Analysis of Vascular Morphogenesis in Zebrafish

Abstract

Analysis of cardiovascular development in zebrafish embryos has become a major driver of vascular research in recent years. Imaging-based analyses have allowed the discovery or verification of morphologically distinct processes and mechanisms of, e.g., endothelial cell migration, angiogenic sprouting, tip or stalk cell behavior, and vessel anastomosis. In this chapter, we describe the techniques and tools used for confocal imaging of zebrafish endothelial development in combination with general experimental approaches for molecular dissection of involved signaling pathways.

1 Introduction

The essential function of regulating oxygen and nutrient supply demands that the cardiovascular system is one of the first vertebrate organs to form. Its implication in many human diseases is based on its additional role in immune cell and hormone trafficking as well as its heterogeneity, allowing for tissue-specific or regional specializations of different vascular beds that differ in shape and function.

The vascular system emerges from mesoderm-derived precursors in a process called vasculogenesis [1]. Hereafter, new blood vessels emanate from pre-existing vessels, which is referred to as angiogenesis [2]. The invasion of endothelial cells into avascular tissue is a tightly controlled process that includes highly conserved signaling mechanisms that transmit attracting or repelling cues to control blood vessel sprouting [3,4,5,6,7]. Detailed analysis of vascular morphogenesis in zebrafish (Danio rerio) embryos has been used in order to identify potential regulators of angiogenesis [6,7,8,9].

Development of the zebrafish cardiovascular system follows the same or highly homologous morphological and molecular events as in mammals [10,11,12]. The advantages of using zebrafish as a model system are based on its rapid development in translucent embryos and the high offspring count [13]. Moreover, owing to external fertilization, the zebrafish embryo is easily amenable to both developmental observation and experimental manipulation, including pharmacologic and genetic approaches. Genetic manipulation provides a broad variety of possibilities such as gene knockdown by using morpholino antisense oligonucleotides [14], genome editing through CRISPR-Cas9 for gene knockout or knock-in approaches [15,16,17,18], or DNA/RNA overexpression studies [19, 20]. Another advantage is the transparency of the small embryos allowing high-resolution imaging techniques to assess endothelial cell behavior at single-cell resolution in vivo over short as well as multiple-day time spans [21,22,23,24,25]. Additionally, the small size of the embryos enables fast passive oxygen and nutrient diffusion throughout their whole body. Hence, embryos can survive for several days without a functional vascular system [26].

In addition to these advantages for analysis during embryonic development, adult zebrafish have been used to study regenerative processes of heart tissue and fin vasculature [27,28,29].

The vascular anatomy of the developing zebrafish embryo and the underlying cellular processes have been described in detail over the last two decades [6,7,8,9, 24, 30].

In zebrafish, endothelial cells can be visualized by vascular-specific transgenes that result in cell-specific fluorophore expression [31, 32]. Additionally, other cells like perivascular cells or immune cells can be visualized in zebrafish embryos to study the interactions of endothelial cells with their environment [33, 34]. Table 1 lists different transgenic zebrafish lines that visualize endothelial cells and their subcellular compartments [11, 25, 31, 32, 35,36,37,38,39,40,41,42,43,44,45,46].

Table 1 Transgenic zebrafish generated for cardiovascular research

The major components of the first circulatory loop of zebrafish embryos encompass a beating heart, the dorsal aorta, and the cardinal veins [6]. Hereafter, other vascular beds in the trunk, the tail, and the brain develop to form a diversely branched vascular network with different vessel calibers and tissue-specific heterogeneity.

In this chapter, we will focus on the formation of three different vascular beds: the intersegmental vessels (ISVs), the intracerebral central arteries (CtAs), and the common cardinal veins (CCVs) (shown in Figs. 1 and 2). Figure 1 provides a developmental overview of the formation of the three heterogenous vessel types, while Fig. 2 displays examples of the visualization of different subcellular compartments in endothelial cells, as mentioned in Table 1.

Fig. 1

Angiogenesis in three vascular beds of the zebrafish. Schematic overviews and confocal projections of the zebrafish vasculature at different developmental stages between 23 hpf (hours post fertilization) and 2.5 dpf (days post fertilization). Localization of blood vessels is indicated in the schematic overviews of zebrafish embryos at 3 dpf (cartoons were generated using BioRender.com). We show the development of the intersegmental vessels in the tail region, the intracerebral central arteries in the hindbrain, and the common cardinal veins during angiogenesis (between 23 and 32 hpf; white arrow indicates sprouting direction) and at 2.5 dpf, when the vessel pattern is established. The confocal projections display the endothelial cells in white, imaged from transgenic expression of GFP (Tg(kdrl:EGFP)^s843) and erythrocytes in red (Tg(gata1:dsRed)^sd2). Scale bar ISV = 20 μm; CtA lateral (32 hpf) = 30 μm; 2.5 dpf dorsal = 50 μm; CCV = 40 μm. DLAV dorsal longitudinal anastomotic vessel, DA dorsal aorta, PCV posterior cardinal vein, PHBC primordial hindbrain channel, BA basilar artery

Fig. 2

Visualization of subcellular structures in endothelial cells of the intersegmental vessels. Confocal maximal projections show different subcellular structures visualized by vascular transgenes at 24 hpf during ISV formation. For better visualization the images use inverted color display. (a) The Golgi apparatus and the nucleus were visualized by expressing mCherry-tagged beta-1,4-galactosyltransferase 1 (Tg(fli1a:B4GALT1-galT-mCherry)^bns9; shown in red) and EGFP in the nucleus (Tg(kdrl:nls-EGFP)^zf109; shown in black). (b) The actin cytoskeleton (shown in black) was visualized by expressing Lifeact-EGFP in endothelial cells (Tg(fli1a:lifeact-EGFP)^mu240). Labeling the actin cytoskeleton is also widely used for visualization of filopodia [51]. (c) The early endosomes are detected by expressing an mCherry-Rab5c fusion protein (Tg(fli1a:mCherry-Rab5c)^mu227; shown in red), while endothelial cells are simultaneously visualized by EGFP expression (Tg(kdrl:EGFP)^s843; shown in black). Scale bar = 15 μm

Firstly, we will concentrate on the ISVs, which sprout from the dorsal aorta (DA) to migrate dorsally between the segments of the trunk and the tail (Fig. 1a). By connecting with other migrating ISV endothelial cells, they build up the dorsal longitudinal anastomotic vessel (DLAV). Through random reconnection of half of these ISVs with sprouts from the posterior cardinal vein (PCV), these endothelial cells form circulatory loops ensuring blood supply through the somites [24]. Initial angiogenic sprouting starts at around 19 h post fertilization (hpf) in the trunk region and continues in a rostral to caudal progression [6, 7]. After secondary venous sprouting, the final circulatory loop is fully established before 48 hpf [6, 7]. Due to their regular and stereotypic pattern and their superficial position, the ISVs have become the most used vascular bed for the observation of sprouting angiogenesis and vessel development in zebrafish. Naturally, various signaling processes are involved in the formation of the ISVs. The initial endothelial cell migration from the DA is, among others, mainly stimulated through vascular endothelial growth factor a (VEGFa) [6, 7, 24].

Secondly, the CtAs serve as an ideal example of tissue-specific heterogeneity, as these brain vessels form the blood-brain barrier, characterized by distinct junction properties (regulated through tight junction proteins) and distinct transport mechanisms [47]. The intracerebral CtAs (also known as hindbrain capillaries) invade the avascular brain parenchyma at around 32 hpf, originating from the primordial hindbrain channels (PHBC) (Fig. 1a). A single tip cell initially migrates dorsally before the vessel loops back and connects to the medial basilar artery (BA) [47,48,49,50]. Alternatively, CtA cells may also connect laterally to other CtAs [47]. At around 48 hpf, the CtAs are fully developed and carry blood [47,48,49,50]. Mechanistically, it has been shown that Wnt signaling in these endothelial cells is indispensable during CtA formation. Here, Wnt signaling reflects an important function in defining pre-tip cell specifications and later on also vascular patterning and vessel anastomosis [33, 47].

Lastly, we focus on the formation of the CCVs, which start their formation together with the more dorsally located cardinal veins with angioblast migration around 17 hpf. The bilateral CCVs collect blood from the cardinal veins and transport it back to the heart. However, unlike any other vessel of the embryo, the CCVs already transport blood from around 25 hpf onward, although they are still actively undergoing angiogenesis, more specifically cell migration. The formation of these vessels is finished, when CCV endothelial cells connect to the cardiac endothelial cells of the inflow tract and finally close around their lumen, which happens at around 50 hpf. During the formation of the CCV, endothelial cells migrate collectively as a sheet toward the ventral side of the yolk, before extending to the cranioventral side to connect to the heart (Fig. 1a) [25, 51]. This process is different to the other two presented models as there is no single tip cell leading sprouting involved. In this case, communication within the endothelial cell sheet is a major regulator of pattern formation [6, 24, 51]. However, similar to other vascular beds, this mode of endothelial cell migration is also controlled by various signals [25, 51].

Here, we describe the preparation of zebrafish embryos in order to image angiogenesis in the three abovementioned vascular beds. We provide detailed instructions on how to treat embryos for confocal microscopy including mounting techniques. Although several different methods can be used to study vascular morphogenesis in zebrafish, such as in situ hybridization techniques or antibody staining experiments, we will focus on fluorescent in vivo live-imaging approaches [52, 53]. Additionally, we provide three analysis examples, one for each introduced vessel type, on how to quantify endothelial cell behavior in zebrafish embryos.

However, in addition to the here described methods, other recently developed techniques to advance from tissue and cellular observations toward more detailed molecular observations are becoming available. While there are no easy-to-use protocols existing at present, the future will contain applications like the use of Förster resonance energy transfer (FRET) sensor probes [54,55,56]. Using vascular tension sensors, changes in inter- and intracellular tension have been measured at single-molecule resolution, for instance, by using a transgenic zebrafish line expressing a vascular endothelial (VE)-cadherin tension sensor to determine junctional tension between adjacent endothelial cells [57]. However, challenges include the need for biological single-fluorophore controls, a specialized microscopy setup (two-photon excitation), and the respective transgenic lines.

Nevertheless, many important observations were made in zebrafish using basic confocal microscopy in the last decades. As described in this chapter, multicellular structures and subcellular compartments can be visualized in zebrafish without the need for any super-resolution techniques. Thanks to live imaging approaches in zebrafish, processes such as vascular lumen formation could be analyzed in more detail than in mouse models leading to new discoveries in our biological understanding [6].

2 Materials

2.1 General Vascular Assessment

2.1.1 Preparing Embryos

• 1. Fertilized zebrafish embryos.

• 2. Incubator (such as Memmert IPP30).

• 3. 60× zebrafish E3 embryo medium: 0.3 M NaCl, 10 mM KCl, 20 mM CaCl₂⋅2H2O, 20 mM MgSO₄⋅7H₂O

• 4. Plastic Petri dishes.

• 5. Dissection instruments: sharp Dumont forceps (number 55), brush, glass pipette.

• 6. Dissection (stereo) microscope.

• 7. 20× PTU solution: 4 mM N-phenylthiourea in 1× E3

• 8. Fluorescence stereomicroscope.

• 9. 25× tricaine solution: 400 mg tricaine methanesulfonate [MS222], 2.1 mL 1 M Tris pH 9.0, 97.9 mL H₂O. Adjust to pH 7.0.

2.1.2 Mounting and Imaging

• 1. Fertilized zebrafish embryos of specific developmental stages (anesthetized with tricaine solution).

• 2. Dissection instruments: sharp Dumont forceps (number 55), brush, glass pipette.

• 3. Dissection (stereo) microscope.

• 4. Agarose (pegGOLD Universal Agarose; Peqlab).

• 5. Microwave.

• 6. Temperature-controlled water bath.

• 7. Glass-bottom dishes (WillCo Wells BV – GWST-5040).

• 8. Imaging casting molds for agarose (World Precision Instruments, or self-made using a 3D printer).

• 9. Confocal microscope, such as LSM880 (Carl Zeiss Microscopy GmbH), including objective lenses (HC Plan-Apochromat 20×/0.8 air, HC Plan-Apochromat 40×/1.2 water, and HC Plan-Apochromat 63×/1.4 oil) and a heating chamber for time-lapse microscopy.

2.2 Manipulations to Study Cell Migration

2.2.1 Microinjections

• 1. Agarose (pegGOLD Universal Agarose; Peqlab).

• 2. Zebrafish E3 embryo medium.

• 3. Plastic Petri dishes.

• 4. Injection casting mold for agarose (World Precision Instruments).

• 5. Fertilized zebrafish embryos.

• 6. Dissection instruments: sharp Dumont forceps (number 55), brush, glass pipette.

• 7. Glass capillaries (World Precision Instruments).

• 8. Micropipette puller (Sutter Instrument).

• 9. Injection mix including (depending on experiment, see Subheading 3.2): 0.05% phenol red (Sigma), plasmid DNA of interest, CAS9 protein (New England Biolabs) or cas9 mRNA (Sigma), guide RNAs (e.g., order via Integrated DNA Technologies), morpholino oligonucleotides (Gene Tools), tol2 mRNA (Sigma).

• 10. Microloader pipette tips (Eppendorf).

• 11. Micromanipulation device (World Precision Instruments) attached to compressed air system.

2.2.2 Drug Treatments

• 1. Materials mentioned in Subheading 2.1.

• 2. Respective chemicals and solvents; here: IWR-1 (Sigma), dimethyl sulfoxide (DMSO, Sigma).

2.3 Analysis

• 1. Image processing and analyzing software such as ImageJ/FIJI or Imaris software (Bitplane).

• 2. Software for data analysis, statistics, and visualization, such as Microsoft Excel, Prism software (GraphPad), or R Statistics software.

3 Methods

3.1 General Vascular Assessment During Embryonal Development

3.1.1 Preparing Embryos for Live Imaging

• 1. Collect embryos expressing a specific transgene (see Table 1) into an E3-filled plastic petri dish and cultivate at 28.5 °C.

• 2. Around 6 hpf, sort out dead, amorphous, or unfertilized embryos using a stereomicroscope. Furthermore, determine the exact fertilization time by using the published guide to assess the developmental stage of zebrafish embryos [58].

• 3. After gastrulation (latest around 24 hpf) supplement the E3 medium with PTU in order to prevent melanocyte formation [59]. This will ensure embryo transparency for vascular assessment via confocal microscopy.

• 4. Remove the chorion manually under the stereomicroscope by using two sharp Dumont forceps.

• 5. Anesthetize embryos using tricaine solution, and sort embryos under a fluorescence stereomicroscope for their respective vascular transgene using a brush and Dumont forceps (see Note 1). Positive embryos can be used for further analysis by confocal microscopy.

• 6. Prepare mounting solution by dissolving 0.3% agarose in E3 medium. Boil briefly in a microwave. After complete dissolution, supplement the agarose with PTU to inhibit pigmentation and with tricaine solution to anesthetize the embryo and prevent movement during microscopy. Cool down and keep agarose flask in a temperature-controlled water bath at 42 °C.

3.1.2 Mounting and Imaging

• 1. In order to mount embryos for imaging, embed and fix embryos in mounting solution.

• 2. Embryos can be mounted using two different techniques. Agarose wells can be generated by using an imaging casting mold. To do this, pour mounting solution into an imaging glass-bottom dish and add the respective imaging casting mold. After solidification at room temperature (around 1 h, or: prepare ahead and store at 4 °C, then warm up to room temperature), the cast can be removed and embryos can be transferred into separate agarose wells with a glass pipette (see Note 2). We further recommend fixing the embryos using another drop of mounting solution (see Note 3).

Alternatively, embryos can simply be transferred onto an empty glass-bottom dish and then be fixed with a drop of mounting medium (see Note 3). Depending on your microscope setup, make sure to additionally consider the lens position when mounting your embryos. Both upright and inverted microscopes can be used with this mounting technique (see Note 4).

• 3. After the agarose has solidified, make sure to place the lid on the glass-bottom dish to avoid excessive evaporation which might compromise specimen integrity during time-lapse imaging (see Note 5).

• 4. Place the imaging dish in a preheated microscope incubation chamber at 28.5 °C to maintain stable embryo development and to avoid heat-dependent microscope stage drift owing to metal extension upon heating (see Note 6).

• 5. Select the appropriate magnification for imaging the structure of interest (see Subheadings 3.1.3, 3.1.4 and 3.1.5).

• 6. Define the x-/y-position and z-stack of your embryos to get an adequate view of the structure of interest. When running a time-lapse series, make sure to consider that cells may migrate out of the field of imaging when defining the image x-/y-position and z-stack depth. It might be necessary to use the tile scan function to observe cells during migration for a longer period, especially at higher magnifications. Make sure to use a suitable time interval for the imaging in order to obtain the desired time resolution of your developmental process (see Note 7).

3.1.3 ISV Imaging

• 1. Mount the embryo in a lateral position. Image individual embryos from 22 hpf onward to observe initial endothelial cell sprouting and migration.

• 2. For imaging ISVs, use a 20x magnification lens. For analysis of subcellular structures such as filipodia or intracellular compartments, higher magnification might be necessary. Make sure to use suitable lens immersion substrates according to the lens’ manufacturer’s instructions.

• 3. For time-lapse analysis, use an interval of 10–20 min to capture cell migration. Subcellular observations might require shorter imaging intervals (see Note 7).

3.1.4 CtA Imaging

• 1. For imaging the CtAs in the hindbrain of the zebrafish, use two distinct imaging angles. For observations of endothelial sprouting, initial sprout migration, as well as lumen formation, use a lateral point of view. Start imaging from 32 hpf onward to capture the whole CtA angiogenesis.

• 2. Assess vessel patterning when CtA formation has been finished before starting to image angiogenesis in detail. Use a dorsal imaging view at around 48 hpf. Depending on the microscope setup, the embryos should be placed with the dorsal side pointing toward the lens of the microscope. Include the whole depth range from the skin to the primordial hindbrain channels in your z-stack to include the complete CtA pattern (due to tissue depth, imaging the connections to the BA might prove challenging).

• 3. Start imaging with a 20× magnification lens. For imaging of filopodia or intracellular compartments, increasing the magnification might be necessary.

• 4. Use an interval of 10–20 min to capture the migration process when performing time-lapse imaging. Shorter time intervals might be necessary depending on the required resolution (see Note 7).

3.1.5 CCV Imaging

• 1. To image CCV formation, focus on one side of the zebrafish and study the formation of one CCV from a dorsolateral point of view.

• 2. Start observing from 32 hpf. This is preferred even though early CCV endothelial cell migration starts from 20 to 22 hpf due to the curved shape of the yolk and the fact that these vessels migrate on top of the yolk syncytial layer (YSL).

• 3. Slightly turn embryos to a more dorsolateral angle toward the lens, rather than mounting them laterally. The beneficial effect of this mounting technique is a smaller z-stack and a better observation angle to include the entire endothelial cell sheet.

• 4. Use magnification of 20× to monitor collective cell migration and formation of the vessel. Since the size and shape of CCV endothelial cells allow reasonable observation of intracellular compartments as well as lamellipodia at the leading edge, we frequently use 40× and 63× magnification combined with the tile scan function to study this vessel.

• 5. When performing time-lapse imaging, use an interval of 10–15 min when examining cellular migration and vessel formation and an interval of 2–5 min when observing lamellipodia at the front and filopodia at the side of the cell sheet (see Note 7).

3.2 Manipulation Techniques to Study Endothelial Cell Migration

3.2.1 Microinjections

A variety of experimental techniques are based on microinjections into the zebrafish embryo which are performed as detailed elsewhere [60]. Below, we provide a protocol for different experiments that can be performed to genetically manipulate zebrafish embryos. Thus, the effect of specific genes and proteins on endothelial cell behavior can be examined in vivo.

• 1. Prepare injection molds that hold embryos during injection by boiling up 1% agarose in E3 medium using a wedge-shaped injection casting mold and a petri dish. Moreover, prepare your injection needle from glass capillaries by using a micropipette puller.

• 2. Load your needle with your injection mix using microloader pipette tips, and place the needle into the holder of the micromanipulation device (see Note 8). The injection mix should include a visual control such as phenol red in order to control injection accuracy.

• 3. Transfer embryos into the injection mold using a glass pipette and adjust orientation with Dumont forceps.

• 4. Using the micromanipulator, inject either into the embryo’s yolk by passing through the chorion or directly into the cell by entering the cell through the yolk at the back of the embryo (see Note 9). The embryo should be preferably in one-cell stage and maximally in two-cell stage at the time point of injection [58].

• 5. After injection, transfer the embryos into a dish containing E3 medium and place at 28.5 °C. Hereafter, proceed as described above (Subheading 3.1).

3.2.2 Gain of Function/Vascular Specific Overexpression

To create an overexpression scenario, we inject plasmid DNA according to Subheading 3.2.1 using an endothelial cell-specific promoter fragment of the ETS transcription factor fli1a to drive expression of our gene of interest (here a wild-type or a dominant-negative version of Rab5c, fused to mCherry) [32, 61] (see Note 10). We performed these overexpression studies in endothelial cells to assess the effects on tip versus stalk cell selection during ISV formation (see Subheading 3.3.1). Notably, gene expression via DNA plasmids results in a mosaic gain of function outcome, ideally suited to analyze individual cell behavior (tip versus stalk cell) (see Note 11). To obtain ubiquitous gain of function in the whole embryo, in vitro transcribed mRNA can be injected. The injection mix for the experiment should include the respective plasmid, containing genomic integration sites for transgenesis such as the transposase Tol2. The Tol2 elements consist of inverted DNA repeats that are used for genomic integration in a variety of vertebrate systems such es zebrafish, Xenopus, chicken, mouse, and human [62]. In addition to the plasmid DNA, we include tol2 mRNA, which is transcribed in vivo into the transposase performing genomic integration of the plasmid. Lastly, we include phenol red as a visible marker for injection.

• 1. Follow steps 1–3 from Subheading 3.2.1.

• 2. Prepare injection mix for Rab5c-mCherry overexpression: 25 ng/μL plasmid DNA, 1 μL 0.05% phenol red (in water), 250 ng/μL tol2 mRNA. Fill up to 5 μL with water.

• 3. Inject 2 nL per embryo as in step 4 of Subheading 3.2.1.

• 4. Follow step 5 of Subheading 3.2.1 and proceed as mentioned in Subheading 3.1.

3.2.3 Loss of Function/Gene Knockdown or Genetic Knockout

Loss of genetic function can be achieved by performing gene knockdown or genetic knockout experiments For gene knockdown, we use either morpholino antisense oligonucleotides to block translation or splicing of a transcript [14] or a CRISPR/Cas9 combination (several guide RNAs targeting the same gene of interest together with Cas9 protein) and inject into zebrafish embryos according to Subheading 3.2.1 [63]. Respective injection mixes are specified below. Nonetheless, genetic knockdown strategies using morpholinos bear the risk of potential off-target effects, and somatic knockdown strategies using CRISPR/Cas9 result in various genetic modifications with no potential prediction regarding the efficiency of the obtained knockout. Therefore, genetic knockouts are often generated [64,65,66]. The most common strategy nowadays is based on CRISPR/Cas9-dependent gene modification [15,16,17,18]. After injections, the embryos need to be grown into adults and then scored for germline transmission into their F1 offspring, and only after growing the F1 into adulthood, heterozygous carrier fish harboring the same mutation are generated. These are then mated to study homozygous gene knockouts in 25% of the F2 offspring embryos. As an example, we analyzed loss of adhesion protein function (ve-cadherin) on collective endothelial cell migration during CCV formation (see Subheading 3.3.3).

• 1. Follow steps 1–3 from Subheading 3.2.1.

• 2. For morpholino injection, prepare the following injection mix: X ng/μL morpholino (see Note 12), 1 μL 0.05% phenol red. Fill up to 5 μL with water.

  For CRISPR/Cas9 knockdown, prepare this injection mix: 200 pg/nL Cas9 protein (NEB), 500 ng/μL gRNA(s), 1 μL 0.05% phenol red. Fill up to 5 μL with water.

  For CRISPR/Cas9 genome editing (knockout), prepare this injection mix: 300 ng/μL cas9 mRNA, 500 ng/μL gRNA, 1 μL 0.05% phenol red. Fill up to 5 μL with water.

• 3. Inject 2 nL per embryo into the yolk for CRISPR/Cas9 knockdown and 1 nL into the cell for CRISPR/Cas9 genome editing (knockout).

• 4. Follow step 5 of Subheading 3.2.1 and proceed as mentioned in Subheading 3.1.

3.2.4 Drug Treatments

Besides genetic manipulation experiments, the zebrafish is also a great model for applying drug screenings and pharmacological manipulation of protein function. Most drugs can simply be added to the medium. However, solubility, concentration, and timing will affect the efficacy of the treatment (see Note 13). An exemplary treatment is the block of Wnt signaling using IWR-1, to study its effect on hindbrain vasculature formation (see Subheading 3.3.2).

• 1. For the treatment transfer dechorionated embryos in E3 medium into 6-well dishes.

• 2. Replace E3 embryo medium with medium containing freshly DMSO dissolved and in E3 diluted IWR-1 (final concentration of 20 μM) for your incubation period of interest. In our example given below, we incubate the embryos in the inhibitor from 29 to 48 hpf [47].

• 3. Proceed as mentioned in Subheading 3.1 to assess inhibition phenotype.

3.3 Analysis

3.3.1 Analysis of Tip and Stalk Cell Competition During ISV Formation

ISV formation can be used as a model to study different parameters of angiogenesis in the zebrafish. Live endothelial cell migration can be observed by time-lapse imaging of ISV endothelial cells to study migration parameters such as migration speed, cellular polarity, or filopodia formation. Furthermore, some signaling cascades do not affect overall migration speed, but are required to pass certain checkpoints, so that deficiencies result in stalling of ISV development, which is most often observed at the level of the horizontal myoseptum [67,68,69]. Other important parameters can be endothelial cell fate specifications, including arterial/venous vessel fates or tip/stalk cell identity [7, 46, 70]. In this example, we show how to analyze the role of the early endosomal Rab GTPase Rab5C on tip/stalk cell formation [46], using a plasmid driving endothelial cell-specific expression of an mCherry-Rab5C fusion protein via the fli1a promoter, which was injected into Tg(kdrl:GFP)^s843 embryos [11]. Expression of wild-type mCherry-Rab5C was compared to expression of a dominant-negative mCherry-Rab5C mutant and mCherry only (Fig. 3a). As these injections result in mosaic overexpression of mCherry or mCherry-tagged Rab5C in endothelial cells, the mCherry-positive cells are scored for their position within the developing ISV (tip or stalk) in order to assess the effect of Rab5C on tip/stalk cell specification within the developing vessel (Fig. 3b, c).

• 1. Take confocal z-stacks of the ISVs of mCherry-positive embryos around 32 hpf (see Subheadings 3.1.2 and 3.1.3) (see Fig. 3c).

• 2. Open the images with image processing and analyzing software such as ImageJ/FIJI or Imaris to analyze the images. For this analysis, each embryo is observed individually for mCherry-positive cell localizations. Hence, all mCherry-positive endothelial cells must be counted and their localization noted (tip or stalk of the sprout).

• 3. For each embryo, calculate the percentage values of the cellular localization to determine the relative numbers for each localization, e.g.:

  • total mCherry-positive cell number = 5 cells [100%];

  • mCherry-positive cells in the stalk of the sprout = 1 [20%];

  • mCherry-positive cells at the tip of the sprout = 4 [80%].

• 4. Plot percentage values in a graph to visualize mCherry-positive cell contributions (Fig. 3b).

Fig. 3

Single-cell fate analysis during ISV angiogenesis. Mosaic overexpression of either mCherry or mCherry-tagged wild-type (WT) or dominant-negative (DN) Rab5c fusion proteins after DNA injection in vascular endothelial cells leads to altered tip or stalk cell fate, analyzed by the number of mCherry-labeled endothelial cells found in the respective position at 32 hpf. (a) Representation of the DNA constructs injected, with the fli1a promoter fragment [32] driving vascular-specific expression of mCherry or mCherry-Rab5c fusion proteins. Note: due to side effects of fluorophore expression in sensitive stem or precursor cells, we strongly recommend a “fluorophore only” control. (b) Example for quantification of tip/stalk cell numbers in injected embryos, taken from [46]. mCherry-positive cells were scored for their position, the total number of counted cells was set to 100%, and then the relative percentage for tip or stalk position was calculated. It is customary to note the number of embryos (N) as well as the number of cells (n) analyzed for each condition. We show means + SEM. Statistical significance indicates tip cell positioning compared to mCherry control (mCherry: N = 12 embryos, n = 123 cells; mCherry-WT-Rab5C: N = 17, n = 148; mCherry-DN-Rab5C: N = 14, n = 8). ***p < 0.001. Statistical analysis was performed using one-way ANOVA for multiple comparisons. (C) Maximum projections of z-stacks obtained by confocal microscopy (lateral views) of the ISVs with red indicating fli1a-driven mCherry expression after DNA injection and white vascular-specific EGFP expression of Tg(kdrl:EGFP)^s843. Scale bar = 30 μm

3.3.2 Analysis of Sprouting Angiogenesis During CtA Formation

The intracerebral CtAs of the zebrafish are another well-known model to study brain angiogenesis. CtA cells invade the hindbrain parenchyma around 32 hpf to form blood vessels around 48 hpf [48,49,50]. Comparative analyses regarding endothelial cell migration parameters and vessel formation can be assessed in this vascular bed. In this example, we show how to analyze CtA number, endothelial cell connection to the BA, and interconnections between endothelial sprouts in control conditions and in embryos treated with the Wnt-signaling inhibitor IWR-1 (Fig. 4a; see Subheading 3.2.4) [47]. The zebrafish line used in this example expresses actin-fused GFP in endothelial cells to visualize the cytoskeleton and filopodia (Tg(kdrl:Lifeact-GFP)^mu240).

• 1. Take dorsal z-stacks of the CtAs around 48 hpf (see Subheadings 3.1.2 and 3.1.4) (see Fig. 4b).

• 2. Open the images with image processing and analyzing software such as ImageJ/FIJI or Imaris to analyze the images. Here, we started the analysis by counting individual CtA sprouts in separate embryos in order to assess vessel patterning.

• 3. Count the number of sprouts that established a connection to the BA to form a completed and lumenized vessel. Calculate the percentage value of these connections for every embryo, e.g., ten sprouts [100%], of which seven are connected to the BA [70%].

• 4. In order to further assess hindbrain vascular patterning, examine the number of (ectopic) interconnections between individual CtA sprouts. Count the number of interconnections and calculate the percentage value of the total number of sprouts, e.g., ten sprouts [100%], of which five form interconnections [50%].

• 5. Plot the values in individual graphs to visualize the effect of Wnt-signaling inhibition on CtA formation (Fig. 4c).

Fig. 4

Analysis of pattern formation during CtA angiogenesis. Impairment of CtA angiogenesis by pharmacological inhibition of Wnt signaling (IWR-1) during sprouting, migration, anastomosis, and lumen formation between 29 hpf and 48 hpf. (a) Illustration of the experimental timeline. (b) Maximum projections of z-stacks obtained by confocal microscopy (dorsal views) of CtAs after vessel and pattern formation at 48 hpf. The actin cytoskeleton (in white) was visualized by vascular-specific expression of Tg(fli1a:lifeact-EGFP)^mu240. Scale bar = 50 μm. (c) Number of CtAs, the percentage of the present CtAs that formed a connection to the basilar artery (BA), and the percentage of CtAs that formed connections with other CtAs (interconnected), taken from [47]. Values represent means + SD. n = number of analyzed embryos. CtA number (n = 39); proportion of CtAs connecting to the BA (n = 39); proportion of interconnection (n = 33). *p < 0.05, ****p < 0.0001. Statistical analysis was performed using Student’s t-test for single comparisons

3.3.3 Analysis of Collective Endothelial Cell Migration Behavior During CCV Formation

During CCV formation, endothelial cells migrate collectively as a cell sheet which clearly differs from the single tip cell-like sprouting of other blood vessels [6, 24, 51]. In order to assess vascular development in this vascular bed, we quantify the angiogenic parameters that can be measured at individual time points to quantify angiogenic progress. Additionally, performing time-lapse imaging during CCV development allows a vivid evaluation of migration parameters such as endothelial cell migration capacity and directionality. In this example, we analyzed CCV formation in two different genetic backgrounds, wild-type and ve-cadherin mutant zebrafish embryos. VE-cadherin is an important adherens junction molecule which is exclusively expressed in endothelial cells to allow cell-cell adhesion, but it is also involved in regulating directional collective cell migration [51, 71, 72]. For this analysis, we used embryos that express a vascular-specific transgene encoding a nuclear localization signal (nls) fused to GFP and a transgene to visualize actin structures to monitor cellular shape (Tg(kdrl:nls-EGFP)^zf109; (kdrl:Lifeact-GFP)^mu240).

• 1. Perform time-lapse imaging of the developing CCV in wild-type and VE-cadherin-deficient embryos from 32 to 38 hpf (see Subheadings 3.1.2 and 3.1.5) (see Fig. 5a).

• 2. Transfer the time-lapse movies into image processing and analyzing software capable of handling files of bigger size, such as Imaris. Consider that these movies might contain several z-stacks of different embryos with several time points and may therefore amount to a file size of several gigabytes.

• 3. To obtain values for migration track length, speed, and directionality (straightness of the tracks), track individual CCV endothelial cells throughout the complete time-lapse movie. You can either use an automated signal threshold-based identification of cells by targeting the nucleus of a cell, or manually select the nucleus of an individual cell throughout the complete time-lapse series using a Spots tool (e.g., the manual Spot Track function in Imaris) (see Note 14).

• 4. After selecting all cell tracks, export the track length, migration speed, and track displacement (see Note 15). Migration directionality is calculated as the ratio of the track displacement and total track length.

• 5. Plot the values for both conditions together in graphs to visualize migration track length and migration speed to assess migration capacity and the directionality to evaluate migration efficiency (Fig. 5b).

Fig. 5

Analysis of CCV endothelial cell migration from confocal time-lapse movies. Loss of migration directionality upon VE-cadherin deficiency during collective CCV endothelial migration from 32 hpf to 38 hpf. (a) Maximum projections of z-stacks obtained by confocal microscopy (dorsolateral view) of the CCV from 32 hpf to 38 hpf. Migration tracks are superimposed on the images to visualize migration direction. The nuclei and the actin cytoskeleton of endothelial cells are both shown in white and were visualized by vascular-specific expression of Tg(kdrl:nls-EGFP)^zf10 and Tg(fli1a:lifeact-EGFP)^mu240, respectively. Scale bar = 40 μm. (b) Migration track length, migration speed, and migration directionality (straightness of the tracks) [51]. While track length and speed are quantified in order to assess the migration capacity, straightness of the tracks is calculated to qualify the migration behavior. Values represent means + SEM. n = number of analyzed cells (n = 30). ****p < 0.0001. Statistical analysis was performed using Student’s t-test for single comparisons

4 Notes

• 1. Using a higher concentration of tricaine (more than 1 mL in 25 mL E3; i.e., more than the recommended 150 mg/L [73]) will stop the heartbeat of the embryo which might have additional effects on your analysis.

• 2. You can simply use the embryo medium that includes PTU and tricaine to transfer the embryos using a glass pipette. Carefully transfer the embryos to not harm them (we recommend a glass pipette with a rounded tip) and to not destroy the agarose well with the tip of the pipette. Remove well-overflowing liquid to ensure fixed embryo positioning in the wells.

• 3. Before adding the agarose, make sure the embryo position is roughly suitable (e.g., anterior-posterior orientation and lateral/dorsal point of view). Carefully arrange the embryos using a brush or Dumont forceps. Be aware that poking the yolk sac or embryo tissue might harm or kill the embryo. Upon agarose addition, use Dumont forceps to maintain adequate embryo position until agarose solidification.

  When performing time-lapse imaging of developmental stages between 20 and 28 hpf, it is important to consider the growth-dependent embryo changes in body curvature, e.g., straightening of the tail region, so that the desired region can be stably imaged. Therefore, agarose at the embryo tail should be removed to allow embryo bending and growing. After imaging, embryos can be removed from the agarose by using forceps, e.g., to extract DNA and analyze the genotype via polymerase chain reaction techniques.

• 4. It might be necessary to adjust the thickness of the agarose layer on your glass-bottom dish when using the agarose well technique. For inverted microscopes, we recommend a thin layer of agarose in order to ensure that the specimen will stay in optimal focus distance, especially when using higher magnification lenses, which require short focus distance.

• 5. To avoid evaporation, you can also add more mounting solution or E3 embryo medium that includes PTU and tricaine to increase the volume of medium in the imaging dish. However, if you only mount the embryos in single agarose drops, too much liquid medium (that does not contain any agarose) might lead to detachment of the agarose drops including your specimen. In this case, we recommend using agarose-containing mounting solution.

• 6. This might be critical for time-lapse imaging since minor stage drift can lead to noticeable shifts in the z-position of the embryo. You can counteract this drift by additionally using a time-lapse-based autofocus/definite focus function on your confocal microscope.

• 7. Since we use z-stacks to obtain 3D images, scanning time of several embryos might require prolonged imaging time. In this case, total z-stack size, number of z-scans, and the range between two scans in z-direction can be adjusted to decrease scanning time. Furthermore, pixel dwell time can be reduced to allow short imaging time. Although these changes allow to sustain the range of the time-lapse, they may reduce image quality.

• 8. Make sure to work RNAse free when preparing an injection mix containing RNA and when loading it into the needle.

• 9. Depending on the experimental design, you should consider direct injection into the cell. Injecting transgenesis components directly into the cell leads to an earlier integration of the plasmid into the embryo’s DNA. This might lead to a higher number of cells expressing the transgene in the embryo at the time of analysis and to a higher probability of germline integration.

• 10. It is also very common to use the vascular endothelial growth factor receptor 2 (vegfr2/kdrl) promoter when a gene of interest should be expressed in endothelial cells [32, 61].

• 11. Alternatively, cell transplantation studies can be used to further examine whether your gene of interest acts in a cell-autonomous way to control processes such as tip cell fate decision. For instance, cells from a mutant donor embryo can be transplanted into a wild-type host embryo in order to study the cellular fate of wild-type and gene knockout endothelial cells during tip and stalk cell selection or to investigate the significance of a specific signaling pathway during ISV formation [74,75,76].

• 12. The working concentration that is needed actually depends on the morpholino. It is recommended to start with a concentration of 1 ng/nL to start with 2 ng of morpholino. Increase the doses to assess for the desired phenotype and for potential toxicity. We recommend to be cautious when using more than 5 ng of morpholino since this has been shown to cause potential side effects such as p53 upregulation [64, 65].

• 13. In general, dechorionation prior to the treatment is necessary since the chorion provides a natural barrier to protect the embryos. Furthermore, depending on the inhibitor, different solvents have to be used. Pay attention to the maximum concentration as well as the exposure time of these solvents that should be used on zebrafish embryos without disturbing their development. For example, a DMSO concentration of more than 0.1% should not be exceeded, as that disturbs heart development. Moreover, when embryos are mounted for imaging in agarose, make sure to include the drug in the agarose mixture to maintain pharmacological efficiency during imaging.

• 14. In the case of cell divisions, tracks can be split up into two new separate tracks that originate from one track to follow the complete migration track of the original cell.

• 15. The migration speed can also be calculated as the track length divided by the time passing while tracking. The displacement describes the distance between the first and last position of the cell and can be used to obtain the directionality value for the migration.