Combining cell transplantation, cytoskeletal labeling and loss/gain of function approaches, this protocol describes how the migrating zebrafish prospective prechordal plate can be used to analyze the function of a candidate gene in in vivo cell migration.
Cell migration is key to many physiological and pathological conditions, including cancer metastasis. The cellular and molecular bases of cell migration have been thoroughly analyzed in vitro. However, in vivo cell migration somehow differs from in vitro migration, and has proven more difficult to analyze, being less accessible to direct observation and manipulation. This protocol uses the migration of the prospective prechordal plate in the early zebrafish embryo as a model system to study the function of candidate genes in cell migration. Prechordal plate progenitors form a group of cells which, during gastrulation, undergoes a directed migration from the embryonic organizer to the animal pole of the embryo. The proposed protocol uses cell transplantation to create mosaic embryos. This offers the combined advantages of labeling isolated cells, which is key to good imaging, and of limiting gain/loss of function effects to the observed cells, hence ensuring cell-autonomous effects. We describe here how we assessed the function of the TORC2 component Sin1 in cell migration, but the protocol can be used to analyze the function of any candidate gene in controlling cell migration in vivo.
In multicellular organisms, cell migration is essential both for the development of the embryo where it ensures the organization of cells into tissues and organs, and for adult life, where it takes part to tissue homeostasis (wound healing) and immunity. In addition to these physiological functions, cell migration is also involved in diverse pathological situations, including, in particular, cancer metastasis.
Cell migration has been analyzed in vitro for decades, providing an overall understanding of the molecular mechanisms ensuring cell movements on flat surfaces. In vivo however, cells are confronted by a more complex environment. It clearly appeared in the past years that migration within an organism may be influenced by external cues such as the extracellular matrix, neighboring cells or secreted chemokines guiding migration, and that the mechanisms driving cell migration may vary from what has been described in vitro1,2. The mechanisms ensuring in vivo cell migration have received less attention so far, mainly because of the increased technical difficulty, compared to in vitro studies. In vivo analysis of cell migration in particular requires direct optical access to migrating cells, techniques to label unique cells in order to see their dynamics and morphology, as well as gain or loss of function approaches to test the role of candidate genes. So far, only a few model systems harboring these characteristics have been used to dissect in vivo cell migration3.
We recently used the migration of the prospective prechordal plate in early zebrafish embryos as a new convenient model system to assess the function of candidate genes in controlling in vivo cell migration4,5. Prospective prechordal plate (also known as anterior mesendoderm) is a group of cells forming at the onset of gastrulation on the dorsal side of the embryo. During gastrulation this group collectively migrates towards the animal pole of the embryo6-8, to form the prechordal plate, a mesendodermal thickening, anterior to the notochord, and underlying the neural plate. The anterior part of the prechordal plate will give rise to the hatching gland, while its posterior part likely contributes to head mesoderm9. Thanks to the external development and optical clarity of the fish embryo, cell migration can be directly and easily observed in this structure.
Cell transplantation is a very potent technique that allows for the rapid and easy creation of mosaic embryos10. Expressing fluorescent cytoskeletal markers in transplanted cells results in the labeling of isolated cells, the morphology and dynamics of which can be easily observed. Combining this to loss or gain of function approaches permits the analysis of cell-autonomous functions of a candidate gene.
The presented protocol describes how we assessed the function of the TORC2 component Sin1 in controlling cell migration and actin dynamics in vivo5. But, as mentioned in the results and further discussed, it could be used to analyze the potential implication of any candidate gene in controlling cell migration in vivo.
Note: Figure 1 presents the outline of the protocol.
1. Preparation of the Needles for Injection and Transplantation
Note: Needles can be prepared at any time and stored. Keep them in a Petri dish, on a band of modeling clay. Seal the dish with parafilm to protect from dust.
2. Preparation of the Dishes for Injection and Cell Transplantation
3. Collection of Embryos and Injection
4. Preparation of the Embryos for the Cell Transplantation
5. Cell Transplantation
6. (OPTIONAL) Single Cell Transplantation
Note: In the embryo, cells adhere to each other, so that it is difficult to draw only one cell in the transplantation needle. We developed a modified protocol to easily transplant single prechordal plate progenitor cells. The idea is to dissociate cells prior to transplantation. Because isolated prechordal plate progenitor cells tend to lose their identity, we genetically impose them a prospective prechordal plate identity, by activating the Nodal signaling pathway, in absence of the Sox32 transcription factor. We have verified that these induced cells behave like endogenous prechordal plate progenitor cells8. Below are the specific steps to perform single cell transplantations. Cell dissociation is achieved by placing embryos in Calcium-free Ringer11, dissecting an explant and mechanically stirring it.
7. Embryo Mounting
8. Live Imaging
9. Cell Dynamics Analysis
The presented technique was used to analyze the role of Sin1, one of the core components of the Tor complex 2 (TORC2), in controlling in vivo cell migration. The use of cell transplantation permits labeling of isolated cells and analysis of cell-autonomous effects. Movie S1 shows the migration of transplanted prechordal plate progenitor cells. Actin labeling with ABP140 allows the easy visualization of actin-rich cytoplasmic protrusions. We measured their frequency and orientation. Wild-type cells produce frequent large cytoplasmic protrusions oriented in direction of the animal pole, i.e. in the direction of migration (Figure 3B). Loss of function of Sin1 leads to a drastic reduction of the number of protrusions, and randomization of the remaining ones, demonstrating the importance of sin1 in controlling actin-rich protrusion formation. Interestingly, this phenotype can be rescued by expression of a constitutively active form of Rac115, strongly suggesting that TorC2 controls actin dynamics and cell protrusion formation through Rac1 (Figure 3B).
The presented technique was also used to characterize the role of Arpin, a recently identified inhibitor of the Arp2/3 complex, on cell dynamics4. Loss of Arpin function leads to an increase in protrusion frequency (average rate of cells harboring a protrusion at a given time). This could be due either to more frequent protrusion formation or to an increase in protrusion stability. Measuring protrusion lifetime revealed that, in absence of Arpin, protrusion temporal persistence is doubled (Figure 3C). This is consistent with the role of Arpin as an Arp2/3 inhibitor, which would facilitate protrusion retraction, and suggests that Arpin affects protrusion frequency by modulating protrusion stability, rather than protrusion initiation.
Figure 1: Outline of the Procedure. Tg(gsc:gfp) embryos are injected at the 4-cell stage (1 hr post fertilization). After 5 hr at 28 °C, embryos showing ABP140-mCherry positive cells within the shield (GFP +) are selected as donor embryos. Shield cells are drawn up within the transplantation needle and transferred into the shield of a host embryo. After 0.5 hr of recovery, host embryos are mounted and imaged with an epifluorescent microscope (40X objective). hpf: hours post fertilization. Please click here to view a larger version of this figure.
Figure 2: Transplantation Dish and Imaging Chamber. (A) Transplantation dish with individual wells. (B) Mold for cell transplantation dish. (C) Home-made imaging chamber. (D) Schematic drawing of the home-made imaging chamber. Scale bar = 1 cm. Please click here to view a larger version of this figure.
Figure 3: Results from Cell Dynamics Analysis. (A) Scheme of protrusion angle measurement. (B) Analysis of cell protrusion orientation and frequency. In absence of Sin1 (Mo-Sin1), cells emit fewer protrusions and the remaining protrusions are randomly oriented. This phenotype can be rescued by expression of a constitutively active form of the GTPase Rac1 (Mo-Sin1 + CA-Rac1). Modified from Dumortier and David, 20155. (C) Analysis of protrusion lifetime. In absence of Arpin (Mo-Arpin), protrusion frequency is increased. This is due to an increase in protrusion lifetime. Reintroducing an arpin RNA insensitive to the morpholino restores this hyper protrusive phenotype. Modified from Dang et al., 20134. Scale bar = 10 µm. A: Anterior; P: Posterior, L: Left, R: Right. Please click here to view a larger version of this figure.
Movie 1: Sin1 controls formation of cell protrusions, through Rac1 (Right click to download). Transplanted prechordal plate progenitor cells, injected with ABP140-mCherrry RNAs and a control morpholino, or the sin1 morpholino, or the sin1 morpholino and RNAs for the constitutive form of Rac1. Protrusion frequency and orientation were measured on these images.
This protocol presents an easy way to study the role of a candidate gene in cell migration in vivo, by combining the creation of chimeric embryos using cell transplantation with live imaging.
Creation of mosaic embryos
Studying the dynamics of a cell requires the visualization of its contour to analyze cytoplasmic extensions. This can be achieved by labeling isolated cells in an otherwise unlabeled – or differently labeled – environment, thus offering good visual contrast.
An easy way to stochastically label a portion of cells in the embryo is to inject plasmidic DNA at the 1-cell stage16. The injected plasmids form aggregates that are randomly and unequally segregated in cells during cell division. This leads to the expression of the plasmid in a random subset of cells, and thus represents a very fast, easy and non-invasive method for generating mosaic embryos. However, cells expressing the injected plasmid tend to form clusters, and the probability of finding embryos containing few isolated expressing cells in the prospective prechordal plate is quite low. Furthermore, use of plasmidic DNA offers very limited control of the expression level of the injected construct. The creation of chimeric embryos by cell transplantation, although technically more difficult than plasmidic DNA injection, offers a number of advantages: control of the number of transplanted cells, control of the expression level of the constructs (RNA injection), and last but not least, cell transplantations allow to test for cell-autonomous effects, by creating mosaic embryos in which transplanted cells contain loss or gain of function constructs. Conversely, transplantations can also be used to assess the non-autonomous role of the environment by transplanting wild-type cells into embryos that have been modified. This can be of particular interest for testing for instance, the function of extracellular matrix components that are particularly relevant for cell migration analysis.
A detailed protocol for cell transplantation has already been described10. Compared to this protocol, we would like to discuss two main differences in our procedure.
The first difference is related to the stage of donor embryos injection. According to our experience, embryos injected at the 1-cell stage with RNAs of a fluorescent protein do not express homogeneously the fluorescent protein in all cells, due to an incomplete diffusion of the RNAs in the cell. Injection of donor embryos at the 4-cell stage in one of the four cells allows to get a homogeneous clone of cells, which is of particular interest when RNAs of the fluorescent protein are co-injected with other RNAs that are not traceable.
The second main difference is the use of air instead of oil in the transplantation syringe. The main drawback of an air-filled system is the inertia resulting from the elasticity of the air. Oil on the contrary, being incompressible, offers good control of suction and pressure. However, with a little training, and by maintaining the interface between air and embryo medium in the thin, tapered part of the needle, the control of suction and pressure with an air-filled system is sufficient for transplanting cells at the shield stage. For later stages, as cells become smaller and more cohesive, the suction requires a high pressure that needs to be very precisely controlled, which can't be achieved with an air-filled setup. Using air instead of oil eases the setup, as filling the system with oil without any air bubble is tricky. Furthermore, it avoids filling the transplantation needle with oil. This allows transplantation needles to be reused, and hence to be carefully prepared. We feel that a needle without sharp edges and of the correct diameter is key to successful transplantation. This transplantation step is clearly one of the critical steps in the protocol, which requires practice before being easily performed.
We have also proposed a specific procedure for transplanting single cells, which consists in dissociating cells prior to transplantation, in order to draw a unique cell in the transplantation needle. This implies that transient cell dissociation will not modify their identity and/or behavior. For prechordal plate progenitor cells, we genetically impose cell identity, and have carefully checked that these cells behave like non-dissociated ones. However, we cannot formally exclude that dissociation and/or genetic induction induce unnoticed modifications in the cells. Transplanting single cells without dissociating them is feasible. Cells should be drawn up carefully in the transplantation needle. However, at least in our hands, success rate is quite low, either because more than one cell gets into the needle, or because the cell shears when drawn up and dies in the needle or once transplanted.
Epifluorescence versus fast confocal imaging
The use of cell transplantation to create mosaic embryos allows for the labeling of isolated cells, separated from other labeled cells. In this context, good imaging of cell morphology and dynamics can be achieved with standard epifluorescence microscopy, as proposed in the protocol. This has the obvious advantage of being a relatively wide-spread equipment and a low-cost alternative to more expensive confocal imaging systems.
For precise subcellular localizations, confocal imaging can be used to improve resolution, in particular axial resolution. To still get sufficient temporal resolution, fast confocal imaging should be used. From our experience, spinning-disc microscopes offer a good compromise between resolution, speed and photo-toxicity. Fast scanning confocal microscopes (like resonant scanning microscopes) are good options as well, but less frequent and more expensive.
Cytoskeleton labeling
Good labeling of actin filaments and their dynamics is crucial to study mechanisms of actin-based cell migration. Although membrane-bound fluorescent markers are more efficient to analyze the cell outlines, actin labeling allows a much better visualization of the cell protrusions. In particular, if three labeled cells get in contact, it is very difficult to identify a cytoplasmic extension located between two neighboring cells as the outline of the extension and the membrane of the neighboring cells can get mixed up. Labeling actin filaments allows the unambiguous detection of actin-based cell protrusions, on which we focused. If cell morphology was to be quantified, a membrane labeling would be preferable. Another option is to use both labeling, either using 3-color imaging (one for the transgenic line, one for actin and one for membrane), or by GFP labeling the membrane. In the transgenic line, GFP is cytoplasmic, and goosecoid-GFP positive cells can thus be identified, even if the membrane is GFP labeled.
Over the past years, a number of probes have been used to label actin in live samples. The first ones were direct fusions to actin monomers. These presented two major drawbacks. First, they labeled all actin and not specifically polymerized actin. Second, they were frequently toxic. The probes currently in use are filamentous actin binding domains fused to fluorescent reporters. In this protocol, we used ABP14017 (actin binding domain of the yeast actin binding protein 140), which is currently the most frequently used marker for filamentous actin, and labels all filamentous actin. Utrophin18 (calponin homology domain) also labels filamentous actin, but appears to label only stable filaments. This difference has been used to identify dynamic actin filaments, as the ones labeled by ABP140 and not Utrophin19.
Recently it has been reported in fly that ABP140 and Utrophin could have toxic effects, in particular over long expression20. Even though we have not noticed any migration defects in ABP140-labeled cells, we cannot exclude that ABP140 expression may perturb endogenous actin dynamics. A third probe for filamentous actin was recently reported21. F-tractin (actin-binding domain from rat inositol triphosphate 3-kinase) has been described as a faithful reporter of filamentous actin, which would not show toxic effects20,22. To our knowledge, the use of F-tractin has not been reported in zebrafish yet.
In addition to actin, many other cell constituents could be labeled to improve our understanding of the cell migration. This includes other cytoskeletal components like myosin, microtubules, centrosomes, as well as known regulators of cell migration (activated forms of the small GTPases Rho, Rac and Cdc42, fluorescent probes for PIP3…) or proteins involved in cell adhesion (integrins, cadherins…).
Overall, this protocol regroups a number of previously described tools and techniques to propose a rapid and easy system to test the involvement of a candidate gene in controlling cell migration in vivo. We used prospective prechordal plate migration as it is an interesting model of directed collective migration, and because of the available transgenic line labeling it. However, a similar protocol could be adapted to analyze other migration events occurring during development.
The authors have nothing to disclose.
We thank F. Bouallague and the IBENS animal facility for excellent zebrafish care. Research reported in this publication was supported by the Fondation ARC pour la recherche sur le cancer, grants N° SFI20111203770 and N° PJA 20131200143.
Glass capillaries (outside diameter 1.0 mm, inside diameter 0.58 mm) | Harvard Apparatus | 300085 | standard thickness |
Glass capillaries (outside diameter 1.0 mm, inside diameter 0.78 mm) | Harvard Apparatus | 300085 | thin-walled |
Penicillin-Streptomycin | Sigma-Aldrich | P4333 | 10 000 units penicillin and 10 mg streptomycin per ml |
fine tweezers | Dumont Fine Science Tools | 11254-20 | 5F |
glass bottom dishes | MatTek | P35G-0-10-C | |
Air transjector | Eppendorf | 5246 | |
Micro-forge | Narishige | MF-900 | |
Microgrinder | Narishige | EG-44 | |
Micromanipulator (for injection) | Narishige | MN-151 | |
Micromanipulator (for cell transplantation) | Leica | Leica Micromanipulator | |
Hammilton Syringe | Narishige | IM-9B | |
Micropipette puller | David Kopf Instruments | Model 720 | |
Transplantation mold | Adapative Science Tools | PT-1 | |
Needle holder | Narishige | HI-7 | |
Tube connector | Narishige | CI-1 | |
PTFE tubing | Narishige | CT-1 |