1. Developmental Biology
  2. Stem Cells and Regenerative Medicine
Download icon

Late developing cardiac lymphatic vasculature supports adult zebrafish heart function and regeneration

  1. Michael RM Harrison  Is a corresponding author
  2. Xidi Feng
  3. Guqin Mo
  4. Antonio Aguayo
  5. Jessi Villafuerte
  6. Tyler Yoshida
  7. Caroline A Pearson
  8. Stefan Schulte-Merker
  9. Ching-Ling Lien  Is a corresponding author
  1. The Saban Research Institute of Children's Hospital Los Angeles, United States
  2. California State University of San Bernardino, United States
  3. University of Southern California, United States
  4. David Geffen School of Medicine, University of California, Los Angeles, United States
  5. University of Münster, Germany
  6. Keck School of Medicine, University of Southern California, United States
Research Article
  • Cited 11
  • Views 2,846
  • Annotations
Cite this article as: eLife 2019;8:e42762 doi: 10.7554/eLife.42762

Abstract

The cardiac lymphatic vascular system and its potentially critical functions in heart patients have been largely underappreciated, in part due to a lack of experimentally accessible systems. We here demonstrate that cardiac lymphatic vessels develop in young adult zebrafish, using coronary arteries to guide their expansion down the ventricle. Mechanistically, we show that in cxcr4a mutants with defective coronary artery development, cardiac lymphatic vessels fail to expand onto the ventricle. In regenerating adult zebrafish hearts the lymphatic vasculature undergoes extensive lymphangiogenesis in response to a cryoinjury. A significant defect in reducing the scar size after cryoinjury is observed in zebrafish with impaired Vegfc/Vegfr3 signaling that fail to develop intact cardiac lymphatic vessels. These results suggest that the cardiac lymphatic system can influence the regenerative potential of the myocardium.

eLife digest

Human hearts have coronary vessels that supply oxygen and essential nutrients to the heart. When this supply is interrupted, a heart attack can occur. After a heart attack, scar tissue forms that impairs the heart’s ability to pump blood around the body. The heart also has lymphatic vessels that drain excess fluid and remove waste products and damaged cells from the heart. Less is known about the lymphatic vessels and their role in heart disease.

Unlike human hearts, which scar easily, the zebrafish heart can regenerate after injury. Because of this, scientists often study zebrafish to try to find ways to improve healing of heart injuries in humans. However, it is not yet known whether lymphatic vessels contribute to regeneration of zebrafish hearts.

Now, Harrison et al. show that, in the zebrafish heart, lymphatic vessels develop after the coronary arteries. In fact, the coronary arteries provide a scaffold that the lymphatic vessels grow along.

When the zebrafish are genetically modified so that they lack coronary arteries, the lymphatic vessels fail to grow. Further experiments showed that, when the heart was injured by briefly freezing part of it, extra lymphatic vessels grew, but this did not happen when a part of the heart was removed via surgery. This may be because the cold-induced injury causes inflammation, which can trigger the growth of lymphatic vessels. The lymphatic vessels then help battle inflammation, allowing regeneration to proceed. Using genetically engineered zebrafish, Harrison et al. were then able to turn the genes that control lymphatic vessel growth on and off. They showed that zebrafish lacking lymphatic vessels in the heart are less efficient at regenerating heart tissue and develop more scar tissue after injury. This result is supported by the findings of a separate study conducted by Gancz et al.

The results suggest that stimulating the growth of lymphatic vessels or enhancing their activity in the injured heart may aid recovery. More studies may help scientists understand exactly how lymphatic vessels aid regeneration in zebrafish and whether promoting lymphatic vessel growth or activity may aid heart attack recovery in humans.

Introduction

The cardiac vascular system is comprised of blood and lymphatic vessels. Arteries and connected capillaries transport oxygenated, nutrient-rich blood to the myocardium. Cardiac veins drain the blood back into the systemic circulation, and excess fluid and nutrients from the blood are hydrostatically released into tissue during this process. This interstitial fluid, immune cells, debris and waste products are then drained via the cardiac lymphatic vessels. The heart critically requires this continuous cycle through the myocardium for optimal function and interruption has pathological results including myocardial infarction (MI) or cardiac lymphedema (Aspelund et al., 2016; Karunamuni, 2013). Not only are the cardiac vessel systems required for heart function, but they are also likely to have active roles in disease resolution, at least helping to provide a permissive environment for regeneration (Das et al., 2019; Klotz et al., 2015; Marín-Juez et al., 2016). This has yet to be successfully leveraged clinically in part due to the lack of effective therapeutic strategies (Taimeh et al., 2013). A more detailed understanding of these systems, their formation, regeneration and promotion of positive disease environments are critically required.

Zebrafish represent a simple coronary vessel system, many features of which are conserved in the mammalian system. However, unlike mammals, the coronary structure in zebrafish starts to develop relatively late at juvenile stages when endocardial cells that give rise to coronary vessels sprout from the atrioventricular (AV) canal region of the heart to vascularize the ventricle (Harrison et al., 2015). The existence and development of a cardiac lymphatic system to complement this coronary system has yet to be described in zebrafish.

Extensive work using zebrafish embryonic models has yielded much of our understanding of embryonic lymphatic development (Hogan and Schulte-Merker, 2017). During trunk angiogenesis, lateral plate mesoderm-derived angioblasts from the ventral side of the posterior cardinal vein (PCV) migrate to the dorsal myoseptum, lose their connection with the PCV and become lymphangioblasts known as parachordal cells (Hogan et al., 2009a; Nicenboim et al., 2015). These cells then migrate from the midline along arteries under the guidance of Cxcl12b-Cxcr4a/b chemokine signaling to form lymphatic vessels, including the thoracic duct between the dorsal aorta and PCV (Cha et al., 2012). This thoracic duct is connected to a continuous dorsal longitudinal lymphatic vessel via intersegmental lymphatic vessels, which align along the same intersegmental arteries that provided guidance during their formation (Bussmann et al., 2010). This process of trunk lymphangiogenesis is completely blocked in embryos with mutations in genes encoding the secreted vascular endothelial growth factor C (vegfc) or its receptor flt4/vegfr3 (Hogan et al., 2009b; Villefranc et al., 2013).

Vegfc is similarly required for the sprouting of lymphatic endothelial cells in mice (Karkkainen et al., 2004) and, as in zebrafish embryos, these lymphatic endothelial cells are derived primarily from venous endothelium (Hägerling et al., 2013; Yang et al., 2012) with some contribution from non-venous sources to specific lymphatic vasculature beds (Ulvmar and Mäkinen, 2016; Eng et al., 2019). In the case of mouse cardiac lymphatic system, there is an additional contribution to the lymphatic vasculature from yolk sac endothelial cells (Klotz et al., 2015). Together with venous-derived lymphatic endothelial cells, they invade the embryonic heart at embryonic day (E) 12.5 from an extra-cardiac source to populate its base via the sinus venosus and outflow tract (Flaht et al., 2012; Klotz et al., 2015). This lymphatic endothelium then continues to sprout over the surface of the ventricle along Emcn-expressing cardiac veins. Vessels expressing lymphatic endothelial marker genes Lyve-1, Prox1 and Flt4 are identifiable specifically along the side of the cardiac veins by birth and continue expanding over the ventricle until postnatal (P) day 15 (Klotz et al., 2015). After a myocardial injury, the cardiac lymphatic vasculature is reactivated. Despite this expansion of the lymphatic network after myocardial infarction, the mammalian heart still scars rather than regenerates functional tissue. However, a reduction in this scarring can be observed when Vegfc is induced to further enhance lymphangiogenesis in injured adult mice (Klotz et al., 2015).

The cardiac lymphatic system is thought to regulate fluid homeostasis and provide immune system surveillance and clearance which may have important implications for cardiac tissue recovery after insult (Vieira et al., 2018). In contrast to the mammals, the zebrafish has retained a remarkable capacity to regenerate cardiac tissue after tissue damage or resection (Gamba et al., 2014; Poss et al., 2002). After resection or moderate cryoinjury to the apex or ventral wall of the ventricle, fully functional cardiac tissue is regenerated within 30–90 days and there is little or no detectable scar in the majority of injured zebrafish (Chablais et al., 2011; González-Rosa and Mercader, 2012; Poss et al., 2002). The regenerated tissue is also vascularized by blood vessels by this time and this vascularization supports the function of this regenerated tissue as well as the repair process itself (Marín-Juez et al., 2016).

Historical anatomical descriptions of lymphatic vessels surrounding the fish heart (Hewson and Hunter, 1769) suggest that these vessels constitute an ancient feature of jawed vertebrates, but the functional overlap of this vasculature with that of mammalian lymphatic vessels has been questioned (Vogel and Claviez, 1981). As such the existence and function of the cardiac lymphatic vasculature in zebrafish and its potential implication for normal tissue homeostasis as well as natural regeneration remains an open question of significant clinical interest.

We here characterize of the development of zebrafish cardiac lymphatic vasculature, describe its functions and analyze lymphangiogenesis during heart regeneration. Our results suggest that cardiac lymphatic vessels should be a promising therapeutic target for the treatment of heart disease and may help modulate a pro-regenerative immune response.

Results

Lymphatic vessels develop along coronary arteries of adult zebrafish

To begin to investigate the subtype identity of the cardiac lymphatic vasculature in the adult zebrafish, we used the flt4:mCitrine transgenic line that is expressed in both venous and lymphatic endothelial cells (Gordon et al., 2013; van Impel et al., 2014). flt4:mCitrine endothelial expression is detected at either systemic pole of the heart in the sinus venosus (SV) and on the surface of the bulbus arteriosus (BA) (Figure 1A). The population of endothelial cells on the outside of the BA was of particular interest given that they cover the entire BA, but do not extend down on the ventricle until adult stages (after 90dpf; Figure 1A–C). In young adult fish, flt4:mCitrine-positive cells bud from this BA population and appear to migrate down on the ventricle to form a vessel that extends to the apex of the ventricle (Figure 1B–D). To determine if this vasculature is venous or lymphatic we used the prox1:Gal4-UAS:RFP transgenic line that is not expressed in venous endothelium, but in lymphatic endothelial cells (LECs) and neurons (van Impel et al., 2014). Expression in the BA population suggests that this is lymphatic endothelium unlike the flt4:mCitrine-positive, prox1:Gal4-UAS:RFP-negative cells of the sinus venous, (Figure 1E,F, Figure 1—figure supplement 1A, B). The lack of markers exclusively specific to lymphatic endothelial cells has been one of the precluding factors for their visualization and study (Jung et al., 2017). We confirmed the lymphatic identity of this endothelium with four additional reporter lines known to be expressed in lymphatic endothelial cells: lyve1b:GFP, lyve1b:DsRed (Okuda et al., 2012), stab1:YFP (Hogan et al., 2009a) and mrc1a:GFP (Jung et al., 2017) (Figure 1—figure supplement 1C-F). As such the cardiac lymphatic system is identifiable in adult zebrafish and appears to extend from a BA population in older zebrafish. We next addressed the stage at which the BA population develops relative to coronary vessel development at juvenile stages.

Figure 1 with 1 supplement see all
Prolonged development of the cardiac lymphatic system in juvenile and young adult zebrafish.

Whole-mount confocal imaging of non-cleared (A-D, F-I) and cleared (E,J) adult (A-F) and juvenile (G-J) transgenic zebrafish hearts expressing the pan-endothelial marker fli1a:GFP (Green), venous/lymphatic endothelial marker flt4:mCitrine (blue: A-F) and lymphatic endothelial marker prox1:Gal4-UAS:RFP (red: E-J). (A) Through juvenile and young adult stages, flt4:mCitrine-expressing endothelial cells are restricted to the surface of the BA. (B) At 26 mm standard body length, sprouts from this BA population are first observed on the ventricle surface (arrowhead and insert). (C) These sprouts then extend down the ventricle and start to form a cardiac lymphatic vessel that eventually extends the entire length of the ventricle to the apex (D). The cardiac lymphatic vessel (D, inset, blue arrowhead) bridges the blood vessel (D, inset, green arrowhead) it accompanies. (E) Distinct from the venous endothelium of the SV, the BA population and the sprouts extending from it also express prox1:Gal4-UAS:RFP. (F) Both flt4:mCitrine and prox1:Gal4-UAS:RFP are maintained in the formed cardiac lymphatic vessel (arrowhead). (G) Larval and early juvenile zebrafish lack the prox1:Gal4-UAS:RFP expressing lymphatic endothelium on the BA. (H) This population is observed in juvenile zebrafish prior to the onset of and during (I) coronary vessel development. (J) In early adult stages, alignment of single cell sprouts off the lymphatic endothelium (red arrowhead) with blood vasculature (green arrowhead) is observed within the cleft between the BA and ventricle. Scale bars 200 μm (A-J) and 50 μm (insets, B, D, F), n ≥ 3 (A-J).

The formation of a cardiac lymphatic vessel on the ventricle of adult zebrafish from a pre-existing population on the BA occurs very late and after the formation of the coronary vasculature in juvenile zebrafish. In rodent models, cardiac lymphatic vasculature on the ventricle is observed in embryos by E14.5 (Klotz et al., 2015). However in zebrafish we found that the lymphatic endothelium is first established on the BA post-embryonically, before the development of the coronary vasculature (Figure 1G–I). The lymphatic endothelium extends to the base of the BA into the cleft between the BA and the ventricle (Figure 1J). Within this cleft the lymphatic endothelial cells sprout along coronary vessels to extend back up out on the ventricle side (Figure 1J). The sprouting lymphatic cells extend along the coronary arteries and track them as they progress down the ventricle towards the apex of the heart (Figure 2A–F). As the cells extend towards the apex they expand and flatten around the artery to form a vessel. This vessel is in close proximity to the arterial endothelium that expresses higher levels of kdrl:mTurquoise at this stage and shows more specific expression of flt1enh:tdTomato, dll4:GFP and cxcr4a:mCitrine (Figure 2D–G). This association between an artery and lymphatic vessel is also observed as the lymphatic vasculature extends along the aorta and branches off to the brachial aches. Here the lymphatic vessels associate with the brachial arteries and the arterial side of the gill filaments (Figure 2—figure supplement 1A–D).

Figure 2 with 1 supplement see all
Extension of the cardiac lymphatic system along arteries in adult zebrafish.

Whole-mount confocal imaging of adult transgenic zebrafish hearts expressing the pan-endothelial fli1a:GFP (green: A-E, G), venous/lymphatic endothelial marker flt4:mCitrine (blue: A-D, red: F), arterial endothelial cell marker flt1enh:tdTomato (red: A-D, G), lymphatic endothelial marker prox1:Gal4-UAS:RFP (red: E), arterially enriched blood endothelial marker kdrl:mTurquoise (blue: E) and arterial marker dll4:GFP (F). (A-C) flt4:mCitrine-positive sprouts migrate along flt1enh:tdTomato expressing arteries not flt1enh:tdTomato-negative veins (white arrowheads). The extension appears somewhat dynamic and fluid, with gaps or dissociations observed during extension and formation (blue arrowheads). This association continues as the cardiac lymphatic vessel forms on the ventricle (D). The formed vessel endothelium is kdrl:mTurquoise-negative and is in close proximity to high-kdrl:mTurquoise (E), dll4:GFP (F) expressing arteries. flt1enh:tdTomato expressing arteries also express high levels of cxcr4a:mCitrene in young adult zebrafish (G). Scale bars 200 μm (A-G) and 50 μm (insets A, C, E, F and G), n ≥ 3 (A-G).

Development of a cardiac lymphatic system is incomplete in the absence of a coronary artery scaffold

In both zebrafish and mammals, there is a sequential development of the coronary vasculature and the cardiac lymphatic system suggesting that the coronary vessels provide a scaffold or guidance for the lymphatic system. To test this hypothesis, we analyzed the formation of cardiac lymphatic vessels in cxcr4a mutant zebrafish that lack coronary arteries. Unlike intersegmental lymphatic vessels, cxcr4a:mCitrine expression in the adult heart is restricted to the coronary arteries and is absent in migrating LECs and lymphatic vessels that form along them (Figure 3A–C).

Figure 3 with 2 supplements see all
Cardiac lymphatic development requires a coronary artery scaffold.

Whole-mount confocal imaging of adult transgenic zebrafish hearts expressing arterial marker cxcr4a:mCitrine (green: A and B), and lymphatic endothelial markers lyve1:RFP (Red: A and C) and flt4:mCitrine (Blue: D-F), the pan-endothelial marker fli1a:GFP (green: D-F), and the arterially-enriched pan-endothelial marker kdrl:mTurquoise (red: D-F). (A-C, n = 8) cxcr4a:mCitrine expression is restricted to the coronary arteries and not detectable in lyve1:RFP-positive LECs. (D) Cardiac lymphatic vessels in control (cxcr4a+/+, n = 8) zebrafish. (E and F) In cxcr4a mutants (cxcr4a-/-, n = 12) that have abnormal and discontinuous coronary vasculature the cardiac lymphatic extension is stunted. (G) Graph showing significant reductions in vessel density, kdrl expression in blood vessels and flt4 expression on the surface of the cxcr4a-/- ventricle. There is comparable levels of fli1a-expressing cells on the surface of the ventricle between larger/older cxcr4a-/- and control zebrafish, but loss of the coronary structure. (H) Schematic representation of the regulation of cardiac lymphatic developmental guidance by coronary arteries. Scale bars 200 μm.

Zebrafish with mutations in cxcr4a present with disrupted coronary vessel development and largely lack formed coronary vessels in young adult fish (Figure 3—figure supplement 1B). These mutants fail to regenerate after ventricular injury (Harrison et al., 2015). In older adult cxcr4a mutant zebrafish there is some formation of lumenized vessels, but the coronary vasculature structure remains highly disorganized over the ventricle. Often this coronary vasculature presents as a conglomerate of enlarged or dense interconnected by sparse intermediary non-luminized vessels or isolated cells (Figure 3E and F, Figure 3—figure supplement 1A–F). The density of this vasculature increases with time, as observed in wildtype transgenic zebrafish, however it lacks any identifiable coronary arteries (Figure 3—figure supplement 1G–H, Figure 3—figure supplement 2C-M). Analysis of lymphatic markers in cxcr4a mutant zebrafish demonstrated that cardiac lymphatic vessels do not extend down the ventricle in the absence of coronary arteries, but that the pre-forming BA population is unaffected (Figure 3D–H, Figure 3—figure supplement 1A–H, Figure 3—figure supplement 2A-M). There is a significant correlation between the extent of the coronary arterial tree over the ventricle and coverage of the ventricle by the cardiac lymphatic vessels (Figure 3G, Figure 3—figure supplement 2M), suggesting that the later uses the coronary vessels to migrate down the ventricle and that coronary artery derived signals may promote this process during development.

The cardiac lymphatic system responds to injury of the ventricle

The zebrafish ventricle shows an amazing capacity to regenerate following insult or injury and previous work suggested that vegfc expression is elevated during this regeneration process (Lien et al., 2006). In adult hearts vegfc expression is observed on the BA in its junction with the ventricle base, but little detectable signal is observed on the ventricle itself (Figure 4A, Figure 4—figure supplement 1A). After amputation elevated vegfc levels are detectable at the wound site at 3dpa and 7dpa, but expression is reduced by 14dpa and absent thereafter (Figure 4B, Figure 4—figure supplement 1B-G). To test the effect of having remaining necrotic tissue and a pre-existing extracellular matrix after injury, we used the more severe cyroinjury model. Cryoinjury also results in vegfc expression at 3dpc also, however subsequent expression appears to be more prolonged and remains elevated after 42dpc (Figure 4C and D, Figure 4—figure supplement 1H-L and G). Following resection of the ventricle apex there is little expansion of the cardiac lymphatic vasculature, with only some hearts showing extension of the lymphatic vessels into the wound site after 14dpa (Figure 4E–H). If present these lymphatic vessels appear to largely be extensions of the existing lymphatic vessel into the apex and do not appear to be enlarged or expanded (Figure 4E–H,Q and R; 4/7 hearts at 60dpa show wound site lymphatic vessels). In contrast, after cryoinjury migration of LECs into the wound site is observed at 14dpc, but also enlargement and expansion of the vessel on the ventricle (Figure 4I–P,S–V). Particularly at later stages, a highly branched network of cardiac lymphatic vasculature is observed within the wound site (Figure 4L–P,S–V). These regenerated lymphatic vessels express flt4:mCitrine, prox1:Gal4-UAS:RFP, lyve1:RFP, mrc1a:eGFP (Figure 4L,N–P). Quantification of lymphatic vessel coverage within a set area of the ventricle shows significantly more lymphatic vessels in or around the wound site after cryoinjury when compared to the resection model (Figure 4U). In addition post-cryo-injured hearts have a more branched cardiac lymphatic structure over the ventricle (Figure 4V). This suggests that the presence of necrotic tissue and scar formation promotes the formation of more lymphatic vessels after cardiac injury. We next wanted to address if this was related to the function of the cardiac lymphatic system and its potential utilization in a disease setting.

Figure 4 with 1 supplement see all
Lymphangiogenesis in response to injury of the zebrafish ventricle.

Confocal images of RNAScope section in situ hybridization to detect vegfc (white: A-D). Whole-mount confocal imaging of adult transgenic zebrafish hearts expressing the pan-endothelial marker fli1a:GFP (Green: E-O, Q-T) and lymphatic endothelial markers flt4:mCitrine (blue: E-O, Q-T), prox1:Gal4-UAS:RFP (red: H and N), lyve1:RFP (Red: O), mrc1a:eGFP (Green: P) and endothelial marker kdrl:mCherry (Red: P). (E-H) Response to amputation of flt4-expressing lymphatic vasculature during regeneration at 7 (E), 14 (F), 35 (G) and 69dpa (H). Lymphatic vasculature is observed to extent into the wound site at 35dpa to connect to the vessel on the BA (blue arrowhead) with some disconnections observed (white arrowhead). Response of flt4-expressing lymphatic vasculature to cryoinjury at 7 (I), 14 (J), 28 (K), 46 (L), 54 (M) and 60dpc (N). Expression prolife of regenerated lymphatic vasculature shows expression of prox1:Gal4-UAS:RFP (red: H and N), lyve1:RFP (Red: O) and mrc1a:eGFP (Green: P). (Q and S) Lymphatic vessels (blue arrowheads) extend to the apex in uninjured zebrafish. After amputation, there is little lymphangiogenesis and most of the regenerated vasculature lacks lymphatic marker expression with some minor extension into the wound site (R, blue arrowhead). Extensive lymphangiogenesis into and around the wound site (60dpc, T, blue arrowheads) is stimulated following damage to the myocardium with cryoinjury. (U) Quantification of the increase in area of flt4:mCitrine expression within a set 600 μm2 region centered on the wound site of sibling zebrafish subject to severe cryoinjury (n = 9) in comparison to that resulting from amputation (n = 7; *p=0.0132, unpaired t-test) or the uninjured apex (n = 4; *p=0.0407, unpaired t-test). (V) Quantification of increased branching of lymphatic vessels on the ventricle in response to cryoinjury (n = 9) in comparison to uninjured hearts uninjured hearts (n = 6; *p=0.0293, unpaired t-test) and those subjected to amputation (n = 7; *p=0.0140, unpaired t-test). Scale bars 200 μm (A-D) and 50 μm (insets).

Cardiac lymphatic vasculature functionally supports the heart during regeneration and homeostasis

The lymphatic system in mammals is known to maintain fluid homeostasis and modulate immune surveillance and clearance (Alitalo, 2011). Given these functions it has long been considered an integral apparatus in the maintenance of heart function and postulated to encourage a regenerative response to injury (Aspelund et al., 2016; Karunamuni, 2013; Klotz et al., 2015; Vieira et al., 2018). In order to ascertain the function of the zebrafish cardiac lymphatic vasculature and in light of the discontinuous nature of LECs sometimes observed in these vessels (Figures 2 and 4), we established an intra-myocardial injection assay of microspheres (MS) and quantum dots (Qdots). Within 1 hr both were found at the injection site (Figure 5A), however the smaller Qdots (<10 nm diameter) were more dispersed from the injection site as well as being concentrated within the lymphatic lumen (Figure 5A–C, Figure 5—video 1). This demonstrates that, despite the discontinuations in lyve1:RFP marker expression, the ventricular cardiac lymphatic vessel forms a blunt ended contiguous tube alongside the blood vessels. In addition these vessels are pervious to the uptake of fluid and small, but not large, debris as the larger microspheres (200 nm diameter) remained at the injection site, but the smaller Qdots become dispersed within the interstitium and are taken up into the lymphatic vessels.

Figure 5 with 6 supplements see all
Function of the cardiac lymphatic vasculature in regeneration.

Whole-mount confocal imaging of adult transgenic zebrafish hearts expressing lyve1:RFP (White: A and B) and lyve1:GFP (Green: D-G) to mark lymphatic vessels. Additional fluorescence signal from injected Firefli Fluorescent Blue-dyed Microspheres (MA; blue: A and B), Quantum dots (Qdot; magenta: A and B) and immuno-labeled neutrophils (Mpx; white: D-G) and Macrophage (IB4; red:D-G). (A) Intra-myocardial injection of 200 nm microspheres (cyan) and 10 nm Qdots (magenta) results in both being deposited at the injection site (magenta/cyan arrowheads). Qdots, but not MS are also highly concentrated within the lymphatic vessels within the lyve1:RFP positive endothelium (arrowheads, A; uninjected control, B). (C) Schematic representation of nanoparticle uptake by the cardiac lymphatic system. (D, n = 5) Uninjured zebrafish heart labeled with antibodies against GFP (lyve1:GFP) and Mpx and lectinB4 (IB4), showing relatively few Mpx-positive neutrophils or IB4-positive macrophage on the ventricle or BA, including within the lymphatic vessels on the BA (green arrowheads). (E, n = 4) 1 day post-cryo-injured (dpc) heart (mild injury) labeled at the same time shows a massive increase in Mpx-expressing neutrophils and more moderate increases IB4-positive macrophage including within the lymphatic vasculature (green arrowhead (GFP) white arrowhead (Mpx) red arrowhead (IB4). Mpx-positive cells are concentrated along the visibly darker blood vessels (inset, blue arrowhead) away from the wound site (orange arrowhead). (F, n = 5) 7dpc heart showing persistent elevation of Mpx-positive and IB4-positive cells on the ventricle and BA, where they are associated with lymphatic vessels. (G) High resolution imaging of lyve1:GFP -positive vessels (green arrowhead, outlined in blue) on the BA of a 7dpc heart IB4-positive macrophage (red arrowhead) within the vessel and Mpx-positive neutrophil (white arrowhead) on the surface of the vessel (H) Graph showing quantification of increase in IB4-positive cells within the BA lymphatic vessels of 1dpc (n = 4) and 7dpc (n = 5) zebrafish hearts compared to those of uninjured hearts (n = 5). *p<0.05; **p<0.01, unpaired t-test, Mean and SD. Scale bars 200 μm (A-G) and 50 μm (right panels, D-F).

In the uninjured heart, resident macrophages (labeled with IB4) were observed near the vasculature and there were very few neutrophils (Mpx-positive) present on the ventricle (Figure 5—figure supplement 1A). Given the strong lymphangiogenic response to cryoinjury (Figure 4), we decided to investigate the immunological role of the cardiac lymphatic system after a mild cryoinjury. Following injury, macrophages migrate to the wound site and are joined by a large number of neutrophil cells (Figure 5—figure supplement 1B). Clearance of these cells is known to be a key step in the process of regenerating heart tissue (Lai et al., 2017; Vieira et al., 2018). We find lymphatic vessels may provide a conduit for this clearance during zebrafish heart regeneration. After 1 and 7dpc (mild cryoinjury) mpx-positive neutrophil cells and debris became highly enriched within or on the cardiac lymphatic vessels on the BA, a location relatively distant from the more apex ventricular wound site (Figure 5D–H, Figure 5—figure supplement 1C–G). Despite weak expression on the ventricle of lyve1:GFP that precluded antibody detection above background, Mpx-positive cells were observed to align up along the optically dense blood vessels on the ventricle (Figure 5E insert) consistent with their association and uptake into lymphatic vessels. Immune cells were observed within the labeled BA cardiac lymphatic vessels after injury (Figure 5G, Figure 5—figure supplement 1E and F, Figure 5—videos 25). This suggests these cells can be taken up by the cardiac lymphatic vasculature at or near the wound site after injury and then transported away from the heart via the lymphatic vasculature on the aorta.

Vegfc promotes the formation of cardiac lymphatic vasculature after the initiation of coronary vessel development

Given Vegfc’s well established role in lymphatic vasculature development (Aspelund et al., 2016; Joukov et al., 1996; Kaipainen et al., 1995) and its potential role after injury (Figure 4) we decided to test if it was utilized during cardiac lymphatic system development. We inactivated the Vegfc ligands by induced expression of a soluble version of the Flt4/Vegfr3 receptor (serving as a ligand trap; sFlt4) in transgenic juvenile zebrafish during the formation of the ventricular cardiac vessels (Matsuoka et al., 2017). Unexpectedly, in heat shocked non-transgenic control zebrafish we observed an increase in lymphatic vessels on the heart relative to non-heat shocked controls (Figure 6A,B, Figure 6—figure supplement 1A,B,D). This increase in cardiac lymphatic vessels is not attributable to an increase in body size, in fact, heat shocked zebrafish showed a slight reduction in the overall length compared to non-heat shocked zebrafish at the same stage (Figure 6—figure supplement 1E). In stark contrast, the induced transgenic zebrafish that expressed sFlt4 from 35dpf had few or no LECs on the ventricle and no cardiac lymphatic vessels on the ventricle (Figure 6C and D, Figure 6—figure supplement 1C,D). This was also the case with sFlt4 induction during and after coronary vessel induction in older juvenile (from 71dpf) and adult zebrafish respectively (from 91dpf; Figure 6—figure supplement 1G–J). Induction of sFlt4 from 35, 71 or 92dpf did not overtly affect the development of the coronary vasculature, nor did we observe overt edema or malformation (Figure 6B,C, Figure 6—figure supplement 1B, C, F, G-N). Ventricle tissue appears to be normal with sFlt4 induction in juvenile zebrafish and does not significantly affect cardiomyocyte (CM) or non-CM numbers or the overall size of the ventricle (Figure 6—figure supplement 1O-S). This suggests that functional post-embryonic Vegfc signaling is specifically critical for the formation of this relatively late developing structure. We next used these zebrafish that lacked cardiac lymphatic vessels in the presence of coronary arteries to specifically test the effect of this system on cardiac regeneration.

Figure 6 with 5 supplements see all
Cardiac lymphatic development requires Vegfc signaling and helps reduce scar volume in response to cryoinjury.

Whole-mount confocal imaging of adult transgenic zebrafish hearts expressing the pan-endothelial fli1a:GFP (green: A), Arterially and endocardially enriched endothelial marker kdrl:GFP (green: B, C), lymphatic endothelial marker lyve1:RFP (Red: A-C). (A) Heart with normal cardiac lymphatic vessels from stage and size-matched control zebrafish (nonHS, n = 7). (B) Hearts from non-Transgenic zebrafish (nonTg, n = 3) that are exposed to heat shock with cardiac lymphatic vessels on the heart that extend further and cover more of the ventricle. (C) In contrast, hearts from hsp70l:sflt4 zebrafish (sFlt4, n = 7), heat shocked in the same way lack extension of the cardiac vasculature onto the ventricle. (D) Schematic representation of the experimental procedure and regulation of cardiac lymphatic development by inducing sFlt4. AFOG-histology images showing fibrin in red and collagen in blue in developmentally heat shocked (1 to 6mpf) non-transgenic (nonTg; E) and transgenic (hsp70l:sflt4; sFlt4; F and G) zebrafish 60 days following severe cryoinjury (60dpc, 8mpf, E and F) or 30 days after apex resection (30dpa, 7mpf, G). Scar volume was estimated and grouped based on severity for cryoinjured hearts (minimal scar,<0.005 mm3, E; large scar,>0.005 mm3, F) and amputated hearts (minimal scar, G). (H) Graph showing the distribution of tissue scaring between non-transgenic (nonTg, n = 5) and transgenic (hsp70l:sflt4) zebrafish at 60dpc (sFlt4, n = 5) and 30dpa (n = 7), *p<0.05, two-tailed Chi-Squared. Fluorescence signal from immuno-labeled neutrophils (Mpx; Red: I and J) and Macrophage (IB4; green: I and J) in and near the 14dpc woundsite (demarked orange; Post-injury sFlt4 induction, n = 6 per group). (J) Elevated level of Mpx-positive immune cells at the 14dpc woundsite of fish with post-injury induced sFlt4 (n = 6). (K) Separate quantification of Mpx and IB4-positive cells in and within 100 μm of the 14dpc woundsite (average of at least three images through woundsite, individually normalized to woundsite area, n = 6 per group). *p<0.05; **p<0.01; ***p<0.001, unpaired t-test, Mean and SD. Scale bars 200 μm.

Lack of cardiac lymphatic vasculature results in persistence of scar tissue following cryoinjury of the ventricle

To test if the zebrafish cardiac lymphatic system has the potential to influence the varying levels of fibrosis observed after cardiac injury (Lai et al., 2017), we used zebrafish with differing levels of cardiac lymphatic vasculature due to the developmental heat shock and induction of a sFlt4 receptor (Figure 6A–D), and performed cryoinjury and assayed scar size 60 days later without post injury induction of sFlt4 (60dpc, Figure 6D). Scoring these scars on the basis of severity we found large deposits of scar tissue in the transgenic fish (5/5, large scar,>0.005 mm3), in comparison to that of non-transgenic siblings, where the majority of scarring was relatively small (3/5, minimal scar,<0.005 mm3) (Figure 6E and F). Comparison of heart tissue scaring levels shows a significant increase in scarring severity with a reduction in the cardiac lymphatic vasculature (Figure 6H). Hearts lacking ventricular cardiac lymphatic vessels have a significantly larger scar volume (Figure 6—figure supplement 2A). This increase in scar volume could not be attributed to an observable defect in the coronary vasculature or a loss of coronary vessel angiogenesis at the wound site (Figure 6—figure supplement 1A-C, G-J; Figure 6—figure supplement 2B-E).

To test if the induction of vegfc after injury supported this function we inactivated the Vegfc ligands by induced expression of sFlt4/Vegfr3 receptor 2 days prior to and then after cryoinjury (Figure 6—figure supplement 3A). As is the case with induction of sFlt4 during cardiac lymphatic development, we observed only moderate non-significant effects on contrary vessel regeneration (Figure 6—figure supplement 3B-G) and no relative difference in post-injury heart morphology or size with sFlt4 post-injury induction (Figure 6—figure supplement 3H-K). Despite the lack of significant extension into the wound site at this stage, an increase in Mpx-positive neutrophils was observed at the 14dpc wound site with post injury induction of sFlt4 (Figure 6I–K), suggesting that blocking Vegfc-signal could prolong the inflammatory response, potentially by limiting the removal of Mpx-positive cells and debris from the wound site. Indeed, some sFlt4-induced hearts still showed elevated neutrophil levels at later stages of regeneration (Developmental induction, Figure 6—figure supplement 2F-H; Post-injury induction, Figure 6—figure supplement 4A-C).

Post-injury induction to block the response of the cardiac lymphatic vasculature did result in a reduction in the regeneration of the cardiac wall and resulting internalization of the scar tissue (Figure 6—figure supplement 4D-J). However the scar area itself showed no significant increase in these zebrafish (Figure 6—figure supplement 4D-J). There appears to be no inhibition of the induction of cardiomyocyte proliferation or epicaridal activation following injury with overexpression of sFlt4 (Figure 6—figure supplement 4K-P), suggesting instead that during regeneration the lack of ventricular cardiac lymphatic vessels, or perturbing their response to injury, shifts the balance of the wound site from pro-regenerative to pro-fibrotic. Consistent with this hypothesis we observed an inability of zebrafish that lack cardiac lymphatic vessels to modulate the wound environment and to resolve scar specifically after cryoinjury (Figure 6H). The majority of amputation-injured heat shocked transgenic hsp70l:sflt4 zebrafish had only minimal scaring after 30 days (30dpa, Figure 6G and H). The amount of scar tissue in those zebrafish with detectable scars at 30dpa was relatively small and also observed in amputated control zebrafish, suggesting the role of cardiac lymphatic vessels is less critical for regeneration following resection (Figure 6—figure supplement 5A–G).

Discussion

We here show that during zebrafish post-embryonic stages, there is a sequential development of blood vessel network and the cardiac lymphatic system as occurs embryonically in mice (Flaht et al., 2012; Klotz et al., 2015). The juvenile zebrafish heart undergoes significant morphological change and expansion due to the physiological demands of increasing body size on heart output (Gupta and Poss, 2012; Harrison et al., 2015). This expanded myocardium likely requires auxiliary oxygenated blood supply and lymph clearance to function optimally (Harrison et al., 2015; Figure 1). By contrast, in mammals the evolution of the pulmonary system has permitted a much elevated systemic blood pressure and metabolic rate that requires a greater cardiac output from birth (Bettex et al., 2014). As such, the needs of the post-natal compact myocardium have driven the timing of cardiac blood and lymphatic vasculature development embryonically.

The sequential development of cardiac blood and lymphatic vasculature is conserved between mammals and fish and has also been reported within the embryonic zebrafish trunk (Bussmann et al., 2010). However, as in the zebrafish heart, intersegmental arteries, not veins, provide a scaffold for the extension of lymphatic vessel sprouts and are required for their formation (Bussmann et al., 2010). We have identified a population of LECs that migrate down the arteries in the zebrafish heart (Figures 13) and a similar interaction of developing cardiac lymphatic vessels with arteries has been described in the human fetus (Kampmeier, 1928). This is in contrast to mice in which the cardiac lymphatic vessels track cardiac veins rather than coronary arteries (Klotz et al., 2015). In humans, like zebrafish, large arteries (and to lesser extent veins) are located subepicardially; in contrast the mouse heart has subepicardial veins, but deeper intramyocardial arteries (Sharma et al., 2017). Zebrafish lymphatic development may represent the ancestral mechanism of lymphatic guidance which has been entirely conserved or diverged to different degrees across the mammalian class (Ratajska et al., 2014). In each case, the cardiac lymphatic vessels appear to follow the more superficial blood vasculature on the ventricle and our data suggest this is required for or promotes their extension over the heart.

We have demonstrated that loss of the coronary artery scaffold results in a failure of these vessels to extend down the ventricle in cxcr4a mutants that lack coronary arteries. Although Cxcr4-Cxcl12 signaling has been implicated in the arterial association of LECs in the zebrafish trunk (Cha et al., 2012), we do not observe cxcr4a:mCitrine expression in lymphatic endothelial cells, but do not rule out that a similar direct signaling pathway, for example through Cxcr4b, could also occur in the cardiac system and that this may be compromised in cxcr4a mutants.

We have however identified a signaling pathway required for the development of the cardiac lymphatic system in zebrafish. Vegfc/Flt4 signaling has been shown to have important roles in embryonic and perinatal angiogenesis (Hogan et al., 2009a), lymphangiogenesis, (Küchler et al., 2006; Nurmi et al., 2015; Yaniv et al., 2006) and maintenance of mature intestinal lymphatic vessels (Nurmi et al., 2015; Tammela et al., 2008). We here show that Vegfc is critical for the outgrowth of cardiac lymphatic vessels along the arteries of the ventricles in adult zebrafish (Figure 6).

Our experiments also demonstrate that regular heat shock treatment of zebrafish actually promotes cardiac lymphatic development in non-transgenic control zebrafish (Figure 6). Such positive effect on lymphatic development is likely due to the induced stress response and associated demand on cardiac output that is observed during heat shock (Sallin and Jaźwińska, 2016). It also suggests that the natural expansion in adult zebrafish is regulated in part by physiological demands of the cardiac tissue to respond to the increasing body size of the fish. In much the same way, in juvenile hearts we see a requirement for a second source of blood vasculature as the myocardium expands (Harrison et al., 2015). With increased vasculature and heart mass comes increased extravascular fluid and cellular waste products such that the increased interstitial fluid may need an auxiliary conduit from the heart back to the circulation. The cardiac lymphatic system in zebrafish appears to have the capacity to clear fluid and debris from the myocardium (Figure 5).

Only moderate lymphangiogenesis is observed after amputation and the majority of the vasculature in the wound site is blood vasculature (Figure 4). Our analyses have demonstrated that there is a marked effect on the vascular response injury after extensive tissue damage (cyroinjury). Upon cryoinjury, there is a significant amount of lymphangiogenesis in and around the regeneration site that is not observed after amputation (Figure 4). Both injury models lead to increased vegfc levels, although this expression is more prolonged after cyroinjury (Figure 4). Coupled with a strong immune response to cryoinjury that could utilize the cardiac lymphatic system for the removal of cell debris and immune clearance (Figure 5) we postulate that this response can avoid a prolonged inflamed wound site and aid regeneration (Lai et al., 2017; Vieira et al., 2018). Consistent with this we see large scar tissue volume following severe cryoinjury of zebrafish ventricles that lack lymphatic vessels (Figure 6). Scar tissue is detectable in control ventricles at 60dpc, but the size and severity of this scar tissue is less than that observed in hsp70l:flt4 zebrafish without cardiac lymphatic vasculature (Figure 6). We also observe significantly more neutrophils at the 14dpc wound site with post-injury induction sFlt4 (Figure 6). This is early in the response of the lymphatic vessels to injury (Figure 4), but suggests compromised ability of the ventricular cardiac lymphatic vessels to remove the influx of neutrophils to the woundsite (Lai et al., 2017) with blocking Vegfc-signaling. We cannot rule out a direct effect on elevating the immune response with blockade of Vegfc-signal, but the resulting shift has a potent negative effect on the continuing regenerative environment of the tissue, while not directly effecting cardiomyocyte proliferation, epicardial activation, coronary vessel revascularization or heart structure and morphology (Figure 6). Consistent with our studies using sFlt4, recently it has been reported that a vegfd-mutation in a vegfc-hypermorphic background can also limit the extension of the cardiac lymphatic vessels down the ventricle while not affecting coronary vessels (Vivien et al., 2019). Further, almost half of the vegfc/d mutant zebrafish surviving to adulthood fail to complete regeneration by 180dpc (Vivien et al., 2019). Interestingly, these zebrafish also show cardiac hypertrophy, which we did not observe after induction of sflt4 at juvenile stages (Figure 6—figure supplement 1). It remains to be determined if sflt4 induction during embryonic stages results in cardiac hypertrophy. Nonetheless our results suggest that disruption of the cardiac lymphatic system impacts the hearts response to injury, rather than the enlargement of the myocardium.

Shifting the regenerative balance in human hearts for patients that have suffered acute myocardial infarct or insult and related sequelae will likely require the priming of a number of component systems in the heart including that which supplies blood to the injured tissue and removes fluid and debris. Key for this is a detailed understanding of the development of these systems and their interaction in a regenerative environment, which will ultimately help us understand what changes to modulate in patients and how this might be done. Zebrafish has provided an excellent system to understand the development of these systems from a deeper evolutionary understanding through to the ability of these systems to regenerate and play important roles in promoting the regenerative response of adult cardiac tissue to injury.

Materials and methods

Zebrafish lines

Request a detailed protocol

The following zebrafish lines were raised and maintained at Children’s Hospital Los Angeles (CHLA) under standard conditions of care (Aleström et al., 2019) and with CHLA IACUC oversight. IACUC vetted and prior approved all experimental procedures used in this study.

Tg(fli1a:EGFP)y1(Lawson and Weinstein, 2002), Tg(−5.1myl7:DsRed2-NLS)f2 (Mably et al., 2003), Tg(−6.5kdrl:mCherry)ci5(Proulx et al., 2010), Tg(−0.8flt1:RFP)hu5333 (referred to as flt1enh:tdtomato) (Bussmann et al., 2010), TgBAC(flt4:Citrine)hu7135 (referred to as flt4:YFP) (Gordon et al., 2013), TgBAC(prox1aBAC:KalTA4-4xUAS-ADV.E1b:TagRFP)nim5 (referred to as prox1a:Gal4-UAS:RFP) (van Impel et al., 2014), Tg(ubb:LOX2272-LOXP-RFP-LOX2272-CFP-LOXP-YFP)a132 (referred to as ubb:zebrabow) (Pan et al., 2013), Tg(kdrl:Cre-ERT2)fb13(Zhao et al., 2014), Tg(flk1:EGFP)s843 (referred to as kdrl:GFP) (Jin et al., 2005), Tg(gata1a:DsRed)sd2 (Traver et al., 2003), Tg(hsp70l:flt4, cryaa:Cerulean) (Matsuoka et al., 2016), Tg(−5.2lyve1b:DsRed)nz101, Tg(−5.2lyve1b:EGFP)nz151(Okuda et al., 2012), TgBAC(cxcr4a:Citrine)mu104 (Harrison et al., 2015), Tg(stab1:YFP)hu4453 (Hogan et al., 2009a), Tg(mrc1a:EGFP)y251 (Jung et al., 2017), Tg(ubb:LOXP-AmCyan-LOXP-ZsYellow)fb5 (referred to as ubb:CSY) (Zhou et al., 2011), Tg(dll4:EGFP)lcr1 (Sacilotto et al., 2013) transgenic lines have been described previously, as has the cxcr4aum20 mutation (Bussmann et al., 2011).

Tg(hsp70l:sflt4, cryaa:Cerulean) juvenile zebrafish were heat-shocked at 39°C for 1 hr 30 min, three times a week from 35d/1mpf to >6 mpf, or four times a week from 71dpf and 92dpf to >6 mpf (or up to the day prior to surgery, with then no heat-shock 21, 60dpc or 30dpa). Sibling zebrafish (both transgenic and non transgenic) were raised at the same density and heat-shocked at the same time. Non-heat-shocked controls born on the same day were maintained at the same density, but not subjected to heat-shock. Post-injury induction was carried out in a similar fashion. For studies performed at 14 and 56dpc fish were first heat-shocked initially for 2 days prior to injury then four times per week after injury. For studies performed at 3dpc, fish were heat-shocked daily for 10 days (seven prior to injury).

Tg(kdrl:mTurquoise) was generated as outlined (Bussmann and Schulte-Merker, 2011 and unpublished).

Immunohistochemistry and in situ hybridization

Request a detailed protocol

Standard confocal imaging was carried out as described previously (Harrison et al., 2015). Whole mount antibody and cleared native fluorescent zebrafish heart tissue was isolated in the same manner but fixed for 2 hr at room temperature in 4% PFA/PBS before being immunolabeled or mounted in 1% agarose for clearing; further details of which included in a forthcoming publication. Whole-mount immunolabeling was carried out on zebrafish hearts following fixation as follows. Hearts were bleached in Dent’s bleach (DMSO:H2O2:methanol, 1:2.5:40) overnight, rinsed in methanol and then fixed overnight in Dent’s fixative (DMSO:Methanol 1:4) at 4°C and washed in PBS + 0.1%Tween (PBST) for 3 to 8 hr the next day. Hearts were then incubated in primary antibody (see Table 1) in blocking solution (HIHS:DMSO:PBS, 1:4:15) for 3–5 days. After which they were washed with PBST for 6 hr. Hearts were then incubated secondary antibody (see Table 1) in blocking solution overnight, then washed in PBST for 5 hr. Hearts were then mounted in low melting point agarose and imaging was carried out as before (Harrison et al., 2015).

Table 1
Table of antibodies used.
AntigenSpeciesConcentrationSupplier
MpxRabbit1:100GeneTex, GTX128379
IB4 (conjugated protein)n/a (Griffonia Simplicifolia)1:100Vector Laboratories, DL-1207
GFPChicken1:500Aves Labs, GFP-1010
Rabbit IgG-647Goat1:500ThermoFisher Scientific, A21245
Mouse IgG-488Goat1:500ThermoFisher Scientific, A11001
Chicken IgG-FITCGoat1:500Aves Labs, F1005
Raldh2Rabbit1:200GeneTex, GTX124302
Mef2cRabbit1:200Santa Cruz, sc-313
PCNAMouse1:200Vector Labs, VP-P980

Section immunohistochemistry was carried out on de-waxed paraffin sections (Toluene 5 min x2, 100% ethanol 5 min x2, 80% ethanol 5 min, PBS 5 min x2), after antigen retrieval in Unmasking Solution (Vector Laboratories, H-3300). Primary antibody (see Table 1) was applied overnight at 4°C in blocking solution (5% goat serum, 2.5% BSA, 0.3% Triton, 1% DMSO, 0.1% Tween20 in TBS), after incubation for 1 hr at room temperature in this blocking solution. Slides were rinsed in PBST (5 min x3) before and after application of secondary antibody (1 hr room temperature) then imaged using lsm710 and lambda/spectral imaging.

Section in situ hybridization was carried out on de-waxed paraffin sections using RNAscope Multiplex Fluorescent Assay v2 as per manufacturer's instruction (Cat No. 323100, Advanced Cell Diagnostics), vegfc probe (Cat No. 528701, Advanced Cell Diagnostics) was visualized with Opal 690 Reagent Pack (PN: FP1497001KT, Akoya Biosciences).

Quantification of Mpx or IB4-positive cell numbers was carried out in ImageJ though unprojected z-planes, only cells within the lymphatic vessel endothelium were counted. ImageJ was used for quantification of Mpx and IB4-cells in and within 100 μm of the woundsite and averaged across at least three images through woundsite. Similarly quantification of percentage PCNA/Mef2c-positive cells was carried by counting double positive and Mef2c-only cells within a 500 μm2 window proximal to the woundsite, and then averaged across at least three images. Quantification of all double positive and Raldh2-only cells within an entire 708 μm2 imaging window proximal to the wound site was carried out to calculate percentage of Raldh2 cells (epicardium) that are proliferative. Quantification of post-injury angiogenic sprouts was carried out by counting kdrl:GFP-positive sprouts within the wound site and then dividing this by woundsite area. For lyve1b:DsRed-positive, flt4:mCitrine-positive lymphatic vessel coverage and kdrl:GFP-positive coronary vessel coverage intensity thresholds of a projected z-stack were used to determine the area of the lyve1-positive, flt4:mCitrine-positive or kdrl:GFP-positive vasculature on the ventricle in comparison to the total/partial area of the ventricle or woundsite as defined in text. For ventricle size quantification, the largest AFOG/immune-labeled heart section of each zebrafish was imaged and area measured using threshold function in ImageJ. This maximum area was then plotted against, or normalized to, standard body length of the zebrafish. Raw Mef2c/DAPI counts we made using ImageJ cell counter. Chi-squared statistical analysis of counts was completed in Prism 6 (GraphPad). Visualization and analysis of confocal data were carried out in Zen (Zeiss) and ImageJ, with the exception of lymphatic-neutrophil/macrophage analysis in Vision4D (Arivis).

Intramyocardial injection

Request a detailed protocol

A 10 μl volume of filtered microspheres (0.2 μm Fluoro-Max, B200, Thermo Fisher Scientific) and Qdots (Qtracker705 vascular label, Q21061MP, Thermo Fisher Scientific) was prepared in 50:50 ratio and loaded into a micro-injector pulled-capillary syringe. Zebrafish were anesthetized in Tricaine and the chest opened and held opened with forceps. The injection was made using a FemtoJet microinjector (Eppendorf), 1 μl was injected into the ventral myocardium tissue at an acute angle to the plane of the heart.

Resection surgery and cryoinjury

Request a detailed protocol

Resection surgery and AFOG histology carried out as described previously (Poss et al., 2002). Cryoinjury was carried out as described (González-Rosa and Mercader, 2012), but with the modification for two injuries: a severe cryoinjury using a 0.8 mm diameter spherical probe and mild cryoinjury using a 0.6 mm dia spherical probe (10160–13, Fine Scientific Tools). Both are cooled in LN2 prior to application to the ventricle for 7–10 s after which the probe was warmed with water and removed, the post-surgery fish were returned to system water to recover. At least five zebrafish were injured per observation/group to allow for potential variability of the wound response within a single experiment. For comparative studies sibling zebrafish were used when possible and randomly assigned into different injury groups.

Estimation of lymphatic vessel coverage after the injury was calculated in a fixed 600 μm2 square section centered on the wound site or uninjured apical region. Using thresholding of the 514 nm channel in ImageJ area of the lymphatic vessel was calculate and results presented as percentage coverage of the fixed area of measurement. To quantify the relative level of branching after injury, vessels were traced and total length calculated in ImageJ, then total number of bifurcations over the ventricle counted to give normalized bifurcations per unit length of vessel. For cyroinjury scar volume calculations, whole ventricle and scar (collagen and fibrin) regions were traced as regions of interest in imageJ. Intensity thresholding was used to select tissue to calculate the area without the inclusion of intertrabecular space. This area measurement was repeated on every 9th 7 μm consecutive section through the entire heart. The individual area measurements were then multiplied by 63 μm and subsequently added together to give a total estimated volume. Hearts were grouped based on the volume of this scar tissue with a cut off set at 0.05 μm3 above which was defined as large scar. For amputation, hearts were grouped on the basis of collagen and fibrin levels. Hearts with or less than a thin layer of collagen (blue) on the inside of the thickened regenerated tissue were classed as ‘minimally scared’. Fibrin deposition at the wound site (red) was considered indicative of scar tissue presence. Level of regeneration of the cardiac tissue or level of myocardial wall thickening at the wound site was not considered in scar severity assessments. For quantification of regenerated tissue, thickened myocardial wall proximal to the scar was traced and area calculated for the section though the middle of the scar (largest scar area). For estimation of scar internalization the minimal distance of scar tissue from the edge of the tissue (epicaridum) was measured in ImageJ. Scar area was calculated as volume, but in a single section with the largest scar area.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
    Stem Cells, Tissue Engineering and Regenerative Medicine
    1. L Gamba
    2. MR Harrison
    3. C-L Lien
    (2014)
    Heart regeneration and repair: What we have learned from model organisms, Stem Cells, Tissue Engineering and Regenerative Medicine, California, World Scientific Publishing.
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
    The Cardiac Lymphatic System an Overview
    1. G Karunamuni
    (2013)
    Springer.
  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

Decision letter

  1. Paul R Riley
    Reviewing Editor; University of Oxford, United Kingdom
  2. Marianne E Bronner
    Senior Editor; California Institute of Technology, United States
  3. Anna Jazwinska
    Reviewer; University of Fribourg, Switzerland

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Thank you for submitting your article "Late developing cardiac lymphatic vasculature supports adult zebrafish heart function and regeneration" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by a Reviewing Editor and Didier Stainier as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Anna Jazwinska (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This study by Harrison and colleagues describes the development and response to injury of the cardiac lymphatic vasculature in zebrafish. The authors document the initial formation of lymphatics in the heart at surprisingly late stages, ie in young adults (90dpf), utilising multiple transgenic lines expressing flt4- (vegfr3) and prox1- driven citrine and RFP reporters, and confirmed lymphatic endothelial cell identity with additional lyve-1 and mrc1a reporters plus exclusion of the blood endothelial cell marker kdl. They revealed a close association of the forming lymphatics with coronary blood vessels (arteries) and confirmed similar findings in a comparative study of the lymphatics in human embryonic hearts. They demonstrated that normal coronary vessel formation is a pre-requisite for the extension of the lymphatics from the bulbus cordis down into the ventricle by studying cxcr4 mutant fish with impaired coronaries. The authors then went on to map the response of the lymphatics to heart injury and during regeneration; they initially confirmed the angiogenic response of the coronary blood vasculature, as previously reported, and determined that lymphatics did not sprout into the wound area following ventricular amputation. In contrast, they observed a marked increase in lymphatics around the wound site following cryo-injury suggesting a requirement for necrosis and scarring to induce lymphangiogenesis. Functionally the newly formed lymphatics appeared to take up quantum dots, suggesting fluid clearance arising from the injury may be one of the lymphatic functions and there was a close association with neutrophils and macrophages, suggesting the lymphatics may also traffic immune cells as has recently been shown in mouse. Finally, the authors tested a requirement for vegfc in transgenic fish, with heat-shock induction of soluble flt4/vegfr3 receptor, which led to significantly fewer LECs, no lymphatics within the ventricle and correlated with increased severity of scarring in cryo-injured fish.

This is an interesting and novel study, as the first to describe the development and response to injury of the cardiac lymphatics in zebrafish. The data, and in particularly the cleared tissue/whole mount imaging, is of a high standard. The findings are timely, given the recent significant interest in the development and injury-response of organ-based lymphatics across model organisms, as illustrated by several high impact studies. In particular, understanding the response of the cardiac lymphatics to injury is clinically important and the benefit of utilising the zebrafish here is in determining a requirement for lymphatics to facilitate scar resolution and heart regeneration.

That said, there are aspects of the study which have not been developed in sufficient depth to fully support the conclusions and several issues which should be addressed:

Essential revisions:

1) Regarding the relationship of the emerging lymphatics with the coronary vessels, it was not clear whether these were coronary arteries or veins from the images/marker sets utilized. This needs clarifying.

2) The study of human embryos appears somewhat preliminary and superficial relative to the rest of the detailed focus on zebrafish. A suggestion would be to omit the human data here and develop this as a separate study.

3) In the cxcr4 mutant studies, coronary vessels are mis-patterned but clearly still form, as demonstrated by Fli1a staining in Figure 3B. The authors should quantify the density of coronary vessels in these cxcr4a mutants and determine their identity (arterial or venous?). The complete absence of any lymphatics emerging into the ventricle is surprising given the apparent incidence of coronaries. This may reflect the fact that cxcr4 signalling is directly required for LEC guidance as opposed to acting through the coronary vessels. At present the suggestion that the coronaries act as a guide/scaffold for the lymphatics is only correlative; statements such as "lymphatic development requires a coronary artery scaffold" and in the abstract "Zebrafish cardiac lymphatic vessels require coronary arteries to guide their expansion" are not supported by the data. These need to be toned down or further supported by functional data showing either that coronaries are sufficient to provide guidance or that cxcr4 is in fact required more directly to guide the forming lymphatics.

4) Figure 4, the analysis of the angiogenic response to injury is essentially confirming prior work (Marin-Juez et al., 2016) and is largely redundant here. It would be better to focus on the lymphatics, albeit alongside blood vessels, in response to injury. A suggestion would be to omit Figure 4, and the corresponding supplementary figure, and to focus on the lymphatics with further images of sprouting lymphatics into the wound area (alongside the blood vasculature).

5) Some additional analyses of the responding lymphatics would be highly informative, to determine whether the newly forming vessels are different (marker profile, anatomy, organisation) from the pre-existing lymphatics.

6) A more detailed time course of the lymphangiogenic response to injury is required. Analysis at a single time point, of 54 dpa or 60 dpc, as presented fails to address important questions such as, when do the lymphatics first respond and/or is there a transient or delayed response following resection (amputation)?

7) The functional studies in Figure 6 are somewhat premature and insufficient to support the conclusions. The presence of Qdots within lymphatics needs demonstrating unequivocally by improved orthogonal imaging to assess whether the dots are actually within the vessel lumen. In addition, a demonstration of clearance of Qdots to sites outside of the heart (analogous to draining lymph nodes in mammals) would be informative regarding function. If these additional analyses are deemed challenging. then it would be preferable not to include the Qdot data and to omit conclusions on attributing a clearance function to the cardiac lymphatics.

8) For the immune cell studies, it is not clear how the authors can claim that neutrophils are in lymphatic vessels given the reliance on Fli1a:GFP as a marker (Figure 6G); this labels all endothelial cells and ought to be replaced with a lymphatic-specific reporter. The evidence that Mpx+ neutrophils are actually within lymphatic vessels requires more detailed orthogonal imaging and other important immune cell types should be assessed, most notably macrophages which dominate the wound healing response to injury can be marked by Mpeg. Improved imaging of both Mpx+ and Mpeg+ cells should be carried out, with quantification of cell numbers that are correlated with the extent of lymphangiogenesis. A suggestion for optimal quantification is to apply FACS, making use of available Mpx and Mpeg reporter fish (Renshaw et al., 2006; Ellet et al., 2011).

9) In the inducible soluble flt4 (sFlt4)/vegfr3 transgenic fish the authors should analyse the mutant heart phenotype(s). Do these hearts have any local oedema? Are they malformed? Are the blood vessels of the coronary vasculature quantitatively normal (they look abnormal or different from controls in the images provided)? A careful and thorough description of these hearts is needed because the reduced ability to recover from injury may be secondary to phenotypes caused by the earlier absence of lymphatics, or due to other roles for Flt4 on blood vessels, rather than an acute requirement for lymphatic function during repair. In addition, the lack of lymphangiogenesis following cryoinjury in the sFlt4 mutants suggests that vegfc is probably driving the response; however, there is no expression data shown for vegfc or other lymphangiogenic factors following injury. The authors should analyse vegfc mRNA levels (by combined qPCR and section in situ hybridisation) following cryo-injury versus resection. If vegfc expression is increased following cry-injury, but not following resection, this may explain the difference in lymphatic response between the two injury models.

Finally, heat-shock, regardless of carrying the sFlt4 transgene, appears to have reduced coronary blood vessels (Figure 7A-C). Quantification of the blood vessels should be provided across experimental groups, as the reduced regeneration could be due to a blood vascular phenotype and may be related to heat shock rather than the flt4/vegfc signalling. Why do the authors think the heat-shocked controls mounted a lymphatic response? This should be commented on further in the text.

10) The assessment of the effect of sFlt4 (impaired lymphangiogenesis) on scarring, as shown in Figure 7, requires a time course to determine whether the effect is with respect to initial scar deposition versus subsequent scar resolution and how this relates to regeneration per se. What are the effects of sFlt4 on other cell types and regenerative processes, such as activation of epicardium/endocardium and cardiomyocyte proliferation? Moreover, to link to the earlier functional studies the authors should investigate an effect on immune cell numbers (Mpx+ and Mpeg+ cells) in the absence of lymphatics.

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

Thank you for resubmitting your work entitled "Late developing cardiac lymphatic vasculature supports adult zebrafish heart function and regeneration" for further consideration at eLife. Your revised article has been favorably evaluated by Didier Stainier (Senior Editor), a Reviewing Editor, and two reviewers.

The manuscript has been significantly improved, but there are a couple of remaining issues that need to be addressed before acceptance, as outlined below:

1) Some further data is required to demonstrate the specificity of the cardiac phenotype following the transgenic loss of lymphatics after injury and to establish rigorously that the cardiac lymphatics are required for heart regeneration. In the sFlt4 model the coronary blood vasculature and general heart morphology should be investigated post-injury, to determine whether they are relatively normal, and thus place the emphasis more specifically on the loss-of-lymphatics.

2) The role of the lymphatics in immune cell trafficking and specifically the data on neutrophils is correlative at this stage. The images in Figure 5 and Video 2 are not entirely convincing, as to whether the documented Mpx+ cells actually reside within the vessel lumen, and there is no functional validation such as via Mtz-mediated neutrophil depletion. Whilst the latter is recognised to be beyond the current scope, and would require several months of new experimentation, there is a need to tone down statements throughout on immune cell/neutrophil clearance and especially with regards being causative in regeneration versus scarring.

https://doi.org/10.7554/eLife.42762.sa1

Author response

Essential revisions:

1) Regarding the relationship of the emerging lymphatics with the coronary vessels, it was not clear whether these were coronary arteries or veins from the images/marker sets utilized. This needs clarifying.

Thank you for pointing out that this is not to sufficiently clear. We have added additional analysis of markers to help address this. We have shown lymphatic formation along vessels expressing dll4high (Figure 2) cxcr4a (Figure 3), flt1enh (Figure 2) and Kdrlhigh (Figure 2). We also show that cxcr4a overlaps with flt1enh expression (Figure 2) and have previously shown this expression overlaps with dll4high and is consistent with the expression of these markers in thearteries of the fins (Harrison et al., 2015).

2) The study of human embryos appears somewhat preliminary and superficial relative to the rest of the detailed focus on zebrafish. A suggestion would be to omit the human data here and develop this as a separate study.

We agree with the suggestion and have removed the human embryo data at this stage.

3) In the cxcr4 mutant studies, coronary vessels are mis-patterned but clearly still form, as demonstrated by Fli1a staining in Figure 3B. The authors should quantify the density of coronary vessels in these cxcr4a mutants and determine their identity (arterial or venous?). The complete absence of any lymphatics emerging into the ventricle is surprising given the apparent incidence of coronaries. This may reflect the fact that cxcr4 signalling is directly required for LEC guidance as opposed to acting through the coronary vessels. At present the suggestion that the coronaries act as a guide/scaffold for the lymphatics is only correlative; statements such as "lymphatic development requires a coronary artery scaffold" and in the abstract "Zebrafish cardiac lymphatic vessels require coronary arteries to guide their expansion" are not supported by the data. These need to be toned down or further supported by functional data showing either that coronaries are sufficient to provide guidance or that cxcr4 is in fact required more directly to guide the forming lymphatics.

We agree with this comment, and have accordingly toned down the use of “require” (subsection “Development of a cardiac lymphatic system is incomplete in the absence of a coronary artery scaffold”). We are also interested in the potential role of cxc-signaling in the extension of the lymphatics and their association with arteries and agree that this is still a possibility. The fact that this signaling could be more broadly affected in the mutant has been noted in the text regardless of any effect on the lymphatic endothelial cells directly (Discussion section).

We have looked in more detail at the correlation of between kdrlhigh and lymphatic marker expression/coverage, both of which vary within the WT population and between heart sides within individual zebrafish (Figure 3—figure supplement 2). There is a highly significant correlation suggesting the former is required or promotes the extension of lymphatics, but we agree this is a suggestion that needs more independent verification and have reflected this in the text (subsection “Development of a cardiac lymphatic system is incomplete in the absence of a coronary artery scaffold”).

We are also currently looking at other markers of coronary vessel identity in the cxcr4a mutant background, but this work could not be completed within the revision timeframe. However, we have studied higher numbers of mutant and control fish and now provide more quantification of these vessels and the variation observed in older cxcr4a mutant fish (Figure 3—figure supplement 2). We believe this clarifies the point we are making with respect to abnormal (arterial) coronary development, while not precluding a more direct role for cxc-signaling in lymphatic development.

4) Figure 4, the analysis of the angiogenic response to injury is essentially confirming prior work (Marin-Juez et al., 2016) and is largely redundant here. It would be better to focus on the lymphatics, albeit alongside blood vessels, in response to injury. A suggestion would be to omit Figure 4, and the corresponding supplementary figure, and to focus on the lymphatics with further images of sprouting lymphatics into the wound area (alongside the blood vasculature).

We have removed the coronary data and instead focused on the lymphatogenesis response in particular the marker profile (5) and time course (6).

5) Some additional analyses of the responding lymphatics would be highly informative, to determine whether the newly forming vessels are different (marker profile, anatomy, organisation) from the pre-existing lymphatics.

We have characterized the expression of Lyve1:DsRed, Flt4:mCitrene, Mrc1a:GFP, kdrl:mTurquiose and Prox1:RFP in regenerated lymphatic vessels and quantified the branching organization observed in Flt4:mCitrene fish (Figure 4). We now also provide a better description of the responding lymphatics.

6) A more detailed time course of the lymphangiogenic response to injury is required. Analysis at a single time point, of 54 dpa or 60 dpc, as presented fails to address important questions such as, when do the lymphatics first respond and/or is there a transient or delayed response following resection (amputation)?

We have analyzed the lymphangiogenesis response earlier with both injury models finding no such transient response to amputation and overt lymphangiogenesis observed by 28dpc, with more subtle changes observed earlier (Figure 4).

7) The functional studies in Figure 6 are somewhat premature and insufficient to support the conclusions. The presence of Qdots within lymphatics needs demonstrating unequivocally by improved orthogonal imaging to assess whether the dots are actually within the vessel lumen. In addition, a demonstration of clearance of Qdots to sites outside of the heart (analogous to draining lymph nodes in mammals) would be informative regarding function. If these additional analyses are deemed challenging. then it would be preferable not to include the Qdot data and to omit conclusions on attributing a clearance function to the cardiac lymphatics.

We agree that showing clearance to remote sites outside the heart would be informative and we thank the reviewers for the suggestion, however given the uncertain anatomy this has proved difficult for this current study.

It appears that flow in the cardiac lymphatic vessel is up the aorta and joins collection from the gills (Figure 2—figure supplement 1) to extend toward the facial lymphatics. However, imaging this extracardiac region is difficult and so we instead decided to remove this analysis, but include clearer imaging of earlier uptake of the Qdots on the ventricle. In doing so we make a simpler point that the ventricular lymphatics form a functional vessel capable of Qdot uptake (Figure 5). We have made the necessary changes to the text to reflect this conclusion subsection “Cardiac lymphatic vasculature functionally supports the heart during regeneration and homeostasis.” and Discussion section).

8) For the immune cell studies, it is not clear how the authors can claim that neutrophils are in lymphatic vessels given the reliance on Fli1a:GFP as a marker (Figure 6G); this labels all endothelial cells and ought to be replaced with a lymphatic-specific reporter. The evidence that Mpx+ neutrophils are actually within lymphatic vessels requires more detailed orthogonal imaging and other important immune cell types should be assessed, most notably macrophages which dominate the wound healing response to injury can be marked by Mpeg. Improved imaging of both Mpx+ and Mpeg+ cells should be carried out, with quantification of cell numbers that are correlated with the extent of lymphangiogenesis. A suggestion for optimal quantification is to apply FACS, making use of available Mpx and Mpeg reporter fish (Renshaw et al., 2006; Ellet et al., 2011).

Lymphatic vessels we identified based on morphology and position, but we agree this is not ideal and so have repeated the experiment with Lyve1GFP zebrafish and in addition carried out higher resolution orthogonal imaging (Figure 5).

Unfortunately we do not have the suggested Mpx or Mpeg zebrafish in our stocks to carry out the experiment as suggested and crossing these fish to lymphatic reporter lines will need more time than the revision allows, but we did successfully used Isolectin B4 conjugate to identify macrophage population in wholemount hearts and included this analysis. We used both markers to determine if there was a prolonged or heightened immune response in the absence of cardiac lymphatic extension/post-injury lymphatogenesis, finding a significant increase of Mpx-cells at the 14dpc wound site (Figure 6), an elevated but not significant increases later at 56dpc (post-injury, Figure 6—figure supplement 1) and 60dpc (developmental extension, Figure 6—figure supplement 1). We have discussed the implications of this (Discussion section) in the context of lymphatic function.

9) In the inducible soluble flt4 (sFlt4)/vegfr3 transgenic fish the authors should analyse the mutant heart phenotype(s). Do these hearts have any local oedema? Are they malformed? Are the blood vessels of the coronary vasculature quantitatively normal (they look abnormal or different from controls in the images provided)? A careful and thorough description of these hearts is needed because the reduced ability to recover from injury may be secondary to phenotypes caused by the earlier absence of lymphatics, or due to other roles for Flt4 on blood vessels, rather than an acute requirement for lymphatic function during repair. In addition, the lack of lymphangiogenesis following cryoinjury in the sFlt4 mutants suggests that vegfc is probably driving the response; however, there is no expression data shown for vegfc or other lymphangiogenic factors following injury. The authors should analyse vegfc mRNA levels (by combined qPCR and section in situ hybridisation) following cryo-injury versus resection. If vegfc expression is increased following cry-injury, but not following resection, this may explain the difference in lymphatic response between the two injury models.

Finally, heat-shock, regardless of carrying the sFlt4 transgene, appears to have reduced coronary blood vessels (Figure 7A-C). Quantification of the blood vessels should be provided across experimental groups, as the reduced regeneration could be due to a blood vascular phenotype and may be related to heat shock rather than the flt4/vegfc signalling. Why do the authors think the heat-shocked controls mounted a lymphatic response? This should be commented on further in the text.

The reviewers make a number of valid points here. Interestingly we did not find overt edema at 6 months, but it is possible this develops in older adults – we have commented on this and the histology of the sFlt4 hearts (subsection “Vegfc promotes the formation of cardiac lymphatic vasculature after the initiation of coronary vessel development”).

We agree that any effect on coronary vessels should be analyzed more closely. We see only transient expression of Flt4 in the angiogenic sprouts emerging from the AVC, but not after when they extend over the ventricle and when sFlt4 induction was initiated. We have included here that developmental induction from 1mpf (after initial coronary vessel development had begun) appears to not overtly affect the kdrl+ coronary vasculature (Figure 6—figure supplement 1) and certainly not the arterial structure that we believe vital for ventricle lymphatic vessel extension.

We have tested if vegfc is driving the lymphangiogenesis response, both by analyzing expression and inducing sFlt4 specifically after injury (Figure 4, see point 10 below).

We agree that it is interesting that heat shocked zebrafish mount a lymphatic response. It is likely in response to the stress placed on heart during heat shock, including increased heart output and the repercussions of this on myocardial energy/waste production. We have discussed this further in the text and think it is an interesting avenue of investigation (Discussion section).

10) The assessment of the effect of sFlt4 (impaired lymphangiogenesis) on scarring, as shown in Figure 7, requires a time course to determine whether the effect is with respect to initial scar deposition versus subsequent scar resolution and how this relates to regeneration per se. What are the effects of sFlt4 on other cell types and regenerative processes, such as activation of epicardium/endocardium and cardiomyocyte proliferation? Moreover, to link to the earlier functional studies the authors should investigate an effect on immune cell numbers (Mpx+ and Mpeg+ cells) in the absence of lymphatics.

We agree that a better understanding of scar formation and removal and examining other processes of heart regeneration in animals with forced sflt4 expression would be beneficial; we have raised more developmentally induced-sflt4 zebrafish to investigate this further in the future. However, raising these fish took 6 months and the length subsequent post-injury analysis has prevented this work from completion in the revision time period. We instead followed reviewer’s suggestion that post-injury blockade of VEGFC-signal may cause a lack of lymphangiogenesis following cryoinjury in the sFlt4 fish. We have found that post-injury induction sflt4 to block the lymphangiogenic response does inhibit regeneration, but such effects are more subtle (Figure 6—figure supplement 2). In addition, we see no difference in scar size (or tissue regeneration) in sFlt4induced fish at 14dpc, suggesting collagen deposition is not elevated.

We have evaluated the effect of sFlt4 induction on cardiomyocyte proliferation and activation of the epicardium (Figure 6—figure supplement 3). Further we have quantified immune cell number at the wound site in both post-injury and developmentally induced fish to observe is prolonged elevated inflammation in these fish (Figure 6—figure supplement 1 and Figure 6—figure supplement 2). We think this provides an interesting and informative complement to the data previously submitted and have discussed these trends more fully in the text (Discussion section).

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

The manuscript has been significantly improved, but there are a couple of remaining issues that need to be addressed before acceptance, as outlined below:

1) Some further data is required to demonstrate the specificity of the cardiac phenotype following the transgenic loss of lymphatics after injury and to establish rigorously that the cardiac lymphatics are required for heart regeneration. In the sFlt4 model the coronary blood vasculature and general heart morphology should be investigated post-injury, to determine whether they are relatively normal, and thus place the emphasis more specifically on the loss-of-lymphatics.

We have examined and/or quantified the cardiac morphology and coronary vessels using both developmentally induced and post-injury induced sFlt4. We did not find significant edema or cardiac hypertrophy and have included more data and quantification to bolster this.

Developmental: whole-mount brightfield imaging, ventricle size, cardiomyocytes and noncardiomyocytes analysis (Fig6—figure supplement 1).

Post-injury: histology and post-injury ventricle area to body length comparisons (Fig6—figure supplement 3),

In addition to finding the coronary vessels are largely normal with developmental induction of sFlt4 (Figure 6—figure supplement 1) we find that injury of these vessels, or post-injury induction of sFlt4, does not significantly attenuate the coronary vessel response to injury (at 14 and 21dpc, Fig6—figure supplement 2 and Figure 6—figure supplement 3). Together this suggests that sFlt4 overexpression does not affect angiogenesis of coronary vessels.

Indeed, for sFlt4 induction during development, we have induced the transgene at 71 and 92dpf (Fig6—figure supplement 1) to ensure that sFlt4 is only over-expressed late in or after the completion of coronary vessel development and this is still sufficient to strongly inhibit the formation of the cardiac lymphatic vessel. This further suggests the observed sFlt4 phenotype is indeed a specific effect on the flt4-expressing cardiac lymphatic endothelial cells and not explained by a repercussion of sFlt4 induction affecting the coronary vasculature.

This finding is consistent with a recent publication by Vivien et al., in npj Regenerative Medicine.

2) The role of the lymphatics in immune cell trafficking and specifically the data on neutrophils is correlative at this stage. The images in Figure 5 and Video 2 are not entirely convincing, as to whether the documented Mpx+ cells actually reside within the vessel lumen, and there is no functional validation such as via Mtz-mediated neutrophil depletion. Whilst the latter is recognised to be beyond the current scope, and would require several months of new experimentation, there is a need to tone down statements throughout on immune cell/neutrophil clearance and especially with regards being causative in regeneration versus scarring.

We agree this imaging is less than convincing and have provided new imaging data to demonstrate that mpx+ neutrophils reside within the vessel lumen (Figure 5—figure supplement 1, Figure 5Video 4 and Video 5). However, we agree with the reviewers that we observed more macrophages than neutrophils clearly residing within the lumen of the lymphatic vessels, while some of the neutrophils maybe along or enveloped within the vessel membrane. Therefore, we have revised the results as “mpx+ neutrophils associated with cardiac lymphatic vessels” (Figure 5—figure supplement 1) and toned down the conclusions relating to this throughout the manuscript (Abstract, Introduction, subsection “Cardiac lymphatic vasculature functionally supports the heart during regeneration and homeostasis.”, subsection “Vegfc promotes the formation of cardiac lymphatic vasculature after the initiation of coronary vessel development”, Discussion section and subsection “Zebrafish lines”).

https://doi.org/10.7554/eLife.42762.sa2

Article and author information

Author details

  1. Michael RM Harrison

    The Saban Research Institute of Children's Hospital Los Angeles, Los Angeles, United States
    Contribution
    Conceptualization, Formal analysis, Supervision, Investigation, Visualization, Methodology, Writing—original draft
    For correspondence
    mharrison@chla.usc.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-1703-9879
  2. Xidi Feng

    The Saban Research Institute of Children's Hospital Los Angeles, Los Angeles, United States
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  3. Guqin Mo

    The Saban Research Institute of Children's Hospital Los Angeles, Los Angeles, United States
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  4. Antonio Aguayo

    The Saban Research Institute of Children's Hospital Los Angeles, Los Angeles, United States
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  5. Jessi Villafuerte

    1. The Saban Research Institute of Children's Hospital Los Angeles, Los Angeles, United States
    2. Department of Biology, California State University of San Bernardino, San Bernardino, United States
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  6. Tyler Yoshida

    1. The Saban Research Institute of Children's Hospital Los Angeles, Los Angeles, United States
    2. Department of Biological Sciences, Dornsife College of Letters, Arts and Sciences, University of Southern California, Los Angeles, United States
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  7. Caroline A Pearson

    Department of Neurobiology, Eli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, United States
    Contribution
    Visualization, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  8. Stefan Schulte-Merker

    1. Institute of Cardiovascular Organogenesis and Regeneration, Faculty of Medicine, University of Münster, Münster, Germany
    2. CiM Cluster of Excellence (EXC1003 CiM), University of Münster, Münster, Germany
    Contribution
    Resources, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3617-8807
  9. Ching-Ling Lien

    1. The Saban Research Institute of Children's Hospital Los Angeles, Los Angeles, United States
    2. Department of Surgery, Keck School of Medicine, University of Southern California, Los Angeles, United States
    3. Department of Biochemistry and Molecular Medicine, Keck School of Medicine, University of Southern California, Los Angeles, United States
    Contribution
    Conceptualization, Supervision, Funding acquisition, Methodology, Writing—original draft, Project administration
    For correspondence
    clien@chla.usc.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5100-9780

Funding

National Heart, Lung, and Blood Institute (1R01HL130172)

  • Ching-Ling Lien

Children's Hospital Los Angeles (2nd R01 and Team Awards)

  • Ching-Ling Lien

American Heart Association (I81PA34180044)

  • Ching-Ling Lien

California Institute for Regenerative Medicine (EDUC2-08418)

  • Jessi Villafuerte

Deutsche Forschungsgemeinschaft (CRC 1348)

  • Stefan Schulte-Merker

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

Acknowledgements

We thank Drs. Young Hong, Henry Sucov for helpful discussions and suggestions. Drs. Mathias Francois, James Hudson, Ruben Marin-Juez, Enzo Porello and Karina Yaniv for communications prior to the submission. Dr. Elke Ober for sharing the prox1:RFP line, Dr. Didier Stainier for the hsp70l:flt4 line, Dr. Phil Crosier and Dr. Kazuhida Okuda for the lyve1b transgenic lines and Dr. Brant Weinstein for the mrc1a:EGFP line.

Ethics

Animal experimentation: All zebrafish husbandry was performed under standard conditions, and all animal experiments were performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and following the ARRIVE guidelines provided by the National Centre for the Replacement, Refinement and Reduction of Animals in Research, and the FELASA recommendations on zebrafish husbandry (Alestrom et al., 2019). All procedures were carried out as approved by the Children's Hospital Los Angeles (CHLA)institutional animal care and use committee (IACUC) protocols (#201-18 and 212-16).

Senior Editor

  1. Marianne E Bronner, California Institute of Technology, United States

Reviewing Editor

  1. Paul R Riley, University of Oxford, United Kingdom

Reviewer

  1. Anna Jazwinska, University of Fribourg, Switzerland

Publication history

  1. Received: October 12, 2018
  2. Accepted: September 24, 2019
  3. Accepted Manuscript published: November 8, 2019 (version 1)
  4. Version of Record published: November 27, 2019 (version 2)

Copyright

© 2019, Harrison 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

  • 2,846
    Page views
  • 447
    Downloads
  • 11
    Citations

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

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)

Further reading

    1. Developmental Biology
    2. Stem Cells and Regenerative Medicine
    Catherine Pfefferli, Anna Jaźwińska
    Insight

    Experiments on zebrafish show that the regeneration of the heart after an injury is supported by lymphatic vessels.

    1. Developmental Biology
    2. Neuroscience
    Athene Knüfer et al.
    Research Article Updated

    In the vertebrate central nervous system, groups of functionally related neurons, including cranial motor neurons of the brainstem, are frequently organised as nuclei. The molecular mechanisms governing the emergence of nuclear topography and circuit function are poorly understood. Here we investigate the role of cadherin-mediated adhesion in the development of zebrafish ocular motor (sub)nuclei. We find that developing ocular motor (sub)nuclei differentially express classical cadherins. Perturbing cadherin function in these neurons results in distinct defects in neuronal positioning, including scattering of dorsal cells and defective contralateral migration of ventral subnuclei. In addition, we show that cadherin-mediated interactions between adjacent subnuclei are critical for subnucleus position. We also find that disrupting cadherin adhesivity in dorsal oculomotor neurons impairs the larval optokinetic reflex, suggesting that neuronal clustering is important for co-ordinating circuit function. Our findings reveal that cadherins regulate distinct aspects of cranial motor neuron positioning and establish subnuclear topography and motor function.