Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Requirement of vasculogenesis and blood circulation in late stages of liver growth in zebrafish

Requirement of vasculogenesis and blood circulation in late stages of liver growth in zebrafish Background: Early events in vertebrate liver development have been the major focus in previous studies, however, late events of liver organogenesis remain poorly understood. Liver vasculogenesis in vertebrates occurs through the interaction of endoderm-derived liver epithelium and mesoderm-derived endothelial cells (ECs). In zebrafish, although it has been found that ECs are not required for liver budding, how and when the spatio-temporal pattern of liver growth is coordinated with ECs remains to be elucidated. Results: To study the process of liver development and vasculogenesis in vivo, a two-color transgenic zebrafish line Tg(lfabf:dsRed; elaA:EGFP) was generated and named LiPan for liver-specific expression of DsRed RFP and exocrine pancreas-specific expression of GFP. Using the LiPan line, we first followed the dynamic development of liver from live embryos to adult and showed the formation of three distinct yet connected liver lobes during development. The LiPan line was then y1 crossed with Tg(fli1:EGFP) and vascular development in the liver was traced in vivo. Liver vasculogenesis started at 55–58 hpf when ECs first surrounded hepatocytes from the liver bud surface and then invaded the liver to form sinusoids and later the vascular network. Using a novel non-invasive and label-free fluorescence correction spectroscopy, we detected blood circulation in the liver starting at ~72 hpf. To analyze the roles of ECs and blood circulation in liver development, both cloche mutants (lacking ECs) and Tnnt2 morphants (no blood circulation) were employed. We found that until 70 hpf liver growth and morphogenesis depended on ECs and nascent sinusoids. After 72 hpf, a functional sinusoidal network was essential for continued liver growth. An absence of blood circulation in Tnnt2 morphants caused defects in liver vasculature and small liver. Conclusion: There are two phases of liver development in zebrafish, budding and growth. In the growth phase, there are three distinct stages: avascular growth between 50–55 hpf, where ECs are not required; endothelium-dependent growth, where ECs or sinusoids are required for liver growth between 55–72 hpf before blood circulation in liver sinusoids; and circulation-dependent growth, where the circulation is essential to maintain vascular network and to support continued liver growth after 72 hpf. Page 1 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 blood flow act in sequence to sustain the continued Background Hepatogenesis in zebrafish has been intensively studied in growth of liver in late development. recent years and these studies indicate that the early stages are rather similar in mice and zebrafish [1-9]. As detected Results by expression of ceruloplasmin [2] and GFP expression in Liver morphogenesis in two-color LiPan transgenic line the gut-GFP transgenic line [5], the liver anlagen initially The two-color LiPan transgenic line was generated by co- appears as a small compact protrusion of the intestinal injection of the dsRed RFP reporter gene under a liver-spe- rod at ~30 hpf (hours postfertilization) on the left side of cific lfabp promoter [16] and the GFP reporter gene under the embryo. By 2 dpf (days postfertilization), the liver bud an exocrine pancreas-specific elaA promoter [17]. The starts to enlarge and forms the hepatic duct connecting the LiPan transgenic line allowed simultaneous monitoring intestinal bulb primordium to the liver, which marks the of the temporal and spatial development of two major end of the budding phase and the beginning of the growth digestive organs in live embryos/fish, the RFP-expressing phase [5,10]. Despite these progresses, the late events of liver and GFP-expressing exocrine pancreas. Generally, the hepatogenesis and the underlying developmental mecha- red fluorescence appeared in the position of the liver bud nisms are not fully understood. A further analysis of liver from the left side at 48–53 hpf (Figure 1A) and the green development during the growth period will complement fluorescence appeared in position of the exocrine pan- existing studies and will be highly desirable as during this creas at the right side of the dorso-anterior intestine of the period the liver develops its vasculature and becomes embryos at 67–72 hpf with its anterior part (head pan- rd functional. creas) at the level of the 3 somite and its posterior part th (tail pancreas) at the level of the 6 somite. At 72 hpf, the Liver organogenesis in vertebrates coincides with the liver actively grew and further expanded laterally and st appearance of ECs adjacent to the endoderm. In mice, antero-ventrally beyond the 1 somite (the first or left before the formation of functional blood vessels, ECs pro- lobe, Figure 1B), it remained its initial shape and was vide a very early morphogenetic signal to the liver bud. It restricted to the left side of the body. At 80–84 hpf, the -/- has been shown that in Vegfr2/Flk-1 mouse embryos that liver rapidly expanded in several directions: anteriorly lack ECs, the liver undergoes the initial step of hepatic towards the ear vesicle and ventrally across the midline to specification and forms a multi-layered epithelium the right side of the body (Figure 1C), where it formed the anlage, but further liver morphogenesis fails prior to mes- second lobe (the right or gall-bladder lobe, Figure 1D). enchyme invasion [11]. In zebrafish cloche mutant Soon after, hepatocytes of the right lobe established tight embryos in which EC differentiation is disrupted at the contact with the gall bladder (80–96 hpf), and a yellow- stage prior to Vegfr2/Flk-1 expression [12,13], liver bud- greenish substance, probably bile, appeared in the intesti- ding and differentiation have been found to proceed nor- nal bulb (not shown), suggesting that at least some hepa- mally [5]. However, the interaction between vasculature tocytes were sufficiently mature to function in food and liver, as well as the role of ECs and growing vessels in digestion. late hepatogenesis in zebrafish, remain largely unclear. At 120 hpf, the left lobe spread even more anteriorly to the To investigate the late events of liver development and the mid-ear and posteriorly along the gut (Figure 1E). While roles of vasculogenesis and blood flow, in the present anteriorly both lobes were almost at the same A-P level study, a two-color transgenic line Tg(lfabf:ds-Red; and in contact with the pericardial cavity, posteriorly the elaA:EGFP) named LiPan for Liver- and Pancreas-specific right lobe was much shorter due to the presence of the gall transgene expression, was generated and it displayed bladder and pancreas immediately posterior to it (Figure strong expression of Ds-Red RFP (red fluorescent protein) 1F). The size of both liver and exocrine pancreas varied in the liver and GFP in the exocrine pancreas. To visualize slightly, but at this A-P level the liver and pancreas were liver vasculogenesis, the LiPan line was crossed with invariably projected onto opposite sides of the body as y1 Tg(fli1:EGFP) [14]. Liver vasculogenesis and growth was compact and separated organs (Figure 1D, F, G). traced in embryos of compound transgenics on wild type and mutant backgrounds using single- and multi-photon Between 5–10 dpf, the liver continued to grow in size. laser scanning microscopy. By taking the advantage of Around 15 dpf, the ventral most portion of the liver began recently developed non-invasive method of fluorescence expanding posteriorly (Figure 1H, I), leading to the for- correlation spectroscopy (FCS) [15], the initiation of mation of the third flat ventral lobe caudally to the first blood flow in liver vessels in live zebrafish embryos was two lobes. The timing of development of the third lobe determined. Finally, by analyses of the EC-less cloche varied between 15–20 dpf. After 15 dpf, the exocrine pan- mutant and blood flow defective Tnnt2 morphants, we creas was enlarged mostly posteriorly following the demonstrated the interaction of ECs with hepatocytes and looped intestine [17]. We observed the exocrine pancreas in the adult zebrafish to be a less compact structure com- Page 2 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 LiPan lateral 50 hpf 72 hpf nt A B nt First lobe LiPan 96 hpf lateral cross LR C D p sb Second LL RL lobe LiPan 120 hpf lateral dorsal gb nt RL sb pg LL LL in LiPan 120 hpf cross montage lateral view 10 dpf G H sb nt LL LR lateral view 15 dpf pi I sb LL Third in lobe LL in VL Liv Figure 1 er development in LiPan transgenic zebrafish Liver development in LiPan transgenic zebrafish. (A, B) Initial RFP expression in the liver starts at 48–53 hpf (A) and GFP expression in the exocrine pancreas starts at 67–72 hpf (B). (C) Expression of RFP and GFP at 96 hpf. (D) Cross section of a 96 hpf larvae shows RFP-positive liver. GFP-expressing exocrine pancreas is faintly visible (arrow). The dotted line repre- sents the midline of the larva and left (L) and right (R) sides are indicated. (E, F) Expression of RFP and GFP in the larva at 120 hpf: lateral (E) and dorsal view (F). The dotted horizontal line in (F) represents the midline with the right side at the top. The solid vertical line represents the plan of the cross section in (G). (G) A cross section to illustrate morphology of internal organs and expression of transgenes right (R) sides are indicated. (H, I) Lateral view of RFP-expressing liver at 10 dpf (H) and in a 120-hpf larva. The dotted line represents the midline of the larva and left (L) and 15dpf (I). Abbreviations: e, eye; gb, gall bladder; in, intestine; L, liver; LL, left lobe; RL, right lobe; nt, notochord; p, pancreas; pi, principal islet; pg, pigment; s, somite, VL, ventral lobe. In all whole mount images anterior is towards the left. Scale bars, 125 μm. Page 3 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 y1 pared to that in the larvae. GFP-expressing exocrine tissue transgenic line Tg(fli1:EGFP) in which GFP was specifi- y1 was consistently observed only on one side of the intesti- cally expressed in ECs [14]. In LiPan/Tg(fli1:EGFP) embryos, we directly observed events of liver vasculogen- nal mesentery along all three intestinal loops and it never surrounded the intestine (21 male and female LiPan esis from the late liver budding stage and during the liver zebrafish, Figure 2A, B), [17]. While the liver in general growth phase. After the liver budding stage was complete occupied the anterior region of the body cavity, exocrine at 50 hpf, distinct GFP-expressing ECs bordered the liver. pancreas occupied the posterior part of the body cavity These ECs probably represented components of nascent (Figure 2B). In addition, liver-expressed RFP and exocrine branches from the subintestinal vessels [5,18]. At this pancreas-expressed GFP could be observed in live adult stage, hepatocytes on the surface and inside the liver bud LiPan zebrafish under a fluorescent microscope (Figure remained tightly interconnected and showed no obvious 2C, D). morphological organization as observed by confocal microscopy (Figure 3A–F) and histological staining (not Formation of sinusoidal network in the zebrafish liver shown). During the growth phase at 55–58 hpf, GFP- To observe in vivo how and when the ECs developed the expressing ECs edged the liver completely and sprouts of liver vascular network, we crossed the LiPan line with the ECs contacted seemingly less associated superficial hepa- LiPan 1.5 mpf three liver lobes 4 mpf, liver and pancreas LL LL in VL VL in p in RL RL in left side right side Adult LiPan CD LL RL in p p VL in Liv Figure 2 er and pancreas in adult LiPan fish Liver and pancreas in adult LiPan fish. (A, B) Ventral view of the three liver lobes (red) in the dissected LiPan zebrafish at 1.5 mpf (A) and 4 mpf (B). The three lobes of liver are indicated by LL, left lobe; RL, right lobe; and VL, ventral lobe. The exo- crine pancreas (p) is in green. (C, D) lateral view of LiPan adult: the left side (C) and the right side (D). Other abbreviation: in, intestine. Scale bars, 1500 μm. Page 4 of 15 (page number not for citation purposes) B BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 tocytes and penetrated between them (Figure 3G, H), allowed measurement of blood flow in sinusoids of whereas inner hepatocytes remained tightly connected zebrafish liver in vivo during development. While first (Figure 3H, I). sinusoids were already formed in liver at around 68–70 hpf, no blood flow was detected in liver sinusoids within As sprouts of ECs penetrated between the surface hepato- a depth of 70 μm from the liver surface in all analyzed lar- cytes of the liver, some RFP-negative areas appeared vae (n = 7, Figure 4A), but in each of them blood flow was between inner layers of hepatocyte and a distinctive daisy detected in trunk vessels by FCS at this developmental pattern of hepatocytes around RFP-negative areas became stage. Soon after, at 72–75 hpf, we detected initiation of visible in the innermost liver layers (Figure 3J). Subse- blood flow in external sinusoids of dorso-lateral part of quently, ECs penetrate deeper and appeared between liver parenchyma in three out of seven analyzed points internal layers of hepatocytes and RFP-negative areas were (Figure 4B). It is likely that these sinusoids were already extended to near the liver surface layers (Figure 3K). Thus, connected to the supraintestinal artery (SIA) and subintes- the timing of ECs and hepatocyte interaction seems corre- tinal veins. It seemed that the time course of initiation of lated with the daisy organization of hepatocytes around blood flow in liver sinusoids correlated well with the the RFP-negative areas. The RFP-negative areas might rep- development of intestinal vessels [8]. After initiation of resent ductal cells previously identified in zebrafish liver the circulation, the mesh of sinusoids in the liver devel- at around 60 hpf [19]. While the liver extended laterally oped extremely fast and the liver increased significantly in and anteriorly, the daisy pattern of hepatocytes was evi- size between 84–120 hpf. By 120 hpf, blood flow was dent in nearly all liver layers and ECs appeared around the detected in 8 out of 10 points located at different confocal daisy patterned hepatocytes (Figure 3K). However, in planes as deep as 70 μm in both left and right liver lobes some cases, we also observed ECs inside the daisy struc- (Figure 4C). Due to a limit of light penetration, the detec- ture of hepatocytes as well (not shown at this time point tion of blood flow was restricted to 70–80 μm in depth but see Figure 3M) and it seemed that ECs and hepato- from the liver surface. At around 120 hpf, liver vasculo- cytes moved towards each other. Between 68–72 hpf, ECs genesis was almost complete and circulation was present formed the structure of the first sinusoids. The sinusoids in almost the entire liver when measured in left and right were 5–8 μm in diameter and contributed to the substan- liver lobes. tial increase of the liver in size. Up to this stage, ECs were arranged into non-functional vessels as no erythrocytes Collectively, our data suggested three distinct stages of were detected within the liver (Figure 3I, L). liver vasculogenesis: first, establishment of contact between ECs and hepatocytes at approximately 55–58 hpf After initiation of vasculogenesis, the liver grew rapidly (Figure 4D); second, formation of first sinusoids at and by 80–84 hpf it extended across the midline to the around 58–72 hpf (Figure 4E); and third, initiation of right side. Sinusoids were enlarged to 8–10 μm in diame- blood flow in first sinusoids and formation of sinusoidal ter (Figure 3M). At 96 hpf, the liver was extended further. network (Figure 4F). At 120 hpf, confocal 3D projections At 120–130 hpf, the diameter of sinusoids was around revealed a dense penetrating vascular network and forma- 12–16 μm, while hepatocytes of liver parenchyma were tion of the primary hepatic portal vein (HPV), which around 10–14 μm in diameter. Sequential endothelium- drained directly into the liver (18; Figure 4F). epithelium interaction during liver vasculogenesis signifi- cantly altered the structure of liver parenchyma. In both In adult zebrafish, the pattern of sinusoidal network was left and right liver lobes hepatocytes were surrounded by comparable to that in larvae at 5–6 dpf. A confocal analy- endothelia and almost every hepatocyte was associated sis of liver parenchyma in 1.5-mpf live LiPan/ y1 with a sinusoid (Figure 3N, also see Figure 4F). Histologi- Tg(fli:EGFP) fish revealed a dense penetrating vascular cal analysis confirmed the presence of blood cells in sinu- network with long stretches of sinusoids separating two soids at this stage (Figure 3O). rows of hepatocytes (Figure 4G, H). This pattern of sinu- soidal network was comparable in all liver lobes (not Detection of blood flow in the embryonic liver shown). Although we observed an increase of sinusoids in size from the very beginning of their formation up to 120 hpf, The initial liver growth and organization was promoted by it remained unknown when they became functional in contact with ECs but not by blood circulation liver. Therefore, we used a recently described non-invasive Previously, it has been shown that there are two distinct and label-free approach of fluorescence correlation spec- phase of liver development in the zebrafish, budding troscopy (FCS) to determine the timing of blood flow ini- phase from 24 hpf to 50 hpf and growth phase after 50 tiation in vessels of zebrafish embryonic liver [15]. Using hpf. It seems that ECs are not essential for the initiation of y1 the LiPan/Tg(fli1:EGFP) embryos, the FCS method liver budding as the liver buds normally in cloche mutants Page 5 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 Li Figure 3 ver vasculogenesis y1 Liver vasculogenesis. (A-C) Left-lateral confocal live images of liver of LiPan/Tg(fli1:EGFP) larva at 50 hpf. Discrete ECs frame outer layers of the liver bud (A-C) while no ECs are present in the inner layer (D-F). Panels A and D are images under a GFP filter, panels B and E under a RFP filter, and panels C and F are combinations of GFP and RFP images. (G, H) Left-lateral y1 confocal live images of LiPan/Tg(fli1:EGFP) liver at 55 hpf: external liver layers (G) and all liver layers (H). ECs establish contact and enter superficial hepatocytes layer. (I) Hematoxylin and eosin (H&E) staining of a cross section of an embryo at 55 hpf to show tightly packed hepatocytes inside the liver. (J, K, M, N) Confocal sections of internal liver layers at 58 hpf (J), 70 hpf (K), 84 hpf (M) and 120 hpf (N). At 58 hpf, RFP negative areas in the innermost liver layers (arrowhead) and a daisy pattern of hepa- tocytes are formed around this areas (J). At 70 hpf, the daisy pattern of hepatocytes around RFP negative areas extended to the superficial layers and ECs surround the daisy clusters of hepatocytes from outside (K). In some cases, ECs are inside of daisy clusters of hepatocytes (M). By 120 hpf, the size of sinusoids significantly increases (N). (L, O) H&E staining of cross sec- tions of embryos at 70 hpf (L) and 120 hpf (O). Note a daisy pattern of hepatocytes in (L) and erythrocytes with pink cyto- plasm in liver sinusoids (O). Abbreviations: dph, daisy pattern of hepatocytes; ec, endothelial cells; e, erythrocyte; h, hepatocyte; in, intestine; nt, notochord; L, liver; s, sinusoid; y, yolk. In all whole mount images anterior is towards the left. Scale bars represent 625 μm except for Panels (I, L, O), where the scale bars are 125 μm. Page 6 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 FLI left lateral 68 hpf 75 hpf 120 hpf D A B C Blood flow L is absent by FCS l + + LFABP / FLI 55 hpf 75 hpf 120 hpf D E F ec ec l HPV in LFABP / FLI 1.5 mpf G H G’ Figure 4 Assessment of blood circulation in the developing liver of zebrafish Assessment of blood circulation in the developing liver of zebrafish. (A-C) Left lateral 3D confocal sections of liver at the level of sinusoids where blood flow was measured. In the most outer liver layers where the first sinusoids were formed, blood flow was absent at 68 hpf (A). Blood flow was detected in sinusoids of external dorso-lateral part of liver parenchyma at 75 hpf (B) and in all measured sites of liver sinusoids at 120 hpf (C). Red cross, points of measurement at the different focal dis- tance as deep as 70–80 μm from the liver surface; "No" sign, points in sinusoids where no blood flow was detected. The edges of livers are marked by dash lines. Arrows indicate the focused sinusoids for FCS measurement in the picture. (D-F) Confocal projections show three stages of liver vasculogenesis. ECs start to contact the surface layer of hepatocytes in the liver bud at 55 hpf (D), first sinusoids form between external layers of hepatocytes at 75 hpf (E) and have well developed sinusoidal net- work at 120 hpf (F). Arrows are endothelial cells and sinusoids; arrowheads are clusters of hepatocytes. (G, H) Confocal images demonstrate the sinusoidal network (green) in the liver of 1.5-month-old fish. (H) is a 2x blow-up of the area defined by the white box (G') to show the sinusoidal network and two rows of hepatocytes between two neighboring sinusoids as indi- cated by yellow lines. Abbreviations: ec, endothelial cells; in, intestine, L, liver; HPV, hepatic portal vein. In all images anterior is towards the left-hand side. Scale bars are 625 μm in (A-B) and 300 μm in (D-H). Page 7 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 that lack ECs [5]. By taking advantages of LiPan/cloche the arrest of liver growth in LiPan/cloche larvae became mutant, we confirmed that at 50 hpf the liver bud in cloche more obvious (Figure 5D–E). Concurrently, the confocal y1 /cloche did not show is similar to that in wild type siblings (not shown). The analysis of LiPan/Tg(fli1:EGFP) livers in LiPan/cloche embryos expanded normally during GFP-expressing ECs associated with hepatocytes and there early growth phase; at 55 hpf, they were similar to those was no obvious daisy pattern of hepatocyte organization of controls. However, at 60 hpf, while the size of the (Figure 5H) as observed in controls (Figure 5G). Only in LiPan/cloche and wild type embryos was comparable, the the innermost layers of cloche liver did we observe a few size of liver was noticeably smaller in LiPan/cloche RFP-negative areas among RFP-positive hepatocytes (Fig- embryos than that in controls (Figure 5A, B). At 68–72 hpf ure 5H) which could represent ductal cells detected in the when the first sinusoid was formed in wild type embryos, cloche mutants at 70 hpf [19]. Thus, the absence of ECs Role Figure 5 of endothelia in liver development y1 Role of endothelia in liver development. (A-F) Left-lateral views: liver morphogenesis in live LiPan/Tg(fli1:EGFP) larvae; -/- wild type (A, D), LiPan/clo mutants (B, E) and LiPan/Tnnt2 morphants (C, F) at 60 hpf (A-C) and 70 hpf (D-F). Dorsal views of the liver region of the same embryos are shown as inserts in each panel and dash lines indicate the midline with the right side -/- at the top. Note that the liver in LiPan/clo mutants is significantly reduced compared to that in controls and Tnnt2 morphants. The pericardial edemas in both clo-/- mutant and tnnt2 morphant are indicated by arrows (B, C, E, F). (G-I) Left-lateral confo- y1 -/- cal in vivo projections of liver of LiPan/Tg(fli1:EGFP) larvae in wild type (G), clo mutant (H) and Tnnt2 morphant (I) back- grounds. Abbreviations: bd, bile duct; dph, daisy pattern of hepatocytes; ec, endothelial cells; e, ear; h, hepatocytes; l, liver; s, sinusoid. In all images, anterior is towards the left. Scale bars, 125 μm in (A-F) and 625 μm in (G-I). Page 8 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 contact in cloche mutants might significantly affect the size tocytes were tightly attached to each other without obvi- of liver and this effect was even more obvious during the ous morphological organization (n = 8; Figure 6H). The y1 liver of LiPan/Tg(fli:EGFP) /Tnnt2 morphants at the cor- formation of first vessels. Apart from the reduction in size, the structural organization of hepatocytes in cloche liver responding stage displayed similarly poorly organized was less obvious than that in non-cloche siblings (Figure hepatocytes with only remnants of ECs in the external lay- 5G, H). It appears that ECs through their contact with ers of liver (Figure 6I) in contrast to the well developed hepatocytes provided some specific morphogenetic sig- sinusoids and organized hepatocytes in liver of LiPan/ y1 nals for hepatocyte organization and liver growth. Tg(fli:EGFP) siblings (Figure 5G). Thus, by 120 hpf, the absence of circulation in Tnnt2 morphants causes liver To further investigate the role of ECs in liver development, deficiency comparable to that of cloche mutants lacking we introduced the Tnnt2 morphants, which phenocopied ECs (Figure 5G–L), indicating a vital role of blood circula- the zebrafish silent heart (sih) mutant with defects in car- tion in formation of the sinusoidal network in the liver diac contractility and blood circulation [20,21,2]. We and liver growth in late development. found that in LiPan/Tnnt2 morphants, the liver bud formed (as detected by in situ hybridization with the cer- According to Field et al. [5] the liver development in uloplasmin probe, not shown) and during early growth zebrafish consists of two phases – budding and growth. phase was about the same size as that in controls (323/ Our current study further defined the timing of formation 323 morphants; Figure 5A, C). A confocal examination of of three liver lobes during growth phase (Figure 7A). In y1 the liver of LiPan/Tg(fli1:EGFP) /Tnnt2 morphants combination with in vivo analysis of liver vasculogenesis revealed normal progression of the initial stage of EC and determination of initiation of blood flow in the liver interaction with hepatocytes at 55–58 hpf (not shown). of live embryos by FCS, we also defined three phases of Tnnt2 morphants suffered from a cardiac edema. Despite liver vasculogenesis: 1) contact of ECs and hepatocytes, 2) that increased cardiac edema was almost comparable to formation of nascent sinusoids, 3) initiation of blood cir- that of cloche, in contrast to the small liver in cloche, livers culation and formation of the vascular network in liver of Tnnt2 morphants (96%, 310/323) expanded notably (Figure 7B). Furthermore, by analysis of liver vasculogen- up to 70 hpf and had about same size and shape as in con- esis and liver growth in wild type as well as in cloche trols (Figure 5D, F). Based on the examination of ~400 mutant and Tnnt2 morphant larvae, we provided addi- cloche mutants and > 300 morphants, we failed to see any tional details to define three distinct stages of liver growth: correlation between the volume of edema and size of 1) avascular growth, which could be considered a transi- y1 liver. Moreover, the livers in LiPan/Tg(fli1:EGFP) /Tnnt2 tion from the budding phase to the growth phase, 2) morphants had similar morphological organization of endothelium-dependent growth, and 3) circulation- hepatocytes to those in non-morphant controls, in con- dependent growth (Figure 7C). trast to the absence of prominent daisy organization of y1 hepatocytes in LiPan/Tg(fli1:EGFP) /cloche mutants (Fig- Discussion Liver morphogenesis in LiPan transgenic zebrafish ure 5G–I). Thus, it seems that in embryonic zebrafish the initial liver growth and organization is promoted by the Although the early stages of endodermal organogenesis in contact with ECs but not by blood circulation, the zebrafish have been a subject of intense studies [1-9], some of the late developmental events in liver growth and Importance of blood circulation for liver vasculogenesis morphogenesis are still relatively poorly understood. To and growth fill this gap, we re-evaluated liver morphogenesis in the After the initiation of hepatic blood circulation at 72 hpf, zebrafish by focusing mainly on late events. We generated the liver in wild type larvae was enlarged laterally, ven- a two-color transgenic zebrafish line, LiPan, that expresses trally, and across the midline. In LiPan/Tnnt2 morphants, dsRed RFP specifically in the liver and GFP specifically in though the liver crossed the midline and extended moder- the exocrine pancreas to observe how liver growth is coor- ately by 75 hpf, we observed significant reduction of liver dinated with that of pancreas during development in live size in 95% (282/297) morphant larvae by 80 hpf (Figure embryos/larvae and adult zebrafish. The expression pat- 6A–C). It seemed that the liver growth in Tnnt2 mor- terns of the two reporter genes are essentially identical to phants was arrested in late development (Figure 6D, F). a combination of expression patterns in the two previ- The lack of blood circulation in Tnnt2 morphants eventu- ously reported transgenic lines, Tg(lfabp:egfp) [16] and ally resulted in vascular regression. The liver phenotype of Tg(elaA:egfp) [17], both of which express only GFP in the LiPan/Tnnt2 morphants progressively became similar to liver and exocrine pancreas respectively. The LiPan that of LiPan/cloche mutants. A confocal analysis of LiPan/ zebrafish expresses two distinct fluorescent proteins mark- y1 Tg(fli:EGFP) /cloche mutants revealed a compact arrange- ers with high intensity, rendering it an excellent experi- ment of hepatocytes similar to that observed in LiPan/ mental tool for detailed analysis of liver and pancreas y1 Tg(fli:EGFP) siblings at the liver budding stage, i.e. hepa- organogenesis. Page 9 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 Role of Figure 6 circulation in liver development Role of circulation in liver development. (A-F) Left-lateral views: liver morphogenesis in live LiPan wild type (A, D), LiPan/ -/- clo mutants (B, E) and LiPan/Tnnt2 morphants (C, F) at 80 hpf (A-C) and 120 hpf (D-F). Dorsal views of the liver region of the same embryos are shown as inserts in each panel and dash lines indicate the midline with the right side at the top. Livers in clo /- mutants and Tnnt2 morphants are located more medial, lack the anterio-ventral and posterior expansion, and are significantly reduced in size compared to the livers in wild type sibling. The cardiac edema in both mutants and morphants is indicated by y1 -/- arrow. (G-I) Left-lateral confocal in vivo projections of liver of LiPan/Tg(fli1:EGFP) larvae at 120 hpf in wild type (G), clo -/ mutant (H) and Tnnt2 morphant (I) backgrounds. In clo mutants ECs are absent (H), whereas in tnnt2 morphants they are present only between hepatocytes of the outer layer (I). (J-L) High-resolution light micrographs of hepatic parenchyma of zebrafish larvae stained with H&E. In 120-hpf wild type sibling, hepatocyte tubules are separated by sinusoids containing eryth- -/- rocytes (J); in contrast, in 120-hpf clo mutant (K) and Tnnt2 morphant (L), hepatocytes are tightly connected to each other. Note two sinusoids separated by two rows of neighboring hepatocytes as defined by yellow lines. Abbreviations: ec, endothe- lial cells; e, ear; h, hepatocytes; l, liver; s, sinusoid. In all images, anterior is towards the left. Scale bars, 125 μm in (A-F) and 625 μm in (G-L). Page 10 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 first, left lobe forms at ~80 hpf; the second, right lobe at Zebrafish liver development ~96 hpf, and the third, ventral lobe at 15–20 dpf during liver growth phase (Figure 1). This is in contrast to many Budding Growth phase phase second third lobe other Teleostei species such as rainbow trout [24] and Liza first liver lobe formation lobe formation formation saliens Risso, Liza aurata Risso [25], where no lobulation 24 hpf 50 hpf 80 hpf 15 dpf was recognized. In some species such as Chondrichthyes and Dipnoi, two liver lobes were found [23]. Liver vasculogenesis Contact Formation Initiation of circulation and Endothelial cells and nascent sinusoids drive liver growth of ECs of nascent formation of sinusoidal network sinusoids and morphogenesis y1 24 50 55 58 72 120 hpf Using Lipan/Tg(fli1:EGFP) transgenic zebrafish, we traced liver vasculogenesis from the very beginning until the liver became functional and showed sinusoidal net- Stages of liver growth work in adult zebrafish. These transgenic lines in combi- Avascular Endothelium Circulation dependent growth nation with cloche mutant and Tnnt2 morphants, allowed growth dependent growth us to investigate developmental mechanisms involved in hpf liver vasculogenesis. Strong RFP expression in hepatocytes 24 50 55 72 120 and GFP expression in ECs allowed the observation of the cloche dynamic process of vasculogenesis in the liver from dis- TNNT2 (silent heart) crete ECs around the liver bud to the formation of func- tional sinusoids. Our study showed the dynamic Summary an Figure 7 d growth of developmental events during liver vasculogenesis formation of sinusoidal network in zebrafish liver and Summary of developmental events during liver vas- revealed some similarities and differences in liver vasculo- culogenesis and growth. (A) Timing of liver morphogene- genesis between zebrafish and other vertebrates. sis in zebrafish. According to Field et al. [5], the liver morphogenesis in zebrafish consists of two phases, budding and growth. Our current study showed the formation of Based upon morphological characteristics, liver develop- three liver lobs during the growth phase. (B) Timing of liver ment in zebrafish has been separated into two phases, vasculogenesis. Detailed analysis of liver vasculogenesis and budding (24–50 hpf) and growth (after 50 hpf). Hepato- precise determination of blood flow initiation in liver sinu- cytes differentiation takes place at the end of liver budding soids of live embryonic zebrafish have led us to define three phase and before growth phase and lfabp is the molecular phases of liver vasculogenesis: contact of ECs and hepato- marker specific to fully differentiated hepatocytes [5,34]. cytes, formation of nascent sinusoids, initiation of blood cir- It has been previously shown that ECs are adjacent to, but culation and formation of the vascular network in the liver. not encasing, the liver bud at 48 hpf [5]. In our current in (C) Stages of liver growth. Based on our analyses of liver vas- vivo analyses, we found that during avascular growth culogenesis and growth in wild type as well as in cloche stage, discrete ECs rim the liver bud completely and then mutant and Tnnt2 morphant larvae, we propose three stages physically interact with hepatocytes prior to blood vessels of liver growth: avascular growth as a transition to the growth phase, where ECs are not required; endothelium- formation similar to that in mice [11]. In mouse Vegfr2/ -/- dependent, and circulation-dependent growth stage. Approx- Flk-1 embryos that lack mature ECs, the multi-layered imate periods of the stages of liver vasculogenesis and liver epithelia forms but later fail to grow [11]. In growth are represented along the time line in hpf or dpf. The zebrafish cloche mutant embryos, which lack ECs, the size green time lines represent the period of liver vasculogenesis. of liver was found to be comparable to that of wild type siblings during avascular growth stage up to 55 hpf and the growth of liver was arrested during vascular stage (Fig- While morphological investigation of visceral organs in ure 7B, C). It appears that the role of ECs during liver mor- larvae and adult fish is often difficult due to similar colors phogenesis is conserved among vertebrates since ECs in of most of internal organs, the LiPan larvae and adults both zebrafish and mice provide a crucial growth stimulus offer a convenient way to discern not only the liver and to the hepatic tissue before formation and function of pancreas but also other internal organs due to the exclu- local vessels [11]. Previous observation of liver develop- sion of the two fluorescent organs. Liver occupies the cra- ment in the cloche mutant was limited to 48 hpf due to an nial region of the body cavity and the exocrine pancreas increased severity of the cardiac edema [5]. Using live by and large occupies the posterior region of this cavity. It LiPan/cloche mutant we were able to analyze liver devel- is clear that zebrafish, unlike some other fish, have sepa- opment at later time points. This study showed that dur- rated and distinct liver and pancreas organs and do not ing endothelium-dependent growth stage, the size of liver form hepatopancreas (Figure 2) [23]. We consistently in cloche mutants was significantly reduced compared to observed the formation of three distinct liver lobes: the that in controls and supported the previous hypothesis Page 11 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 that the liver of cloche mutant may be affected during Role of functional sinusoidal network for liver growth growth phase if the endothelium is important for liver After the initiation of blood flow in the first liver vessels, development in zebrafish [5]. Furthermore, we also circulation became essential for further liver growth. From showed that at the cellular level, liver morphogenesis in 72 to 120 hpf, the liver enlarges 8–10 times in size. The cloche mutant was also affected and a prominent daisy pat- developing sinusoidal network directly contributes to the tern of hepatocytes was absent. increased liver volume. Before the initiation of blood cir- culation in the liver, the size of sinusoids is 5–8 μm in To eliminate the possibility that the arrest of liver growth diameter, while with the blood flow in the liver, sinusoids in cloche mutants was due to the lack of blood circulation progressively increased in size up to 12–16 μm. Before the that may transport certain factors required for liver initiation of circulation, liver size and its early structural growth, we analyzed Tnnt2 morphants that lack blood cir- organization depend on the presence of ECs. However, culation. In contrast to the reduced and unorganized liver after its initiation, circulation becomes an essential factor in cloche mutants, the size of liver and the pattern of ECs for subsequent liver development and growth, as evident y1 and hepatocytes interaction in Tnnt2 morphants embryos by comparative analyses of LiPan/Tg (fli1:EGFP) /Tnnt2 y1 /cloche mutant (Fig- were similar to that of wild type siblings during both avas- morphant and LiPan/Tg(fli1:EGFP) cular and endothelium-dependent.growth stages. Thus, ure 5J–L). Therefore, the blood circulation stimulates ECs in zebrafish may provide some morphogenic signals development of the vascular network and this network important not only for liver growth but also for structural remarkably enlarges liver size because of its own volume organization of hepatocytes. as well as delivery of nutrients to support liver growth by cell proliferation. Further analysis of mutants with defects Despite similarities there are some differences in liver of vascular patterning and vessels maintenance could development between zebrafish and mice. In mice, where uncover additional critical factors involved in liver vascu- the liver is an early hematopoietic organ, endothelial- logenesis [28]. endoderm interaction is initiated during the early phase of liver budding, whereas in zebrafish this interaction The development of the hepatic vascular architecture is a starts much later and during the growth phase. Moreover, multistep process through the interaction of the two tis- -/- allele causes embry- mouse homozygous mutant Flk-1 sues, hepatocytes of endodermal origin and endothelia of onic lethality by E10.5 (1 day after initiation of hepato- mesoderm origin. Further to the model on early liver cytes migration into the surrounding septum transversum development proposed by Field et al. [5], we, based on mesenchyma) [26], but homozygous cloche mutants lack- new observation of developmental events during the liver ing ECs survive 6 days (almost 4 days after initiation of growth phase, proposed to divide the liver growth phase liver vasculogenesis). This time of development (5–6 dpf) into three distinct stages: avascular growth as a transition corresponds with the appearance of the adult form of the to the growth phase, where ECs are not required and liver heart and the transition from diffusive to convective oxy- extended by proliferation of hepatocytes; endothelium- gen supply. dependent growth, when liver grew due to proliferation of hepatocytes and ECs, and circulation-dependent growth In mice and human, ECs provide stimuli for hepatocytes stage, when the blood circulation stimulates development to outgrow towards septum transversum mesenchyme of the vascular network which increases liver size because [27,11]. Previously it has been proposed that vasculogen- of its own volume and releases nutrients to support cell esis in zebrafish is achieved by endothelial invasion of growth and proliferation (Figure 7). This model could be liver [5]. In contrast, our data suggested that both ECs and useful as a roadmap to design further experiments liver grow towards each other. As ECs rim the liver com- addressing the role of the key factors required for liver vas- pletely and contact seemingly less associated surface lay- culature development, the functions of signaling path- ers of hepatocytes, the liver extends simultaneously, ways and interactions between them during intriguing effectively moving hepatocytes towards endothelia. In events of liver vasculogenesis. mice, after growing into the septum transversum mesen- chyme, hepatocytes organize around already formed sinu- Conclusion soids. In zebrafish, discrete ECs contact liver and then In the present study, a two-color transgenic zebrafish line form nascent sinusoids. Formation of the daisy pattern of (LiPan) with RFP expression in the liver and GFP expres- hepatocytes and nascent sinusoids proceeds at the same sion in the exocrine pancreas was generated and the LiPan time. This interaction significantly changes the liver mor- line allowed us to analyze detailed liver development in phology, but molecular mechanisms underlying this live embryos and larvae. By crossing the LiPan line with y1 developmental process remain to be elucidated. Tg(fli1:EGFP) , we found that liver vasculogenesis started at 55–58 hpf when ECs first surrounded the hepatocytes Page 12 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 from the liver bud surface and then invaded the liver to Homozygous LiPan zebrafish were viable and had no vis- form sinusoids and later the vascular network. Fluores- ible phenotype. The LiPan line has been maintained in cence correction spectroscopy detected blood circulation our laboratories for over eight generations and co-expres- in the liver first at ~72 hpf. Analysis of cloche mutants and sion of RFP in the liver and GFP in the exocrine pancreas Tnnt2 morphants led us to conclude that both ECs and is always observed. Thus the two injected DNA constructs blood circulation are required for continued liver growth are likely co-integrated into the same chromosomal locus. and morphogenesis. Collectively, we propose to divide the growth phase of liver development in zebrafish into Microscopy three distinct stages: avascular growth between 50–55 hpf, To facilitate visualization of liver and exocrine pancreas of where ECs are not required; endothelium-dependent larval zebrafish in whole-mount preparations, pigmenta- growth, where ECs or sinusoids are required for liver tion of skin was inhibited by raising embryos and larvae growth from 55 hpf to 72 hpf before blood circulation in in egg water [29] containing 0.2 mM 1-phenyl-2-thiourea the liver sinusoids; circulation-dependent growth where (Sigma, USA). Microscopic observations and photogra- the circulation is essential to maintain vascular network phy of live embryos were performed using the dissecting and to support the continued liver growth and morpho- fluorescent microscope SZX12 (Olympus, Japan), com- genesis after 72 hpf. pound microscope Zeiss Axioscope 2 and confocal micro- scope Zeiss LSM510 (Zeiss, Germany). Three images were Methods taken at the same focal plane, using a DIC filter for trans- Zebrafish maintenance mitted light for the first, epifluorescence with a Rhod for Zebrafish were maintained in the fish facilities at the the second and FITC filter for the third. These three images Department of Biological Sciences, National University of were then superimposed using Zeiss AxioVision software Singapore (NUS) and the Institute of Molecular and Cell or Photoshop (Adobe, USA). Three-dimensional confocal Biology (IMCB) of Singapore according to established projections were generated using Zeiss LSM510 software protocols [29] and in compliance with Institutional Ani- (Zeiss, Germany). In all confocal studies, at each time mal Care and Use Committee (IACUC) guidelines. Devel- point, 5–8 embryos/larvae from random pairs were exam- opmental stages are presented in hour post fertilization ined. (hpf), day post fertilization (dpf) or month post fertiliza- tion (mpf). Blood flow detection by fluorescence correlation spectroscopy Microinjection and establishment of Tg(lfabf:ds-Red; Fluorescence correlation spectroscopy (FCS) is a single- elaA:EGFP) zebrafish line (LiPan) molecule sensitive fluorescence technique which can pro- Isolation of elaA promoter (1.9 kb) and construction of vide information about diffusion coefficient, concentra- the chimeric plasmid pElaA-EGFP have been described tion, microfluidic flow, etc. [30,31]. It is based on an previously [17] The liver-specific promoter (2.8 kb) autocorrelation analysis of fluorescence fluctuations from derived from the zebrafish liver fatty acid binding protein a small focal volume in the specimen that is defined by a gene was provided by Dr. G.-M. Her and was inserted into high numerical aperture objective and a small pinhole. pDsRed-Express-1 (Clontech, USA) to make the chimeric Autocorrelation functions are derived for different cases plasmid pLFABP-RFP. Both plasmids were linearized, such as 3D diffusion and microfluidic flow, and the mod- mixed with 0.25% phenol red solution (1:1:1) at a final els can be used to fit the experimental data. The two-flow concentration of 100 ng/μl of each plasmid. Microinjec- model to extract the diastolic and systolic blood flow tion was carried out at the 1–2 cell stage. The DNA solu- velocity at the same blood vessel in zebrafish larva was tion was injected into the boundary between the yolk and developed and the detailed setup was described previ- blastodisc. After microinjection, the embryos were main- ously [32]. The blood flow velocity in the liver sinusoids tained in egg water [29] with ~0.0005% methylene blue in of zebrafish larvae was obtained by FCS measurement at a 28.5°C incubator. Transgenic founders were screened by the points of interest after the confocal image acquisition observation of F1 embryos for RFP and GFP expression. which helps to locate the position of sinusoids in the liver. 444 injected embryos were raised to adult and 66 of them In this work, each larva was anesthetized with freshly pre- were screened for transgenics. Two of them were found to pared 0.05–0.1 mg/ml Tricaine (ethyl m-aminoboen- produce F1 embryos with strong liver-specific RFP expres- zoate, Sigma, Singapore) dissolved in egg water [29], sion and exocrine pancreas-specific GFP expression. Thus, immobilized in 1.5% low-melting-temperature agarose two stable transgenic lines were established and both (Invitrogen, Singapore) (agarose was dissolved in 0.05 showed standard Mendelian inheritance from F2 genera- mg/ml of Tricaine in egg water), in WillCo-dish glass bot- tion onwards. Since identical reporter gene expression tom dish (GW-3512, WillCo-Wells, The Netherlands), patterns were observed in the two lines, only one line, placed in a temperature-controlled environment and named LiPan, was used for further characterization. immediately proceed for measurements. For each larva, Page 13 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 blood flow was measured in a trunk vessel as a control Authors' contributions and then in the vessels of liver parenchyma at 7–10 ran- SK – made genetic crosses and selection, systematic anal- domly selected points at different depths of liver tissue up yses of organogenesis and vasculogenesis in transgenic to 80 μm. For each stage 3–5 larvae were measured. With and mutant fish, and wrote the manuscript; XP – designed this technique, we were able to measure blood flow inside and made the LiPan transgenic line; MGL – made confocal the liver parenchyma as deep as 70 μm from the liver sur- microscopic images; CLM – made genetic crosses; XP-ana- face. lyzed blood flow in liver; TW- developed blood flow measurement method; VK – developed the concept of the Generation of double (LiPan) and triple [Tg(lfabf:ds-Red; project, wrote and approved the manuscript; ZG – devel- elaA:EGFP; fli1: EGFP)] transgenic cloche mutants oped the concept of the project, designed the LiPan trans- To visualize vascular development in the liver in live genics, wrote and approved the manuscript. embryos, LiPan homozygotes were mated with y1 Tg(fli1:EGFP) homozygotes and triple transgenic Acknowledgements We are thankful to Dr. G.M. Her for the plasmid pLFABP-EGFP, Dr. B. embryos were generated. To analyze effect of endothelia y1 Weinstein for Tg(fli1:EGFP) transgenic line (through Drs. Z. Wen and R. on liver development and growth, we used cloche mutants Ge), Dr. D. Stainier for cloche mutant, Dr. K. Osborne for careful reading s5 (clo [point mutation allele]) which lack almost all and comments, and personnel of NUS and IMCB fish and histology facilities endothelial cells [5,12,33]. Both LiPan and LiPan/ for technical help and discussion. The financial support for this project came y1 Tg(fli1:EGFP) transgenic fish were crossed with cloche from Biomedical Research Council of Singapore. V.K. laboratory in the heterozygotes to transfer the transgenes into the cloche IMCB has been supported by the Agency for Science, Technology and mutants. After their progeny reached maturity, these Research of Singapore. fishes were crossed randomly to identify cloche heterozy- gotes that carry the transgenes. These fishes were crossed References 1. Warga R, Nüsslein-Volhard C: Origin and development of the to obtain homozygous mutants with double (LiPan) or zebrafish endoderm. Development 1999, 126:827-838. y1 triple [LiPan/Tg(fli1:EGFP) ] transgenic background and 2. Korzh S, Emelyanov A, Korzh V: Developmental analysis of cer- their development was monitored. uloplasmin gene and liver formation in zebrafish. Mech. Dev 2001, 103:137-139. 3. Duncan SA: Mechanisms controlling early development of the Tnnt2 morphants liver. Mech Dev 2003, 120:19-33. To analyze the role of circulatory defects in liver develop- 4. Wallace KN, Pack M: Unique and conserved aspects of gut development in zebrafish. Dev Biol 2003, 255:12-29. ment, Tnnt2 antisense morpholino oligonucleotide (5'- 5. Field HA, Ober EA, Roeser T, Stainier DY: Formation of the diges- CATGTTTGCTCTGATCTGACACGCA), which is targeted tive system in zebrafish. I. Liver morphogenesis. Dev Biol 2003, 253:279-290. the Tnnt2 translation start codon and phenocopies the 6. Ober EA, Holly AF, Stainier DYR: From endoderm formation to silent heart (sih) mutation [22], was obtained from Gene liver and pancreas development in zebrafish. Mech Dev 2003, Tools (USA). A total of 4 ng of Tnnt2 morpholino oligo- 120:5-18. 7. David N, Mourrain P, Rosa F: Endoderm formation in zebrafish. nucleotide was injected into 1–2 cell stage of LiPan or In Molecular Aspects of Fish and Marine Biology, Fish Development and y1 LiPan/Tg(fli1:EGFP) embryos and a stable non-contrac- Genetics: zebrafish and medaka Volume 2. Edited by: Gong Z, Korzh V. Singapore: World Scientific; 2004:424-453. tile heart phenotype was observed in 98% of injected 8. Dong PD, Munson CA, Norton W, Crosnier C, Pan X, Gong Z, Neu- embryos and larvae from 24 hpf to 7 dpf (their last day of mann CJ, Stainier DY: Fgf10 regulates hepatopancreatic ductal survival). Morphants without obvious motility defect and system patterning and differentiation. Nat Genet 2007, 39:397-402. lacking blood circulation (with pericardial edema) were 9. Li Z, Korzh V, Gong Z: Localized rbp4 expression in the yolk used for analysis of liver growth and vasculogenesis. syncytial layer plays a role in yolk cell extension and early liver development. BMC Dev Biol 2007, 7:e117. 10. Pack M, Solnica-Krezel L, Malicki J, Neuhauss SC, Schier AF, Stemple Histology DL, Driever W, Fishman MC: Mutations affecting development For cryosection, zebrafish embryos were ice-chilled and of zebrafish digestive organs. Development 1996, 123:321-328. 11. Matsumoto K, Yoshitomi H, Rossant J, Zaret J: Liver organogene- fixed with ice-cold 4% paraformaldehyde (PFA) in phos- sis promoted by endothelial cells prior to vascular function. phate-buffered saline (PBS) at 4°C overnight. After fixa- Science 2001, 294:559-563. tion, embryos were embedded in 1.5% Bacto agar 12. Liao W, Bisgrove BW, Sawyer H, Hug B, Bell B, Peters K, Grunwald DJ, Stainier DYR: The zebrafish gene cloche acts upstream of a containing 5% sucrose and incubated in 30% sucrose at flk-1 homologue to regulate endothelial cell differentiation. 4°C overnight. The embedded embryos were oriented and Development 1997, 124:381-389. 13. Thompson MA, Ransom DG, Pratt SJ, MacLennan H, Kieran MW, sectioned with a cryostat microtome (10 μm thickness). Detrich HW, Vail B, Huber TL, Paw B, Brownlie AJ, Oates AC, Fritz For paraffin sections, the embryos and larvae were fixed A, Gates MA, Amores A, Bahary N, Talbot WS, Her H, Beier DR, with either 4% PFA in PBS or Bouin's fixative, followed by Postlethwait JH, Zon LI: The cloche and spadetail genes differ- entially affect hematopoiesis and vasculogenesis. Dev Biol paraffin embedding and sectioning (4 μm). Serial sections 1998, 197:248-269. (from at least 3–5 embryos or larvae at each time point) 14. Lawson ND, Weinstein BM: In vivo imaging of embryonic vascu- were de-paraffinized, stained with hematoxylin-eosin, lar development using transgenic zebrafish. Dev Biol 2002, 248:307-318. dehydrated, and examined. Page 14 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 15. Pan X, Yu H, Shi X, Korzh V, Wohland T: Characterization of flow direction in microchannels and zebrafish blood vessels by scanning fluorescence correlation spectroscopy. J Biomed Opt 2007, 12:14034. 16. Her GM, Chiang CC, Chen WY, Wu JL: In vivo studies of liver- type fatty acid binding protein (L-FABP) gene expression in liver of transgenic zebrafish (Danio rerio). FEBS Lett 2003, 538:125-133. 17. Wan H, Korzh S, Li Z, Mudumana SP, Korzh V, Jiang YJ, Lin S, Gong Z: Analyses of pancreas development by generation of gfp transgenic zebrafish using an exocrine pancreas-specific elastaseA gene promoter. Exp Cell Res 2006, 312:1526-1539. 18. Isogai S, Horiguchi M, Weinstein BM: The vascular anatomy of the developing zebrafish: an atlas of embryonic and early lar- val development. Dev Biol 2001, 230:278-301. 19. Lorent K, Yeo SY, Oda T, Chandrasekharappa S, Chitnis A, Mattews R, Pack M: Inhibition of Jagged-mediated Notch signaling dis- rupts zebrafish biliary development and generates multi- organ defects compatible with an Alagille syndrome pheno- copy. Development 2004, 131:5753-5766. 20. Chen JN, Haffter P, Odenthal J, Vogelsang E, Brand M, van Eeden FJ, Furutani-Seiki M, Granato M, Hammerschmidt M, Heisenberg CP, Jiang YJ, Kane DA, Kelsh RN, Mullins MC, Nüsslein-Volhard C: Muta- tions affecting the cardiovascular system and other internal organs in zebrafish. Development 1996, 123:293-302. 21. Stainier DY, Fouquet B, Chen JN, Warren KS, Weinstein BM, Meiler SE, Mohideen MA, Neuhauss SC, Solnica-Krezel L, Schier AF, Zwartkruis F, Stemple DL, Malicki J, Driever W, Fishman MC: Muta- tions affecting the formation and function of the cardiovas- cular system in the zebrafish embryo. Development 1996, 123:285-292. 22. Sehnert AJ, Huq A, Weinstein BM, Walker C, Fishman M, Stainier DY: Cardiac troponin T is essential in sarcomere assembly and cardiac contractility. Nat Genet 2002, 31:106-110. 23. Gonzales G, Crespo S, Brusle J: Histo-cytological study of the liver in the cabrilla sea bass, Serranus cabrilla (Teleostei, Ser- ranidae) an available model for marine fish experimental studies. J Fish Biol 1993, 43:364-373. 24. Robertson OH, Wexler BC: Histopathological changes in the organs and tissues of migrating and spawning Pacific salmon (genus Oncorhunchus). Endocrinology 1960, 66:222-239. 25. Biagianti-Risbourg S: Fine structure of hepatocytes in juvenile grey mullets: Liza saliens Risso, L. ramada Risso and L. aurata Risso (Teleostei, Mugilidae). J Fish Biol 1991, 39:687-703. 26. Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu XF, Breit- man ML, Schuh AC: Failure of blood-island formation and vas- culogenesis in Flk-1-deficient mice. Nature 1995, 376:62-66. 27. Elias H, Sherrick JC: Morphology of the liver NY and London: Academic Press; 1969. 28. Jin SW, Herzog W, Santoro MM, Mitchell TS, Frantsve J, Jungblut B, Beis D, Scott IC, D'Amico LA, Ober EA, Verkade H, Field HA, Chi NC, Wehman AM, Baier H, Stainier DY: A transgene-assisted genetic screen identifies essential regulators of vascular development in vertebrate embryos. Dev Biol 2007, 307:29-42. 29. Westerfield M: The zebrafish book, a guide for the laboratory use of zebrafish (Danio rerio). Eugene, OR, University of Oregon Press; 2000. 30. Magde D, Elson E, Webb WW: Thermodynamic fluctuations in a reacting system – measurement by fluorescence correla- tion spectroscopy. Phys Rev Lett 1972, 29:705-708. 31. Magde D, Webb W, Elson E: Fluorescence correlation spectros- Publish with Bio Med Central and every copy. III. Uniform translation and laminar flow. Biopolymers scientist can read your work free of charge 1978, 17(2):361-376. 32. Pan X, Fok M, Foo W, Lim W, Liu P, Yu H, Maruyama I, Wohland T: "BioMed Central will be the most significant development for Multifunctional fluorescence correlation microscope for disseminating the results of biomedical researc h in our lifetime." intracellular microfluidic measurements. Rev Sci Instrum 2007, Sir Paul Nurse, Cancer Research UK 78:53711. 33. Stainier DYR, Weinstein BM, Detrich HW, Zon LI, Fishman MC: Your research papers will be: cloche, an early acting zebrafish gene, is required by both the available free of charge to the entire biomedical community endothelial and hematopoietic lineages. Development 1995, 121:3141-3150. peer reviewed and published immediately upon acceptance 34. Chen J, Ruan H, Ng SM, Gao C, Soo HM, Wu W, Zhang Z, Wen Z, cited in PubMed and archived on PubMed Central Lane DP, Peng J: Loss of function of def selectively up-regulates {Delta}113p53 expression to arrest expansion growth of yours — you keep the copyright digestive organs in zebrafish. Genes Dev 2005, BioMedcentral 19(23):2900-2911. Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 15 of 15 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Developmental Biology Springer Journals

Requirement of vasculogenesis and blood circulation in late stages of liver growth in zebrafish

Download PDF
15 pages

Loading next page...
 
/lp/springer-journals/requirement-of-vasculogenesis-and-blood-circulation-in-late-stages-of-Q0ncfoqGrx

References (40)

Publisher
Springer Journals
Copyright
Copyright © 2008 by Korzh et al; licensee BioMed Central Ltd.
Subject
Life Sciences; Developmental Biology; Animal Models; Life Sciences, general
eISSN
1471-213X
DOI
10.1186/1471-213X-8-84
pmid
18796162
Publisher site
See Article on Publisher Site

Abstract

Background: Early events in vertebrate liver development have been the major focus in previous studies, however, late events of liver organogenesis remain poorly understood. Liver vasculogenesis in vertebrates occurs through the interaction of endoderm-derived liver epithelium and mesoderm-derived endothelial cells (ECs). In zebrafish, although it has been found that ECs are not required for liver budding, how and when the spatio-temporal pattern of liver growth is coordinated with ECs remains to be elucidated. Results: To study the process of liver development and vasculogenesis in vivo, a two-color transgenic zebrafish line Tg(lfabf:dsRed; elaA:EGFP) was generated and named LiPan for liver-specific expression of DsRed RFP and exocrine pancreas-specific expression of GFP. Using the LiPan line, we first followed the dynamic development of liver from live embryos to adult and showed the formation of three distinct yet connected liver lobes during development. The LiPan line was then y1 crossed with Tg(fli1:EGFP) and vascular development in the liver was traced in vivo. Liver vasculogenesis started at 55–58 hpf when ECs first surrounded hepatocytes from the liver bud surface and then invaded the liver to form sinusoids and later the vascular network. Using a novel non-invasive and label-free fluorescence correction spectroscopy, we detected blood circulation in the liver starting at ~72 hpf. To analyze the roles of ECs and blood circulation in liver development, both cloche mutants (lacking ECs) and Tnnt2 morphants (no blood circulation) were employed. We found that until 70 hpf liver growth and morphogenesis depended on ECs and nascent sinusoids. After 72 hpf, a functional sinusoidal network was essential for continued liver growth. An absence of blood circulation in Tnnt2 morphants caused defects in liver vasculature and small liver. Conclusion: There are two phases of liver development in zebrafish, budding and growth. In the growth phase, there are three distinct stages: avascular growth between 50–55 hpf, where ECs are not required; endothelium-dependent growth, where ECs or sinusoids are required for liver growth between 55–72 hpf before blood circulation in liver sinusoids; and circulation-dependent growth, where the circulation is essential to maintain vascular network and to support continued liver growth after 72 hpf. Page 1 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 blood flow act in sequence to sustain the continued Background Hepatogenesis in zebrafish has been intensively studied in growth of liver in late development. recent years and these studies indicate that the early stages are rather similar in mice and zebrafish [1-9]. As detected Results by expression of ceruloplasmin [2] and GFP expression in Liver morphogenesis in two-color LiPan transgenic line the gut-GFP transgenic line [5], the liver anlagen initially The two-color LiPan transgenic line was generated by co- appears as a small compact protrusion of the intestinal injection of the dsRed RFP reporter gene under a liver-spe- rod at ~30 hpf (hours postfertilization) on the left side of cific lfabp promoter [16] and the GFP reporter gene under the embryo. By 2 dpf (days postfertilization), the liver bud an exocrine pancreas-specific elaA promoter [17]. The starts to enlarge and forms the hepatic duct connecting the LiPan transgenic line allowed simultaneous monitoring intestinal bulb primordium to the liver, which marks the of the temporal and spatial development of two major end of the budding phase and the beginning of the growth digestive organs in live embryos/fish, the RFP-expressing phase [5,10]. Despite these progresses, the late events of liver and GFP-expressing exocrine pancreas. Generally, the hepatogenesis and the underlying developmental mecha- red fluorescence appeared in the position of the liver bud nisms are not fully understood. A further analysis of liver from the left side at 48–53 hpf (Figure 1A) and the green development during the growth period will complement fluorescence appeared in position of the exocrine pan- existing studies and will be highly desirable as during this creas at the right side of the dorso-anterior intestine of the period the liver develops its vasculature and becomes embryos at 67–72 hpf with its anterior part (head pan- rd functional. creas) at the level of the 3 somite and its posterior part th (tail pancreas) at the level of the 6 somite. At 72 hpf, the Liver organogenesis in vertebrates coincides with the liver actively grew and further expanded laterally and st appearance of ECs adjacent to the endoderm. In mice, antero-ventrally beyond the 1 somite (the first or left before the formation of functional blood vessels, ECs pro- lobe, Figure 1B), it remained its initial shape and was vide a very early morphogenetic signal to the liver bud. It restricted to the left side of the body. At 80–84 hpf, the -/- has been shown that in Vegfr2/Flk-1 mouse embryos that liver rapidly expanded in several directions: anteriorly lack ECs, the liver undergoes the initial step of hepatic towards the ear vesicle and ventrally across the midline to specification and forms a multi-layered epithelium the right side of the body (Figure 1C), where it formed the anlage, but further liver morphogenesis fails prior to mes- second lobe (the right or gall-bladder lobe, Figure 1D). enchyme invasion [11]. In zebrafish cloche mutant Soon after, hepatocytes of the right lobe established tight embryos in which EC differentiation is disrupted at the contact with the gall bladder (80–96 hpf), and a yellow- stage prior to Vegfr2/Flk-1 expression [12,13], liver bud- greenish substance, probably bile, appeared in the intesti- ding and differentiation have been found to proceed nor- nal bulb (not shown), suggesting that at least some hepa- mally [5]. However, the interaction between vasculature tocytes were sufficiently mature to function in food and liver, as well as the role of ECs and growing vessels in digestion. late hepatogenesis in zebrafish, remain largely unclear. At 120 hpf, the left lobe spread even more anteriorly to the To investigate the late events of liver development and the mid-ear and posteriorly along the gut (Figure 1E). While roles of vasculogenesis and blood flow, in the present anteriorly both lobes were almost at the same A-P level study, a two-color transgenic line Tg(lfabf:ds-Red; and in contact with the pericardial cavity, posteriorly the elaA:EGFP) named LiPan for Liver- and Pancreas-specific right lobe was much shorter due to the presence of the gall transgene expression, was generated and it displayed bladder and pancreas immediately posterior to it (Figure strong expression of Ds-Red RFP (red fluorescent protein) 1F). The size of both liver and exocrine pancreas varied in the liver and GFP in the exocrine pancreas. To visualize slightly, but at this A-P level the liver and pancreas were liver vasculogenesis, the LiPan line was crossed with invariably projected onto opposite sides of the body as y1 Tg(fli1:EGFP) [14]. Liver vasculogenesis and growth was compact and separated organs (Figure 1D, F, G). traced in embryos of compound transgenics on wild type and mutant backgrounds using single- and multi-photon Between 5–10 dpf, the liver continued to grow in size. laser scanning microscopy. By taking the advantage of Around 15 dpf, the ventral most portion of the liver began recently developed non-invasive method of fluorescence expanding posteriorly (Figure 1H, I), leading to the for- correlation spectroscopy (FCS) [15], the initiation of mation of the third flat ventral lobe caudally to the first blood flow in liver vessels in live zebrafish embryos was two lobes. The timing of development of the third lobe determined. Finally, by analyses of the EC-less cloche varied between 15–20 dpf. After 15 dpf, the exocrine pan- mutant and blood flow defective Tnnt2 morphants, we creas was enlarged mostly posteriorly following the demonstrated the interaction of ECs with hepatocytes and looped intestine [17]. We observed the exocrine pancreas in the adult zebrafish to be a less compact structure com- Page 2 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 LiPan lateral 50 hpf 72 hpf nt A B nt First lobe LiPan 96 hpf lateral cross LR C D p sb Second LL RL lobe LiPan 120 hpf lateral dorsal gb nt RL sb pg LL LL in LiPan 120 hpf cross montage lateral view 10 dpf G H sb nt LL LR lateral view 15 dpf pi I sb LL Third in lobe LL in VL Liv Figure 1 er development in LiPan transgenic zebrafish Liver development in LiPan transgenic zebrafish. (A, B) Initial RFP expression in the liver starts at 48–53 hpf (A) and GFP expression in the exocrine pancreas starts at 67–72 hpf (B). (C) Expression of RFP and GFP at 96 hpf. (D) Cross section of a 96 hpf larvae shows RFP-positive liver. GFP-expressing exocrine pancreas is faintly visible (arrow). The dotted line repre- sents the midline of the larva and left (L) and right (R) sides are indicated. (E, F) Expression of RFP and GFP in the larva at 120 hpf: lateral (E) and dorsal view (F). The dotted horizontal line in (F) represents the midline with the right side at the top. The solid vertical line represents the plan of the cross section in (G). (G) A cross section to illustrate morphology of internal organs and expression of transgenes right (R) sides are indicated. (H, I) Lateral view of RFP-expressing liver at 10 dpf (H) and in a 120-hpf larva. The dotted line represents the midline of the larva and left (L) and 15dpf (I). Abbreviations: e, eye; gb, gall bladder; in, intestine; L, liver; LL, left lobe; RL, right lobe; nt, notochord; p, pancreas; pi, principal islet; pg, pigment; s, somite, VL, ventral lobe. In all whole mount images anterior is towards the left. Scale bars, 125 μm. Page 3 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 y1 pared to that in the larvae. GFP-expressing exocrine tissue transgenic line Tg(fli1:EGFP) in which GFP was specifi- y1 was consistently observed only on one side of the intesti- cally expressed in ECs [14]. In LiPan/Tg(fli1:EGFP) embryos, we directly observed events of liver vasculogen- nal mesentery along all three intestinal loops and it never surrounded the intestine (21 male and female LiPan esis from the late liver budding stage and during the liver zebrafish, Figure 2A, B), [17]. While the liver in general growth phase. After the liver budding stage was complete occupied the anterior region of the body cavity, exocrine at 50 hpf, distinct GFP-expressing ECs bordered the liver. pancreas occupied the posterior part of the body cavity These ECs probably represented components of nascent (Figure 2B). In addition, liver-expressed RFP and exocrine branches from the subintestinal vessels [5,18]. At this pancreas-expressed GFP could be observed in live adult stage, hepatocytes on the surface and inside the liver bud LiPan zebrafish under a fluorescent microscope (Figure remained tightly interconnected and showed no obvious 2C, D). morphological organization as observed by confocal microscopy (Figure 3A–F) and histological staining (not Formation of sinusoidal network in the zebrafish liver shown). During the growth phase at 55–58 hpf, GFP- To observe in vivo how and when the ECs developed the expressing ECs edged the liver completely and sprouts of liver vascular network, we crossed the LiPan line with the ECs contacted seemingly less associated superficial hepa- LiPan 1.5 mpf three liver lobes 4 mpf, liver and pancreas LL LL in VL VL in p in RL RL in left side right side Adult LiPan CD LL RL in p p VL in Liv Figure 2 er and pancreas in adult LiPan fish Liver and pancreas in adult LiPan fish. (A, B) Ventral view of the three liver lobes (red) in the dissected LiPan zebrafish at 1.5 mpf (A) and 4 mpf (B). The three lobes of liver are indicated by LL, left lobe; RL, right lobe; and VL, ventral lobe. The exo- crine pancreas (p) is in green. (C, D) lateral view of LiPan adult: the left side (C) and the right side (D). Other abbreviation: in, intestine. Scale bars, 1500 μm. Page 4 of 15 (page number not for citation purposes) B BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 tocytes and penetrated between them (Figure 3G, H), allowed measurement of blood flow in sinusoids of whereas inner hepatocytes remained tightly connected zebrafish liver in vivo during development. While first (Figure 3H, I). sinusoids were already formed in liver at around 68–70 hpf, no blood flow was detected in liver sinusoids within As sprouts of ECs penetrated between the surface hepato- a depth of 70 μm from the liver surface in all analyzed lar- cytes of the liver, some RFP-negative areas appeared vae (n = 7, Figure 4A), but in each of them blood flow was between inner layers of hepatocyte and a distinctive daisy detected in trunk vessels by FCS at this developmental pattern of hepatocytes around RFP-negative areas became stage. Soon after, at 72–75 hpf, we detected initiation of visible in the innermost liver layers (Figure 3J). Subse- blood flow in external sinusoids of dorso-lateral part of quently, ECs penetrate deeper and appeared between liver parenchyma in three out of seven analyzed points internal layers of hepatocytes and RFP-negative areas were (Figure 4B). It is likely that these sinusoids were already extended to near the liver surface layers (Figure 3K). Thus, connected to the supraintestinal artery (SIA) and subintes- the timing of ECs and hepatocyte interaction seems corre- tinal veins. It seemed that the time course of initiation of lated with the daisy organization of hepatocytes around blood flow in liver sinusoids correlated well with the the RFP-negative areas. The RFP-negative areas might rep- development of intestinal vessels [8]. After initiation of resent ductal cells previously identified in zebrafish liver the circulation, the mesh of sinusoids in the liver devel- at around 60 hpf [19]. While the liver extended laterally oped extremely fast and the liver increased significantly in and anteriorly, the daisy pattern of hepatocytes was evi- size between 84–120 hpf. By 120 hpf, blood flow was dent in nearly all liver layers and ECs appeared around the detected in 8 out of 10 points located at different confocal daisy patterned hepatocytes (Figure 3K). However, in planes as deep as 70 μm in both left and right liver lobes some cases, we also observed ECs inside the daisy struc- (Figure 4C). Due to a limit of light penetration, the detec- ture of hepatocytes as well (not shown at this time point tion of blood flow was restricted to 70–80 μm in depth but see Figure 3M) and it seemed that ECs and hepato- from the liver surface. At around 120 hpf, liver vasculo- cytes moved towards each other. Between 68–72 hpf, ECs genesis was almost complete and circulation was present formed the structure of the first sinusoids. The sinusoids in almost the entire liver when measured in left and right were 5–8 μm in diameter and contributed to the substan- liver lobes. tial increase of the liver in size. Up to this stage, ECs were arranged into non-functional vessels as no erythrocytes Collectively, our data suggested three distinct stages of were detected within the liver (Figure 3I, L). liver vasculogenesis: first, establishment of contact between ECs and hepatocytes at approximately 55–58 hpf After initiation of vasculogenesis, the liver grew rapidly (Figure 4D); second, formation of first sinusoids at and by 80–84 hpf it extended across the midline to the around 58–72 hpf (Figure 4E); and third, initiation of right side. Sinusoids were enlarged to 8–10 μm in diame- blood flow in first sinusoids and formation of sinusoidal ter (Figure 3M). At 96 hpf, the liver was extended further. network (Figure 4F). At 120 hpf, confocal 3D projections At 120–130 hpf, the diameter of sinusoids was around revealed a dense penetrating vascular network and forma- 12–16 μm, while hepatocytes of liver parenchyma were tion of the primary hepatic portal vein (HPV), which around 10–14 μm in diameter. Sequential endothelium- drained directly into the liver (18; Figure 4F). epithelium interaction during liver vasculogenesis signifi- cantly altered the structure of liver parenchyma. In both In adult zebrafish, the pattern of sinusoidal network was left and right liver lobes hepatocytes were surrounded by comparable to that in larvae at 5–6 dpf. A confocal analy- endothelia and almost every hepatocyte was associated sis of liver parenchyma in 1.5-mpf live LiPan/ y1 with a sinusoid (Figure 3N, also see Figure 4F). Histologi- Tg(fli:EGFP) fish revealed a dense penetrating vascular cal analysis confirmed the presence of blood cells in sinu- network with long stretches of sinusoids separating two soids at this stage (Figure 3O). rows of hepatocytes (Figure 4G, H). This pattern of sinu- soidal network was comparable in all liver lobes (not Detection of blood flow in the embryonic liver shown). Although we observed an increase of sinusoids in size from the very beginning of their formation up to 120 hpf, The initial liver growth and organization was promoted by it remained unknown when they became functional in contact with ECs but not by blood circulation liver. Therefore, we used a recently described non-invasive Previously, it has been shown that there are two distinct and label-free approach of fluorescence correlation spec- phase of liver development in the zebrafish, budding troscopy (FCS) to determine the timing of blood flow ini- phase from 24 hpf to 50 hpf and growth phase after 50 tiation in vessels of zebrafish embryonic liver [15]. Using hpf. It seems that ECs are not essential for the initiation of y1 the LiPan/Tg(fli1:EGFP) embryos, the FCS method liver budding as the liver buds normally in cloche mutants Page 5 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 Li Figure 3 ver vasculogenesis y1 Liver vasculogenesis. (A-C) Left-lateral confocal live images of liver of LiPan/Tg(fli1:EGFP) larva at 50 hpf. Discrete ECs frame outer layers of the liver bud (A-C) while no ECs are present in the inner layer (D-F). Panels A and D are images under a GFP filter, panels B and E under a RFP filter, and panels C and F are combinations of GFP and RFP images. (G, H) Left-lateral y1 confocal live images of LiPan/Tg(fli1:EGFP) liver at 55 hpf: external liver layers (G) and all liver layers (H). ECs establish contact and enter superficial hepatocytes layer. (I) Hematoxylin and eosin (H&E) staining of a cross section of an embryo at 55 hpf to show tightly packed hepatocytes inside the liver. (J, K, M, N) Confocal sections of internal liver layers at 58 hpf (J), 70 hpf (K), 84 hpf (M) and 120 hpf (N). At 58 hpf, RFP negative areas in the innermost liver layers (arrowhead) and a daisy pattern of hepa- tocytes are formed around this areas (J). At 70 hpf, the daisy pattern of hepatocytes around RFP negative areas extended to the superficial layers and ECs surround the daisy clusters of hepatocytes from outside (K). In some cases, ECs are inside of daisy clusters of hepatocytes (M). By 120 hpf, the size of sinusoids significantly increases (N). (L, O) H&E staining of cross sec- tions of embryos at 70 hpf (L) and 120 hpf (O). Note a daisy pattern of hepatocytes in (L) and erythrocytes with pink cyto- plasm in liver sinusoids (O). Abbreviations: dph, daisy pattern of hepatocytes; ec, endothelial cells; e, erythrocyte; h, hepatocyte; in, intestine; nt, notochord; L, liver; s, sinusoid; y, yolk. In all whole mount images anterior is towards the left. Scale bars represent 625 μm except for Panels (I, L, O), where the scale bars are 125 μm. Page 6 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 FLI left lateral 68 hpf 75 hpf 120 hpf D A B C Blood flow L is absent by FCS l + + LFABP / FLI 55 hpf 75 hpf 120 hpf D E F ec ec l HPV in LFABP / FLI 1.5 mpf G H G’ Figure 4 Assessment of blood circulation in the developing liver of zebrafish Assessment of blood circulation in the developing liver of zebrafish. (A-C) Left lateral 3D confocal sections of liver at the level of sinusoids where blood flow was measured. In the most outer liver layers where the first sinusoids were formed, blood flow was absent at 68 hpf (A). Blood flow was detected in sinusoids of external dorso-lateral part of liver parenchyma at 75 hpf (B) and in all measured sites of liver sinusoids at 120 hpf (C). Red cross, points of measurement at the different focal dis- tance as deep as 70–80 μm from the liver surface; "No" sign, points in sinusoids where no blood flow was detected. The edges of livers are marked by dash lines. Arrows indicate the focused sinusoids for FCS measurement in the picture. (D-F) Confocal projections show three stages of liver vasculogenesis. ECs start to contact the surface layer of hepatocytes in the liver bud at 55 hpf (D), first sinusoids form between external layers of hepatocytes at 75 hpf (E) and have well developed sinusoidal net- work at 120 hpf (F). Arrows are endothelial cells and sinusoids; arrowheads are clusters of hepatocytes. (G, H) Confocal images demonstrate the sinusoidal network (green) in the liver of 1.5-month-old fish. (H) is a 2x blow-up of the area defined by the white box (G') to show the sinusoidal network and two rows of hepatocytes between two neighboring sinusoids as indi- cated by yellow lines. Abbreviations: ec, endothelial cells; in, intestine, L, liver; HPV, hepatic portal vein. In all images anterior is towards the left-hand side. Scale bars are 625 μm in (A-B) and 300 μm in (D-H). Page 7 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 that lack ECs [5]. By taking advantages of LiPan/cloche the arrest of liver growth in LiPan/cloche larvae became mutant, we confirmed that at 50 hpf the liver bud in cloche more obvious (Figure 5D–E). Concurrently, the confocal y1 /cloche did not show is similar to that in wild type siblings (not shown). The analysis of LiPan/Tg(fli1:EGFP) livers in LiPan/cloche embryos expanded normally during GFP-expressing ECs associated with hepatocytes and there early growth phase; at 55 hpf, they were similar to those was no obvious daisy pattern of hepatocyte organization of controls. However, at 60 hpf, while the size of the (Figure 5H) as observed in controls (Figure 5G). Only in LiPan/cloche and wild type embryos was comparable, the the innermost layers of cloche liver did we observe a few size of liver was noticeably smaller in LiPan/cloche RFP-negative areas among RFP-positive hepatocytes (Fig- embryos than that in controls (Figure 5A, B). At 68–72 hpf ure 5H) which could represent ductal cells detected in the when the first sinusoid was formed in wild type embryos, cloche mutants at 70 hpf [19]. Thus, the absence of ECs Role Figure 5 of endothelia in liver development y1 Role of endothelia in liver development. (A-F) Left-lateral views: liver morphogenesis in live LiPan/Tg(fli1:EGFP) larvae; -/- wild type (A, D), LiPan/clo mutants (B, E) and LiPan/Tnnt2 morphants (C, F) at 60 hpf (A-C) and 70 hpf (D-F). Dorsal views of the liver region of the same embryos are shown as inserts in each panel and dash lines indicate the midline with the right side -/- at the top. Note that the liver in LiPan/clo mutants is significantly reduced compared to that in controls and Tnnt2 morphants. The pericardial edemas in both clo-/- mutant and tnnt2 morphant are indicated by arrows (B, C, E, F). (G-I) Left-lateral confo- y1 -/- cal in vivo projections of liver of LiPan/Tg(fli1:EGFP) larvae in wild type (G), clo mutant (H) and Tnnt2 morphant (I) back- grounds. Abbreviations: bd, bile duct; dph, daisy pattern of hepatocytes; ec, endothelial cells; e, ear; h, hepatocytes; l, liver; s, sinusoid. In all images, anterior is towards the left. Scale bars, 125 μm in (A-F) and 625 μm in (G-I). Page 8 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 contact in cloche mutants might significantly affect the size tocytes were tightly attached to each other without obvi- of liver and this effect was even more obvious during the ous morphological organization (n = 8; Figure 6H). The y1 liver of LiPan/Tg(fli:EGFP) /Tnnt2 morphants at the cor- formation of first vessels. Apart from the reduction in size, the structural organization of hepatocytes in cloche liver responding stage displayed similarly poorly organized was less obvious than that in non-cloche siblings (Figure hepatocytes with only remnants of ECs in the external lay- 5G, H). It appears that ECs through their contact with ers of liver (Figure 6I) in contrast to the well developed hepatocytes provided some specific morphogenetic sig- sinusoids and organized hepatocytes in liver of LiPan/ y1 nals for hepatocyte organization and liver growth. Tg(fli:EGFP) siblings (Figure 5G). Thus, by 120 hpf, the absence of circulation in Tnnt2 morphants causes liver To further investigate the role of ECs in liver development, deficiency comparable to that of cloche mutants lacking we introduced the Tnnt2 morphants, which phenocopied ECs (Figure 5G–L), indicating a vital role of blood circula- the zebrafish silent heart (sih) mutant with defects in car- tion in formation of the sinusoidal network in the liver diac contractility and blood circulation [20,21,2]. We and liver growth in late development. found that in LiPan/Tnnt2 morphants, the liver bud formed (as detected by in situ hybridization with the cer- According to Field et al. [5] the liver development in uloplasmin probe, not shown) and during early growth zebrafish consists of two phases – budding and growth. phase was about the same size as that in controls (323/ Our current study further defined the timing of formation 323 morphants; Figure 5A, C). A confocal examination of of three liver lobes during growth phase (Figure 7A). In y1 the liver of LiPan/Tg(fli1:EGFP) /Tnnt2 morphants combination with in vivo analysis of liver vasculogenesis revealed normal progression of the initial stage of EC and determination of initiation of blood flow in the liver interaction with hepatocytes at 55–58 hpf (not shown). of live embryos by FCS, we also defined three phases of Tnnt2 morphants suffered from a cardiac edema. Despite liver vasculogenesis: 1) contact of ECs and hepatocytes, 2) that increased cardiac edema was almost comparable to formation of nascent sinusoids, 3) initiation of blood cir- that of cloche, in contrast to the small liver in cloche, livers culation and formation of the vascular network in liver of Tnnt2 morphants (96%, 310/323) expanded notably (Figure 7B). Furthermore, by analysis of liver vasculogen- up to 70 hpf and had about same size and shape as in con- esis and liver growth in wild type as well as in cloche trols (Figure 5D, F). Based on the examination of ~400 mutant and Tnnt2 morphant larvae, we provided addi- cloche mutants and > 300 morphants, we failed to see any tional details to define three distinct stages of liver growth: correlation between the volume of edema and size of 1) avascular growth, which could be considered a transi- y1 liver. Moreover, the livers in LiPan/Tg(fli1:EGFP) /Tnnt2 tion from the budding phase to the growth phase, 2) morphants had similar morphological organization of endothelium-dependent growth, and 3) circulation- hepatocytes to those in non-morphant controls, in con- dependent growth (Figure 7C). trast to the absence of prominent daisy organization of y1 hepatocytes in LiPan/Tg(fli1:EGFP) /cloche mutants (Fig- Discussion Liver morphogenesis in LiPan transgenic zebrafish ure 5G–I). Thus, it seems that in embryonic zebrafish the initial liver growth and organization is promoted by the Although the early stages of endodermal organogenesis in contact with ECs but not by blood circulation, the zebrafish have been a subject of intense studies [1-9], some of the late developmental events in liver growth and Importance of blood circulation for liver vasculogenesis morphogenesis are still relatively poorly understood. To and growth fill this gap, we re-evaluated liver morphogenesis in the After the initiation of hepatic blood circulation at 72 hpf, zebrafish by focusing mainly on late events. We generated the liver in wild type larvae was enlarged laterally, ven- a two-color transgenic zebrafish line, LiPan, that expresses trally, and across the midline. In LiPan/Tnnt2 morphants, dsRed RFP specifically in the liver and GFP specifically in though the liver crossed the midline and extended moder- the exocrine pancreas to observe how liver growth is coor- ately by 75 hpf, we observed significant reduction of liver dinated with that of pancreas during development in live size in 95% (282/297) morphant larvae by 80 hpf (Figure embryos/larvae and adult zebrafish. The expression pat- 6A–C). It seemed that the liver growth in Tnnt2 mor- terns of the two reporter genes are essentially identical to phants was arrested in late development (Figure 6D, F). a combination of expression patterns in the two previ- The lack of blood circulation in Tnnt2 morphants eventu- ously reported transgenic lines, Tg(lfabp:egfp) [16] and ally resulted in vascular regression. The liver phenotype of Tg(elaA:egfp) [17], both of which express only GFP in the LiPan/Tnnt2 morphants progressively became similar to liver and exocrine pancreas respectively. The LiPan that of LiPan/cloche mutants. A confocal analysis of LiPan/ zebrafish expresses two distinct fluorescent proteins mark- y1 Tg(fli:EGFP) /cloche mutants revealed a compact arrange- ers with high intensity, rendering it an excellent experi- ment of hepatocytes similar to that observed in LiPan/ mental tool for detailed analysis of liver and pancreas y1 Tg(fli:EGFP) siblings at the liver budding stage, i.e. hepa- organogenesis. Page 9 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 Role of Figure 6 circulation in liver development Role of circulation in liver development. (A-F) Left-lateral views: liver morphogenesis in live LiPan wild type (A, D), LiPan/ -/- clo mutants (B, E) and LiPan/Tnnt2 morphants (C, F) at 80 hpf (A-C) and 120 hpf (D-F). Dorsal views of the liver region of the same embryos are shown as inserts in each panel and dash lines indicate the midline with the right side at the top. Livers in clo /- mutants and Tnnt2 morphants are located more medial, lack the anterio-ventral and posterior expansion, and are significantly reduced in size compared to the livers in wild type sibling. The cardiac edema in both mutants and morphants is indicated by y1 -/- arrow. (G-I) Left-lateral confocal in vivo projections of liver of LiPan/Tg(fli1:EGFP) larvae at 120 hpf in wild type (G), clo -/ mutant (H) and Tnnt2 morphant (I) backgrounds. In clo mutants ECs are absent (H), whereas in tnnt2 morphants they are present only between hepatocytes of the outer layer (I). (J-L) High-resolution light micrographs of hepatic parenchyma of zebrafish larvae stained with H&E. In 120-hpf wild type sibling, hepatocyte tubules are separated by sinusoids containing eryth- -/- rocytes (J); in contrast, in 120-hpf clo mutant (K) and Tnnt2 morphant (L), hepatocytes are tightly connected to each other. Note two sinusoids separated by two rows of neighboring hepatocytes as defined by yellow lines. Abbreviations: ec, endothe- lial cells; e, ear; h, hepatocytes; l, liver; s, sinusoid. In all images, anterior is towards the left. Scale bars, 125 μm in (A-F) and 625 μm in (G-L). Page 10 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 first, left lobe forms at ~80 hpf; the second, right lobe at Zebrafish liver development ~96 hpf, and the third, ventral lobe at 15–20 dpf during liver growth phase (Figure 1). This is in contrast to many Budding Growth phase phase second third lobe other Teleostei species such as rainbow trout [24] and Liza first liver lobe formation lobe formation formation saliens Risso, Liza aurata Risso [25], where no lobulation 24 hpf 50 hpf 80 hpf 15 dpf was recognized. In some species such as Chondrichthyes and Dipnoi, two liver lobes were found [23]. Liver vasculogenesis Contact Formation Initiation of circulation and Endothelial cells and nascent sinusoids drive liver growth of ECs of nascent formation of sinusoidal network sinusoids and morphogenesis y1 24 50 55 58 72 120 hpf Using Lipan/Tg(fli1:EGFP) transgenic zebrafish, we traced liver vasculogenesis from the very beginning until the liver became functional and showed sinusoidal net- Stages of liver growth work in adult zebrafish. These transgenic lines in combi- Avascular Endothelium Circulation dependent growth nation with cloche mutant and Tnnt2 morphants, allowed growth dependent growth us to investigate developmental mechanisms involved in hpf liver vasculogenesis. Strong RFP expression in hepatocytes 24 50 55 72 120 and GFP expression in ECs allowed the observation of the cloche dynamic process of vasculogenesis in the liver from dis- TNNT2 (silent heart) crete ECs around the liver bud to the formation of func- tional sinusoids. Our study showed the dynamic Summary an Figure 7 d growth of developmental events during liver vasculogenesis formation of sinusoidal network in zebrafish liver and Summary of developmental events during liver vas- revealed some similarities and differences in liver vasculo- culogenesis and growth. (A) Timing of liver morphogene- genesis between zebrafish and other vertebrates. sis in zebrafish. According to Field et al. [5], the liver morphogenesis in zebrafish consists of two phases, budding and growth. Our current study showed the formation of Based upon morphological characteristics, liver develop- three liver lobs during the growth phase. (B) Timing of liver ment in zebrafish has been separated into two phases, vasculogenesis. Detailed analysis of liver vasculogenesis and budding (24–50 hpf) and growth (after 50 hpf). Hepato- precise determination of blood flow initiation in liver sinu- cytes differentiation takes place at the end of liver budding soids of live embryonic zebrafish have led us to define three phase and before growth phase and lfabp is the molecular phases of liver vasculogenesis: contact of ECs and hepato- marker specific to fully differentiated hepatocytes [5,34]. cytes, formation of nascent sinusoids, initiation of blood cir- It has been previously shown that ECs are adjacent to, but culation and formation of the vascular network in the liver. not encasing, the liver bud at 48 hpf [5]. In our current in (C) Stages of liver growth. Based on our analyses of liver vas- vivo analyses, we found that during avascular growth culogenesis and growth in wild type as well as in cloche stage, discrete ECs rim the liver bud completely and then mutant and Tnnt2 morphant larvae, we propose three stages physically interact with hepatocytes prior to blood vessels of liver growth: avascular growth as a transition to the growth phase, where ECs are not required; endothelium- formation similar to that in mice [11]. In mouse Vegfr2/ -/- dependent, and circulation-dependent growth stage. Approx- Flk-1 embryos that lack mature ECs, the multi-layered imate periods of the stages of liver vasculogenesis and liver epithelia forms but later fail to grow [11]. In growth are represented along the time line in hpf or dpf. The zebrafish cloche mutant embryos, which lack ECs, the size green time lines represent the period of liver vasculogenesis. of liver was found to be comparable to that of wild type siblings during avascular growth stage up to 55 hpf and the growth of liver was arrested during vascular stage (Fig- While morphological investigation of visceral organs in ure 7B, C). It appears that the role of ECs during liver mor- larvae and adult fish is often difficult due to similar colors phogenesis is conserved among vertebrates since ECs in of most of internal organs, the LiPan larvae and adults both zebrafish and mice provide a crucial growth stimulus offer a convenient way to discern not only the liver and to the hepatic tissue before formation and function of pancreas but also other internal organs due to the exclu- local vessels [11]. Previous observation of liver develop- sion of the two fluorescent organs. Liver occupies the cra- ment in the cloche mutant was limited to 48 hpf due to an nial region of the body cavity and the exocrine pancreas increased severity of the cardiac edema [5]. Using live by and large occupies the posterior region of this cavity. It LiPan/cloche mutant we were able to analyze liver devel- is clear that zebrafish, unlike some other fish, have sepa- opment at later time points. This study showed that dur- rated and distinct liver and pancreas organs and do not ing endothelium-dependent growth stage, the size of liver form hepatopancreas (Figure 2) [23]. We consistently in cloche mutants was significantly reduced compared to observed the formation of three distinct liver lobes: the that in controls and supported the previous hypothesis Page 11 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 that the liver of cloche mutant may be affected during Role of functional sinusoidal network for liver growth growth phase if the endothelium is important for liver After the initiation of blood flow in the first liver vessels, development in zebrafish [5]. Furthermore, we also circulation became essential for further liver growth. From showed that at the cellular level, liver morphogenesis in 72 to 120 hpf, the liver enlarges 8–10 times in size. The cloche mutant was also affected and a prominent daisy pat- developing sinusoidal network directly contributes to the tern of hepatocytes was absent. increased liver volume. Before the initiation of blood cir- culation in the liver, the size of sinusoids is 5–8 μm in To eliminate the possibility that the arrest of liver growth diameter, while with the blood flow in the liver, sinusoids in cloche mutants was due to the lack of blood circulation progressively increased in size up to 12–16 μm. Before the that may transport certain factors required for liver initiation of circulation, liver size and its early structural growth, we analyzed Tnnt2 morphants that lack blood cir- organization depend on the presence of ECs. However, culation. In contrast to the reduced and unorganized liver after its initiation, circulation becomes an essential factor in cloche mutants, the size of liver and the pattern of ECs for subsequent liver development and growth, as evident y1 and hepatocytes interaction in Tnnt2 morphants embryos by comparative analyses of LiPan/Tg (fli1:EGFP) /Tnnt2 y1 /cloche mutant (Fig- were similar to that of wild type siblings during both avas- morphant and LiPan/Tg(fli1:EGFP) cular and endothelium-dependent.growth stages. Thus, ure 5J–L). Therefore, the blood circulation stimulates ECs in zebrafish may provide some morphogenic signals development of the vascular network and this network important not only for liver growth but also for structural remarkably enlarges liver size because of its own volume organization of hepatocytes. as well as delivery of nutrients to support liver growth by cell proliferation. Further analysis of mutants with defects Despite similarities there are some differences in liver of vascular patterning and vessels maintenance could development between zebrafish and mice. In mice, where uncover additional critical factors involved in liver vascu- the liver is an early hematopoietic organ, endothelial- logenesis [28]. endoderm interaction is initiated during the early phase of liver budding, whereas in zebrafish this interaction The development of the hepatic vascular architecture is a starts much later and during the growth phase. Moreover, multistep process through the interaction of the two tis- -/- allele causes embry- mouse homozygous mutant Flk-1 sues, hepatocytes of endodermal origin and endothelia of onic lethality by E10.5 (1 day after initiation of hepato- mesoderm origin. Further to the model on early liver cytes migration into the surrounding septum transversum development proposed by Field et al. [5], we, based on mesenchyma) [26], but homozygous cloche mutants lack- new observation of developmental events during the liver ing ECs survive 6 days (almost 4 days after initiation of growth phase, proposed to divide the liver growth phase liver vasculogenesis). This time of development (5–6 dpf) into three distinct stages: avascular growth as a transition corresponds with the appearance of the adult form of the to the growth phase, where ECs are not required and liver heart and the transition from diffusive to convective oxy- extended by proliferation of hepatocytes; endothelium- gen supply. dependent growth, when liver grew due to proliferation of hepatocytes and ECs, and circulation-dependent growth In mice and human, ECs provide stimuli for hepatocytes stage, when the blood circulation stimulates development to outgrow towards septum transversum mesenchyme of the vascular network which increases liver size because [27,11]. Previously it has been proposed that vasculogen- of its own volume and releases nutrients to support cell esis in zebrafish is achieved by endothelial invasion of growth and proliferation (Figure 7). This model could be liver [5]. In contrast, our data suggested that both ECs and useful as a roadmap to design further experiments liver grow towards each other. As ECs rim the liver com- addressing the role of the key factors required for liver vas- pletely and contact seemingly less associated surface lay- culature development, the functions of signaling path- ers of hepatocytes, the liver extends simultaneously, ways and interactions between them during intriguing effectively moving hepatocytes towards endothelia. In events of liver vasculogenesis. mice, after growing into the septum transversum mesen- chyme, hepatocytes organize around already formed sinu- Conclusion soids. In zebrafish, discrete ECs contact liver and then In the present study, a two-color transgenic zebrafish line form nascent sinusoids. Formation of the daisy pattern of (LiPan) with RFP expression in the liver and GFP expres- hepatocytes and nascent sinusoids proceeds at the same sion in the exocrine pancreas was generated and the LiPan time. This interaction significantly changes the liver mor- line allowed us to analyze detailed liver development in phology, but molecular mechanisms underlying this live embryos and larvae. By crossing the LiPan line with y1 developmental process remain to be elucidated. Tg(fli1:EGFP) , we found that liver vasculogenesis started at 55–58 hpf when ECs first surrounded the hepatocytes Page 12 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 from the liver bud surface and then invaded the liver to Homozygous LiPan zebrafish were viable and had no vis- form sinusoids and later the vascular network. Fluores- ible phenotype. The LiPan line has been maintained in cence correction spectroscopy detected blood circulation our laboratories for over eight generations and co-expres- in the liver first at ~72 hpf. Analysis of cloche mutants and sion of RFP in the liver and GFP in the exocrine pancreas Tnnt2 morphants led us to conclude that both ECs and is always observed. Thus the two injected DNA constructs blood circulation are required for continued liver growth are likely co-integrated into the same chromosomal locus. and morphogenesis. Collectively, we propose to divide the growth phase of liver development in zebrafish into Microscopy three distinct stages: avascular growth between 50–55 hpf, To facilitate visualization of liver and exocrine pancreas of where ECs are not required; endothelium-dependent larval zebrafish in whole-mount preparations, pigmenta- growth, where ECs or sinusoids are required for liver tion of skin was inhibited by raising embryos and larvae growth from 55 hpf to 72 hpf before blood circulation in in egg water [29] containing 0.2 mM 1-phenyl-2-thiourea the liver sinusoids; circulation-dependent growth where (Sigma, USA). Microscopic observations and photogra- the circulation is essential to maintain vascular network phy of live embryos were performed using the dissecting and to support the continued liver growth and morpho- fluorescent microscope SZX12 (Olympus, Japan), com- genesis after 72 hpf. pound microscope Zeiss Axioscope 2 and confocal micro- scope Zeiss LSM510 (Zeiss, Germany). Three images were Methods taken at the same focal plane, using a DIC filter for trans- Zebrafish maintenance mitted light for the first, epifluorescence with a Rhod for Zebrafish were maintained in the fish facilities at the the second and FITC filter for the third. These three images Department of Biological Sciences, National University of were then superimposed using Zeiss AxioVision software Singapore (NUS) and the Institute of Molecular and Cell or Photoshop (Adobe, USA). Three-dimensional confocal Biology (IMCB) of Singapore according to established projections were generated using Zeiss LSM510 software protocols [29] and in compliance with Institutional Ani- (Zeiss, Germany). In all confocal studies, at each time mal Care and Use Committee (IACUC) guidelines. Devel- point, 5–8 embryos/larvae from random pairs were exam- opmental stages are presented in hour post fertilization ined. (hpf), day post fertilization (dpf) or month post fertiliza- tion (mpf). Blood flow detection by fluorescence correlation spectroscopy Microinjection and establishment of Tg(lfabf:ds-Red; Fluorescence correlation spectroscopy (FCS) is a single- elaA:EGFP) zebrafish line (LiPan) molecule sensitive fluorescence technique which can pro- Isolation of elaA promoter (1.9 kb) and construction of vide information about diffusion coefficient, concentra- the chimeric plasmid pElaA-EGFP have been described tion, microfluidic flow, etc. [30,31]. It is based on an previously [17] The liver-specific promoter (2.8 kb) autocorrelation analysis of fluorescence fluctuations from derived from the zebrafish liver fatty acid binding protein a small focal volume in the specimen that is defined by a gene was provided by Dr. G.-M. Her and was inserted into high numerical aperture objective and a small pinhole. pDsRed-Express-1 (Clontech, USA) to make the chimeric Autocorrelation functions are derived for different cases plasmid pLFABP-RFP. Both plasmids were linearized, such as 3D diffusion and microfluidic flow, and the mod- mixed with 0.25% phenol red solution (1:1:1) at a final els can be used to fit the experimental data. The two-flow concentration of 100 ng/μl of each plasmid. Microinjec- model to extract the diastolic and systolic blood flow tion was carried out at the 1–2 cell stage. The DNA solu- velocity at the same blood vessel in zebrafish larva was tion was injected into the boundary between the yolk and developed and the detailed setup was described previ- blastodisc. After microinjection, the embryos were main- ously [32]. The blood flow velocity in the liver sinusoids tained in egg water [29] with ~0.0005% methylene blue in of zebrafish larvae was obtained by FCS measurement at a 28.5°C incubator. Transgenic founders were screened by the points of interest after the confocal image acquisition observation of F1 embryos for RFP and GFP expression. which helps to locate the position of sinusoids in the liver. 444 injected embryos were raised to adult and 66 of them In this work, each larva was anesthetized with freshly pre- were screened for transgenics. Two of them were found to pared 0.05–0.1 mg/ml Tricaine (ethyl m-aminoboen- produce F1 embryos with strong liver-specific RFP expres- zoate, Sigma, Singapore) dissolved in egg water [29], sion and exocrine pancreas-specific GFP expression. Thus, immobilized in 1.5% low-melting-temperature agarose two stable transgenic lines were established and both (Invitrogen, Singapore) (agarose was dissolved in 0.05 showed standard Mendelian inheritance from F2 genera- mg/ml of Tricaine in egg water), in WillCo-dish glass bot- tion onwards. Since identical reporter gene expression tom dish (GW-3512, WillCo-Wells, The Netherlands), patterns were observed in the two lines, only one line, placed in a temperature-controlled environment and named LiPan, was used for further characterization. immediately proceed for measurements. For each larva, Page 13 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 blood flow was measured in a trunk vessel as a control Authors' contributions and then in the vessels of liver parenchyma at 7–10 ran- SK – made genetic crosses and selection, systematic anal- domly selected points at different depths of liver tissue up yses of organogenesis and vasculogenesis in transgenic to 80 μm. For each stage 3–5 larvae were measured. With and mutant fish, and wrote the manuscript; XP – designed this technique, we were able to measure blood flow inside and made the LiPan transgenic line; MGL – made confocal the liver parenchyma as deep as 70 μm from the liver sur- microscopic images; CLM – made genetic crosses; XP-ana- face. lyzed blood flow in liver; TW- developed blood flow measurement method; VK – developed the concept of the Generation of double (LiPan) and triple [Tg(lfabf:ds-Red; project, wrote and approved the manuscript; ZG – devel- elaA:EGFP; fli1: EGFP)] transgenic cloche mutants oped the concept of the project, designed the LiPan trans- To visualize vascular development in the liver in live genics, wrote and approved the manuscript. embryos, LiPan homozygotes were mated with y1 Tg(fli1:EGFP) homozygotes and triple transgenic Acknowledgements We are thankful to Dr. G.M. Her for the plasmid pLFABP-EGFP, Dr. B. embryos were generated. To analyze effect of endothelia y1 Weinstein for Tg(fli1:EGFP) transgenic line (through Drs. Z. Wen and R. on liver development and growth, we used cloche mutants Ge), Dr. D. Stainier for cloche mutant, Dr. K. Osborne for careful reading s5 (clo [point mutation allele]) which lack almost all and comments, and personnel of NUS and IMCB fish and histology facilities endothelial cells [5,12,33]. Both LiPan and LiPan/ for technical help and discussion. The financial support for this project came y1 Tg(fli1:EGFP) transgenic fish were crossed with cloche from Biomedical Research Council of Singapore. V.K. laboratory in the heterozygotes to transfer the transgenes into the cloche IMCB has been supported by the Agency for Science, Technology and mutants. After their progeny reached maturity, these Research of Singapore. fishes were crossed randomly to identify cloche heterozy- gotes that carry the transgenes. These fishes were crossed References 1. Warga R, Nüsslein-Volhard C: Origin and development of the to obtain homozygous mutants with double (LiPan) or zebrafish endoderm. Development 1999, 126:827-838. y1 triple [LiPan/Tg(fli1:EGFP) ] transgenic background and 2. Korzh S, Emelyanov A, Korzh V: Developmental analysis of cer- their development was monitored. uloplasmin gene and liver formation in zebrafish. Mech. Dev 2001, 103:137-139. 3. Duncan SA: Mechanisms controlling early development of the Tnnt2 morphants liver. Mech Dev 2003, 120:19-33. To analyze the role of circulatory defects in liver develop- 4. Wallace KN, Pack M: Unique and conserved aspects of gut development in zebrafish. Dev Biol 2003, 255:12-29. ment, Tnnt2 antisense morpholino oligonucleotide (5'- 5. Field HA, Ober EA, Roeser T, Stainier DY: Formation of the diges- CATGTTTGCTCTGATCTGACACGCA), which is targeted tive system in zebrafish. I. Liver morphogenesis. Dev Biol 2003, 253:279-290. the Tnnt2 translation start codon and phenocopies the 6. Ober EA, Holly AF, Stainier DYR: From endoderm formation to silent heart (sih) mutation [22], was obtained from Gene liver and pancreas development in zebrafish. Mech Dev 2003, Tools (USA). A total of 4 ng of Tnnt2 morpholino oligo- 120:5-18. 7. David N, Mourrain P, Rosa F: Endoderm formation in zebrafish. nucleotide was injected into 1–2 cell stage of LiPan or In Molecular Aspects of Fish and Marine Biology, Fish Development and y1 LiPan/Tg(fli1:EGFP) embryos and a stable non-contrac- Genetics: zebrafish and medaka Volume 2. Edited by: Gong Z, Korzh V. Singapore: World Scientific; 2004:424-453. tile heart phenotype was observed in 98% of injected 8. Dong PD, Munson CA, Norton W, Crosnier C, Pan X, Gong Z, Neu- embryos and larvae from 24 hpf to 7 dpf (their last day of mann CJ, Stainier DY: Fgf10 regulates hepatopancreatic ductal survival). Morphants without obvious motility defect and system patterning and differentiation. Nat Genet 2007, 39:397-402. lacking blood circulation (with pericardial edema) were 9. Li Z, Korzh V, Gong Z: Localized rbp4 expression in the yolk used for analysis of liver growth and vasculogenesis. syncytial layer plays a role in yolk cell extension and early liver development. BMC Dev Biol 2007, 7:e117. 10. Pack M, Solnica-Krezel L, Malicki J, Neuhauss SC, Schier AF, Stemple Histology DL, Driever W, Fishman MC: Mutations affecting development For cryosection, zebrafish embryos were ice-chilled and of zebrafish digestive organs. Development 1996, 123:321-328. 11. Matsumoto K, Yoshitomi H, Rossant J, Zaret J: Liver organogene- fixed with ice-cold 4% paraformaldehyde (PFA) in phos- sis promoted by endothelial cells prior to vascular function. phate-buffered saline (PBS) at 4°C overnight. After fixa- Science 2001, 294:559-563. tion, embryos were embedded in 1.5% Bacto agar 12. Liao W, Bisgrove BW, Sawyer H, Hug B, Bell B, Peters K, Grunwald DJ, Stainier DYR: The zebrafish gene cloche acts upstream of a containing 5% sucrose and incubated in 30% sucrose at flk-1 homologue to regulate endothelial cell differentiation. 4°C overnight. The embedded embryos were oriented and Development 1997, 124:381-389. 13. Thompson MA, Ransom DG, Pratt SJ, MacLennan H, Kieran MW, sectioned with a cryostat microtome (10 μm thickness). Detrich HW, Vail B, Huber TL, Paw B, Brownlie AJ, Oates AC, Fritz For paraffin sections, the embryos and larvae were fixed A, Gates MA, Amores A, Bahary N, Talbot WS, Her H, Beier DR, with either 4% PFA in PBS or Bouin's fixative, followed by Postlethwait JH, Zon LI: The cloche and spadetail genes differ- entially affect hematopoiesis and vasculogenesis. Dev Biol paraffin embedding and sectioning (4 μm). Serial sections 1998, 197:248-269. (from at least 3–5 embryos or larvae at each time point) 14. Lawson ND, Weinstein BM: In vivo imaging of embryonic vascu- were de-paraffinized, stained with hematoxylin-eosin, lar development using transgenic zebrafish. Dev Biol 2002, 248:307-318. dehydrated, and examined. Page 14 of 15 (page number not for citation purposes) BMC Developmental Biology 2008, 8:84 http://www.biomedcentral.com/1471-213X/8/84 15. Pan X, Yu H, Shi X, Korzh V, Wohland T: Characterization of flow direction in microchannels and zebrafish blood vessels by scanning fluorescence correlation spectroscopy. J Biomed Opt 2007, 12:14034. 16. Her GM, Chiang CC, Chen WY, Wu JL: In vivo studies of liver- type fatty acid binding protein (L-FABP) gene expression in liver of transgenic zebrafish (Danio rerio). FEBS Lett 2003, 538:125-133. 17. Wan H, Korzh S, Li Z, Mudumana SP, Korzh V, Jiang YJ, Lin S, Gong Z: Analyses of pancreas development by generation of gfp transgenic zebrafish using an exocrine pancreas-specific elastaseA gene promoter. Exp Cell Res 2006, 312:1526-1539. 18. Isogai S, Horiguchi M, Weinstein BM: The vascular anatomy of the developing zebrafish: an atlas of embryonic and early lar- val development. Dev Biol 2001, 230:278-301. 19. Lorent K, Yeo SY, Oda T, Chandrasekharappa S, Chitnis A, Mattews R, Pack M: Inhibition of Jagged-mediated Notch signaling dis- rupts zebrafish biliary development and generates multi- organ defects compatible with an Alagille syndrome pheno- copy. Development 2004, 131:5753-5766. 20. Chen JN, Haffter P, Odenthal J, Vogelsang E, Brand M, van Eeden FJ, Furutani-Seiki M, Granato M, Hammerschmidt M, Heisenberg CP, Jiang YJ, Kane DA, Kelsh RN, Mullins MC, Nüsslein-Volhard C: Muta- tions affecting the cardiovascular system and other internal organs in zebrafish. Development 1996, 123:293-302. 21. Stainier DY, Fouquet B, Chen JN, Warren KS, Weinstein BM, Meiler SE, Mohideen MA, Neuhauss SC, Solnica-Krezel L, Schier AF, Zwartkruis F, Stemple DL, Malicki J, Driever W, Fishman MC: Muta- tions affecting the formation and function of the cardiovas- cular system in the zebrafish embryo. Development 1996, 123:285-292. 22. Sehnert AJ, Huq A, Weinstein BM, Walker C, Fishman M, Stainier DY: Cardiac troponin T is essential in sarcomere assembly and cardiac contractility. Nat Genet 2002, 31:106-110. 23. Gonzales G, Crespo S, Brusle J: Histo-cytological study of the liver in the cabrilla sea bass, Serranus cabrilla (Teleostei, Ser- ranidae) an available model for marine fish experimental studies. J Fish Biol 1993, 43:364-373. 24. Robertson OH, Wexler BC: Histopathological changes in the organs and tissues of migrating and spawning Pacific salmon (genus Oncorhunchus). Endocrinology 1960, 66:222-239. 25. Biagianti-Risbourg S: Fine structure of hepatocytes in juvenile grey mullets: Liza saliens Risso, L. ramada Risso and L. aurata Risso (Teleostei, Mugilidae). J Fish Biol 1991, 39:687-703. 26. Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu XF, Breit- man ML, Schuh AC: Failure of blood-island formation and vas- culogenesis in Flk-1-deficient mice. Nature 1995, 376:62-66. 27. Elias H, Sherrick JC: Morphology of the liver NY and London: Academic Press; 1969. 28. Jin SW, Herzog W, Santoro MM, Mitchell TS, Frantsve J, Jungblut B, Beis D, Scott IC, D'Amico LA, Ober EA, Verkade H, Field HA, Chi NC, Wehman AM, Baier H, Stainier DY: A transgene-assisted genetic screen identifies essential regulators of vascular development in vertebrate embryos. Dev Biol 2007, 307:29-42. 29. Westerfield M: The zebrafish book, a guide for the laboratory use of zebrafish (Danio rerio). Eugene, OR, University of Oregon Press; 2000. 30. Magde D, Elson E, Webb WW: Thermodynamic fluctuations in a reacting system – measurement by fluorescence correla- tion spectroscopy. Phys Rev Lett 1972, 29:705-708. 31. Magde D, Webb W, Elson E: Fluorescence correlation spectros- Publish with Bio Med Central and every copy. III. Uniform translation and laminar flow. Biopolymers scientist can read your work free of charge 1978, 17(2):361-376. 32. Pan X, Fok M, Foo W, Lim W, Liu P, Yu H, Maruyama I, Wohland T: "BioMed Central will be the most significant development for Multifunctional fluorescence correlation microscope for disseminating the results of biomedical researc h in our lifetime." intracellular microfluidic measurements. Rev Sci Instrum 2007, Sir Paul Nurse, Cancer Research UK 78:53711. 33. Stainier DYR, Weinstein BM, Detrich HW, Zon LI, Fishman MC: Your research papers will be: cloche, an early acting zebrafish gene, is required by both the available free of charge to the entire biomedical community endothelial and hematopoietic lineages. Development 1995, 121:3141-3150. peer reviewed and published immediately upon acceptance 34. Chen J, Ruan H, Ng SM, Gao C, Soo HM, Wu W, Zhang Z, Wen Z, cited in PubMed and archived on PubMed Central Lane DP, Peng J: Loss of function of def selectively up-regulates {Delta}113p53 expression to arrest expansion growth of yours — you keep the copyright digestive organs in zebrafish. Genes Dev 2005, BioMedcentral 19(23):2900-2911. Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 15 of 15 (page number not for citation purposes)

Journal

BMC Developmental Biology – Springer Journals

Published: Sep 16, 2008

There are no references for this article.