Application of Infrared Laser to the Zebrafish Vascular SystemSignificance
Gene Induction, Tracing, and Ablation of Single Endothelial Cells
Objective—Infrared laser–evoked gene operator is a new microscopic method optimized to heat cells in living organisms without causing photochemical damage. By combining the promoter system for the heat shock response, infrared laser–evoked gene operator enables laser-mediated gene induction in targeted cells. We applied this method to the vascular system in zebrafish embryos and demonstrated its usability to investigate mechanisms of vascular morphogenesis in vivo.
Approach and Results—We used double-transgenic zebrafish with fli1:nEGFP to identify the endothelial cells, and with hsp:mCherry to carry out single-cell labeling. Optimizing the irradiation conditions, we finally succeeded in inducing the expression of the mCherry gene in single targeted endothelial cells, at a maximum efficiency rate of 60%. In addition, we indicated that this system could be used for laser ablation under certain conditions. To evaluate infrared laser–evoked gene operator, we applied this system to the endothelial cells of the first intersegmental arteries, and captured images of the connection between the vascular systems of the brain and spinal cord.
Conclusions—Our results suggest that the infrared laser–evoked gene operator system will contribute to the elucidation of the mechanisms underlying vascular morphogenesis by controlling spatiotemporal gene activation in single endothelial cells, by labeling or deleting individual vessels in living embryos.
The vascular system is necessary for blood supply, thereby providing oxygen and nutrition to all organs. Using time-lapse analysis of living zebrafish embryos transgenic for fli1:EGFP, in which the endothelial cells specifically expressed enhanced green fluorescent protein, we studied the formation of the vascular system of the brain during early ontogeny. Our morphological data have indicated that vascular morphogenesis proceeded with a regular time course to form a uniform structure, which led to the subsequent interest in understanding the underlying mechanisms. Although many studies have investigated the morphogenesis of a developing embryo, little is known about the mechanisms underlying the formation of the vascular system. At present, vasculogenesis and angiogenesis have been advocated as being part of vascular development. In vasculogenesis, the primary vascular network was first formed from the lateral plate mesoderm, and then, further blood vessels were generated by both sprouting and nonsprouting angiogenesis. The vascular plexus was rapidly remodeled, and flow dynamics were believed to play critical roles in determining the processes underlying vascular morphogenesis, including the differentiation of arteries and veins.1 However, the formation of the early vascular network in the developing trunk progressed without blood flow in zebrafish2; similar data have been obtained for the vascular development of the brain (unpublished data). These results suggest that vascular morphogenesis is regulated by genetic cues, indicating that precise analysis of gene function is required for its complete understanding.
Several methods have been used to analyze specific gene functions associated with morphogenesis in developing embryos. Recently, the use of conventional knockout mice using the Cre-loxP system3 or microelectroporated chicken embryos4 has become common technique. To a certain extent, spatiotemporal gene regulation can be achieved by these methods, whereas strict gene regulation in a single cell is obligatory for the precise understanding of its function, especially in a noncell autonomous system, such as morphogenesis. In contrast, the heat shock protein (hsp) promoter has frequently been used for ectopic gene induction in vivo. A combination of the hsp promoter and laser irradiation enables the spatiotemporal regulation of gene induction in targeted cells.5,6 The heat shock response is widely conserved in all organisms,7 and is thus easily applied to both model and nonmodel organisms in the absence of specific DNA enhancers. The guidance mechanism of spinal motor axons by semaphorin 3a1 has been reported using sublethal laser–induced transgene expression in targeted single neurons.8
Because this method used a coumarin dye laser (440 nm), it may be suitable to irradiate superficial tissue, but not deep-lying tissue, such as the vasculature. In addition, the mechanism underlying the activation of the hsp promoter is still unclear, although the DNA damage induced by laser irradiation might be involved. We thus focused on infrared (IR) laser (1480 nm), which has a superior ability to heat water and can efficiently access deep-lying tissues. The IR laser–evoked gene operator (IR-LEGO) system is a new microscopic method optimized to heat cells by using an IR laser. We confirmed that this system was efficient in regulating spatiotemporal gene expression, and reported IR laser–mediated gene induction in single targeted cells in nematodes (Caenorhabditis elegans),9 and the local gene expression in various tissues, such as the muscle, notochord, and retina in some living vertebrates, for example, zebrafish (Danio rerio) and medaka (Oryzias latipes).10
In this study, we applied this IR-LEGO system to the vascular system in zebrafish and established, to the best of our knowledge, for the first time, an excellent method to induce laser-mediated gene expression in single targeted endothelial cells in vivo. We optimized the irradiation conditions, which resulted in raising the efficiency of laser-mediated gene induction up to 60%. Furthermore, we applied this method to the endothelial cells of the first intersegmental arteries (SeAs) to evaluate the system, and revealed their contribution in connecting the vascular systems of the brain and spinal cord. Our data indicate that the IR-LEGO system is a useful method for the investigation of vascular morphogenesis in vivo.
Materials and Methods
Materials and Methods are available in the online-only Supplement.
IR Laser–Mediated Gene Expression in Targeted Single Endothelial Cells
Previously, we reported IR laser–mediated local gene induction in various tissues in some vertebrates and higher plants.10 In this study, we attempted to control gene expression in the single endothelial cells in zebrafish. We identified each single endothelial cell by using zebrafish transgenic for fli1:nEGFP, and we used the IR laser to irradiate them individually, and induced the expression of mCherry by using the IR-LEGO (Figure 1A–1C). The efficiency of IR laser in targeting the cell was confirmed by observing the reduction of fluorescence intensity (Figure 1D–1F). To optimize the irradiation conditions, we applied various irradiation powers (17.4, 15.6, 13.8, 12.1, and 10.8 mW) to each single SeA endothelial cell for 1 s, respectively. After 18 hours, we evaluated laser-mediated gene induction by using confocal microscopy (Figure 2). Two types of cells expressed mCherry fluorescence after irradiation with various IR laser-power sources: one type included the nontargeted muscle cells that were superficial to the irradiated endothelial cells (Figure 2, arrows), and the other type included the targeted single endothelial cells (Figure 2, arrowheads). To investigate the optimal irradiation conditions, we compared IR laser power–dependent gene induction more precisely; 20 single endothelial cells were irradiated with each laser power condition. The efficiency of gene induction in the targeted and nontargeted cells is summarized in Table 1. The 1-s-long irradiation with IR laser at 17.4, 15.6, 12.8, 12.1, and 10.8 mW could induce gene expression in the targeted single endothelial cells; IR laser irradiation at 12.1 mW achieved the highest efficiency of 30% (6/20). In contrast, gene expression in the nontargeted cells, such as muscle cells, was also confirmed in each irradiated embryo. Suppressing gene induction in the nontargeted cells was difficult under conditions of higher power, such as 17.4 and 15.6 mW (Figure 2A and 2B, and Table 1). However, reducing the irradiation power decreased the efficiency in nontargeted cells (Table 1), and led to the weakening of the fluorescent intensity (Figure 2C–2E). In this case, the 1-s-long irradiation with the 10.8-mW IR laser had the least influence on the nontargeted cells, but the efficiency to induce the expression of mCherry in the targeted cells was inadequate. We, therefore, determined that the 1-s-long irradiation with the 12.1-mW IR laser was the optimal condition to induce gene expression in the endothelial cells, with a lesser influence on the nontargeted cells (Figure 2D and Table 1). Furthermore, we studied the time course of laser-mediated gene expression. We applied various irradiation powers (17.4, 15.6, 13.8, 12.1, and 10.8 mW) to single SeA endothelial cells in the same embryo for 1 s each; mCherry expression was then observed by fluorescence microscopy every 2 hours after irradiation (Figure IIA–IID in the online-only Data Supplement), and 18 hours after irradiation by confocal microscopy (Figure IIE–IIG in the online-only Data Supplement). The first 2 irradiations at 17.4 and 15.6 mW resulted in mCherry induction only in the nontargeted muscle cells (Figure IIF and IIG in the online-only Data Supplement, arrows), whereas the fourth irradiation at 12.1 mW succeeded in inducing laser-mediated gene expression in the targeted single endothelial cell (Figure IIF and IIG in the online-only Data Supplement, arrowheads). The expression of mCherry in the nontargeted muscle cells was mild at 4 hours after the irradiation (Figure IIB in the online-only Data Supplement, arrows), but grew stronger with time (Figure IIC and IID in the online-only Data Supplement, arrows). The targeted endothelial cell also expressed mCherry 6 hours after irradiation, which also grew stronger (Figure IIC and IID in the online-only Data Supplement, arrowheads). These observations will be valuable to assess the effect of the expression of particular genes in endothelial cells by using this system.
Efficiency of IR Laser–Mediated Gene Induction
As we succeeded in optimizing the irradiation conditions to induce laser-mediated gene expression in single targeted endothelial cells with minimum effect on nontargeted cells, we next investigated its efficiency by using this optimized condition (12 mW for 1 s). We selected 1 endothelial cell from each SeA, and in total, 70 single endothelial cells from the SeAs at 2 dpf were irradiated. As a result, the gene expression of mCherry was induced in 25 irradiated single endothelial cells with an efficiency of 35.7% (25/70). The expression in nontargeted cells was also observed, and its efficiency was 12.9% (9/70). We then decreased the irradiation power from 12 mW to 10.8 mW and increased the frequency from 1 time to 3 times (pulsed irradiation) to further improve efficiency. Under these conditions, we used the IR laser to completely irradiate 30 single endothelial cells. As a result, the gene expression of mCherry was induced in 18 irradiated endothelial cells, that is, the efficiency increased to 60% (18/30). Accordingly, the gene expression in nontargeted cells also increased to 33.3% (10/30). These results are summarized in Table 2, and select images of irradiated embryos in 2 different conditions have been presented as examples (Figures III and IV in the online-only Data Supplement). Comparison of these results suggests an inconvenient relationship, whereby increasing the frequency of irradiation significantly improves the efficiency of laser-mediated gene induction in the targeted cells, whereas the induction in nontargeted cells increased at the same time. However, the maximum induction rate of 60% was an adequate achievement with respect to controlling spatiotemporal gene expression in vivo. Our data suggest that the IR-LEGO system will be a powerful tool for the functional analysis of a particular gene in vivo associated with vascular morphogenesis, especially for a system that satisfies the requirement of complete deletion of gene induction in nontargeted cells.
Tracing the Targeted Single Endothelial Cells of the First Intersegmental Arteries
The induction of mCherry in the targeted endothelial cells enabled us to trace them throughout the ontogeny within the period of mCherry expression. We used this advantage to analyze the formation of the connecting portion of the vascular systems between the brain and spinal cord. Morphogenesis in this region has previously been reported using microangiography, and it showed that the primordial hindbrain channel (PHBC) and basilar artery (BA) extended caudally and connected with the dorsal longitudinal anastomotic vessel (DLAV) at 1.5 to 2.0 dpf, which was formed by the longitudinal anastomosis of the branches of the intersegmental vessels in the dorsal region of the neural tube (schematic illustrations are indicated in Figure 3A and 3B).11 However, this method allowed the observation of only the lumen of the vessels, not the growing vessels before tube formation, and therefore, it is still unknown how these vessels were connected. We focused on the first SeAs as the candidate connecting these vascular systems, because their contribution to this connection had not been revealed. They could be observed obviously in the fli1:nEGFP transgenic zebrafish embryos (Figure 3C), although they could not be identified by microangiography (Figure 3A). To investigate their role in this connection, we labeled and traced them after irradiation with the IR laser. After 16 to 18 hours, we confirmed the localization of the green/red fluorescence-labeled endothelial cells dorsally and laterally after these vessels became connected (Figure 3D and 3E). We applied 2 irradiating conditions for these labeling experiments; a 1-s-long single pulse from a 12-mW laser and three 1-s-long pulses with a 10.8-mW laser (pulse irradiation). The efficiency labeling of the first SeA of the former condition was 16.7% (3/18), and that of the latter was 20% (2/10). In total, we succeeded in labeling 5 targeted endothelial cells of the first SeAs, and the mCherry-labeled cells from the first SeAs were located in the connecting portion of the BA, PHBC, and DLAV, suggesting that the endothelial cells of first SeAs bridge the vascular systems of the brain and spinal cord. Our labeling data for the first SeAs are summarized in Figure 3F and 3G, and we assume that the first SeAs could not be represented by microangiography because of the lack of connection with the dorsal aorta.
Ablation of the Endothelial Cells in the First Segmental Arteries
Because IR-laser is capable of heating water, irradiation overdose can cause injury to the cells. We used this property to ablate particular vessels in the formation of the vascular system. For the first target of the ablation study, we again selected the first SeA. We applied a high-power flash irradiation of IR laser (70 mW for 8-ms) to the endothelial cells, and evaluated its influence over the morphogenesis of the vascular systems between the brain and spinal cord. Whether this attempt to ablate the targets succeeded was judged by the disappearance of nucleus-localized enhanced green fluorescent protein. The change in the connecting portion of the BA, PHBC, and DLAV was observed by confocal microscopy 16 to 18 hours after ablation, in both lateral and dorsal views (Figure 4A and 4B). Although PHBC usually extends obliquely from the caudal head region to the dorsal end of the second SeA, PHBC on the irradiated side did not connect to the second SeA; that is, their extension stopped around the ablated first SeA (Figure 4, black arrows). Furthermore, the connection of the BA and DLAV on the irradiated side diminished (Figure 4, white arrowheads), and alternatively, the collateral blood vessel formed between the BA and PHBC (Figure 4, black arrowheads), whereas the BA in the nonirradiated side connected to the DLAV bending laterally (Figure 4, white arrows). We totally ablated 7 embryos (with the following condition: 70 mW, 8 ms) and observed the effects. The PHBC and DLAV of the ablated side were not connected in all the embryos (Figure 4, black arrows), whereas the bridge between the BA and DLAV (Figure 4, white arrowheads) diminished in 6 of the embryos. (In the rest, the ectopic vessel from the collateral vessel between the BA to PHBC connected to the DLAV.) To reconfirm the vascular formation of the ablated and nonablated first SeAs, we performed confocal microangiography. As a result, the disconnection between the BA, PHBC, and DLAV was clearly demonstrated on the ablated side (Figure V in the online-only Data Supplement). These ablation data for the first SeA are summarized in Figure 4C and 4D. To examine the influence of the overdose of irradiation to the surrounding tissues, we observed the targeted single endothelial cells by both bright-field and fluorescent microscopy before and after the irradiation (Figure VI in the online-only Data Supplement). After the ablation, the nuclear-localized fluorescence in the targeted cell weakened and finally diminished. Burn injury of the ablated cell was also confirmed in the bright-field image, whereas obvious defects to the cells in the surrounding tissues, such as the somites and neural tubes, could not be observed (we confirmed diminishment of the connecting vessels in this embryo). Furthermore, we stained some ablated embryos by 4',6-diamidino-2-phenylindole 16 to 18 hours after the ablation and observed it using confocal microscopy (Figure VII in the online-only Data Supplement). The obvious difference between the ablated and nonablated sides could not be confirmed. Taken together, the vessels diminished by the ablation of the first SeAs were well corresponded to the labeled vessel in Figure 3; therefore, these data are also useful to investigate the contribution of the first SeAs in the vascular morphogenesis between the brain and spinal cord.
We applied the IR-LEGO system to the zebrafish vascular system to establish spatiotemporal gene regulation. Previous studies using electroporation,4 point laser,6 and the IR-LEGO system10 enabled only local gene regulation in vertebrates. In this study, we isolated each endothelial cell by using nucleus-localized enhanced green fluorescent protein and optimized the irradiation conditions needed to decrease gene induction in nontargeted cells. Thus, we achieved precise control of gene expression in single endothelial cells, which were extreme thin and contiguously surrounded by pericytes and muscle cells (Figure 2). Another single-cell gene expression system in zebrafish using a high-power pulse laser has recently been reported.12 This system uses physical impulses of pulse irradiation for transferring injected material to the target cells from the surroundings, hence requires DNA or mRNA microinjection to the neighboring area for gene expression. This system does not require transgenic strains; hence, it is suitable for a nonmodel animal, but is not appropriate for our objectives. Recently, we revealed how the vasculature in the brain was developed during early ontogeny by time-lapse analysis (unpublished data). Gene regulation in the targeted single endothelial cells in vivo will enable the analysis of the influence of a particular gene over the morphogenesis of each vessel in these regions.
Repeating the irradiation several times improved the efficiency of gene induction and led to a maximum induction rate of 60% (Table 2). However, the induction rate of nontargeted cells also rose by 3-fold. When the IR laser was used for irradiation, the endothelial cells shrunk as if evading the heat stress. The surrounding muscle cells may also contract after this stress and cause the irradiated endothelial cells to transfer. These movements seemed to influence the leakage of heat stress and made it difficult to diminish gene induction in nontargeted cells. Furthermore, applying the irradiating IR laser to the tissue resulted in intense heating of the elliptic region elongated vertically just before the focus,9 suggesting that the muscle cells located in the vertical region of the optical axis were easily influenced by the laser. We attempted some modification of the IR-LEGO system by interrupting the center of the irradiated laser before focusing and tried to reduce the leakage to the surrounding cells in the vertical region. Eliminating ectopic heating will require further improvements to the optical system in the IR-LEGO system. In contrast, we were able to achieve deletion of gene expression in the nontargeted cells from another viewpoint by using tissue-/cell-specific promoters and a recombination system. We then attempted to use the Cre-loxP system3 to reduce the minor leakage in nontargeted cells and establish double-transgenic zebrafish line with hsp:Cre and kdrl:loxP:nlsGFP:loxP:mCherry (the promoter of the kdrl gene also induces gene expression specifically in the endothelial cells). Using the embryos from these transgenic organisms, we noted that laser-mediated induction of the Cre gene only influenced the endothelial cells, in which the downstream effecter of loxP sequence could be expressed by genome recombination of Cre. These improvements will elevate the value of IR-LEGO system as the precise in vivo gene regulation method in near future.
Because we succeeded in identifying the optimized condition of irradiation, we tried to label and trace the single endothelial cells of the first SeAs to elucidate their contribution to the connection between the vascular systems of the brain and spinal cord (Figure 3D and 3E). The tracing experiment with IR-LEGO system revealed that these endothelial cells bridged the BA to DLAV, and suggested that this system was valuable for fate-mapping analysis. We were able to trace strong mCherry expression in the targeted cells for >48 hours after induction. Lineage tracing analysis with time-lapse imaging of the fli1:nEGFP transgenic zebrafish was used to determine the source of the lymphatic endothelial cells to the thoracic duct.13 Although tracing backward in time the particular nucleus in time-lapse sequences was a useful method, the range of tracing the targeted cells was restricted to the imaging view. Transplantation of labeled cells or the labeling of individual cells by injecting heritable dye has been also used for tracing analysis and has provided useful products.14 However, these methods have been applied to the embryos only up to the mid-to-late blastula stages. In contrast, our method using the IR-LEGO system has no limitations for tracing the range and developmental stage, and therefore, it will be a powerful tool to map the fate of various cell types at later stages and to analyze the distribution or function of tissue-specific stem cells in various tissues, such as the digestive system.
Using the water heating property of the IR laser effectively, an overdose of laser irradiation was able to ablate the targeted cells. We ablated the endothelial cells of the first SeAs and assessed its influence on the connecting vessels between the brain and spinal cord. On the irradiated side, the connection between the BA and DLAV diminished, and the PHBC did not connect to the DLAV (Figure 4A and 4B). This suggested the primary role of the first SeA in bridging BA and PHBC to DLAV, and the ablation experiment was a valuable method for analyzing the influence of the particular vessel on vascular morphogenesis.
By tracing and ablating the cells of the first SeAs, we clearly indicated their contribution to the connection between the vascular systems of the brain and spinal cord in zebrafish. In contrast, Padget15 precisely described how these vascular systems were constructed by reconstructing serial sections of human embryos. In human embryos, the PHBC was identified at the 20-somite stage; the bilateral longitudinal neural arteries subsequently formed in the ventromedial region. Bilateral longitudinal neural arteries were medially fused to form the BA and were subsequently connected to the first SeAs. A zebrafish does not have a neck, and the final vascular systems in this region may greatly differ from those in human beings. However, the early processes, during which these vasculatures form, progress in a manner highly similar to that observed in humans; therefore, we believe that zebrafish can be considered a useful model to analyze the vascular formation between the brain and spinal cord. We are now very interested in the guidance mechanisms surrounding vascular formation in this region and are focusing on the perivascular tissues, such as the pericytes and mesenchymal cells, as well as the neural tissue. The IR-LEGO system will help us address this issue.
We demonstrated that the IR-LEGO system is a powerful tool with the potential to contribute to developmental biology. However, there are some limitations to this system. The most critical limitation is the depth of targeted cells. We achieved a maximum induction rate of 60% for the targeted single endothelial cells in the intersegmental vessels, which are usually located at a depth of 50 to 60 μm. The endothelial cells of the first SeA were localized at a depth of ≈70 μm, which might cause a decrease in the efficiency of inducing gene expression in the tracing experiment. We assume that the limitation of gene induction in the single cells is at a depth of ≈100 μm. The IR-LEGO system can be applied to various tissues at later developmental stages; however, the targeted region is restricted to a depth of 100 μm. We also indicated that the IR-LEGO system is useful for the ablation of single cells. Combining the ablation study with time-lapse imaging enables us to observe the regeneration process of ablated tissues. Induction of regeneration-associated genes after ablation will also contribute to understanding the underlying mechanisms.
In this study, we applied the IR-LEGO system to vascular biology and succeeded in demonstrating its usability in zebrafish. Our method enabled gene induction in single endothelial cells, tracing of targeted cells by fluorescent proteins, and ablation of the targeted vessel. Inducing the genes associated with vascular formation will help us to investigate the regulation of vascular connections. Furthermore, combined with small interfering RNA-mediated gene suppression, IR-LEGO will enable in vivo gene manipulation with vertebrate model organisms, and it leads to the prospect of uncovering the mechanisms underlying vascular morphogenesis.
We thank Dr Brant Weinstein (National Institutes of Health) for the transgenic zebrafish strain and Dr Ryozo Nagai (Jichi Medical University) for proofreading the article. We also thank Mutsumi Kurosawa, Taro Ando, Yohei Sawa, Yuki Matsumoto, and Motohiko Koizumi (Iwate Medical University) for the technical support offered for the screening and bleeding of the transgenic zebrafish, and we thank Misako Saida-Taniguchi (National Institute for Basic Biology) for offering technical support in optics.
Sources of Funding
This work was carried out under the National Institute for Basic Biology Cooperative Research Program (11-336) for E. Kimura and was supported by Grants-in-Aid for Scientific Research (KAKENHI) of Japan Society for the Promotion of Science (JSPS) grant number 23590222 (E. Kimura).
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.112.300602/-/DC1.
- Received October 5, 2012.
- Accepted March 10, 2013.
- © 2013 American Heart Association, Inc.
- Isogai S,
- Lawson ND,
- Torrealday S,
- Horiguchi M,
- Weinstein BM
- Lakso M,
- Sauer B,
- Mosinger B Jr.,
- Lee EJ,
- Manning RW,
- Yu SH,
- Mulder KL,
- Westphal H
- Halloran MC,
- Sato-Maeda M,
- Warren JT,
- Su F,
- Lele Z,
- Krone PH,
- Kuwada JY,
- Shoji W
- Sato-Maeda M,
- Tawarayama H,
- Obinata M,
- Kuwada JY,
- Shoji W
- Kimmel CB,
- Warga RM
- Padget DH
In vivo spatiotemporal gene regulation in a single cell is obligatory to analyze specific gene functions associated with morphogenesis in developing embryos. The infrared laser–evoked gene operator system is a new microscopic method optimized to heat cells by using an IR laser, which has a superior ability to heat water and can efficiently access deep-lying tissues. We applied this method to the vascular system in zebrafish embryos and first succeeded in gene induction driven by heat shock promoter in the targeted endothelial cells. Inducing the particular genes using this method will help us to investigate the particular gene function in vivo. Additionally, combined with small interfering RNA-mediated gene suppression, infrared laser–evoked gene operator will enable in vivo gene manipulation in the future. Our achievement in this study will lead us to the prospect of uncovering the mechanisms underlying morphogenesis of vertebrate model organisms.