Multiple modes of proepicardial cell migration require heartbeat
© Plavicki et al.; licensee BioMed Central Ltd. 2014
Received: 10 December 2013
Accepted: 6 May 2014
Published: 15 May 2014
Skip to main content
© Plavicki et al.; licensee BioMed Central Ltd. 2014
Received: 10 December 2013
Accepted: 6 May 2014
Published: 15 May 2014
The outermost layer of the vertebrate heart, the epicardium, forms from a cluster of progenitor cells termed the proepicardium (PE). PE cells migrate onto the myocardium to give rise to the epicardium. Impaired epicardial development has been associated with defects in valve development, cardiomyocyte proliferation and alignment, cardiac conduction system maturation and adult heart regeneration. Zebrafish are an excellent model for studying cardiac development and regeneration; however, little is known about how the zebrafish epicardium forms.
We report that PE migration occurs through multiple mechanisms and that the zebrafish epicardium is composed of a heterogeneous population of cells. Heterogeneity is first observed within the PE and persists through epicardium formation. Using in vivo imaging, histology and confocal microscopy, we show that PE cells migrate through a cellular bridge that forms between the pericardial mesothelium and the heart. We also observed the formation of PE aggregates on the pericardial surface, which were released into the pericardial cavity. It was previously reported that heartbeat-induced pericardiac fluid advections are necessary for PE cluster formation and subsequent epicardium development. We manipulated heartbeat genetically and pharmacologically and found that PE clusters clearly form in the absence of heartbeat. However, when heartbeat was inhibited the PE failed to migrate to the myocardium and the epicardium did not form. We isolated and cultured hearts with only a few epicardial progenitor cells and found a complete epicardial layer formed. However, pharmacologically inhibiting contraction in culture prevented epicardium formation. Furthermore, we isolated control and silent heart (sih) morpholino (MO) injected hearts prior to epicardium formation (60 hpf) and co-cultured these hearts with “donor” hearts that had an epicardium forming (108 hpf). Epicardial cells from donor hearts migrated on to control but not sih MO injected hearts.
Epicardial cells stem from a heterogeneous population of progenitors, suggesting that the progenitors in the PE have distinct identities. PE cells attach to the heart via a cellular bridge and free-floating cell clusters. Pericardiac fluid advections are not necessary for the development of the PE cluster, however heartbeat is required for epicardium formation. Epicardium formation can occur in culture without normal hydrodynamic and hemodynamic forces, but not without contraction.
The proepicardium is a cluster of cardiac progenitor cells that develops adjacent to the heart and migrates onto the heart to form the outermost layer, the epicardium [1, 2]. After the epicardium has formed, a subset of epicardial cells undergo epithelial-to-mesenchymal transitioning (EMT) and contribute to the development and maturation of many cardiac cell types, such as cardiac fibroblasts, endothelial cells, and vascular smooth muscle cells [3–7]. Disruptions in epicardial development are associated with defects in endocardial valve development, heart looping, cardiomyocyte proliferation and alignment, development of the coronary vasculature, cardiac conduction system maturation, and cardiac regeneration (reviewed in [8–12]).
PE development and migration have primarily been studied using the chick and mouse models. In chick, the PE forms asymmetrically on the right sinus horn and migrates to the dorsal surface of the ventricular myocardium via an extracellular matrix bridge, which connects the PE and myocardium [13–16]. Epicardial coverage proceeds over the myocardium in a sheet-like manner . Studies of Xenopus and the axolotl find that PE cell migration in amphibians also occurs via a bridge [18, 19]. However, it has been debated whether murine PE cell migration occurs through a mechanism involving direct contact between the PE and myocardium or, alternatively, through free-floating PE-cell aggregates. In the latter model, aggregates are released into the pericardial space and attach at various sites on the myocardium creating “epicardial islands” . Epicardial islands spread out and are ultimately stitched together to form an epicardial sheet covering the myocardium. Work by Rogers et al.  argues that the mouse epicardium forms, as in the in chick, through villi that protrude from the mouse PE and contact the myocardium directly. Movement of the beating heart transfers the PE villi onto the myocardium. In the same study, PE cell aggregates were also observed, indicating more than one mode of transfer occurs during epicardial development, which was also suggested in an earlier study by Komiyama et al. .
Zebrafish form a PE on the pericardial wall, adjacent to the atrioventricular (AV) junction [1, 22]. However, in zebrafish, how epicardial progenitor cells migrate onto the zebrafish myocardium remains poorly understood. In this work, we show that PE cells migrate to the heart using both direct contact and the release of free-floating aggregates. We find that a PE cluster located at the AV junction forms a cellular bridge between the pericardial mesothelium and the heart. Additional PE clusters form near the venous pole, are released into the pericardial space, and subsequently attach to the heart.
Although it has previously been reported that pericardial fluid forces acting on the mesothelium are required to induce the formation of PE clusters and direct epicardial morphogenesis , we found that PE clusters clearly form without a heartbeat. However, without a heartbeat, the PE cells failed to migrate onto and across the heart. To determine if specific pericardial fluid forces or hemodynamic forces were necessary for epicardium formation, we isolated hearts just as the first epicardial progenitors had attached, and grew these hearts in culture. Starting from only a few pioneer progenitors, a complete epicardial layer formed in vitro, thus indicating the pericardial fluid forces and hemodynamic forces are not necessary for directing epicardial development.
To examine if heartbeat was need for epicardial cell migration, we developed an in vitro epicardial cell migration assay to test whether epicardial cells can migrate from a donor heart onto a younger recipient heart, which had not yet formed an epicardium. Indeed, epicardial cells were able to migrate onto control recipient hearts, but not onto recipient hearts in which heartbeat was inhibited. Together our results show the critical importance of myocardial contraction for PE migration and epicardium formation.
The murine PE and epicardium are composed of heterogeneous populations of cells that have divergent roles during heart development . As shown in Figure 1C, we observed evidence for similar heterogeneity during zebrafish epicardial development. A close-up confocal z-stack and orthogonal slice of the PE at 72 hpf (Figure 1D) shows that not all cells within the PE cluster expressed tcf21. An optical slice generated along the line indicated by the “X”, orthogonal to the plane of the image, shows a cross-section of the PE with both tcf21+ and tcf21- cells. These results indicate that differences between PE cells exist prior to reaching the myocardium.
Heartbeat is necessary for important steps in heart development, including valve cushion formation . We manipulated heartbeat genetically by using sih MOs to completely and specifically blocking heart contractions to determine whether a heartbeat is needed for epicardium development. As an alternative approach, we pharmacologically inhibited contractions using BDM . BDM treatment dramatically reduced heart contractility when present in the water, but once removed, heart contractions resumed.
To confirm that the cells seen adjacent to the heart in the brightfield images were in fact specified PE cells, we injected sih MOs into the tcf21:DsRed2 reporter line (Figure 5D and E). We examined control and sih MO-injected larvae at 72 hpf and could clearly identify tcf21+ PE clusters in 15 out of 15 larvae in each group (Figure 5D and E). Together, our results demonstrate that heartbeat is not necessary for PE specification or cluster formation.
Although the PE developed in the absence of heartbeat when we examined the sih MO-injected tcf21:DsRed2 larvae at 96 hpf we found that the epicardium had not formed (Figure 5F and G). Consistent with our finding at 72 hpf, we found PE-like clusters of tcf21:DsRed2+ cells at the venous pole at 96 hpf (arrowhead in Figure 5G).
In this work we describe zebrafish PE migration and epicardium development in detail. As in other vertebrates, a PE forms, the progenitors migrate to the heart, and cells envelop the myocardium. We find that in zebrafish that PE migration occurs through both a cellular bridge and the release of PE aggregates into the pericardial cavity. Similar to murine epicardium development, we observe heterogeneous gene expression in the PE  and continue to observe heterogeneous gene expression in the developing larval and juvenile epicardium. We find the epicardium first forms over the ventricle and then begins to cover the atrium. Formation of the epicardium begins after the basic organization of the heart has been established and the heart is beating. We found that, as with valve development , heartbeat was essential for epicardium formation. The PE formed in the absence of the heartbeat, but epicardial progenitors failed to migrate to the myocardium, and the epicardium did not form.
In the chick embryo, the PE migrates towards the myocardium by an extracellular matrix bridge . In the mouse there is evidence for PE cell migration through both direct contact as well as free-floating PE cell aggregates . In our experiments, the zebrafish PE protrudes towards the heart and makes direct contact with the myocardium forming a cellular bridge. Our finding with zebrafish suggests that teleosts develop a similar proepicardial bridge as reported in avians , amphibians [18, 19, 29], lamprey, dogfish , sturgeon  and the rat . We also observed the release of PE cell aggregates during epicardium development, which indicates that multiple modes of PE migration occur during zebrafish epicardium development. Dual mechanisms have also been observed in mouse , axolotls , and dogfish .
In mice, the lineages of subsets of cells originating in the PE have been traced during development and found to contribute to different cell types in the heart including cardiac fibroblasts, smooth muscle, pericytes, subepicardial EPDCs, and perivascular cells such as the smooth muscle of the outflow tract. [16, 24, 33–35]. However, whether epicardial cells contribute to formation of cardiac muscle has been controversial [25, 36–40]. The most significant evidence that the epicardium does not contribute to muscle comes from tracing epicardial cells with the tcf21 marker. Our work clearly shows that not all epicardial cells are tcf21+. We observed cellular heterogeneity beginning within the PE and persisting through epicardium development. The tcf21- cells observed in the epicardium potentially descended from the tcf21- cells originally detected in the PE. Lineage tracing of the tcf21- cells, to complement what we know about the tcf21+ cells in the epicardium, will be needed to fully understand how epicardial progenitors contribute to zebrafish heart structures.
Inhibiting heartbeat prevented the epicardium from forming, however it did not prevent the specification of PE cells or the development of PE clusters. If the heartbeat was stopped later in development, established epicardial cells remained on the heart surface but did not migrate and expand further across the heart.
Our findings are consistent with the report that PE cells are specified in sih morphants . However, our work contradicts the conclusion from Peralta et al.  that PE cluster formation requires heartbeat induced pericardial fluid forces. We note that we used different techniques and markers, and scored for PE cluster formation later in development, which may explain the differences between the two reports. We initially used brightfield microscopy as well as video microscopy to identify PE cluster formation in live embryos suspended in methylcellulose. PE formation was clearly evident in samples both with and without heartbeat. After observing PE cluster formation in these samples, we injected sih MO into a known PE marker line, tcf21:DsRed2 and confirmed that the observed clusters were composed of specified PE cells. Together, our findings demonstrate that heartbeat is not necessary for PE cluster formation. However, our results are in agreement with Peralta et al.  that the heartbeat is needed for epicardium formation.
Although it has been proposed that heartbeat has an indirect impact on the epicardium through hydrodynamic forces generated within the pericardium, our results with hearts cultured in vitro lead us to doubt this. Clearly the heartbeat itself is needed, but none of the specific flow patterns of blood or pericardial fluid are present in vitro, nor are factors found in the pericardial space that might promote adhesion and proliferation. Yet the epicardial layer formed nonetheless: as long as the beat continued. We were also able to show that the transfer of epicardial cells from a donor to a recipient heart was heartbeat dependent. Since epicardial cell migration does not occur in the absence of heartbeat, we speculate that the cardiomyocyte surface may be altered due to the loss of regular beating.
In addition, inhibiting heartbeat may alter the expression of signaling molecules and/or their receptors, which may be needed for PE migration. In chick, bone morphogenetic proteins (bmp) signals emanating from the myocardium direct PE protrusion and attachment . Planar cell polarity also known to play an important role in PE cell migration and it may be that signals from the myocardium induce polarity in migrating PE cells. We consistently observed that the initially spherical PE cells acquired an oblong, planar shape as they encountered the myocardial surface. When contractility was blocked in vivo, PE cells adjacent to the myocardium remained rounded. In mice, loss of the planar cell polarity protein Par3 results in failure of PE cell migration . In zebrafish, Serluca showed that knockdown of the cell polarity genes heart and soul (has/aPKC/PRKCl) and nagie oko (nok) resulted in defects in PE morphogenesis .
We show that PE migration occurs through multiple modes, including a cellular bridge that forms between the pericardial wall and the heart near the AV junction. Consistent with Peralta et al. , we observe the development of multiple PEs and the release of progenitor cell aggregates into the pericardial space. We find that the epicardium first envelops the ventricle before moving across the atrium. The presence of a heartbeat is not required for PE formation, but it is necessary for expansion of the epicardium across the myocardium. The formation of PE clusters in the absence of a heartbeat and the finding that epicardial growth and expansion occurs across hearts in vitro, rule out pericardiac fluid advections as a critical requirement for epicardium development. Heart contraction, however, was required for epicardial formation in vitro and in vivo.
Adult zebrafish lines were maintained and zebrafish embryos were reared and housed according to procedures described by . The AB wild type line was used unless otherwise indicated. Transgenic lines used: pard3:EGFP [ET(krt4:EGFP) sqet27 ] , tcf21:DsRed2 [Tg(tcf21:DsRed2) pd37 ] , and cmlc2:EGFP [Tg(cmlc2:EGFP) f1 ] . All procedures involving animals were approved by the Animal Care and Use Committee of the University of Wisconsin-Madison, and adhered to the National Institutes of Health's “Guide for the Care and Use of Laboratory Animals”.
Antibody staining was performed as previously described [46, 47]. Primary antibodies were used at the following dilutions: mouse anti- activated leukocyte cell adhesion molecule (ALCAM/zn5; ZIRC) 1:50, rabbit anti-DsRed (AnaSpec, Fremont, CA) at 1:200 in phosphate buffered saline with 0.03% triton and 4% bovine serum albumin (PBT). Secondary anti mouse antibodies (Alexa 488, Alexa 568; Invitrogen) were used at 1:200 dilution in PBT. Embryos were mounted in Vectashield or Vectashield with Dapi (Vector Laboratories). Confocal images were collected on an Olympus Fluoview FV1000 microscope. Brightest point projections were made using Olympus Fluoview software and images were processed using Adobe Photoshop. Optical sections in z-series were collected at 0.52 μm intervals.
Live embryos were imaged as previously described . Briefly, fish were imaged at 72 hpf in 3% methylcellulose using a Nikon TE300 inverted microscope attached to a Princeton Instruments Micromax CCD camera. Videos were collected using MotionsScope software and analyzed using Metamorph software.
All morpholino oligonucleotides (MOs; Gene Tools, LLC) were used as previously reported [48, 49]. The TNNT2 MO (5’ CAT GTT TGC TCT GAT CTG ACA CGC A 3') was designed to block the translational start site of zebrafish cardiac troponin 2 (tnnt2; silent heart, sih). The standard Gene Tools Control MO (5’-CCT CTT ACC TCA GTT ACA ATT TAT A-3’) was used to control for non-specific responses. One-cell embryos were injected with 2 ng of MO. 2,3-Butanedione 2-monoxime (BDM, Sigma Aldrich) was used at a final concentration of 10 mM embryo water.
Hearts from cmlc2:GFP and tcf21:DsRed2 larvae were isolated at 60 hpf and 108 hpf, respectively, as previously described . Hearts were mixed in a 1:2 ratio of tcf21:DsRed2 to cmlc2:GFP hearts and placed in culture plates with matrigel (thin gel coating method, BD Biosciences). Heart cultures were incubated at 28°C with 5% CO2 in cell culture media containing Leibovitz’s L-15 (Fisher) with 10% fetal bovine serum (Sigma) and 4x penicillin/streptomycin (Invitrogen). Cultures were monitored daily and media was refreshed every other day. Images were obtained with an Olympus DP72 camera mounted on an Olympus SZX16 epifluorescence stereo microscope with cellSens software. On Day 7 in culture, heart clusters were removed and fixed in 4% paraformaldehyde and immediately prepped for immunohistochemistry.
We thank Dorothy Nesbit for execellent oversight and care of the zebrafish facility, and Drs. Vladamir Korzh, Kenneth Poss, and C. Geoffrey Burns for providing the pard3:EGFP, tcf21:DsRed2, and cmlc2:GFP transgenic zebrafish lines respectively. We thank Maggie Shuda, Erica Christensen, John K. Samson and Paul Forslyn for technical support.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.