Myocardial Angiopoietin-1 Controls Atrial Chamber Morphogenesis by Spatiotemporal Degradation of Cardiac Jelly

Article

Graphical Abstract Authors Kyun Hoo Kim, Yoshikazu Nakaoka, Hellmut G. Augustin, Gou Young Koh

In Brief Kim et al. demonstrate that myocardium- derived Angpt1 is indispensable for atrial chamber morphogenesis. Myocardial Angpt1 activates endocardial Tie2 signaling and positively regulates ADAMTS-dependent degradation of cardiac jelly. Proper distribution of cardiac jelly is critical for the regulation of myocardial growth and chamber morphogenesis in the atrium.

Highlights d Loss of myocardial Angpt1 impairs atrial chamber development d Loss of myocardial Angpt1 leads to excessive deposition of cardiac jelly d Angpt1/Tie2 signaling promotes ADAMTS-dependent degradation of cardiac jelly d Regulation of cardiac jelly is essential for atrial chamber morphogenesis

Kim et al., 2018, Cell Reports 23, 2455–2466 May 22, 2018 ª 2018 The Author(s). https://doi.org/10.1016/j.celrep.2018.04.080 Cell Reports Article

Myocardial Angiopoietin-1 Controls Atrial Chamber Morphogenesis by Spatiotemporal Degradation of Cardiac Jelly

Kyun Hoo Kim,1 Yoshikazu Nakaoka,2,3 Hellmut G. Augustin,4,5 and Gou Young Koh1,6,7,* 1Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 34141, Republic of Korea 2Department of Vascular Physiology, National Cerebral and Cardiovascular Center Research Institute, Osaka 565-8565, Japan 3Department of Cardiovascular Medicine, Osaka University Graduate School of Medicine, Osaka 565-0871, Japan 4Division of Vascular Oncology and Metastasis, German Cancer Research Center Heidelberg (DKFZ-ZMBH Alliance), 69120 Heidelberg, Germany 5Department of Vascular Biology and Tumor Angiogenesis, Medical Faculty Mannheim (CBTM), Heidelberg University, 68167 Mannheim, Germany 6Center for Vascular Research, Institute for Basic Science (IBS), Daejeon 34141, Republic of Korea 7Lead Contact *Correspondence: [email protected] https://doi.org/10.1016/j.celrep.2018.04.080

SUMMARY and expansion to develop a defined four-chamber structure. Further chamber maturation is achieved by development of the The four-chamber structure of the mammalian heart septum and the atrioventricular valves and proliferation of the is established during embryonic development. While ventricular myocardium (Buckingham et al., 2005). During car- key regulators for ventricular development are well diac chamber morphogenesis, the myocardium interacts with studied, regulatory mechanisms for atrial chamber the overlying endocardium. Endocardium, a vascular bed of morphogenesis remain poorly understood. Here, the heart, senses biophysical forces conveyed by blood flow or we found that angiopoietin-1 (Angpt1), a vascular contraction of cardiac muscle and provides morphogenetic cues to the myocardium (Haack and Abdelilah-Seyfried, 2016; maturation factor, is highly and specifically ex- Tian and Morrisey, 2012). Analysis of the cloche mutant zebra- pressed in atrial myocardium during heart develop- fish, which lacks endocardium, has shown a dysmorphic heart ment. Loss of myocardial Angpt1 in mouse embryo with dilated atrium and collapsed ventricle, demonstrating the led to severe impairment in atrial chamber morpho- importance of endocardium in cardiac morphogenesis (Reischa- genesis. We revealed that Angpt1 deficiency results uer et al., 2016; Stainier et al., 1995). Molecular mapping for in excessive deposition of cardiac jelly, which dis- ventricular trabeculation through close interactions between turbs regulation of myocardial growth, thereby endocardium and myocardium has been intensely studied in impairing maturation of atrial chambers. Mechanisti- recent decades. Coordination of multiple signaling pathways, cally, myocardial Angpt1 activates endocardial Tie2 including Notch1, neuregulin (NRG)/erbb receptor tyrosine ki- and positively regulates expression of ADAMTS pro- nase 2 and 4 (ErbB2/ErbB4), and fibroblast growth factor teases, which is crucial for proper degradation of signaling, plays critical roles in ventricular trabeculation (D’Am- ato et al., 2016; Grego-Bessa et al., 2007; Haack and Abdeli- cardiac jelly. Accordingly, loss of Tie2 also impairs lah-Seyfried, 2016; Lai et al., 2010; Lavine et al., 2005; Liu ADAMTS-mediated degradation of cardiac jelly in et al., 2010; Tian and Morrisey, 2012; Zhang et al., 2010). How- atrium. Collectively, myocardial Angpt1/endocardial ever, the role of endocardial-myocardial interactions in estab- Tie2 signaling in atrium promotes spatiotemporal lishing the atrial chambers remains largely unknown, despite degradation of cardiac jelly during early cardiac the importance of the atrium in blood circulation and the cardiac development and is therefore indispensable for atrial conduction system. chamber morphogenesis. In the early tubular heart, endocardium is separated from myocardium by a layer of extracellular matrix (ECM), called cardiac jelly (Lockhart et al., 2011). Major components of cardiac INTRODUCTION jelly include glycosaminoglycan (GAG) proteins such as versican and hyaluronan (Lockhart et al., 2011; Stainier, 2001). Depletion Morphogenesis of the cardiac chamber begins in the mouse on of versican or hyaluronan synthase-2 in mouse embryo leads to embryonic day 8.0 (E8.0) with formation of a heart tube severe defects in cardiac chamber formation and loss of trabe- composed of endothelium and surrounding myocardium (Buck- culation, indicating a critical role of cardiac jelly in myocardial ingham et al., 2005). After looping of the heart tube at E8.5, up to growth and maturation (Camenisch et al., 2002; Camenisch E10.5, the primitive atrium and ventricle undergo segmentation et al., 2000; Hatano et al., 2012; Yamamura et al., 1997). On

Cell Reports 23, 2455–2466, May 22, 2018 ª 2018 The Author(s). 2455 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). the other hand, proper degradation of cardiac jelly by secreted atrium was high—roughly equivalent to that of megakaryocytes proteases such as ADAMTS (a disintegrin and metallopeptidase in bone marrow, which is highest among adult mouse tissues with thrombospondin motif) family proteins is also required for (Figure S1B) (Park et al., 2017; Zhou et al., 2015a). myocardial remodeling and valve formation (Dupuis et al., At E11.5 and E15.5, high expression of Angpt1 was also 2011; Kern et al., 2006; Kern et al., 2010; Stankunas et al., identified in the inner trabecular layer of the ventricles, but it 2008). A recent study indicated that cerebral cavernous malfor- was not detected in the myocardium of the outer compact layer mation signaling in endocardium negatively regulates expres- and interventricular septum (Figures 1B and 1C). After birth, sion of ADAMTS4/5 by inhibiting the MEKK3 signaling pathway expression of Angpt1 remained only at the luminal side of the (Zhou et al., 2015b). However, other key signaling pathways ventricular myocardium (Figure S1A), presumably because of regulating expression of ADAMTS genes remain largely the coalescence process of the trabecular layer into compact unknown. myocardium (Sedmera et al., 2000). Moreover, sagittal section Angiopoeitin-1 (Angpt1) is a major angiopoietin family protein analysis indicated that robust Angpt1 expression was limited to that acts as an agonistic ligand for Tie2, a tyrosine kinase recep- the roof of atrium and ventricular trabeculae and was not tor in endothelial cells. Genetic studies of mice lacking Angpt1 or detected in the atrioventricular canal or outflow tract myocar- Tie2 and pharmacological studies with Angpt1 ligand have indi- dium at E10.5 (Figure 1D). Thus, Angpt1 expression shows cated essential roles of Angpt1/Tie2 signaling in vascular remod- spatiotemporal pattern in the heart during cardiac chamber eling, maturation, integrity, and stability (Augustin et al., 2009; morphogenesis. Jeansson et al., 2011; Saharinen et al., 2017; Suri et al., 1996). At variance with the original and prevailing concept (Armulik Loss of Angpt1 Causes Defective Cardiac Chamber et al., 2011; Augustin et al., 2009; Quaegebeur et al., 2011; Tei- Morphogenesis chert et al., 2017), we recently found that pericytes in retina and To investigate the role of Angpt1 in cardiovascular development, brain vessels show no Angpt1 expression, while pericytes in we crossed Angpt1GFP/+ heterozygous mice to produce Angpt1 choroidal vessels and Schlemm’s canal have substantial expres- null (Angpt1GFP/GFP) embryos. Consistent with findings from pre- sion (Kim et al., 2017; Park et al., 2017; Thomson et al., 2017), vious studies (Jeansson et al., 2011; Suri et al., 1996), Angpt1 null indicating heterogeneous sources of Angpt1 in its vascular ac- embryos exhibited modest defects in vascular branching and tions. Of note, myocardium-derived Angpt1 regulates coronary remodeling in the head and intersomitic regions at E10.5 vein formation in the embryonic heart (Arita et al., 2014), which compared with wild-type (WT) embryos (Figure 2A). We noted is in line with recent study (Chu et al., 2016) revealing the role that the hearts of Angpt1 null embryos had substantially small at- of Tie2 in venous specification. Moreover, Angpt1 has been pro- ria, a less complex endocardial plexuses in both ventricles, and posed as a crucial factor for cardiac development (Jeansson abnormal patterning of cardinal veins connecting to the atria et al., 2011; Suri et al., 1996). Upregulation of Angpt1 expression (Figures 2B and S2A). Moreover, Angpt1 null embryos exhibited could be implicated in the NRG1/ErbB signaling-induced cardiac congested blood and hemorrhage at E11.5 (Figure S2B), result- morphogenesis (Nakaoka et al., 2007). However, the precise role ing in embryonic lethality at E12.5 (Figure S2C). of Angpt1 and its underlying molecular mechanisms have been To pinpoint the exact morphological changes, we dissected remained elusive. the hearts from the embryos. Hearts of Angpt1 null embryos at In this study, we show that myocardium-derived Angpt1 posi- E10.5 exhibited impaired expansion and segmentation of atrial tively regulates the expression of ADAMTS proteases in endo- chambers compared with those of WT embryos (Figure 2C). At cardium via activation of Tie2 signaling, which in turn promotes E11.5, we found that atrial chamber morphogenesis was prom- degradation of cardiac jelly. Our findings demonstrate that this inently impaired in the hearts of Angpt1 null embryos (Figure 2D). cascade of reactions induced by Angpt1 contributes to atrial Right atrial chambers of Angpt1 null embryos had 30% reduced chamber morphogenesis. chamber width but no changes in chamber height, while left atrial chambers had 42% and 50% reduced chamber width and RESULTS height, respectively, compared with those of WT embryos (Fig- ures 2E and 2F). On the other hand, ventricles of Angpt1 null em- Expression of Angpt1 Is Robust and Selective in Atrial bryos showed a mild or no reduction in chamber width or height Myocardium and Ventricular Trabeculae during Cardiac compared with those of WT embryos (Figures S2D and S2E). Chamber Morphogenesis Sectional analyses of Angpt1 null embryos at E11.5 further re- To gain insight into the role of Angpt1 in cardiovascular develop- vealed indistinct atrial septum and lack of venous valves in the ment, we examined its expression during embryonic develop- right atrium (Figure 2G). These findings suggest that the expan- ment using Angpt1GFP knockin mice (Zhou et al., 2015a). At sion and segmentation of atrial chambers are severely impaired E9.5, Angpt1 expression was found mainly in the cardiac region, in Angpt1 null embryos. Meanwhile, ventricles of Angpt1 null em- but it spread widely into mesenchymal tissues at E10.5 (Fig- bryos at E11.5 had less complex trabeculae compared with ure 1A). Further detailed analysis revealed robust and selective those of WT embryos (Figure 2G), consistent with the findings expression of Angpt1 in the myocardium of the primitive atrium in whole-mount embryos (Figure 2B). but only marginal expression in the primitive ventricle at E8.5 To delineate the specific contribution of myocardium-derived (Figures 1B and 1C). Of interest, expression of Angpt1 in the Angpt1 for cardiac chamber morphogenesis, we generated an atrium persisted throughout the embryonic, postnatal, and adult Angpt1DNkx2.5 mouse by crossing the Angpt1flox/flox mouse periods (Figures 1B and S1A). The level of Angpt1 expression in (Arita et al., 2014; Lee et al., 2013) with the Nkx2.5IRESCre mouse

2456 Cell Reports 23, 2455–2466, May 22, 2018 Figure 1. Robust and Selective Expression of Angpt1 in Atrial Myocardium and Ventric- ular Trabeculae during Heart Development (A) Images of Angpt1-GFP and CD31 in whole- mount WT embryo at E9.5 and Angpt1GFP/+ embryos at E9.5 and E10.5. Note the robust Angpt1 expression in the heart region (arrows) and high Angpt1 expression in the mesenchymal tissue regions along the intersomitic vessels (arrowheads) and the forelimbs (bracket). Scale bars, 200 mm. (B) Images of Angpt1-GFP and CD31 in transverse sections of Angpt1GFP/+ hearts at E8.5, E11.5, and E15.5. Scale bars, 200 mm. (C) Images of Angpt1-GFP and a-Actinin in transverse sections of Angpt1GFP/+ hearts at E8.5 and E11.5. Arrows indicate Angpt1 expression in ventricular trabeculae. Scale bars, 100 mm. (D) Image of Angpt1-GFP and CD31 in sagittal sections of Angpt1GFP/+ hearts at E10.5. Note that Angpt1 is not expressed in atrioventricular canal (AVC) and outﬂow tract (OFT) myocardium. Scale bar, 200 mm. A, atrium; V, ventricle. See also Figure S1.

Angpt1DNkx2.5 embryos had 30% reduced chamber width but no changes in chamber height, while left atrial chambers had 41% and 50% reduced chamber width and height, respectively, compared with those of WT embryos (Figures 3B–3D). However, the ventricles of Angpt1DNkx2.5 embryos showed only a slight or no reduction in the chamber width or height (Figures S3B and S3C). Sectional analyses of hearts of Angpt1DNkx2.5 embryos revealed indistinct atrial septum and right venous valves and less complex trabeculae in ventricles compared with those of WT embryos (Figure 3E). Analysis of inner cavity of the hearts of Angpt1DNkx2.5 embryos at E11.5 clearly showed that the venous valves surrounding the open- ing of sinus venosus in right atrium were completely absent, and the structure of atrial septum was not obvious compared with WT embryos (Figure 3F). Sagittal section analysis also showed the absence of venous valves between the right atrium and sinus venosus in Angpt1DNkx2.5 embryos (Figure 3G). Of note, Angpt1DNkx2.5 (Stanley et al., 2002), which allowed depletion of Angpt1 expres- embryos showed congested blood in large vessels at E11.5 (Fig- sion in Nkx2.5+ cells and their derivatives from the cardiac cres- ure S3D), and all Angpt1DNkx2.5 embryos were dead by E13.5 cent stage (Figure S3A). Analysis of Angpt1DNkx2.5 embryos at (Figures S3D and S3E), with signs of severe hemorrhage and E11.5 revealed relatively normal development of the vascular pericardial effusion (Figures S3D). These ﬁndings indicate that plexus in the head and intersomitic regions (Figure 3A). However, impaired formation of valves and atrial septum causes severe these embryos had barely recognizable atrial chambers and less disturbance in cardiovascular functions, ultimately leading to complex endocardial plexus in the ventricles compared embryonic death (Figure S3E). Thus, myocardium-derived with those of WT embryos (Figure 3A). Consistent with the ﬁnd- Angpt1 plays a crucial role in proper cardiac chamber formation ings from the Angpt1 null embryos, right atrial chambers of during embryonic development.

Cell Reports 23, 2455–2466, May 22, 2018 2457 Figure 2. Lack of Angpt1 Impairs Cardiac Chamber Morphogenesis (A) Images of CD31 in whole-mount WT and Angpt1GFP/GFP embryos at E10.5. Magnified images of head and intersomite region are shown below. Magnified images of the left side of hearts (yellow boxes) are shown in (B). Scale bars, 500 mm. (B) Images of left and right sides of hearts from whole-mount WT and Angpt1GFP/GFP embryos at E10.5. Note the substantially smaller atrial chambers (yellow arrowheads) and less vascularized endocardial plexuses in ventricles (yellow arrow) of Angpt1GFP/GFP embryos. Scale bars, 500 mm. (C) Images of dissected hearts from WT and Angpt1GFP/GFP embryos at E10.5. Dashed lines outline right atrium (RA) and left atrium (LA). Arrows indicate hypoplastic atrial chambers. Scale bars, 500 mm. (D) Images of dissected hearts of WT and Angpt1GFP/GFP embryos at E11.5. Dashed lines outline RA and LA. Arrows indicate hypoplastic LA. Double-arrow lines indicate the width and height of atrial chambers. Scale bars, 500 mm. W, width; H, height. (E and F) Comparisons of the width (E) and height (F) of RA and LA of WT and Angpt1GFP/GFP embryos at E11.5. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. (G) Images of transverse sections of the hearts from WT and Angpt1GFP/GFP embryos at E11.5. Magnified images of indicated regions are shown below. Arrows indicate venous valves in RA of WT embryos, which are absent in Angpt1GFP/GFP embryos. Brackets indicate the atrial septum. Scale bars, 200 mm. LA, left atrium; LV, left ventricle; RA, right atrium. See also Figure S2.

widened spaces between the endocardium and myocardium in the hearts of Angpt1DNkx2.5 embryos (Figure 4A, arrowheads). Alcian blue staining revealed that the space was abnormally ﬁlled with GAG-rich cardiac jelly (Figure 4D, arrowheads). Cardiac jelly in the atrial cham- Loss of Angpt1 Causes Excessive Deposition of Cardiac bers of WT embryos was limited to the region of the venous Jelly valve in the right atrium and to the tip of the atrial septum, To determine the cause of defective chamber morphogenesis, which is connected with the endocardial cushion (Figure 4D). we further performed histologic analyses of the hearts of WT However, in Angpt1DNkx2.5 embryos, we noted excessive and Angpt1DNkx2.5 embryos at E11.5. H&E staining of the accumulation of cardiac jelly over the roof of the right atrium, sectioned hearts of Angpt1DNkx2.5 embryos showed irregularly a presumptive site of atrial septum, and a part of left atrial thickened myocardium with widened interstitial spaces in both wall (Figure 4D). Furthermore, compared with WT embryos, atria, while WT embryos had evenly thickened, compact thicker deposition of cardiac jelly was found surrounding the myocardium in both atria (Figure 4A). In fact, Angpt1DNkx2.5 ventricular trabeculae of Angpt1DNkx2.5 embryos, which led to embryos showed increased myocardial wall thickness in increased space between the endocardium and myocardium, the right atrium (2.5-fold) and left atrium (4.3-fold), as well by 1.7- and 2-fold in the right and left ventricles, respectively as increased trabecular thickness in the right ventricle (Figures 4D and 4E). These results indicate that loss of (1.4-fold) and left ventricle (1.7-fold) compared with WT em- myocardial Angpt1 causes excessive deposition of cardiac bryos (Figures 4B and 4C). Moreover, we found abnormally jelly during cardiac chamber formation.

2458 Cell Reports 23, 2455–2466, May 22, 2018 Figure 3. Myocardial Angpt1 Is Crucial for Cardiac Chamber Morphogenesis (A) Images of CD31 in whole-mount WT and Angpt1DNkx2.5 embryos at E11.5. Magniﬁed images of indicated regions are shown below. Note the small left atrium (yellow arrowheads) and less vascularized endocardial plexus in ventricle (yellow arrow) of Angpt1DNkx2.5 embryos. Scale bars, 500 mm. (B) Images of dissected hearts of WT and Angpt1DNkx2.5 embryos at E11.5. Dashed-lines outline RA and LA. Arrow indicates hypoplastic LA. Double-arrow lines indicate the width and height of atrial chambers. Scale bars, 500 mm. H, height; W, width. (C and D) Comparisons of the width (C) and height (D) of RA and LA of WT and Angpt1DNkx2.5 embryos at E11.5. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. (E) Images of transverse sections of the hearts from WT and Angpt1DNkx2.5 embryos at E11.5. Arrows indicate venous valves in RA of WT embryo. Brackets indicate the atrial septum. Scale bars, 200 mm. (F) Images of inner aspects of the hearts of trans- versely sectioned and DAPI-stained WT and Angpt1DNkx2.5 embryos at E11.5. Lower panels show magniﬁed images of deeper region of hearts revealing openings of sinus venosus (SV) in RA (yellow asterisks). Yellow arrows indicate venous valves in RA of WT embryo. Yellow brackets indicate the atrial septum. Scale bars, 200 mm. (G) Images of sagittal sections of the hearts from WT and Angpt1DNkx2.5 embryos at E11.5. Double- arrow lines indicate chamber space in the LA. Arrows indicate venous valves in RA of WT embryo. Scale bars, 200 mm. LA, left atrium; RA, right atrium; SV, sinus venosus. See also Figure S3.

loss of Angpt1, we analyzed mRNA expression of versican (Vcan) and of hyaluronan synthase 2 (Has2) and UDP- glucose dehydrogenase (Ugdh), two en- zymes that are positively involved in the biosynthesis of hyaluronan (Toole, 2004; Walsh and Stainier, 2001). While Vcan Loss of Angpt1/Tie2 Signaling Reduces ADAMTS- expression was unchanged, expression of Has2 and Ugdh Mediated Degradation of Cardiac Jelly was reduced by 60% and 52%, respectively, in the hearts of To delineate the causes of defective chamber morphogenesis Angpt1DNkx2.5 embryos compared with WT embryos (Figure 5C). and excessive deposition of cardiac jelly in Angpt1DNkx2.5 em- Nevertheless, excessive deposition of versican was detected bryos, we analyzed and compared gene expressions in the in the atria and ventricles of Angpt1DNkx2.5 embryos (Figures 5B whole hearts of WT and Angpt1DNkx2.5 embryos at E10.5. and S4A). Expression of the genes involved in myocardial proliferation These ﬁndings led us to hypothesize that loss of myocardial and trabeculation such as Jag1, Nrg1, and Bmp10 were reduced Angpt1 might hinder ADAMTS-mediated degradation of versi- by 35%–45%, while those of Dll4 and Fgf9 were unaltered (Fig- can (Dupuis et al., 2011; Stankunas et al., 2008; Zhou et al., ure 5A). Of note, at E10.5, Angpt1DNkx2.5 embryos showed 2015b). Analysis of mRNA levels of the ADAMTS family re- reduced trabecular length in the right and left ventricles by vealed that expression of Adamts1, Adamts4, Adamts5, and 32% and 45%, respectively, compared with those of WT em- Adamts9, which show versicanase activity (Porter et al., bryos (Figure 5B, S4A, and S4B). Next, to determine whether 2005), was reduced by 33%–61% in the hearts of Angpt1DNkx2.5 the synthesis of GAG proteins in cardiac jelly is increased upon embryos compared with those of WT embryos. Expression of

Cell Reports 23, 2455–2466, May 22, 2018 2459 Figure 4. Loss of Angpt1 Results in Myocar- dial Wall Thickening and Excessive Deposi- tion of Cardiac Jelly (A) Images of H&E staining of transverse sections of the hearts from WT and Angpt1DNkx2.5 embryos at E11.5. Magniﬁed images of indicated regions are shown in the right panels. Note widened interstitial space within atrial myocardium (arrows), indistinct atrial septum (brackets), and increased space between endocardium and myocardium (arrowheads) in Angpt1DNkx2.5 embryos. Scale bars, 200 mm. (B) Comparisons of wall thickness of each atrial chambers between WT and Angpt1DNkx2.5 embryos. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. (C) Comparisons of trabecular thickness in each ventricles of WT and Angpt1DNkx2.5 embryos. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. (D) Images of Alcian blue and nuclear fast red staining of transverse sections of the hearts from WT and Angpt1DNkx2.5 embryos at E11.5. Note widened interstitial space within atrial myocardium (arrows), indistinct atrial septum (brackets), and excessive cardiac jelly between endocardium and myocardium (arrowheads) in Angpt1DNkx2.5 embryos. Blue bars indicate distance between endocardium and myocardium. Scale bars, 200 mm. (E) Comparisons of distance between endocardium and myocardium in each ventricles of WT and Angpt1DNkx2.5 embryos. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. AS, atrial septum; LA, left atrium; LV, left ventricle; RA, right atrium; RV, right ventricle.

Angpt1DNkx2.5 embryos despite excessive deposition of versican (Figure 5E). Fully supporting these findings, immunoblot analysis revealed that the total amount of cleaved versican in the whole heart was reduced by 72% in Angpt1DNkx2.5 embryos compared with WT embryos (Figures 5F and 5G). Because Angpt1-induced activation of Tie2 signaling is a major pathway for key Adamts2 and Adamts3, which show procollagen N-proteinase cellular activities in endothelial cells (Augustin et al., 2009; Sahar- activity (Porter et al., 2005), was unchanged or slightly inen et al., 2017), we next asked whether Tie2 signaling is reduced (Figure 5D). Of note, expression of tissue inhibitor of involved in the regulation of cardiac jelly degradation. To answer metalloproteinase 3 (Timp3), an endogenous inhibitor of this question, we first crossed the Tie2flox/+ mouse (Savant et al., ADAMTS, was unaltered in the hearts of Angpt1DNkx2.5 embryos 2015) with the Zp3-Cre mouse (de Vries et al., 2000) to obtain a (Figure 5D). Tie2null/+ mouse harboring a null allele, and we then crossed To assess changes in ADAMTS-mediated degradation of Tie2null/+ mice to produce Tie2 null (Tie2null/null) embryos. Consis- versican upon loss of Angpt1, we measured levels of cleaved tent with previous reports (Dumont et al., 1994; Sato et al., 1995), versican (DPEAAE), which is produced by proteolytic cleavage Tie2 null embryos exhibited severe vascular defects and growth of versican by ADAMTS1/4 (Sandy et al., 2001). In WT embryos, restriction at E10.5 (Figure 6A). Dissected hearts of Tie2 null em- cleaved versican was detected abundantly in atrial septum re- bryos at E10.5 exhibited immature chamber morphology, similar gions, the endocardial cushion, and the ventricular trabeculae, to that of Angpt1 null embryos at E10.5 (Figures 2B and 6B). Of showing a correlation with the distribution of intact special note, compared with WT embryos at E10.5, Tie2 null versican. However, it was rarely observed in those regions in embryos had a substantially reduced atrial chamber cavity,

2460 Cell Reports 23, 2455–2466, May 22, 2018 Figure 5. Lack of Angpt1 Causes Decreased ADAMTS Expression and Reduced Degra- dation of Versican (A, C, and D) Comparisons of relative expressions of genes associated with myocardial proliferation and trabeculation (A), cardiac jelly protein synthesis (C), and cardiac jelly protein degradation (D) in the whole hearts of WT and Angpt1DNkx2.5 embryos at E10.5 (n = 5 for WT, n = 7 for Angpt1DNkx2.5). Error bars represent mean ± SD. *p < 0.05, **p < 0.01. (B and E) Images of intact versican (B) and cleaved versican (DPEAAE) (E) in transverse sections of the hearts from WT and Angpt1DNkx2.5 embryos at E10.5. Magniﬁed images of indicated regions are shown in the right panels. For comparative analysis, we performed immunostaining for intact and cleaved versican in adjacent sections of the hearts from WT and Angpt1DNkx2.5 embryos. White bars indicate trabecular length. Arrowheads in (B) indicate intact versican, and arrows in (E) indicate cleaved versican. AS, atrial septum; LA, left atrium; LV, left ventricle. Scale bars, 100 mm. (F and G) Images of representative immunoblot analysis (F) and comparison of cleaved versican (DPEAAE) (G) in whole lysate of the hearts from E10.5 WT and Angpt1DNkx2.5 embryos. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. See also Figure S4.

Based on these results, we evaluated the effect of Tie2 activation on the expression of Adamts1, a major ADAMTS prote- ase in vascular endothelial cells and endocardium (Fu et al., 2011; Hatipoglu et al., 2009; Stankunas et al., 2008; Xu et al., 2006), using primary cultured human umbilical vein endothelial cells (HUVECs). Treatment with COMP-Ang1 (500 ng/mL), a soluble and potent Angpt1 variant (Cho et al., 2004), induced 1.7-fold and GAG-rich cardiac jelly filled the widened space between the upregulation of Adamts1 expression. This upregulation was endocardium and myocardium (Figure 6C, arrow). However, completely blocked by the phosphatidylinositol 3-kinase (PI3K) cleaved versican was rarely detected in atrial chamber of Tie2 inhibitor (LY294002) but unaffected by the MEK1/2 inhibitor null embryos at E10.5, despite the excessive deposition of versi- (U0126) (Figure 6F), indicating that Angpt1 positively regulates can (Figures 6D and 6E). This indicates that Tie2 signaling is also Adamts1 expression through an intracellular PI3K signaling critical for the regulation of cardiac jelly degradation. pathway. Together, these findings suggest that myocardial Previous studies revealed that skeletal muscle satellite cells Angpt1 activates endocardial Tie2 to enhance expression of and pericytes have autocrine Angpt1/Tie2 signaling loop Adamts1, which induces degradation of cardiac jelly. (Abou-Khalil et al., 2009; Teichert et al., 2017). These reports led us to hypothesize that autocrine Angpt1/Tie2 signaling might Regulation of Cardiac Jelly by Angpt1 Is Crucial for Atrial directly regulate the synthesis of versican in myocardium. To Morphogenesis address this hypothesis, we crossed the Tie2null/flox mouse We next questioned how the impaired regulation of cardiac with the Nkx2.5IRESCre mouse (Stanley et al., 2002) to produce jelly eventually leads to defective atrial morphogenesis in a Tie2DNkx2.5 mouse, which allowed depletion of Tie2 expression Angpt1DNkx2.5 embryos. Given that bulging of atrial chambers in Nkx2.5+ cells and their derivatives. Compared with WT em- in the primitive atrium occurs between E9.5 and E10.5 (Anderson bryos, Tie2DNkx2.5 embryos did not exhibit morphological defects et al., 2006), we focused on analyzing the hearts of WT and in the heart, indicating an absence of autocrine Angpt1/Tie2 Angpt1DNkx2.5 embryos at E9.5. We noted that compared with signaling in Nkx2.5+ cardiac progenitors (Figures S5A and S5B). WT embryos, Angpt1DNkx2.5 embryos had excessive versican

Cell Reports 23, 2455–2466, May 22, 2018 2461 Figure 6. Loss of Tie2 Impairs Degradation of Versican and Cardiac Chamber Morpho- genesis (A) Images of CD31 in whole-mount WT and Tie2null/null embryos at E10.5. Magnified images of the hearts are shown below. Note small atrial chamber (yellow arrowheads) and less vascularized endocardial plexus in ventricle (yellow arrow) of Tie2null/null embryos. Scale bars, 500 mm. (B) Images of dissected hearts from WT and Tie2null/null embryos at E10.5. Scale bars, 200 mm. (C) Images of Alcian blue and nuclear fast red staining of transverse sections of the hearts from WT and Tie2null/null embryos at E10.5. Magnified images of the left atrium (LA) are shown below. Note substantially decreased atrial chamber cavity (blue line) and abundant cardiac jelly (arrows). Scale bars, 200 mm. (D and E) Images of intact versican (D) and cleaved versican (DPEAAE) (E) in transverse sections of the hearts from WT and Tie2null/null embryos at E10.5. Magnified images of the LA are shown below. For comparative analysis, we performed immunostaining for intact and cleaved versican in adjacent sections of the hearts from WT and Tie2null/null embryos. Note substantially decreased atrial chamber cavity (yellow line in D), excessive deposition of versican (arrowheads in D), and lack of cleaved versican (arrows in E) in Tie2null/null embryos. Scale bars, 200 mm. LA, left atrium. (F) Comparisons of relative expression of Adamts1 in HUVECs treated with or without COMP-Ang1 (500 ng/mL), PI3K inhibitor (LY294002, 20 nM), and MEK1/2 inhibitor (U0126, 10 nM). n = 4 per group. Error bars represent mean ± SD. *p < 0.05. NS, nonsignificant. See also Figure S5.

embryos at E10.5 (Figures 7D, right panels). Given that chamber expansion requires proper elongation of cardiomyocytes (Auman et al., 2007), these findings suggest that early proliferation and thickening of myocardium in Angpt1DNkx2.5 embryos may disturb bulging of atrial accumulation along the internal surface of the primitive atrium chambers. Indeed, Angpt1DNkx2.5 embryos at E10.5 showed se- and thickened atrial myocardium (Figure 7A, arrowheads and vere impairment in the elongation and polarization of myocardial bars). EdU-incorporation assay at E9.5 revealed that the actin filaments, particularly in the left atrium, compared with WT portion of proliferative cardiomyocytes in the primitive atrium embryos (Figure 7F). Furthermore, we examined the myocardial of Angpt1DNkx2.5 embryos was increased by 24% compared expression of YAP (Yes-associated protein 1), a mechanosensi- with that of WT embryos (Figures 7B and 7C). Moreover, immu- tive transcription factor that shuttles into the nucleus in response nostaining analysis for VCAM1 (a surface marker of cardiomyo- to mechanical stretch (Dupont et al., 2011). Of note, compared cytes; Giacomelli et al., 2017; Uosaki et al., 2011) at E10.5 with the hearts of WT embryos, nuclear expression of YAP was revealed that the number of cell layers in wall myocardium was markedly reduced in the left atrial myocardium of Angpt1DNkx2.5 increased in the right atrium (1.8-fold) and left atrium (3.0-fold) embryos at E10.5, while it was unchanged in the right atrium and of Angpt1DNkx2.5 embryos compared with those of WT embryos ventricles (Figures 7G, 7H, S6A, and S6B), indicating that (Figures 7D and 7E). These findings suggest that excessive ver- myocardial stretching was particularly impaired in the left atrium. sican in the primitive atrium may contribute to enhanced prolifer- Taken together, these findings suggest that Angpt1 plays an ation of cardiomyocytes. essential role in atrial chamber formation by regulating cardiac Cardiomyocytes in the atrial myocardium of Angpt1DNkx2.5 em- jelly distribution in the primitive atrium, thereby coordinating bryos failed to elongate properly compared with those of WT myocardial proliferation and chamber expansion.

2462 Cell Reports 23, 2455–2466, May 22, 2018 Figure 7. Regulation of Cardiac Jelly Is Crucial for Atrial Morphogenesis (A) Images of transverse and sagittal sections of the hearts from WT and Angpt1DNkx2.5 embryos at E9.5. Magnified images of LA are shown at right. Yellow bars indicate thickness of myocardium. Arrowheads indicate accumulated versican (arrowheads) in the atrium of Angpt1DNkx2.5 embryos. Scale bars, 100 mm. (B) Images of EdU incorporation (arrowheads) in the left atrial myocardium of WT and Angpt1DNkx2.5 embryos at E9.5. Scale bars, 100 mm. (C) Comparison of the percentage of EdU+/total cardiomyocytes (CM) in the atrial myocardium of WT and Angpt1DNkx2.5 embryos at E9.5. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. (D) Images of VCAM1-stained myocardium of RA and LA of WT and Angpt1DNkx2.5 embryos at E10.5. Yellow bars indicate myocardial thickness. Magnified images of indicated regions are shown at right. Note the shape of yellow colored cardiomyocytes in WT (rectangular) and Angpt1DNkx2.5 (round) hearts. Scale bars, 25 mm. (E) Comparisons of the number of VCAM1+ cardiomyocyte (CM) layer in the walls of RA and LA of WT and Angpt1DNkx2.5 embryos at E10.5. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. (F) Images of F-actin in the myocardium of RA and LA of WT and Angpt1DNkx2.5 embryos at E10.5. White arrows indicate well-aligned and elongated actin filament in WT embryo. Scale bars, 50 mm. (G) Images of YAP in the myocardium of RA and LA of WT and Angpt1DNkx2.5 embryos at E10.5. Scale bars, 50 mm. (H) Comparisons of nuclear localization of YAP on a-Actinin+/DAPI+ CM nuclei in the myocardium of RA and LA of WT and Angpt1DNkx2.5 embryos at E10.5. n = 4 embryos per group. Error bars represent mean ± SD. *p < 0.05. LA, left atrium; RA, right atrium. See also Figure S6.

E10.5. Compared with the findings from our previous study that utilized a-MHC- Cre driver to deplete myocardium- DISCUSSION derived Angpt1 (Arita et al., 2014), ablation of Angpt1 using Nkx2.5IRESCre mice results in earlier and severe manifestation In this study, we investigated the role of Angpt1 in cardiac devel- of atrial defects. However, considering the broad contribution opment and clarified the underlying molecular mechanisms. of Nkx2.5+ progenitors beyond myocardium, future studies Although previous studies have shown spatial patterns of that enable myocardium-specific deletion of Angpt1 from the Angpt1 expression in atrial and trabecular myocardium by using earliest stage of cardiac development are required. Detailed an- in situ hybridization or Angpt1mCherry reporter mice (Arita et al., alyses of the hearts of Angpt1DNkx2.5 embryos revealed that lack 2014; Davis et al., 1996; Maisonpierre et al., 1997; Suri et al., of Angpt1 leads to abnormal growth of myocardium and exces- 1996), their findings were unclear regarding the detection of dy- sive deposition of cardiac jelly. Versican and hyaluronan have namic changes in Angpt1 expression during cardiac develop- been proposed to be involved in the regulation of cellular prolif- ment. In this study, taking advantage of the Angpt1GFP knockin eration (Henderson and Copp, 1998; Passer et al., 2016; Sheng mice (Zhou et al., 2015a), we could observe spatiotemporal et al., 2005), but the underlying molecular mechanisms remain expression of Angpt1 in developing hearts in detail. largely unknown. A recent study demonstrated that hyaluronan Analyses of Angpt1 null and Angpt1DNkx2.5 embryos revealed in cardiac jelly indirectly contributes to the polarized division of severely compromised atrial chamber formation as early as cardiomyocytes during trabeculation by modulating endocardial

Cell Reports 23, 2455–2466, May 22, 2018 2463 signaling (Passer et al., 2016). Of note, versican is initially ex- (Lee et al., 2013) and Tie2flox/flox (Savant et al., 2015) mice were transferred pressed in the myocardium of the tubular heart, and its expres- and bred in the SPF animal facilities at KAIST. All mice were maintained in sion is rapidly downregulated in the walls of atrial and ventricular the C57BL/6 background. Mice were fed with free access to water and stan- dard diet (PMI Lab). Embryos and tissue samples were harvested after mice chambers after looping but maintained in the developing struc- were anesthetized and sacrificed. The anesthetic agent was prepared as a tures, such as the atrial septum, venous valves in the right atrium, combination of ketamine (80 mg/kg) and xylazine (12 mg/kg), which was intra- and ventricular trabeculae (Henderson and Copp, 1998). Taken peritoneally administered. Embryonic ages are indicated in the figures and together, such restricted distribution of versican may regulate figure legends. The genders of the embryos were unknown. Adult tissue sam- the directional growth of those structures perpendicular to the ples were harvested from 8- to 12-week-old male mice. wall myocardium. However, excessive deposition of versican D Statistical Analyses in hearts of Angpt1 Nkx2.5 embryos disrupts proper regulation Values are presented as mean ± SD. Statistical significance was determined of myocardial proliferation in chamber walls and shaping of the by the Mann-Whitney U test between 2 groups or the Kruskal-Wallis test atrial septum, venous valves, and trabeculae (Figure S7). followed by Tukey’s HSD (honest significant difference) test with ranks for Furthermore, our findings indicate that myocardial thickening in multiple-group comparison. Statistical analyses were performed using primitive atrium of Angpt1DNkx2.5 embryos disturbs chamber PASW statistics 18 (SPSS). Statistical significance was set to p < 0.05. expansion, particularly that of left atrium. It is possible that he- modynamic forces generated by venous flow into the right atrium SUPPLEMENTAL INFORMATION was not fully delivered to the left atrium in Angpt1DNkx2.5 embryos Supplemental Information includes Supplemental Experimental Procedures, due to immature venous valves and atrial septum, which are seven figures, and one table and can be found with this article online required for preventing backflow. at https://doi.org/10.1016/j.celrep.2018.04.080. A previous study showed that Angpt1 inhibits FOXO1 (Fork- head box protein O1) target genes, such as MMP7 (matrix metal- ACKNOWLEDGMENTS loproteinases 7), that are involved in matrix remodeling (Daly et al., 2004). That result is consistent with in vivo findings that We thank Intae Park for proofreading the manuscript and Sujin Seo, Hyun-Tae Kim, and Tae-Chang Yang for technical assistance. We also thank Dr. Sean Angpt1 stabilizes the close association between endothelial cells J. Morrison (University of Texas Southwestern Medical Center) for providing and the surrounding ECM (Suri et al., 1996). However, our results the Angpt1GFP mice. This study was supported by the Institute for Basic Sci- demonstrated that Angpt1 can also promote ADAMTS-depen- ence funded by the Ministry of Science and ICT, Republic of Korea (IBS- dent degradation of the ECM, thereby controlling the microenvi- R025-D1, G.Y.K.). ronment of developing hearts. Another earlier study found that whole-body deletion of Angpt1 after E13.5 did not lead to any AUTHOR CONTRIBUTIONS overt phenotypes (Jeansson et al., 2011). This result implies K.H.K. designed and performed the experiments and analyzed the data; that Angpt1 is dispensable for cardiac development after K.H.K. and G.Y.K. generated the figures and wrote and edited the manuscript; E13.5. Moreover, we showed that myocardial Angpt1 expression Y.N. and H.G.A. provided mice and critical comments on this study; and emerged in the primitive atrium around E8.5 and that the loss of G.Y.K. directed and supervised the project. Angpt1 led to excessive deposition of cardiac jelly and myocardial thickening at E9.5; this suggests that Angpt1-induced regu- DECLARATION OF INTERESTS lation of cardiac jelly distribution in the atrium occurs during the The authors declare no competing interests. looping of tubular heart (E8.5–E9.5), the most dynamic stage of cardiac development. Received: January 10, 2018 Taken together, the results of this study delineate the role Revised: March 9, 2018 of myocardial Angpt1/endocardial Tie2 signaling in the spatio- Accepted: April 17, 2018 temporal degradation of cardiac jelly, revealing a mechanism Published: May 22, 2018 involved in the regulation of atrial chamber morphogenesis. REFERENCES Given that Tie2 mutations that cause hereditary cutaneomucosal venous malformation are also associated with ventricular septal Abou-Khalil, R., Le Grand, F., Pallafacchina, G., Valable, S., Authier, F.J., Rud- defects (Wouters et al., 2010), further investigation identifying nicki, M.A., Gherardi, R.K., Germain, S., Chretien, F., Sotiropoulos, A., et al. Angpt1 mutations in patients with congenital septal defects will (2009). Autocrine and paracrine angiopoietin 1/Tie-2 signaling promotes mus- provide additional insights into the pathogenesis of congenital cle satellite cell self-renewal. Cell Stem Cell 5, 298–309. cardiac defects. Anderson, R.H., Brown, N.A., and Moorman, A.F. (2006). Development and structures of the venous pole of the heart. Dev. Dyn. 235, 2–9. Arita, Y., Nakaoka, Y., Matsunaga, T., Kidoya, H., Yamamizu, K., Arima, Y., Ka- EXPERIMENTAL PROCEDURES taoka-Hashimoto, T., Ikeoka, K., Yasui, T., Masaki, T., et al. (2014). Myocar- dium-derived angiopoietin-1 is essential for coronary vein formation in the Mice developing heart. Nat. Commun. 5, 4552. Animal care and experimental procedures were performed under the approval from the Institutional Animal Care and Use Committee (no. KA2013-39) of Armulik, A., Genove´ , G., and Betsholtz, C. (2011). Pericytes: developmental, KAIST. Specific-pathogen-free (SPF) C57BL/6J, Nkx2.5IRESCre (Stanley et al., physiological, and pathological perspectives, problems, and promises. Dev. 2002), ZP3-Cre (de Vries et al., 2000), and ROSA26-tdTomato (C57BL/6J) Cell 21, 193–215. (Madisen et al., 2010) mice were purchased from The Jackson Laboratory. Augustin, H.G., Koh, G.Y., Thurston, G., and Alitalo, K. (2009). Control of Angpt1GFP mice (Zhou et al., 2015a) were generously provided by Dr. Sean vascular morphogenesis and homeostasis through the angiopoietin-Tie sys- Morrison (University of Texas Southwestern Medical Center). Ang1flox/flox tem. Nat. Rev. Mol. Cell Biol. 10, 165–177.

2464 Cell Reports 23, 2455–2466, May 22, 2018 Auman, H.J., Coleman, H., Riley, H.E., Olale, F., Tsai, H.J., and Yelon, D. ventricular septal formation subsequent to cardiac atrioventricular cushion (2007). Functional modulation of cardiac form through regionally confined development. Glycobiology 22, 1268–1277. cell shape changes. PLoS Biol. 5, e53. Hatipoglu, O.F., Hirohata, S., Cilek, M.Z., Ogawa, H., Miyoshi, T., Obika, M., Buckingham, M., Meilhac, S., and Zaffran, S. (2005). Building the mammalian Demircan, K., Shinohata, R., Kusachi, S., and Ninomiya, Y. (2009). ADAMTS1 heart from two sources of myocardial cells. Nat. Rev. Genet. 6, 826–835. is a unique hypoxic early response gene expressed by endothelial cells. J. Biol. Camenisch, T.D., Spicer, A.P., Brehm-Gibson, T., Biesterfeldt, J., Augustine, Chem. 284, 16325–16333. M.L., Calabro, A., Jr., Kubalak, S., Klewer, S.E., and McDonald, J.A. (2000). Henderson, D.J., and Copp, A.J. (1998). Versican expression is associated Disruption of hyaluronan synthase-2 abrogates normal cardiac morphogen- with chamber specification, septation, and valvulogenesis in the developing esis and hyaluronan-mediated transformation of epithelium to mesenchyme. mouse heart. Circ. Res. 83, 523–532. J. Clin. Invest. 106, 349–360. Jeansson, M., Gawlik, A., Anderson, G., Li, C., Kerjaschki, D., Henkelman, M., Camenisch, T.D., Schroeder, J.A., Bradley, J., Klewer, S.E., and McDonald, and Quaggin, S.E. (2011). Angiopoietin-1 is essential in mouse vasculature J.A. (2002). Heart-valve mesenchyme formation is dependent on hyalur- during development and in response to injury. J. Clin. Invest. 121, 2278–2289. onan-augmented activation of ErbB2-ErbB3 receptors. Nat. Med. 8, 850–855. Kern, C.B., Twal, W.O., Mjaatvedt, C.H., Fairey, S.E., Toole, B.P., Iruela- Cho, C.-H., Kammerer, R.A., Lee, H.J., Steinmetz, M.O., Ryu, Y.S., Lee, S.H., Arispe, M.L., and Argraves, W.S. (2006). Proteolytic cleavage of versican dur- Yasunaga, K., Kim, K.-T., Kim, I., Choi, H.-H., et al. (2004). COMP-Ang1: a de- ing cardiac cushion morphogenesis. Dev. Dyn. 235, 2238–2247. signed angiopoietin-1 variant with nonleaky angiogenic activity. Proc. Natl. Kern, C.B., Wessels, A., McGarity, J., Dixon, L.J., Alston, E., Argraves, W.S., Acad. Sci. USA 101, 5547–5552. Geeting, D., Nelson, C.M., Menick, D.R., and Apte, S.S. (2010). Reduced ver- Chu, M., Li, T., Shen, B., Cao, X., Zhong, H., Zhang, L., Zhou, F., Ma, W., Jiang, sican cleavage due to Adamts9 haploinsufficiency is associated with cardiac H., Xie, P., et al. (2016). Angiopoietin receptor Tie2 is required for vein specifi- and aortic anomalies. Matrix Biol. 29, 304–316. cation and maintenance via regulating COUP-TFII. eLife 5, e21032. Kim, J., Park, D.Y., Bae, H., Park, D.Y., Kim, D., Lee, C.K., Song, S., Chung, D’Amato, G., Luxa´ n, G., del Monte-Nieto, G., Martıńez-Poveda, B., Torroja, C., T.Y., Lim, D.H., Kubota, Y., et al. (2017). Impaired angiopoietin/Tie2 signaling Walter, W., Bochter, M.S., Benedito, R., Cole, S., Martinez, F., et al. (2016). compromises Schlemm’s canal integrity and induces glaucoma. J. Clin. Sequential Notch activation regulates ventricular chamber development. Invest. 127, 3877–3896. Nat. Cell Biol. 18, 7–20. Lai, D., Liu, X., Forrai, A., Wolstein, O., Michalicek, J., Ahmed, I., Garratt, A.N., Daly, C., Wong, V., Burova, E., Wei, Y., Zabski, S., Griffiths, J., Lai, K.M., Lin, Birchmeier, C., Zhou, M., Hartley, L., et al. (2010). Neuregulin 1 sustains the H.C., Ioffe, E., Yancopoulos, G.D., and Rudge, J.S. (2004). Angiopoietin-1 gene regulatory network in both trabecular and nontrabecular myocardium. modulates endothelial cell function and gene expression via the transcription Circ. Res. 107, 715–727. factor FKHR (FOXO1). Genes Dev. 18, 1060–1071. Lavine, K.J., Yu, K., White, A.C., Zhang, X., Smith, C., Partanen, J., and Ornitz, Davis, S., Aldrich, T.H., Jones, P.F., Acheson, A., Compton, D.L., Jain, V., D.M. (2005). Endocardial and epicardial derived FGF signals regulate myocar- Ryan, T.E., Bruno, J., Radziejewski, C., Maisonpierre, P.C., and Yancopoulos, dial proliferation and differentiation in vivo. Dev. Cell 8, 85–95. G.D. (1996). Isolation of angiopoietin-1, a ligand for the TIE2 receptor, by Lee, J., Kim, K.E., Choi, D.K., Jang, J.Y., Jung, J.J., Kiyonari, H., Shioi, G., secretion-trap expression cloning. Cell 87, 1161–1169. Chang, W., Suda, T., Mochizuki, N., et al. (2013). Angiopoietin-1 guides direc- de Vries, W.N., Binns, L.T., Fancher, K.S., Dean, J., Moore, R., Kemler, R., and tional angiogenesis through integrin avb5 signaling for recovery of ischemic Knowles, B.B. (2000). Expression of Cre recombinase in mouse oocytes: a retinopathy. Sci. Transl. Med. 5, 203ra127. means to study maternal effect genes. Genesis 26, 110–112. Liu, J., Bressan, M., Hassel, D., Huisken, J., Staudt, D., Kikuchi, K., Poss, K.D., Dumont, D.J., Gradwohl, G., Fong, G.-H., Puri, M.C., Gertsenstein, M., Auer- Mikawa, T., and Stainier, D.Y. (2010). A dual role for ErbB2 signaling in cardiac bach, A., and Breitman, M.L. (1994). Dominant-negative and targeted null mu- trabeculation. Development 137, 3867–3875. tations in the endothelial receptor tyrosine kinase, tek, reveal a critical role in vasculogenesis of the embryo. Genes Dev. 8, 1897–1909. Lockhart, M., Wirrig, E., Phelps, A., and Wessels, A. (2011). Extracellular matrix and heart development. Birth Defects Res. A Clin. Mol. Teratol. 91, 535–550. Dupont, S., Morsut, L., Aragona, M., Enzo, E., Giulitti, S., Cordenonsi, M., Zan- conato, F., Le Digabel, J., Forcato, M., Bicciato, S., et al. (2011). Role of YAP/ Madisen, L., Zwingman, T.A., Sunkin, S.M., Oh, S.W., Zariwala, H.A., Gu, H., TAZ in mechanotransduction. Nature 474, 179–183. Ng, L.L., Palmiter, R.D., Hawrylycz, M.J., Jones, A.R., et al. (2010). A robust and high-throughput Cre reporting and characterization system for the whole Dupuis, L.E., McCulloch, D.R., McGarity, J.D., Bahan, A., Wessels, A., Weber, mouse brain. Nat. Neurosci. 13, 133–140. D., Diminich, A.M., Nelson, C.M., Apte, S.S., and Kern, C.B. (2011). Altered versican cleavage in ADAMTS5 deficient mice; a novel etiology of myxoma- Maisonpierre, P.C., Suri, C., Jones, P.F., Bartunkova, S., Wiegand, S.J., Rad- tous valve disease. Dev. Biol. 357, 152–164. ziejewski, C., Compton, D., McClain, J., Aldrich, T.H., Papadopoulos, N., et al. (1997). Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angio- Fu, Y., Nagy, J.A., Brown, L.F., Shih, S.-C., Johnson, P.Y., Chan, C.K., Dvorak, genesis. Science 277, 55–60. H.F., and Wight, T.N. (2011). Proteolytic cleavage of versican and involvement of ADAMTS-1 in VEGF-A/VPF-induced pathological angiogenesis. Nakaoka, Y., Nishida, K., Narimatsu, M., Kamiya, A., Minami, T., Sawa, H., J. Histochem. Cytochem. 59, 463–473. Okawa, K., Fujio, Y., Koyama, T., Maeda, M., et al. (2007). Gab family proteins are essential for postnatal maintenance of cardiac function via neuregulin-1/ Giacomelli, E., Bellin, M., Sala, L., van Meer, B.J., Tertoolen, L.G.J., Orlova, ErbB signaling. J. Clin. Invest. 117, 1771–1781. V.V., and Mummery, C.L. (2017). Three-dimensional cardiac microtissues composed of cardiomyocytes and endothelial cells co-differentiated from hu- Park, D.Y., Lee, J., Kim, J., Kim, K., Hong, S., Han, S., Kubota, Y., Augustin, man pluripotent stem cells. Development 144, 1008–1017. H.G., Ding, L., Kim, J.W., et al. (2017). Plastic roles of pericytes in the blood- Grego-Bessa, J., Luna-Zurita, L., del Monte, G., Bolo´ s, V., Melgar, P., Ara- retinal barrier. Nat. Commun. 8, 15296. ndilla, A., Garratt, A.N., Zang, H., Mukouyama, Y.S., Chen, H., et al. (2007). Passer, D., van de Vrugt, A., Atmanli, A., and Domian, I.J. (2016). Atypical pro- Notch signaling is essential for ventricular chamber development. Dev. Cell tein kinase C-dependent polarized cell division is required for myocardial tra- 12, 415–429. beculation. Cell Rep. 14, 1662–1672. Haack, T., and Abdelilah-Seyfried, S. (2016). The force within: endocardial Porter, S., Clark, I.M., Kevorkian, L., and Edwards, D.R. (2005). The ADAMTS development, mechanotransduction and signalling during cardiac morpho- metalloproteinases. Biochem. J. 386, 15–27. genesis. Development 143, 373–386. Quaegebeur, A., Lange, C., and Carmeliet, P. (2011). The neurovascular link in Hatano, S., Kimata, K., Hiraiwa, N., Kusakabe, M., Isogai, Z., Adachi, E., Shi- health and disease: molecular mechanisms and therapeutic implications. nomura, T., and Watanabe, H. (2012). Versican/PG-M is essential for Neuron 71, 406–424.

Cell Reports 23, 2455–2466, May 22, 2018 2465 Reischauer, S., Stone, O.A., Villasenor, A., Chi, N., Jin, S.-W., Martin, M., Lee, Teichert, M., Milde, L., Holm, A., Stanicek, L., Gengenbacher, N., Savant, S., M.T., Fukuda, N., Marass, M., Witty, A., et al. (2016). Cloche is a bHLH-PAS Ruckdeschel, T., Hasanov, Z., Srivastava, K., Hu, J., et al. (2017). Pericyte-ex- transcription factor that drives haemato-vascular speciﬁcation. Nature 535, pressed Tie2 controls angiogenesis and vessel maturation. Nat. Commun. 8, 294–298. 16106.

Saharinen, P., Eklund, L., and Alitalo, K. (2017). Therapeutic targeting of the Thomson, B.R., Souma, T., Tompson, S.W., Onay, T., Kizhatil, K., Siggs, O.M., angiopoietin-TIE pathway. Nat. Rev. Drug Discov. 16, 635–661. Feng, L., Whisenhunt, K.N., Yanovitch, T.L., Kalaydjieva, L., et al. (2017). Sandy, J.D., Westling, J., Kenagy, R.D., Iruela-Arispe, M.L., Verscharen, C., Angiopoietin-1 is required for Schlemm’s canal development in mice and Rodriguez-Mazaneque, J.C., Zimmermann, D.R., Lemire, J.M., Fischer, humans. J. Clin. Invest. 127, 4421–4436. J.W., Wight, T.N., and Clowes, A.W. (2001). Versican V1 proteolysis in human Tian, Y., and Morrisey, E.E. (2012). Importance of myocyte-nonmyocyte inter- aorta in vivo occurs at the Glu441-Ala442 bond, a site that is cleaved by reactions in cardiac development and disease. Circ. Res. 110, 1023–1034. combinant ADAMTS-1 and ADAMTS-4. J. Biol. Chem. 276, 13372–13378. Sato, T.N., Tozawa, Y., Deutsch, U., Wolburg-Buchholz, K., Fujiwara, Y., Gen- Toole, B.P. (2004). Hyaluronan: from extracellular glue to pericellular cue. Nat. dron-Maguire, M., Gridley, T., Wolburg, H., Risau, W., and Qin, Y. (1995). Rev. Cancer 4, 528–539. Distinct roles of the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel Uosaki, H., Fukushima, H., Takeuchi, A., Matsuoka, S., Nakatsuji, N., Yama- formation. Nature 376, 70–74. naka, S., and Yamashita, J.K. (2011). Efficient and scalable purification of car- Savant, S., La Porta, S., Budnik, A., Busch, K., Hu, J., Tisch, N., Korn, C., Valls, diomyocytes from human embryonic and induced pluripotent stem cells by A.F., Benest, A.V., Terhardt, D., et al. (2015). The orphan receptor tie1 controls VCAM1 surface expression. PLoS ONE 6, e23657. angiogenesis and vascular remodeling by differentially regulating Tie2 in tip Walsh, E.C., and Stainier, D.Y. (2001). UDP-glucose dehydrogenase required and stalk cells. Cell Rep. 12, 1761–1773. for cardiac valve formation in zebrafish. Science 293, 1670–1673. Sedmera, D., Pexieder, T., Vuillemin, M., Thompson, R.P., and Anderson, R.H. (2000). Developmental patterning of the myocardium. Anat. Rec. 258, Wouters, V., Limaye, N., Uebelhoer, M., Irrthum, A., Boon, L.M., Mulliken, J.B., 319–337. Enjolras, O., Baselga, E., Berg, J., Dompmartin, A., et al. (2010). Hereditary Sheng, W., Wang, G., Wang, Y., Liang, J., Wen, J., Zheng, P.-S., Wu, Y., Lee, cutaneomucosal venous malformations are caused by TIE2 mutations with V., Slingerland, J., Dumont, D., and Yang, B.B. (2005). The roles of versican V1 widely variable hyper-phosphorylating effects. Eur. J. Hum. Genet. 18, and V2 isoforms in cell proliferation and apoptosis. Mol. Biol. Cell 16, 1330– 414–420. 1340. Xu, Z., Yu, Y., and Duh, E.J. (2006). Vascular endothelial growth factor upregu- Stainier, D.Y. (2001). Zebrafish genetics and vertebrate heart formation. Nat. lates expression of ADAMTS1 in endothelial cells through protein kinase C Rev. Genet. 2, 39–48. signaling. Invest. Ophthalmol. Vis. Sci. 47, 4059–4066. Stainier, D.Y., Weinstein, B.M., Detrich, H.W., 3rd, Zon, L.I., and Fishman, M.C. Yamamura, H., Zhang, M., Markwald, R.R., and Mjaatvedt, C.H. (1997). A heart (1995). Cloche, an early acting zebrafish gene, is required by both the endothe- segmental defect in the anterior-posterior axis of a transgenic mutant mouse. lial and hematopoietic lineages. Development 121, 3141–3150. Dev. Biol. 186, 58–72. Stankunas, K., Hang, C.T., Tsun, Z.Y., Chen, H., Lee, N.V., Wu, J.I., Shang, C., Zhang, Y., Li, S., Yuan, L., Tian, Y., Weidenfeld, J., Yang, J., Liu, F., Chokas, Bayle, J.H., Shou, W., Iruela-Arispe, M.L., and Chang, C.P. (2008). Endocardial A.L., and Morrisey, E.E. (2010). Foxp1 coordinates cardiomyocyte proliferation Brg1 represses ADAMTS1 to maintain the microenvironment for myocardial through both cell-autonomous and nonautonomous mechanisms. Genes Dev. morphogenesis. Dev. Cell 14, 298–311. 24, 1746–1757. Stanley, E.G., Biben, C., Elefanty, A., Barnett, L., Koentgen, F., Robb, L., and Harvey, R.P. (2002). Efficient Cre-mediated deletion in cardiac progenitor cells Zhou, B.O., Ding, L., and Morrison, S.J. (2015a). Hematopoietic stem and conferred by a 3’UTR-ires-Cre allele of the homeobox gene Nkx2-5. Int. J. Dev. progenitor cells regulate the regeneration of their niche by secreting Biol. 46, 431–439. Angiopoietin-1. eLife 4, e05521. Suri, C., Jones, P.F., Patan, S., Bartunkova, S., Maisonpierre, P.C., Davis, S., Zhou, Z., Rawnsley, D.R., Goddard, L.M., Pan, W., Cao, X.J., Jakus, Z., Zheng, Sato, T.N., and Yancopoulos, G.D. (1996). Requisite role of angiopoietin-1, H., Yang, J., Arthur, J.S., Whitehead, K.J., et al. (2015b). The cerebral a ligand for the TIE2 receptor, during embryonic angiogenesis. Cell 87, cavernous malformation pathway controls cardiac development via regulation 1171–1180. of endocardial MEKK3 signaling and KLF expression. Dev. Cell 32, 168–180.

2466 Cell Reports 23, 2455–2466, May 22, 2018