To EndoMT or not to EndoMT: Zebrafish heart valve development

In the developing vertebrate heart, valves form in the primitive heart tube to ensure unidirectional blood flow. Mammals and birds have four-chambered hearts with two atrioventricular (AV) valves between the atria and ventricles, as well as semilunar valves between each ventricle and great vessel. Mammalian heart valve development begins with the formation of endocardial cushions through endothelial-to-mesenchymal transition (EndoMT) followed by expansion, elongation, and remodeling of valve primordia into mature leaflets.¹ Valve remodeling continues postnatally, and the mature valves consist of extracellular matrix layers interspersed with interstitial cells and surrounded by endocardial endothelial cells (Figure). In contrast, fish have two-chambered hearts separated by a single AV valve with a single outflow valve. Zebrafish are a well-established genetic model for the early stages of heart development as the transparency of early embryos facilitates in vivo visualization² However, less is known about the later stages of valve development and heart chamber maturation. In this issue, a study by Gunawan et al.³ reports for the first time that zebrafish heart valve interstitial cells (VICs) are generated by EndoMT dependent on the transcription factors Nfatc1 and Twist1b.

Figure: Comparison of heart valve development in zebrafish and mice.

Major events of zebrafish (top panels) and mouse (bottom panels) atrioventricular valve development are shown. Note that embryonic zebrafish valve cells do not fully detach from endocardial progenitors, representing an incomplete EndoMT, within 96 hpf. Abbreviations: hpf, hours post fertilization; dpf, days post fertilization; E, Embryonic day; EndoMT, Endothelial-to-Mesenchymal Transition; VEC, Valve Endothelial Cell; VIC, Valve Interstitial Cell; NCC, Neural Crest Cell.

The earliest stages of zebrafish heart valve formation have been defined as an outgrowth of endocardial endothelial cells in the AV canal to form a two-layered structure lacking mesenchymal cells.⁴ This structure forms by 96 hours after fertilization with a stereotypic arrangement of individual endothelial and adjoined cells. The formation of these endothelial outgrowths is dependent on blood flow and is regulated by mechanosensitive factors, such as ion channels and the transcription factor Klf2, as well as integrins and focal adhesions.^4–6 This process of endothelial rearrangement resulting in valve outgrowth has been termed collective migration. Since endothelial cell contacts are maintained, the initial stages of valvulogenesis do not recapitulate complete EndoMT as it occurs in birds and mammals. The later stages of zebrafish valvulogenesis have been obscured by lack of larval transparency and inadequate tools for genetic manipulations in older fish.

The study by Gunawan et al focuses on the regulation of zebrafish heart valvulogenesis by the transcription factor Nfatc1 and takes advantage of newly generated genetic tools and imaging strategies to examine VIC formation and leaflet morphogenesis in juvenile fish.³ In contrast to most previous reports, this study includes analysis of the maturation of heart valves 20–90 days post fertilization (dpf) and also assesses mature heart valve function through echocardiography. In addition, cell lineage tracing was used to demonstrate the presence of fully detached endothelial-derived VICs by 20 dpf, which is in contrast to the timing of early embryonic endocardial cushion formation and EndoMT in the primitive heart tube of mammals. Interestingly, a small population of VICs derived from neural crest cells was identified, resembling a similar population found in the mature semilunar valves of the four-chambered heart. For the first time, this study revealed how, similar to mammalian valves, the majority of zebrafish heart VICs arise from the EndoMT of endocardial endothelial cells. However, completion of EndoMT, evident in full detachment of VICs from endocardial endothelial cells, occurs many days after initial chamber specification in the primitive heart, a critical difference from mammalian valvulogenesis (Figure).

The gene regulatory networks that control generation of zebrafish VICs were also examined using new transgenic lines for tracking valve cell lineages and manipulation of valve gene expression. The NF-kB-related transcription factor Nfatc1 is expressed specifically in valve endothelial cells and is required for elongation and remodeling of heart valve primordia in mice.^{7, 8} In zebrafish, Nfatc1 is also expressed in valve endothelial cells, and loss of Nfatc1 results in decreased numbers of VICs and valvular insufficiency evident in regurgitant blood flow in adult animals.³ Gene expression profiling of nfatc1^−/− hearts demonstrated reduced expression of several critical regulators of EndoMT, including twist1b. Additional studies confirmed that twist1b gene expression is induced by Nfatc1 and that expression of a dominant negative form of Twist1 prevents VIC formation and also leads to valvular insufficiency. Together, these studies define a critical regulatory interaction between Nfatc1 and Twist1b in EndoMT and VIC formation during heart valve maturation in zebrafish.

Valve formation in four-chambered hearts is first evident at three days of development in birds and at Embryonic day (E)9.5 in mice with the formation of endocardial cushions in the AV canal and outflow tract.¹ Endocardial cushion formation begins with expansion of the hyaluronan-rich cardiac jelly between the heart muscle and endothelial cells, leading to swellings of the endocardial cushions followed by robust EndoMT and formation of mesenchymal VIC progenitors within 24 hours. The initiation of endocardial cushion swellings is dependent on Bone Morphogenetic Protein (BMP) and Notch signaling in the AV canal, which also regulates EndoMT as endocardial cells transform into VIC progenitor cells that express mesenchymal genes including twist1, snai1 and msx2.⁹ Interestingly, the loss of Nfatc1 in mammalian valves does not affect VIC progenitor formation but does affect outgrowth of the valve primordia and remodeling of valve leaflets.^{7, 8, 10} Similarly, Notch and BMP signaling are active in the AV canal of zebrafish at around 48 hpf. However, zebrafish VIC formation is dependent on Nfatc1 and definitive detached VICs embedded in ECM are not apparent until 20–30 dpf. Thus, regulatory pathways involved are conserved between the two-chambered and four-chambered heart, but the timing, notably when VICs fully detach from the endocardium relative to AV canal specification, is delayed in zebafish relative to mice.

The study by Gunawan et al. also demonstrates evidence of valve leaflet ECM remodeling at 20 dpf.³ In mice and humans, heart valve remodeling and ECM stratification occurs after birth concurrent with the separation of pulmonic and systemic circulation and increased cardiac demand.¹ It is not clear how the ECM maturation in month-old zebrafish compares to the perinatal valve maturation of mammals, since zebrafish body size is variable and can be dependent on environmental conditions. During early stages of zebrafish heart development, the role of fluid forces in valve development has been studied extensively, as blood flow can be manipulated by drugs or genetic means without causing lethality.^4–6 These studies have defined mechanosensitive regulatory mechanisms that govern early outgrowth of endothelial cells during the initiation of valvulogenesis. Similarly, nfatc1^−/− mutant zebrafish are viable as adults with severely regurgitant valves. Thus, zebrafish may be a useful model for examining the specific effects of altered fluid flow on mature valve organization and function that are inaccessible in mammals due to their strict dependence on cardiac function.

In addition to the Gunawan study, the Stainier group also recently reported the ability of adult zebrafish to regenerate heart valves after genetic ablation of Nfatc1-expressing valve endothelial and interstitial cells.¹¹ By 21 days after ablation, new valve cells are generated from endothelial cells and infiltrating kidney marrow-derived cells, leading to improved function and ECM expression. The formation of new valve tissue is dependent on Transforming Growth Factor (TGF)β signaling, which promotes valve cell proliferation and differentiation after injury. The resulting valve is thickened with disorganized and ectopic ECM relative to the uninjured valve. Interestingly, myxomatous valve degeneration in humans, dogs, pigs, and mice includes TGFβ activation and macrophage infiltration, in addition to thickening and disorganization of the valve leaflet ECM.¹² In mammals this has been characterized as a pathologic sterile inflammatory response. Since valve remodeling after cell ablation in zebrafish shares many of these features, it remains to be seen whether this valve regenerative process could be harnessed to promote repair in human heart valve disease.

In the past few years, zebrafish have emerged as a valuable animal model for studies of heart valve development, in particular in valve maturation processes related to fluid dynamics. Although there are differences in the timing of EndoMT and in ECM remodeling of mature valves, many of the critical regulatory mechanisms governing valvulogenesis are conserved between zebrafish and mammals. In addition, the recent reports of valve maturation and regeneration in adult zebrafish heart valves suggest that they could be used in studies of adult valve homeostasis and disease. Additional studies are needed to determine if major valve pathogenic processes, including calcification and myxomatous degeneration, occur in adult zebrafish. However, if appropriate disease models are generated, it would open up new avenues of investigation using classic tools of small molecule and genetic screens for discovery of new therapeutics for heart valve disease. These high throughput in vivo approaches for testing potential valve therapeutics would represent a significant new tool in drug discovery efforts for some of the most common types of cardiovascular disease.