Anthrax lethal toxin induces cell death-independent permeability in zebrafish vasculature
- Robert E. BolcomeIII*,†,
- Sarah E. Sullivan*,
- René Zeller*,
- Adam P. Barker†,
- R. John Collier†,‡, and
- Joanne Chan*,‡
- *Vascular Biology Program, Children's Hospital Boston and Department of Surgery, Harvard Medical School, Boston, MA 02115; and
- †Department of Microbiology and Molecular Genetics, Harvard Medical School, 200 Longwood Avenue, Boston, MA 02115
-
Contributed by R. John Collier, December 28, 2007 (received for review November 4, 2007)
Abstract
Vascular dysfunction has been reported in human cases of anthrax, in mammalian models of Bacillus anthracis, and in animals injected with anthrax toxin proteins. To examine anthrax lethal toxin effects on intact blood vessels, we developed a zebrafish model that permits in vivo imaging and evaluation of vasculature and cardiovascular function. Vascular defects monitored in hundreds of embryos enabled us to define four stages of phenotypic progression leading to circulatory dysfunction. We demonstrated increased endothelial permeability as an early consequence of toxin action by tracking the extravasation of fluorescent microspheres in toxin-injected embryos. Lethal toxin did not induce a significant amount of cell death in embryonic tissues or blood vessels, as shown by staining with acridine orange, and endothelial cells in lethal toxin-injected embryos continued to divide at the normal rate. Vascular permeability is strongly affected by the VEGF/vascular permeability factor (VPF) signaling pathway, and we were able to attenuate anthrax lethal toxin effects with chemical inhibitors of VEGFR function. Our study demonstrates the importance of vascular permeability in anthrax lethal toxin action and the need for further investigation of the cardiovascular component of human anthrax disease.
Disruption of vascular integrity has been consistently observed in human anthrax disease, in mammalian studies by infection with Bacillus anthracis or by i.v. injection of anthrax toxin proteins (1–8). The importance of blood vessels and endothelial cells in anthrax toxicity has been difficult to investigate because of the inability to observe progressive vascular changes without sacrificing the mammalian host. Anthrax toxin is an ensemble of three proteins: two enzymatic moieties, edema factor (EF), and/or lethal factor (LF), which act on cytosolic substrates, and a receptor–binding, pore-forming moiety, protective antigen (PA), which binds EF and/or LF and delivers them to the cytosol. The combination of PA and LF is known as lethal toxin (LeTx), and the combination of PA and EF is known as edema toxin (EdTx). LeTx induces rapid death in experimental animals, and its action is associated with vascular defects and pleural effusions (4, 5). Early reports indicated that EdTx did not produce significant mortality (9), but recently robust effects have been reported, including lethality in rodents (10, 11). However, the ability to induce loss of vascular integrity and leakage has been consistently associated with LeTx (7, 10–12).
Two mammalian anthrax toxin receptors (ANTXRs) are reported to bind PA: tumor endothelial marker 8 (TEM8, also known as ANTXR1) (13) and capillary morphogenesis gene 2 (CMG2, also known as ANTXR2) (14, 15). Both receptors mediate anthrax toxin internalization and intracellular delivery of LF, and are expressed in many cell types including endothelial cells (12, 16, 17).
To evaluate the action of LeTx on intact blood vessels, we developed a zebrafish model that permits in vivo imaging of the vasculature. Zebrafish embryos are transparent allowing real-time observation of blood flow, which begins from 24 to 26 hpf (hours postfertilization) (18). In our assays, LeTx was delivered into the embryonic circulation, and cardiovascular function was monitored over 20 h using transgenic zebrafish lines (19–21). We found that LeTx induced an increase in vascular permeability that was not due to cell death, because individual endothelial cells could be counted (20). Furthermore, LeTx did not generate widespread cell death in other tissues.
Vascular function is tightly regulated by the vascular endothelial growth factor (VEGF) signaling pathway. VEGF was first identified as the vascular permeability factor (VPF), because its ability to induce vascular leakage is unique among angiogenic growth factors (22). Using chemical inhibitors of VEGFR, we demonstrated attenuation of anthrax toxicity in our zebrafish model. Anti-VEGF therapy is currently in clinical use so that approved drugs, as well as drugs under development, could be further investigated as anti-anthrax therapeutics.
Results
LeTx Effects and Vascular Leakage in the Zebrafish.
Conservation of genes, signaling pathways, and biological processes, has made the zebrafish a useful system to define gene function in vertebrates (23, 24). The zebrafish embryo develops rapidly so that by ≈24 hpf the primordia of the brain and organ systems are formed, the heart starts to beat, and blood flow begins in the major axial vessels (18, 25). To develop a zebrafish model for anthrax toxin action, we reasoned that introduction of the large anthrax toxin proteins (>83 kDa) into the vasculature of zebrafish embryos would closely mirror systemic toxin challenges in rodents. After confirming that zebrafish have conserved orthologs for the ANTXR2s [supporting information (SI) Fig. 5], we introduced LeTx into the circulation of embryos at 48 hpf (Fig. 1 A). The transparency of the zebrafish embryo and the availability of transgenic lines with fluorescently labeled endothelial (19, 20) and blood cells (21) permitted direct visualization of toxin-induced phenotypes.
LeTx effects in zebrafish embryos. (A) Diagram of zebrafish embryo at 48 hpf (time point of toxin protein microinjection, at blue arrow), functional vasculature in red. (B–D) Images were taken at 68 hpf. (B) An inert phenol red dye was injected as vehicle in all control embryos and a WT phenotype was observed at 68 hpf, or 20 hpi. (C) 1× LeTx (defined below) generated pericardial edema and narrowed vessels (C′). (D) Embryo was treated with 2.5 μM CI-1040 at 48 hpf for 6 h, then washed out. Black arrow indicates heart and pericardial edema in A–D. (Scale bar, 250 μm.) (B′–D′) Enlarged images showing LeTx and CI-1040 induced narrowing of ISVs in Tg(fli1:EGFP)y (19) embryos, indicated by arrows. (Scale bar, 80 μm.) (E–G) LeTx-injected Tg(gata1:dsRED) (21) embryos exhibited outflow tract lumen size reduction from 20 μm to <10 μm between LeTx phenotype stages 2 (F) and 3 (G). (Scale bar, 50 μm.) DA, dorsal aorta; PCV, posterior cardinal vein; CCV, common cardinal vein; ISV, intersegmental vessel; BA, bulbous arteriosus; OT, outflow tract; A, atrium; V, ventricle, for all figures. 1× LeTx is defined as 37 fmol of LF and 25 fmol of PA.
Pericardial edema and trapping of blood cells in the heart by 20 hpi (hours postinjection, see SI Movie 1), a reproducible endpoint of observed phenotypes, were observed in >90% of injected embryos (n > 600; Fig. 1 B and C). These effects were specific to LeTx, because single injections of either LF or PA alone did not induce changes in cardiovascular function or in embryo morphology over the course of our assays (SI Fig. 6A). LF is known to cleave and inactivate MEKs (26), and we were able to generate a phenocopy of LeTx effects using a chemical MEK1/2 inhibitor, CI-1040, one of the most selective of available MEK1/2 inhibitors, which has been used in whole animal and cellular studies (number of experiments (N) = 2; n = 20; Fig. 1 D) (27). Whereas LF is also known to cleave MEK3/4/6/7, no selective inhibitors are available for these molecules. Our data suggest the importance of MEK1/2 for the vascular effects observed after LeTx exposure.
Zebrafish toxin experiments included the injection of an inert visual dye (vehicle) to confirm delivery of toxins and a second, “blinded” investigator to score resulting phenotypes. To minimize variation, toxin delivery occurred within a 2-h period on 200–300 embryos fertilized within 30 min. Initially, six repeats of >200 embryos per experiment were used to define LeTx phenotypes. We observed highly reproducible, quantifiable distinctions such as the progressive narrowing of the outflow tract of the heart (Fig. 1 E–G), which restricted the amount of blood flow in the embryo at the end stage of our assay, at 20 hpi or 68 hpf (SI Fig. 6). At this time point, we were able to separate all injected embryos into three phenotypic classes: severe, mild, or WT (normal) appearance. We defined the severe toxin defect as having no circulation and the mild LeTx phenotype as having slow circulation. These phenotypic classes were directly correlated with LeTx dosage (below), with the toxin dose of 37 fmol LF and 25 fmol PA as a standard (1× LeTx). This dose facilitated analysis of a potential shift in scoring of LeTx effects between severe and mild phenotypes (Fig. 2 A).
Conservation of the anthrax toxin internalization pathway and attenuation of LeTx effects in zebrafish. (A) Dose–response curves for LeTx demonstrated that all embryos displayed toxin effects at high doses. Conditions for each lane are as indicated on the x axis. LeTx phenotypes (severe, mild, or WT appearance) were converted to percentages in all panels. No toxin phenotypes were observed when WT LF was replaced with a catalytic mutant, LF Y728F (31). (B) Comparison of inhibitor efficacies conducted in a single representative experiment. Embryos were injected with LeTx and/or protein inhibitors as indicated at 48 hpf and scored at 20 hpi. 1× LeTx was used (37 fmol of LF and 25 fmol of PA). Treatments in each lane were as follows: lane 1, LeTx, n = 25; lane 2, LeTx with 6 pmol of LFN (low), n = 28 (P = 0.912); lane 3, LeTx with 12 pmol of LFN (high), n = 28 (P = 0.007); lane 4, LeTx with 6 pmol of sol-CMG2 (labeled low, because this dose matches the low dose for LFN), n = 23 (P < 0.001); lane 5, injection of 37 fmol of LF alone, n = 20 (P < 0.001); lane 6, 37 fmol of LF and 12.5 fmol of PA (50%), with 12.5 fmol of PA F427A (50%), n = 32 (P < 0.001); lane 7, 37 fmol of LF and 25 fmol of PAF427A, n = 30 (P < 0.001); lane 8, 25 fmol of PA, n = 31 (P < 0.001); lane 9, 25 fmol of PAF427A, n = 30 (P < 0.001); lane 10, uninjected control embryos, n = 30 (P < 0.001). (C) SU11652, ZM323881, and PTK/ZK, selective inhibitors against VEGFR, demonstrated protection against LeTx effects (P < 0.001 for 1.5 μM SU11652; P < 0.001 for 1 μM ZM323881; P < 0.001 for 500 nM PTK/ZK). A 2× LeTx dose was used (75 fmol of LF and 50 fmol of PA). Statistics were completed by using the χ2 test.
Inhibitors of toxin action and inactive forms of the toxin proteins functioned in our zebrafish assay as predicted from earlier, in vitro studies (Fig. 2 A and B). The soluble, extracellular region of the human CMG2/ANTXR2 receptor, sol-CMG2 (14, 28) inhibited LeTx action strongly, whereas LFN, the isolated N-terminal domain of LF, which competes with LF and EF for binding to PA, provided mild protection (29). The transport-defective mutant, PA F427A (30), or a catalytically inactive form, LF Y728F (31), each led to 100% normal embryos (Fig. 2 A and B). Finally, we showed that LFNDTA, a fusion of LFN and DTA, the catalytic domain of the cytocidal toxin, diphtheria toxin, induced cell death throughout the embryo (below).
Recent studies using endothelial cells in culture have reported LeTx induced apoptosis or barrier dysfunction, depending on the particular endothelial cell type and treatment (12, 16, 17). Our model exposes endothelial cells to LeTx within the context of intact, lumenized, and functional blood vessels. We probed whether the toxin was inducing cell death using a vital dye, acridine orange (32) (Fig. 3 A–D) and found very little cell death throughout the zebrafish embryo as compared with those injected with LFNDTA plus PA (Fig. 3 A–D). In addition, we examined endothelial cell numbers using a transgenic line with nuclear localization of EGFP (enhanced green fluorescent protein, Fig. 3 E and F) (20, 32), and found that LeTx did not induce significant cell death in endothelial cells. Endothelial cell damage seemed to be minimal as cell divisions continued to occur after introduction of toxins into the vasculature. The number of endothelial cell nuclei was the same in control and LeTx-injected embryos over a defined set of eight ISVs (intersegmental vessels) in the trunk region of the zebrafish embryo (N = 3, n = 6, Fig. 3 E and F). In contrast, LFNDTA plus PA generated a progressive reduction in endothelial cell numbers over the same time course (SI Fig. 7).
LeTx does not induce generalized cell death or alter endothelial cell numbers in intact zebrafish embryonic vasculature. (A–D) Embryos were stained with acridine orange to examine cell death in dye-injected (A), LeTx-injected embryo at stage 2 or end stage of 20 hpi (B and C), or LFNDTA plus PA-injected (D) embryos. No significant difference was observed between dye and LeTx-injected embryos (A–C), whereas overall cell death was observed in those injected with LFNDTA plus PA (D) (each panel representative of n > 20 embryos). (A′–D′) Enlarged views of boxed regions. (E and F) Nuclei counts using the TG(fli1:nEGFP)y7 (20) line demonstrated equal endothelial cell numbers over the eight ISVs anterior to the cloacae between dye-injected controls (E), and embryos displaying a severe LeTx phenotype (F). There was no difference in the numbers of nuclei in this region between severe or mild phenotype and WT embryos (N = 3, n = 6 per condition; P = 0.378 and 0.887 respectively). Statistics were completed by using the Holm-Sidak method. (Scale bars, 80 μm.)
To examine the progression of endothelial changes leading to edema, we monitored vascular effects in real-time over 20 h. Four distinct stages of LeTx phenotypic progression, each lasting ≈2 h were defined (Fig. 4 and SI Fig. 6A). Variation in the time of onset of toxin effects was observed (from 4 to 12 hpi), but all phenotypes stabilized by 20 hpi (N = 4, n > 30). To quantify vascular leakage, we tracked the localization of microinjected fluorescent microspheres at each stage after LeTx exposure. We focused on the heart as fluid seemed to accumulate here. Permeability changes leading to microsphere extravasation in the heart facilitated photography, because beads were trapped between its endocardial (endothelial in origin) and myocardial layers. We note that fluorescent microspheres were also detected in extravascular spaces along embryonic trunk vessels, including the dorsal aorta (DA), ISVs, and dorsal longitudinal anastomotic vessel (DLAV) (Fig. 4 A–H).
Microsphere leakage in the zebrafish blood vessels and heart. The Tg(fli1:EGFP)y1 line was used for EGFP-labeled endothelial cells; fluorescent microspheres (100 nm, blue; 500 nm, red) were injected at the beginning of stage 2 and leakage was monitored at 20 hpi. (A–D) Panels indicate WT distribution of microspheres within blood vessels (blue arrows). (E–H) Panels indicate LeTx induced microsphere leakage (red arrows). (I–N) The heart was chosen for ease of photography as microspheres are trapped between the endocardial and myocardial layers. (I) A schematic of a control embryonic heart. (J and K) Control embryos injected with dye injection did not show microsphere leakage. (L) A schematic drawing indicates a thickened heart wall because of vascular leakage. (M and N) Embryos previously injected with 2× LeTx (75 fmol of LF and 50 fmol of PA) displayed leakage of 100-nm blue microspheres at the beginning of stage 2 (red arrows, M), and 500-nm red microspheres at the end of this stage (red arrows, N). A dotted line indicates the endocardial layer (M′ and N′). E, endocardium; M, myocardium; DLAV, dorsal longitudinal anastomotic vessel. (Scale bar, 50 μm.)
In the first stage of phenotypic progression, blood cells accumulated at the inflow tract of the heart with slight blood flow regurgitation, leading to enlarged volume within both cardiac chambers. Increased permeability was first detected by the beginning of stage 2, because 100-nm blue fluorescent microspheres were extravasated between the endocardial and myocardial layers (Fig. 4 I–N). By the end of this stage, 500-nm red fluorescent microspheres had leaked into this space in 58% of LeTx-injected embryos (n = 12). In addition, the ISVs began to collapse until they could no longer support blood flow (Fig. 1 B′–D′). The lumen of the common cardinal vein, a large vessel that empties into the heart, also became progressively narrowed (SI Fig. 6A). By stage 3, the outflow tract of the heart narrowed from ≈20 μm to <10 μm in diameter, as measured by blood cells and 500 nm microspheres (Fig. 1 E–G). Flow became restricted primarily to the major axial blood vessels. The most distinctive features at the end of this stage were pericardial edema, pooling of blood at the inflow tract, and toggling of blood cells between the two chambers of the heart.
In the final stage of LeTx phenotype development, defects from the previous stages led to a cessation of blood flow in embryos displaying a severe toxin phenotype. Blood that had pooled at the inflow tract of embryos at stage 3 entered and became trapped within the heart chambers of all embryos. An hour into this stage, 75% of embryos displayed microsphere extravasation (500-nm microspheres, n = 12). In the mild form of the LeTx phenotype, the outflow tract was not completely closed, permitting slow circulation in the axial vessels that could be quantitated by counting blood cell passage because the toxin did not induce changes in heart rate (SI Fig. 6 B–D). By the end stage time point in our experiments, 20 hpi, embryos with severe LeTx effects exhibited 500-nm microsphere extravasation (93%, n = 14).
Chemical VEGFR Blockade Counteracts LeTx-Induced Vascular Leakage.
Although studies in mammals have shown vascular defects toward the final stages of LeTx toxicity (5–7), our live imaging of LeTx effects revealed that it is an important early consequence of toxicity in our zebrafish model (Fig. 4). Because the VEGF-VEGFR signaling pathway is known to be a major regulator of endothelial permeability, we tested the ability of chemical VEGFR inhibitors to attenuate LeTx phenotypes. We selected three structurally distinct VEGFR kinase inhibitors: ZM323881, PTK787/ZK222584 (PTK/ZK), and SU11652. ZM323881 is a highly selective inhibitor of VEGFR2 tyrosine kinase activity (IC50 < 2 nM), as demonstrated by its inhibition of VEGFR2 phosphorylation in frog lung tissue at nanomolar levels (33). The affinity of ZM323881 toward the next closest receptor, VEGFR1, as well as toward other receptor tyrosine kinases is significantly lower (IC50 > 50 μM) (33). PTK/ZK inhibits the kinase activities of the VEGFRs at nanomolar concentrations, as well as PDGFR, c-Kit, and c-Fms at higher doses (e.g., VEGFR2: IC50 = 42 nM; for PDGFRβ:IC50 = 490 nM) (34). It was shown to be effective in treating pleural effusions and in perturbing angiogenesis in zebrafish adults and embryos (34–37). SU11652 inhibits the VEGFRs, as well as FGFR, and Kit family members. It has been used successfully to block endometrial endothelial cell proliferation in C57/CBA mice (38, 39). PTK/ZK is the best characterized of these three inhibitors. It was shown in an unbiased study to be one of the most selective among 20 kinase inhibitors tested against 119 kinases (40).
Each VEGFR2 inhibitor attenuated LeTx action, as observed by a reduced percentage of embryos having severe toxin effects and an increase in mild or normal phenotypes. Over a range of doses (Fig. 2 C; SI Fig. 8), we found SU11652 to be most effective at 1.5 μM, PTK/ZK at 500 nM, and ZM323881 at 1 μM. Selective inhibitors that work at nanomolar levels in cell based assays are typically used at 50- to 100-fold higher concentrations in zebrafish embryos, possibly because of the need to penetrate many cell layers in a developing embryo (41).
To examine the selectivity for VEGFR2 inhibition, we evaluated commercially available, characterized inhibitors for several other kinases (SI Table 1). We and others have used numerous small molecule inhibitors and found that they inhibit the same targets in mammals and zebrafish (41). Kinase inhibitors such as AG-1296, ZM449829, Y-27632, and 3-(1-methyl-1H-indol-3-yl-methylene)-2-oxo-2,3-dihydro-1H-indol-5-sulfonamide were all tested. ZM449829 inhibits JAK3, JAK1, EGFR, and STAT5 (42, 43). AG-1296 is selective for PDGFR and Kit family members (44, 45). Y-27632 is a widely used inhibitor of Rho-associated protein kinases (46). 3-(1-Methyl-1H-indol-3-yl-methylene)-2-oxo-2,3-dihydro-1H-indol-5-sulfonamide is a Syk inhibitor (47). None of these kinase inhibitors attenuated LeTx effects as observed with VEGFR selective compounds. Collectively, these experiments underscore the importance of the endothelial cell as a major target cell type for anthrax toxicity, and the potential for using a chemical approach to attenuate toxin effects.
Discussion
Our study demonstrates the need for further investigation of LeTx induced vascular permeability. Recent clinical reports have placed emphasis on the presence of pleural effusions as a diagnostic indication for human anthrax disease, because it has occurred with high frequency (nearly 100%) among known cases (3). LeTx alone can induce death in mammalian models, with concomitant vascular damage and pleural effusions (4–7). Our study focused on the endothelial effects of LeTx, to examine the connection between toxin induced vascular leakage, blood vessel dysfunction, and the role of host endothelial cell signaling pathways.
Taking advantage of the transparency of zebrafish embryos and transgenic lines, we determined functional changes in endothelial cells upon toxin challenge. Using sized fluorescent microspheres, we found that vascular LeTx delivery led to the leakage of 100-nm and 500-nm beads. A similar assay was used to examine vascular permeability in adult mice, where the sizes and extent of extravasated microspheres correlated with increased leakage (48). Using endothelially labeled transgenic lines (19, 20), we determined that LeTx did not affect endothelial cell proliferation in developing blood vessels in vivo, because cell numbers continued to increase at the same rate as controls. Collectively, these LeTx effects reduced functional blood flow in the whole embryo that could be scored visually. We were able to demonstrate that LeTx did not cause endothelial cell death or induce widespread cell death in other tissues. The rapid response time, the short duration, and consistency of our assay provide a facile in vivo model for LeTx studies.
The progression of anthrax disease likely involves several major target cell types such as macrophages and dendritic cells, as well as contributions from EdTx and other bacterial components. However, LeTx has been shown to induce vascular leakage in experimental models and from clinical data (1, 5–7, 12, 17). Endothelial cells line the entire vascular tree in the human body, covering ≈100 m2 of surface area (49), making it an important potential target for anthrax action. The VEGFRs are preferentially expressed on the surface of endothelial cells, and are master regulators of endothelial cell growth, survival, and permeability throughout the life of an organism, in normal physiology, and in disease (22, 50). Our data provide a possible link between LeTx action and VEGF signaling in endothelial cells, suggesting that these pathways could interact to modulate vascular permeability. Activation of the VEGFR2 could recruit kinases such as MAPK, AKT, and Src (50). Endothelial Src and AKT activation have been linked to the regulation of vascular permeability, whereas MEK-MAPK has been more closely associated with proliferation (50–54). LF could dampen the MEK-MAPK proliferative signal without altering other VEGFR effectors, resulting in increased permeability. By modulating the VEGFR signaling pathway, it may be possible to control vascular permeability and prolong the duration of effectiveness for antibiotic and antitoxin therapies.
We attempted LeTx toxin assays in a zebrafish VEGFR2 mutant line, but the early developmental vascular defects in these embryos were too severe to allow comparison with WT controls. However, chemical VEGFR inhibitors provided precise control over the duration and extent of receptor inhibition that is currently not possible with a genetic line. The ability of three structurally distinct VEGFR inhibitors to attenuate LeTx effects suggests that a chemical approach could be used in further studies in mammalian models of toxin challenge.
A number of pathogens induce vascular permeability defects leading to hemorrhagic fever and shock-like symptoms including those resulting from the infections of Dengue virus, Ebola virus, Marburg virus, Hanta virus, and B. anthracis. Further studies into the vascular mechanisms of additional infectious agents may lead to an improved understanding of the role of the endothelial cell and the vasculature in mediating whole organism host responses to pathogens.
Materials and Methods
Animals.
All animal protocols were approved by the Institutional Animal Care and Use Committee of Children's Hospital Boston. Zebrafish lines were maintained at 28.5°C on a 14-h light/10-h dark cycle. Embryos were collected by natural spawning, and raised in 10% Hanks' saline at 32°C.
Microinjection of Toxin Proteins and Use of Inhibitors.
Microinjections into the zebrafish vasculature at 48 hpf were carried out as described by Weinstein et al (55)., with minor modifications (41). We prepared PA as described in ref. 15. LF was purchased from List Laboratories, Inc. LFNDTA was a kind gift from S. Juris. LF Y728F was cloned and produced as described in ref. 31. Combinations of LF, PA, LFNDTA, LF Y728F, LFN, PA F427A, and sol-CMG2 were prepared immediately before, and kept at 4°C until injection. Injected amounts are indicated in the figure legends for each experiment. Phenol red (0.05%) was added for visibility during microinjection. Volumes of 40 nl or less were delivered into the common cardinal vein of embryos anesthetized with tricaine (Sigma) at 48 hpf using a gas driven microinjector (Medical Systems Corp.). After injection, embryos were transferred into fresh medium for recovery, maintained at 32°C, and blind scored for toxin action at time points indicated in the text. Small molecule inhibitors were added to embryo medium as described in the text and figure legends. PTK/ZK was provided by Novartis Pharma; CI-1040 was provided by Pfizer. All other inhibitors were purchased commercially. For CI-1040 experiments, dechorionated 48 hpf embryos were placed in 5 ml of 2.5 μM CI-1040 or DMSO control and maintained at 32°C. The drug was washed out 6 h later (54 hpf) with three washes of 10% Hanks' saline and embryos were scored for phenotypes at 68 hpf.
Microinjection of Microspheres.
Microsphere mixtures were prepared as 1 part each of 100-nm (diameter) blue and 500-nm red fluorescent bead suspensions (Duke Scientific Corp.), stored at 4°C, and microinjected at ≤20 nl at time points indicated in the text. For confocal imaging of the heart: 15 minutes after bead injection, embryos were fixed in 2% PFA for 5–10 min; then embedded in 2% low melt agarose (Bio-Rad) on 35-mm glass bottom microwell dishes (MatTek Corp.). For fluorescence microscopy of bead leakage from the dorsal aorta, DLAV, and ISVs, embryos were anesthetized with tricaine (Sigma), then mounted in 4% methylcellulose.
Acridine Orange Staining.
Dechorionated embryos were placed in 5 ml of a 50 μg/ml solution of acridine orange (acridinium chloride hemi-[zinc chloride], Sigma) in 10% Hanks' saline solution at room temperature. After 30 min of staining while protected from light, embryos were washed three times for 10 min with 10% Hanks' saline. After the third wash, embryos were mounted on glass slides for fluorescence microscopy.
Statistics.
Statistical analysis was by χ2 for LeTx studies in which populations consisting of severe phenotype, mild phenotype, and WT appearance embryos were compared. The Holm-Sidak method was used for all other experiments. Sigma Stat 3.0 software was used for all statistical analysis. P < 0.05 was considered significant.
Isolation and Characterization of Zebrafish ANTXR cDNA Clones.
Zebrafish ANTXR2A and B cDNAs (GenBank accession no. XP_689332.1 and no. XM_684240.1, respectively), were isolated as described in SI Materials and Methods.
Acknowledgments
This article is dedicated to the memory of Judah Folkman (1933–2008), a pioneer in angiogenesis research. We thank T. Roberts, J. Folkman, L. Zon, A. Goldstein, J. Mably, D. Cowan, R. Melnyk, and S. Juris for helpful discussions and/or critical reading of the manuscript. We thank S. Juris (Univeristy of Michigan, Mount Pleasant, MI) for the generous gift of LFNDTA; L. Zon (Children's Hospital Boston), B. Roman (University of Pittsburgh, Pittsburgh, PA), and N. Lawson (University of Massachusetts, North Worcester, MA) for providing transgenic and mutant zebrafish lines; Novartis Pharma AG, Basel, Switzerland, for the use of the PTK/ZK VEGFR inhibitor; and Pfizer Pharmaceuticals for the use of the CI-1040 MEK inhibitor. We acknowledge R. Pimental, K. Bellavance, M. Lin, and J. Abrams, for technical assistance. This work was supported in part by National Institute of Allergy and Infectious Diseases and New England Regional Center of Excellence Grant AI057159, and also by National Institutes of Health Grants AI056134 and CA111564 (to J.C.) and AI022021 (to R.J.C.).
Footnotes
- ‡To whom correspondence may be addressed at: Vascular Biology Program, Children's Hospital Boston, Karp Family Research Laboratories, 12th floor, Room 12.217, 300 Longwood Avenue, Boston, MA 02115-5737. E-mail: joanne.chan{at}childrens.harvard.edu or jcollier{at}hms.harvard.edu
-
Author contributions: R.E.B., R.J.C., and J.C. designed research; R.E.B., S.E.S., R.Z., and J.C. performed research; A.P.B. and R.J.C. contributed new reagents/analytic tools; R.E.B., S.E.S., R.Z., and J.C. analyzed data; and R.E.B., R.J.C., and J.C. wrote the paper.
-
Conflict of interest statement: R.J.C. is cofounder of and an equity holder in PharmAthene, Inc., and consults for CombinatoRx, Inc.
-
Data deposition: The sequences reported in this paper have been deposited in the GenBank database [accession nos. DQ415957 (CMG2A) and EF591979 (CMG2B)].
-
This article contains supporting information online at www.pnas.org/cgi/content/full/0712195105/DC1.
-
Freely available online through the PNAS open access option.
- © 2008 by The National Academy of Sciences of the USA