A Crucial Role of Caldesmon in Vascular Development in Vivo

Home , Angiogenesis, Anterior cardinal vein, Aortic arches, Dorsal aorta, Posterior cardinal vein, Vasculogenesis

Cardiovascular Research (2009) 81, 362–369 doi:10.1093/cvr/cvn294

A crucial role of caldesmon in vascular development in vivo

Ping-Pin Zheng1, Lies-Anne Severijnen2, Marcel van der Weiden1, Rob Willemsen2†, and Johan M. Kros1*†

1Department of Pathology, Erasmus Medical Center, JNI Room 230-c, Dr Molewaterplein 50, PO Box 1738, 3000 DR Rotterdam, The Netherlands; and 2Department of Clinical Genetics, Erasmus Medical Center, Rotterdam, The Netherlands Downloaded from https://academic.oup.com/cardiovascres/article/81/2/362/286189 by guest on 24 September 2021

Received 24 July 2008; revised 28 October 2008; accepted 29 October 2008; online publish-ahead-of-print 3 November 2008

Time for primary review: 13 days

KEYWORDS Aims We explored the in vivo effects of knockdown of caldesmon on vascular development in zebrafish. Caldesmon; Methods and results We investigated the effects of caldesmon knockdown on the vascular development Vascular development; in a zebrafish model with special attention for the trunk and head vessels including the aortic arches. We Zebrafish model; examined the developing fishes at various time points. The vascular abnormalities observed in the cal- Vasculogenesis; desmon morphants were morphologically and functionally characterized in detail in fixed and living Angiogenesis embryos. The knockdown of caldesmon caused serious defects in vasculogenesis and angiogenesis in zebrafish morphants, and the vascular integrity and blood circulation were concomitantly impaired. Conclusion The data provide the first functional assessment of the role of caldesmon in vascular development in vivo, indicating that this molecule plays a crucial role in vasculogenesis and angiogenesis in vivo. Interfering with caldesmon opens new therapeutic avenues for anti-angiogenesis in cancer and ischaemic cardiovascular disease.

1. Introduction primordial hindbrain channel, the anterior cardinal vein, and the basilar artery in the head.8 The formation of most Caldesmon (CaD) is evolutionally conserved among ver- 1 of the subsequent vessels in the embryo occurs by sprouting tebrates. The zebrafish homologue is similar to mammalian from pre-existing vessels in a process known as angiogenesis. low-molecular-weight caldesmon (l-CaD). Previously, we Many (presumably angiogenic) blood vessels that sub- reported the specific upregulation of this protein (l-CaD) in sequently develop in the zebrafish have orthologues in glioma neovasculature and its association with migration other vertebrates and among these are the central cranial and proliferation of endothelial cells (ECs) and endothelial 2–7 arteries and the massive network of microvessels in the progenitor cells (EPCs) in human tumours. The findings head, the dorsal longitudinal anastomotic vessels (DLAVs), triggered us to explore the effects of this protein on the the intersegmental vessels (ISVs), the subintestinal veins development of blood vessels in vivo for the design of new (SIVs), the caudal vessel plexus (CVP), the parachordal therapeutic strategies. Here we explored the effects of vessels, the vertebral vessels in the trunk, and more.9 The knockdown of CaD on the vascular development in zebrafish embryology of the aortic arch (AA) system in zebrafish is embryos. Zebrafish embryos can survive several days very similar to that of birds and mammals.10 Six pairs of without a functioning circulatory system, allowing detailed vessels, connecting the ventral aorta to the lateral DA, analysis of the animals with severe cardiovascular defects. emerge in a cranial-to-caudal sequence, each of which is The development of the vascular system in vertebrates embedded in its respective pharyngeal arch which is collec- occurs by two distinct processes: vasculogenesis and angio- tively known as the branchial AAs.10 The AA primordia arise genesis. The same primary vasculogenic vessels that estab- by vasculogenesis and extend via angiogenesis.10 At the mol- lish the initial circulatory circuits in other vertebrate ecular level, several important genes, including VEGF,Flk-1/ embryos are also present in the zebrafish. These vessels KDR, Fli-1, Flt-1, Tie-1, and Tie-2, have been cloned in include the dorsal aorta (DA) and posterior cardinal vein zebrafish and show expression patterns similar to those (PCV) in the trunk and the internal carotid artery, the in mammals.11,12 The striking conservation of vascular anatomy and the expression pattern of the associated genes across the vertebrate phyla indicate similar vasculo- * Corresponding author. Tel: þ31 10 7043905; fax: þ31 10 7043905. E-mail address: [email protected] genic and angiogenic signalling pathways for blood vessel † These authors contributed equally to this work. formation and patterning.

2. Methods (3 10 min), followed by incubating with horseradish peroxidase- conjugated secondary antibody (1:1000) for 2 h at RT and washed 2.1 Morpholino injections and verification (3 10 min). The target protein spots were visualized by enhanced of the specificity chemiluminescence (Amersham Biosciences Corp., Piscataway, NJ, USA). The films were scanned for analysis and imaging. The caldesmon antisense morpholino oligonucleotides (MOs) and 5-base mismatch controls were purchased from Gene Tools (Philo- math, OR, USA). The caldesmon antisense MO1 5-AGTAAAGTCTCTTA 2.3 Morpholino rescue experiment TTCTTCAACGC-3 and MO2 5-TAAGAGTTCATCCTGTAGAGTGATG-3 were designed to inhibit translation of the caldesmon RNA (gene: The cDNA of human CaD served as a template containing a T7 RNA ENSDARG00000032052; transcript: ENSDART00000067366; trans- polymerase promoter. RNA was in vitro synthesized using the mMES- lation: ENSDARP00000067365) and a 5-base mismatch control 5- AGA SAGE mMACHINE kit (Ambion, Austin, TX, USA) according to the AAACTCTCTTATTGTTGAAGGC-3 was used. First, in a titration exper- manufacturer’s protocol and co-injected with the morpholinos. iment, the MOs were injected into the yolk sac of zebrafish embryos Downloaded from https://academic.oup.com/cardiovascres/article/81/2/362/286189 by guest on 24 September 2021 between one- and two-cell stages at different concentrations (2, 4, 2.4 Immunostaining of whole mount and sections and 8 ng/embryo), and the embryos were raised at 28.58C until analysis under standard laboratory conditions. The concentration Briefly, embryos were fixed by 4% PFA at RT for minimal 3 h following (4 ng/embryo) was used in all the subsequent experiments, standard procedures. Embryos were treated with 1 M NH4Cl at RT for because the survival of the embryos was satisfactory (.86%). The 3 h to quench autofluorescence, permeabilized, blocked by 5% goat use of zebrafish embryos was approved by the Institutional Review serum in PBST for 1 h at RT, and incubated with the selected primary Board for experimental animals. antibodies: VEGFR2/Flk1 (Lab Vision), VEGFR1/Flt1 (Lab Vision), There are several options to determine whether a phenotype is endothelial nitric oxide synthase (eNOS) (Lab Vision), CD105 (Lab the result of knocking down a gene-of-interest by blocking trans- Vision), Glut-1 (Dako), occludin (Zymed), ZO-1 (Zymed), and Tie-2 lation with morpholinos (MO): (a) quantification of the target (R&D System) at dilution 1:50 to 100 for two to three overnights protein by an antibody (check if the translation is blocked); (b) at 48C. After post-incubation washing, embryos were incubated RNA rescue experiments; (c) a control morpholino with 5 bp mis- with FITC- or rhodamine-conjugated goat-anti-rabbit or goat-anti- match (discussed earlier); and (d) application of a second MO with mouse (Jackson ImmunoResearch Laboratories, Inc.) at a dilution similar blocking effects (discussed earlier). We verified the speci- of 1:100 for two overnights at 48C. After thoroughly washing, fluor- ficity of the MO phenotype in our experiments by combining the escence images were recorded by a fluorescence microscope and/or above-mentioned methods. confocal laser scanning microscopy (CLSM). Embryos at identical developmental stages, processed without primary antibody, were used as controls for each experiment. 40,6-diamidino-2-phenylindole 2.2 Quantification of the homologue protein (DAPI) was used for nuclei counterstaining. of caldesmon Immunohistochemical analysis of sections was performed by standard methods. Briefly, embryos were fixed, dehydrated, embedded Quantification of the homologue protein was performed by vision in paraffin, and sectioned (5 mm). Sections were deparaffinized, assay (see whole-mount staining and section immunohistochemis- blocked, antigen-retrieved, incubated with the selected primary try), dot blot, and whole-mount enzyme-linked immunosorbent antibodies: EP050852 (Eurogentec) at 1:100 and Glut-1 (Dako) at assay (ELISA) by using the antibody EP050852 (Eurogentec, 1:150, washed, and stained by AP-conjugated secondary antibody. Belgium) against the protein. Western blotting was impossible AP-based substrate was used for visualization. because the antibody was not working in a reducing status. The cross-species reactivity of the antibodies used was confirmed In whole-mount ELISA, zebrafish embryos were fixed in 4% paraf- by immunohistochemistry, western blotting, and dot blot, unless ormaldehyde (PFA) at least 3 h, washed and permeabilized by TBST already tested by others,13–15 or the antibody was specifically (0.05% Triton X-100 in TBS), and treated with alkaline phosphatase raised in zebrafish (Tie-2). (AP) suppressor (0.5 M EDTA) to inhibit endogenous AP for 1 h. Next, the embryos were routinely blocked and incubated with EP050852 (1:50) for two to three overnights at 48C, washed with 2.5 Whole-mount endogenous alkaline phosphatase TBST, and incubated with AP-conjugated secondary antibody (EAP) staining (1:100) for two overnights at 48C. After washing, the embryos EAP staining was applied following fixation and permeabilization of were placed in a 96-well microplate, one embryo per well. the embryos at a defined stage in 4% PFA and 0.05% Triton X-100 in p-nitrophenyl phosphate substrate (Sigma-Aldrich) was used as an PBS (PBST). The staining procedures were as described previously.16 enzyme substrate. The protein was quantified by measuring the optical density (OD) of the enzymatic end-product at 405 nm using a microplate reader (Thermo Multiskan Ascent). The embryos 2.6 Whole-mount lectin staining without primary antibody incubation were processed to estimate non-specific background, the mean value of which was subtracted Briefly, embryos were fixed with 4% PFA, permeabilized with PBST, from each 405 measurement. The mean value of the normal and washed by PBS, followed by incubation at 48C overnight with FITC-conjugated BSI-B4 (Sigma, 10 mg/mL) in PBS. After subsequent embryos was as the ODcontrol in the below formula. The MO inhibition effect of the protein was calculated by the following formula: washes, fluorescent images were recorded by fluorescence microscopy and/or CLSM. %inhibition¼[(ODcontrol2ODMO)/ODcontrol100%]. For the dot blot assay, a nitrocellulose membrane (NM) was used for spotting the samples. The extracts were made by identical 2.7 Parameters for assessment of the numbers of control embryos, wild type (WT), and caldesmon mor- cardiovascular system and the circulation phants (CaD-MOs) at an identical developmental stage, which were homogenized by an identical volume of lysis buffer. The The parameters for assessing the cardiovascular system included: protein concentration was adjusted at 2 mg/mL. The extract (2 mL) (a) DA and PCV; (b) ISV; (c) DLAV; (d) SIV; (e) CVP; (f) cranial from controls, WT, and CaD-MOs was spotted on the NM, respect- vessels (CVs); and (g) AA. The parameters for monitoring the circu- ively. The NM was air-dried and blocked in blocking solution (10% lation included: (a) relative number of circulating red blood cells non-fat dry milk) for 1 h at room temperature (RT). The NM was (RBCs); (b) axial circulation; (c) ISV and DLAV circulation; and (d) incubated with EP050852 (1:100) for 2 h at RT, washed in PBS circulation shunting. 364 P.-P. Zheng et al.

Table 1 Alignment of human CALD1 isoforms to the zebraﬁsh homologue

Protein Accession no. Protein ID MW Designation EMBOSS Align Results (Blosum62)

Method Identity Similarity Isoform 1 NM_033138 NP_149129.2 h-CaD h-CaD Global 34.6% 50.1% Isoform 2 NM_004342 NP_004333.1 l-CaD WI-38 l-CaD II Global 33.7% 46.1% Isoform 3 NM_033157 NP_149347.1 l-CaD WI-38 l-CaD I Global 34.7% 47.0% Isoform 4 NM_033139 NP_149130.1 l-CaD Hela l-CaD I Global 34.6% 46.5% Isoform 5 NM_033140 NP_149131.1 l-CaD Hela l-CaD II Global 33.5% 45.6% C-terminal Functional domain Local 47.8% 67.5% N-terminal Functional domain Local 42.0% 55.6% Downloaded from https://academic.oup.com/cardiovascres/article/81/2/362/286189 by guest on 24 September 2021

2.8 Staining of circulating red blood cells the homologue (Table 1). According to domain mapping of the human CALD1, there are two major functional domains: N-terminal Non-fixed embryos were stained by o-dianisidine for 15 min in the 17 myosin/calmodulin-binding domains and C-terminal calmodulin/ dark, as described previously. actin-binding domains, which are conserved in each isoform.28–30 The functional domains were further aligned with the homologue, 2.9 General strategy of labelling of the developing respectively (Table 1). vasculature Quantification of mRNA does not necessarily provide information about the amount of active proteins in a cell because of nonsense-mediated 3. Results and discussion 18 mRNA decay and microRNA-mediated mRNA silencing. These mech- The evolutionarily conserved similarities of the amino acid 19 anisms are conserved in vertebrates. Therefore, in this study, we sequences at the two major functional domains between preferentially used protein-based approaches for fine characterization the zebrafish homologue and the human CaD reach 67.5 of the effects on the various blood vessels in the CaD-MOs. and 55.6% (Table 1), respectively. Therefore, the homologue For the overall screening of the structural defects of the developing vasculature, we used whole-mount EAP and BS-l isolectin B4 staining. may be regarded as an orthologue of the human caldesmon RBC expression of globulin was used to examine the functional integ- gene (CALD1). The expression pattern of CaD in 2dpf, 3dpf, rity of the circulatory system. In vivo imaging was used for monitoring and 4dpf is similar. It is predominantly expressed in the AA the dynamics of the circulation. At the molecular level, we used the (Figure 1A1, C1 and E1); the jaw primordium (Figure 1A1); molecular markers VEGFR2/Flk1, VEGFR1/Flt1, eNOS, CD105, occlu- the subintestinal vessels (Figure 1B1); DA; PCV; ISVs; and din, Glut-1, ZO-1, and Tie-2 to specifically label the angiogenic ECs/ other regions such as somites (Figure 1D1 and E2). At EPCs and their interaction as well as the regionally specialized earlier stages, CaD is predominantly expressed in DA and vessels according to their labelling peculiarity. VEGFR2/Flk1 and PCV (Figure 1F). VEGFR1/Flt1, encoding receptor-type protein tyrosine kinases, are We compared the ATG- and the control-MO-injected specific markers for angiogenic EC/EPC.12,20 The expression of embryos at various developmental time points. ATG MOs VEGFR2/Flk1 diminishes in the late developmental stages. For example, by 26hpf, the expression of VEGFR2/Flk1 can be seen in knockdown of CaD resulted in severe and reproducible phe- the trunk in the domains, where the ISVs will form and the expression notypes. Morphological changes were apparent by visual is still detected in newly forming ISVs after the onset of the blood cir- inspection at 2dpf and became even more pronounced at culation.21 By 36hpf, the VEGF2 expression is weak or not detectable 3dpf through later time points. The phenotypes observed in newly formed vessels.21 In contrast, VEGFR1/Flt1 retains its in early stages were still present in 5dpf CaD-MOs. The mor- expression until later stages.22 Clearly, the expression of the various pholino phenotype was highly penetrant as representatively markers is dependent on the specific developmental stage. eNOS demonstrated by 2dpf (88%; n ¼ 96), 3dpf (86%; n ¼ 96), and CD105 are well-characterized endothelial-lineage markers, pre- and 4.5/5dpf (92%; n ¼ 126) vs. the control-injected dominantly expressed in relatively differentiated EPCs and mature embryos 2dpf (6%; n ¼ 98) and 3dpf (5%; n ¼ 86). All ECs.23,24 Occludin and ZO-1 are major components of tight junctions paired comparisons were highly significant (P , 0.001). (TJ) of endothelium,15 and the molecules serve well as a monitor of vascular integrity. Glut-1, a marker for the integrity of the brain– The specificity of the morpholino phenotype was verified blood barrier (BBB), is expressed in the ECs of microvessels with a by the following approaches. Reduced or absent expression barrier property.25,26 Tie-2 is an endothelium-specific receptor tyro- of this protein was observed in the CaD-MOs by immunohis- sine kinase.27 tochemistry (Figure 1A2, B2, C2, and D2). Confirmation of the reduction was obtained by whole-mount ELISA and dot 2.10 Sequence identification and alignment of the blot assay (Figure 1G). The protein was inhibited 86% in zebrafish homologue to human CALD1 CaD-MOs as measured by whole-mount ELISA. These results are evidence that the targeted translation was successfully Based on a reciprocal BLAST analysis of the Ensembl zebrafish blocked and proof the specificity of the knockdown. genomic sequence database, the following gene (Gene ID: Further support for the specificity of the knockdown was ENSDARG00000032052; peptide ID: ENSDARP00000067365) was the similar effects of a second morpholino when compared identified as a homologue (a putative orthologue) to human CALD1 in the Ensembl Gene Report. The human caldesmon (CALD1) with the first one and the insignificant penetrance (6%) sequences were acquired from public databases (NCBI). We of the 5-base mismatch control morpholino. The specificity aligned the homologue against the peptide sequences (each was further confirmed by the RNA rescue experiments. isoform) of the human CALD1. The sequence comparisons resulted Co-injection of the ATG MOs with the RNA resulted in a sig- in a nearly equal identity and similarity for each isoform against nificant rescue (89%, n ¼ 66, P , 0.001). Caldesmon in vascular development 365 Downloaded from https://academic.oup.com/cardiovascres/article/81/2/362/286189 by guest on 24 September 2021

Figure 1 Expression of the caldesmon (CaD) homologue in zebraﬁsh embryos and its knockdown in the caldesmon morphants (CaD-MOs). (A1) CaD expression in aortic arch (AA) (square) and JP (arrows) in a 3dpf wild-type (WT). (A2) A reduction of CaD expression in this region was seen in the corresponding CaD-MOs (square). (B1) CaD are expressed in the subintestinal vessel in a 3dpf WT (arrow). (B2) Loss of this protein from this region is seen in the corresponding CaD-MO (arrow). (C1) CaD is expressed in AA (square) in a 4dpf control. (C2) Massive loss of expression in this region is seen in the corresponding CaD-MO. (D1) CaD is expressed in dorsal aorta (DA), posterior cardinal vein (PCV) (arrowheads), intersegmental vessels, and other regions such as somites in 4dpf control. (D2) Reduced expression of this protein in these regions is seen in the corresponding CaD-MO. (E1) CaD expression in AA (arrow) in a 2dpf WT. (E2) Expression pattern of CaD in the trunk of a 2dpf WT. The expression pattern is similar to that in 3dpf and 4dpf. (F) CaD is expressed in DA and PCV (arrows) in 1.5dpf+WT. (G) An overall reduction of the protein expression in the CaD-MOs was assessed by the dot blot assay. (H ) The mismatch-morpholino (MM)-injected embryos did not show a reduction in the expression. (A1) to (B2): section staining; (C1) to (E2): whole-mount staining, lateral views, anterior towards the left, and dorsal is up; (F): whole-mount staining, whole embryo view. Scale bar, 50 mm.

The vasculogenic axial vessels (DA and PCV) and the angio- angiogenesis including cellular proliferation and migration.33 genic trunk vessels (ISV, DLAV, SIV, and CVP) in the CaD-MOs Thus, the development of the ISVs serves as an ideal model were either completely or partially missing or abnormally for studying EC migration during vascular development.21 patterned (Figure 2). Concomitant impairment of the circu- The ISVs are generated by synchronously collective migration lation (Table 2 and Figure 2) and vascular integrity (Figure 2) of ECs in a two-step process.32 The ﬁrst step requires the such as decreased RBCs in the circulation system, disrupted sprouting and migration of ECs from the DA to form a TJs, and haemorrhages were recorded in the morphants. primary network of the ISV segments, whereas the second Functionally, in live imaging, the control embryos showed step encompasses the sprouting of ECs from the PCV interface vigorously circulating blood cells throughout the length of with this primary network to form the ISV network.32 Anatomi- the body, including the axial vessels, ISV, and DLAV. In the cally, the ISVs are typically along the vertical somite bound- CaD-MOs, we found a signiﬁcantly reduced number of circu- aries (myoseptal boundaries).34 A schematic illustration for lating blood cells moving sluggishly through the axial vessels the network of the ISVs as well as its relationship to DA, with absent or reduced circulating blood cells in the ISV and PCV, DLAV, and somites is shown (Figure 3). A well-formed DLAV (Table 2). The circulatory dysfunction and vascular DA and PCV are prerequisites for the formation of an intact immaturity are the sequel of abnormal vessels, rather than primary and secondary network of the ISV segments. In the a defect in RBC development, as normal looking RBCs were completed primary ISV network, ECs are located at the still seen in the lumina of the vessels of the CaD-MOs. DLAV–primary segment junction (ECD), at the level of the Other major circulatory defects in the CaD-MOs included parachordal vessels (ECp), and at the DA–primary segment 32 misconnection of arteries and veins and shunting (Figure 2 junction (ECA) (Figure 3). The normal pattern of ISVs is out- and Table 2). linedbyFlt1(Figure 2B1), occludin (Figure 2C1), Tie-2 Despite the fact that ISVs are generated by sprouting (Figure 2D1), and lectin (Figure 2H1). The pattern is indicative angiogenesis, it seems more appropriate to describe this of the proper alignment of the ECs. In the primary network of process as vasculogenesis type II. There is almost exclusive the ISVs, a defect in the formation of the DA results in an migration of angioblasts followed by tubular formation and initially defective ECA with complete loss of the sprouts of little or no cell division,31,32 rather than traditional the ISVs (Figure 2B3, C2, D2,andH2). The absence of ECp 366 P.-P. Zheng et al. Downloaded from https://academic.oup.com/cardiovascres/article/81/2/362/286189 by guest on 24 September 2021

Figure 2 Defects of developing trunk vasculature and circulation in the caldesmon morphants (CaD-MOs). (A1) VEGFR2/Flk1 labelling in a 3dpf control. The intense-stained sites are located in the tip endothelial cells (ECs) (green arrowhead) and the budding regions (blue arrowheads). (A2) The tip cells are aligned on the bottom of the tissue section (square). (A3) The signal of VEGFR2/Flk1 is perturbed in the corresponding CaD-MO, as reflected by a significant loss of the tip and budding EC population instead of the dot-patchy patterns of dorsal aorta (DA) and posterior cardinal vein (PCV) (arrowheads). (B1) A normally patterned intersegmental vessel (ISV) is revealed by VEGFR1/Flt1 labelling in a 3dpf control. (B2) A complete or partial loss of ISVs with an interrupted DA and loss of dorsal longitudinal anastomotic vessel (DLAV) is seen in the corresponding CaD-MO. (B3) Bright red dots (the cell bodies of ECs) stained by VEGFR1/Flt1 are aligned in the midline without formation of ISVs in a 5dpf CaD-MO. (C1) Occludin is expressed in the trunk vasculature of a 2dpf control. (C2) The signal of occludin is absent in the corresponding regions of the CaD-MO. (D1) Tie-2 outlines the trunk vasculature in a 2dpf control. The expression pattern is similar to that of VEGFR1/Flt1 in (B1). (D2) The Tie-2 signal almost completely disappeared in the CaD-MO. (E1) The caudal vessel plexus (CVP) shows normal patterning as revealed by endothelial nitric oxide synthase (eNOS) staining in a 5dpf control. (E2) A less-developed CVP region with reduced endothelial sheets is seen in the corresponding CaD-MO. (F1) The endothelium of PCV labelled by eNOS is seen as a monolayer in a 5dpf control. (F2) A multilayered staining pattern is seen in the corresponding CaD-MO. (G1) A normal patterning of subintestinal vein (SIV) is shown in a 3dpf control. (G2) An abortive SIV is seen in the corresponding CaD-MO. (H1) Lectin staining outlines the trunk vasculature in a 5dpf control. (H2) Bright green dots (the cell bodies of ECs) labelled by lectin are chaotically distributed. Most dots are out of the normal vascular areas and some are aggregated in clusters. (I1) Red blood cells (RBCs) travel through the body length of a 5dpf control as revealed by RBC staining. (I2) In the corresponding CaD-MO, a remarkable reduction in the circulating RBCs is shown in the axial vessels (DA and PCV), ISV, and DLAV. Extravasation of RBCs from the heart (I2, circle) or haemorrhage (I3). (J1) The PCV is precisely anchored at the ventromost part of the mid- trunk in a 5dpf control, as revealed by the RBC staining. The distance between DA and PCV (height) is precisely determined. (J2) The PCV in the corresponding region of the corresponding CaD-MO is detached from the ventromost part by dorsal shifting of the axial vein as revealed by RBC staining. The linear pattern of the blood flow is lost, while widening is noticed in the CaD-MO. In addition, the flow in the PCV is conducted erroneously into the ISV (arrow). (K1) The normal pattern of the caudal circulation is noticed by the continuous blood flow from DA into the caudal artery and its return via the caudal vein in a 5dpf control. (K2) The flow is interrupted in the CaD-MO. Instead, there is direct return into the PCV by shunting (arrows). All panels: whole-mount staining, lateral views, anterior towards the left, dorsal is up. Scale bar, 100 mm. Caldesmon in vascular development 367

Table 2 Frequencies (%) of circulation defects at 2, 3, and 4.5dpf caldesmon morphants

Parameters Methods Frequency (%) Number

Circulating RBCs RBC staining Reduced 69 Rare 26 Axial circulation Living imaging Sluggish 52 Absent 42 ISV þ DLAV circulation Living imaging Absent þ slow 96

Circulation shunting RBC staining þ live imaging 62 Downloaded from https://academic.oup.com/cardiovascres/article/81/2/362/286189 by guest on 24 September 2021 Total embryo 186

Figure 3 Relationship of the trunk vasculature with somites. The scheme illus- trates the anatomical relationship of dorsal aorta, posterior cardinal vein, intersegmental vessel, dorsal longitudinal anastomotic vessel, and somites.

and ECD in the CaD-MOs most likely results from the failure of migration of the ECs, which are committed to become ECp and ECD. The sequel is partial loss of dorsal sprouts of ISVs (Figure 2B2). The ISV formation is completed before 3dpf during normal development.35 However, in 5dpf CaD-MOs, ECs are still aligned in the midline (Figure 2B3 and H2), additional evidence of impaired EC migration. These results proof a role for CaD in the regulation of adhesion-dependent signalling of cytoskeletal organization and cell migration, con- sistent with various in vitro studies.36–39 The large CVs were replaced by disorganized vascular channels, and the cranial microvasculature was signiﬁcantly reduced (Figure 4). The developing AAs were structurally and functionally damaged in the CaD-MOs (Figure 5). AA defects may associate with other defects in pharyngeal arch-derived structures. However, defects in aortic and pharyngeal arches may also occur separately,10 because the pharyngeal arch structure is highly heterogeneous and receives contributions from endoderm, mesoderm, ecto- derm, and neural crest.10 Dissociated defect of aortic and pharyngeal arches is seen in the CaD-MOs (Figure 5G1–G3). In most vertebrates, TJs between adjacent ECs are major components of the BBB,15 including tetraspanning trans- membrane proteins such as occludin and claudin or cytoplasmic-anchoring proteins such as ZO-1.40 The BBB in zebraﬁsh is revealed by the expression of ZO-1 in the Figure 4 Defects of cranial vasculature in the caldesmon morphants (CaD-MOs). (A1) Functional cephalic vessels are revealed by actively passing through red blood cells (RBCs) in a 5dpf control. (A2) Normal pat- stain microvessels or capillaries and the signals overlap well. (D1) Glut-1 is terned vessels are greatly reduced and replaced by dilated, aneurysmatic expressed in the blood–brain barrier capillaries of the normal brain in a channels (arrows) in the corresponding CaD-MO. (B1) Well-developed 3dpf control. (D2) Loss of Glut-1 expression is observed in the corresponding cranial microvasculature is seen by CD105 staining in a 5dpf control. (B2) CaD-MO. (C1), (C2), (D1), and (D2): section staining; all others: whole-mount The microvasculature is greatly decreased in the corresponding CaD-MO. staining; ventral view; anterior towards top [(A1) and (A2)]; and ventrolateral (C1) and (C2) On frozen sections of 3dpf wild-type, double labelling of ZO-1 view and anterior towards the left [(B1) and (B2)]. Scale bar, 50 mm. and Glut1 was performed. The image shows brain area. Both antibodies 368 P.-P. Zheng et al.

crucial for a proper vascular development in vivo, and this molecule deserves to be further explored as a therapeutic target for anti-angiogenesis or be used as a stimulator of neoangiogenesis in ischaemic diseases.

Acknowledgements The authors thank W.C. Hop (Department of Biostatistics, Erasmus Medical Center, Rotterdam, The Netherlands) for his assistance with the statistical analysis.

Conﬂict of interest: none declared. Downloaded from https://academic.oup.com/cardiovascres/article/81/2/362/286189 by guest on 24 September 2021

References

1. Haruna M, Hayashi K, Yano H, Takeuchi O, Sobue K. Common structural and expressional properties of vertebrate caldesmon genes. Biochem Biophys Res Commun 1993;197:145–153. 2. Zheng PP, Luider TM, Pieters R, Avezaat CJ, van den Bent MJ, Sillevis Smitt PA et al. Identification of tumor-related proteins by proteomic analysis of cerebrospinal fluid from patients with primary brain tumors. J Neuropathol Exp Neurol 2003;62:855–862. 3. Zheng PP, Sieuwerts AM, Luider TM, van der Weiden M, Sillevis-Smitt PA, Figure 5 Defects of aortic arches (AAs) in the caldesmon morphants Kros JM. Differential expression of splicing variants of the human caldes- (CaD-MOs). (A1) AA is seen in a 3dpf control by endogenous alkaline phospha- mon gene (CALD1) in glioma neovascularization versus normal brain tase (EAP) staining (circle). (A2) AA is mostly lost in the corresponding microvasculature. Am J Pathol 2004;164:2217–2228. CaD-MO. (B1) Glut-1 is expressed in AA region in a 3dpf control. (B2) Glut-1 4. Zheng PP,van der Weiden M, Kros JM. Differential expression of Hela-type signal is largely lost in the corresponding CaD-MO. (C1) Glut-1 is expressed caldesmon in tumour neovascularization: a new marker of angiogenic in the AA region in a tissue section of a 3dpf wild-type (arrow). (C2) The endothelial cells. J Pathol 2005;205:408–414. Glut-1 signal is completely lost in the corresponding CaD-MO (arrow). (D1) 5. Zheng PP, Hop WC, Sillevis Smitt PA, van den Bent MJ, Avezaat CJ, A more developed AA is seen in a 5dpf control by EAP staining (circle). (D2) Luider TM et al. Low-molecular weight caldesmon as a potential serum AA is lost in the corresponding CaD-MO. (E1) Occludin is expressed in the marker for glioma. Clin Cancer Res 2005;11:4388–4392. AA region in the 5dpf control (circle). (E2) The occludin signal is completely 6. Zheng PP, van der Weiden M, Sillevis Smitt PA, Luider TM, Kros JM. Hela lost in the corresponding CaD-MO. (F1) Functional AA is revealed by the red /-CaD undergoes a DNA replication-associated switch in localization blood cell staining in a 5dpf control (square). (F2) Loss of normal functional from the cytoplasm to the nuclei of endothelial cells/endothelial pro- AA is seen in the corresponding CaD-MO. [(G1)to(G3)] Glut-1 signal is lost genitor cells in human tumor vasculature. Cancer Biol Ther 2007;6: in the CaD-MO (G1), but retains the pharyngeal arches as counterstained by 886–890. DAPI (G2). (C1) and (C2): section staining; all others: whole-mount staining; 7. Zheng PP, van der Weiden M, Kros JM. Hela l-CaD is implicated in the ventrolateral view and anterior towards the left. Scale bar, 100 mm. migration of endothelial cells/endothelial progenitor cells in human neo- plasms. Cell Adhes Migrat 2007;1:84–91. 8. Isogai S, Horiguchi M, Weinstein BM. The vascular anatomy of the devel- cerebral microvessels.15 The BBB in zebrafish is functionally oping zebrafish: an atlas of embryonic and early larval development. Dev Biol 2001;230:278–301. and structurally similar to that of higher vertebrates such as 9. Weinstein BM. Plumbing the mysteries of vascular development using the 15 humans, which is illustrated by overlapping expression of zebrafish. Semin Cell Dev Biol 2002;13:515–522. Glut-1 and ZO-1 (Figure 4C1 and C2). In the CaD-MOs, the 10. Anderson MJ, Pham VN, Vogel AM, Weinstein BM, Roman BL. Loss of development of the BBB was interrupted (Figure 4D2). unc45a precipitates arteriovenous shunting in the aortic arches. Dev The structural integrity and functional maturity of the Biol 2008;318:258–267. 11. Lyons MS, Bell B, Stainier D, Peters KG. Isolation of the zebrafish homo- blood vessels are not only determined by the presence of logues for the tie-1 and tie-2 endothelium-specific receptor tyrosine normally functioning ECs, but also involve interendothelial kinases. Dev Dyn 1998;212:133–140. junctions and mural cells (smooth muscle cells or pericytes), 12. Habeck H, Odenthal J, Walderich B, Maischein H, Schulte-Merker S. which are generally conserved among vertebrates. However, Analysis of a zebrafish VEGF receptor mutant reveals specific disruption of angiogenesis. Curr Biol 2002;12:1405–1412. some interspecies differences exist. For instance, because 13. Jensen AM, Westerfield M. Zebrafish mosaic eyes is a novel FERM protein there are no additional layers of elastic lamina or mural required for retinal lamination and retinal pigmented epithelial tight cells in the vascular walls of the early developing trunk vas- junction formation. Curr Biol 2004;14:711–717. culature of zebrafish embryos, vascular integrity is predomi- 14. Fritsche R, Schwerte T, Pelster B. Nitric oxide and vascular reactivity in nantly determined by intact endothelium and associated developing zebrafish, Danio rerio. Am J Physiol Regul Integr Comp 41,42 Physiol 2000;279:R2200–R2207. TJs. The junction-associated molecules may well serve 15. Jeong JY, Kwon HB, Ahn JC, Kang D, Kwon SH, Park JA et al. Functional as sensitive markers of normal or disturbed function of the and developmental analysis of the blood–brain barrier in zebrafish. developing vasculature (Figure 2C1). The haemorrhages Brain Res Bull 2008;75:619–628. (Figure 2I2 and I3), the regional and systemic arrest of 16. Serbedzija GN, Flynn E, Willett CE. Zebrafish angiogenesis: a new model for drug screening. Angiogenesis 1999;3:353–359. blood circulation (Figure 2I2 and Table 2), and the shunting 17. Detrich HW III, Kieran MW, Chan FY, Barone LM, Yee K, Rundstadler JA of blood (Figure 2J2 and K2 and Table 2) all are conse- et al. Intraembryonic hematopoietic cell migration during vertebrate quences of vascular dysfunction and/or immaturity. development. Proc Natl Acad Sci USA 1995;92:10713–10717. The present study provides the first functional assessment 18. Shyu AB, Wilkinson MF, van Hoof A. Messenger RNA regulation: to trans- of CaD in embryonic development in vivo by showing that late or to degrade. EMBO J 2008;27:471–481. 19. Muhlemann O, Mock-Casagrande CS, Wang J, Li S, Custodio N, knockdown of its expression severely interferes with vascu- Carmo-Fonseca M et al. Precursor RNAs harboring nonsense codons lar development in zebrafish. The data indicate that CaD is accumulate near the site of transcription. Mol Cell 2001;8:33–43. Caldesmon in vascular development 369

20. Fong GH, Rossant J, Gertsenstein M, Breitman ML. Role of the Flt-1 recep- 32. Isogai S, Lawson ND, Torrealday S, Horiguchi M, Weinstein BM. Angiogenic tor tyrosine kinase in regulating the assembly of vascular endothelium. network formation in the developing vertebrate trunk. Development Nature 1995;376:66–70. 2003;130:5281–5290. 21. Blum Y, Belting HG, Ellertsdottir E, Herwig L, Luders F, Affolter M. 33. Risau W. Mechanisms of angiogenesis. Nature 1997;386:671–674. Complex cell rearrangements during intersegmental vessel sprouting 34. Roman BL, Weinstein BM. Building the vertebrate vasculature: research is and vessel fusion in the zebrafish embryo. Dev Biol 2008;316:312–322. going swimmingly. Bioessays 2000;22:882–893. 22. Dumont DJ, Fong GH, Puri MC, Gradwohl G, Alitalo K, Breitman ML. Vas- 35. Song M, Yang H, Yao S, Ma F, Li Z, Deng Y et al. A critical role of vascular cularization of the mouse embryo: a study of flk-1, tek, tie, and vascular endothelial growth factor D in zebrafish embryonic vasculogenesis and endothelial growth factor expression during development. Dev Dyn 1995; angiogenesis. Biochem Biophys Res Commun 2007;357:924–930. 203:80–92. 36. Helfman DM, Levy ET, Berthier C, Shtutman M, Riveline D, Grosheva I 23. Duda DG, Fukumura D, Jain RK. Role of eNOS in neovascularization: NO et al. Caldesmon inhibits nonmuscle cell contractility and interferes for endothelial progenitor cells. Trends Mol Med 2004;10:143–145. with the formation of focal adhesions. Mol Biol Cell 1999;10:3097–3112. 24. Fox SB, Harris AL. Histological quantitation of tumour angiogenesis. 37. Numaguchi Y, Huang S, Polte TR, Eichler GS, Wang N, Ingber DE. APMIS 2004;112:413–430. Caldesmon-dependent switching between capillary endothelial cell 25. Pardridge WM. Advances in cell biology of blood–brain barrier transport. growth and apoptosis through modulation of cell shape and contractility. Downloaded from https://academic.oup.com/cardiovascres/article/81/2/362/286189 by guest on 24 September 2021 Semin Cell Biol 1991;2:419–426. Angiogenesis 2003;6:55–64. 26. Abbott NJ. Astrocyte–endothelial interactions and blood–brain barrier 38. Yokouchi K, Numaguchi Y, Kubota R, Ishii M, Imai H, Murakami R et al. permeability. J Anat 2002;200:629–638. l-Caldesmon regulates proliferation and migration of vascular smooth 27. Nishishita T, Lin PC. Angiopoietin 1, PDGF-B, and TGF-beta gene regu- muscle cells and inhibits neointimal formation after angioplasty. Arter- lation in endothelial cell and smooth muscle cell interaction. J Cell ioscler Thromb Vasc Biol 2006;26:2231–2237. Biochem 2004;91:584–593. 39. Zheng PP, van der Weiden M, Kros JM. Hela l-CaD is implicated in the 28. Sobue K, Sellers JR. Caldesmon, a novel regulatory protein in smooth migration of endothelial cells/endothelial progenitor cells in human neo- muscle and nonmuscle actomyosin systems. J Biol Chem 1991;266: plasms. Cell Adhes Migrat 2007;1:84–91. 12115–12118. 40. Persidsky Y, Ramirez SH, Haorah J, Kanmogne GD. Blood-brain barrier: 29. Hayashi K, Fujio Y, Kato I, Sobue K. Structural and functional relation- structural components and function under physiologic and pathologic ships between h- and l-caldesmons. J Biol Chem 1991;266:355–361. conditions. J Neuroimmune Pharmacol 2006;1:223–236. 30. Wang CL, Chalovich JM, Graceffa P, Lu RC, Mabuchi K, Stafford WF. A long 41. De Maziere A, Parker L, Van Dijk S, Ye W, Klumperman J. Egfl7 knockdown helix from the central region of smooth muscle caldesmon. J Biol Chem causes defects in the extension and junctional arrangements of endo- 1991;266:13958–13963. thelial cells during zebrafish vasculogenesis. Dev Dyn 2008;237:580–591. 31. Childs S, Chen JN, Garrity DM, Fishman MC. Patterning of angiogenesis in 42. Miano JM, Georger MA, Rich A, De Mesy Bentley KL. Ultrastructure of zeb- the zebrafish embryo. Development 2002;129:973–982. rafish dorsal aortic cells. Zebrafish 2006;3:455–463.