J-M Garnier for the cloning of the piezo1:nls-Venus construct. We thank the IGBMC fish facility (S Pajot and C Moebs) and the IGBMC imaging center, in particular B Gurchenkov, D Hentsch, E Guiot and E Grandgirard. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme: GA N°682939, the Fondation pour la Recherche Médicale: DEQ29553, Agence Nationale de la Recherche: ANR-15-CE13-0015-01, ANR-10-IDEX-0002–02, ANR-12-ISV2-0001-01 and ANR-10-LABX-0030-INRT and the European Molecular Biology Organization Young Investigator Program. HV was supported by the IGBMC International PhD program: ANR-10-LABX-0030-INRT. ALD was supported by a post doctoral fellowship from the Lefoulon-Delalande Foundation.

Ethics

Animal experimentation: Animal experiments were approved by the Animal Experimentation Committee of the Institutional Review Board of the IGBMC.(reference numbers MIN APAFIS#4669-2016032411093030 v4 and MIN 4669-2016032411093030 v4-detail of entry 1).

Senior Editor

  1. Didier Y Stainier, Max Planck Institute for Heart and Lung Research, Germany

Reviewing Editor

  1. Holger Gerhardt, Max Delbrück Center for Molecular Medicine, Germany

Reviewer

  1. Holger Gerhardt, Max Delbrück Center for Molecular Medicine, Germany

Publication history

  1. Received: December 24, 2018
  2. Accepted: September 14, 2019
  3. Accepted Manuscript published: September 16, 2019 (version 1)
  4. Version of Record published: October 7, 2019 (version 2)

Copyright

© 2019, Duchemin et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 981
    Page views
  • 282
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)