Lmo2 modulates Sphk1 and promotes endothelial cell migration

Abstract

Objective

Lim-domain only (Lmo)2 transcription factor is involved in hematopoiesis and vascular remodeling. Sphingosine kinase (Sphk)1 phosphorylates sphingosine to sphingosine-1-phosphate (S1P). We hypothesized that Lmo2 regulates Sphk1 to promote endothelial cell (EC) migration and vascular development.

Approach and Results

Lmo2 and Sphk1 knockdown (KD) were performed in Tg(fli1:EGFP)^y1 zebrafish and in HUVEC. Rescue of phenotypes or overexpression of these factors were achieved using mRNA encoding Lmo2 or Sphk1. Endothelial cell proliferation in vivo was assessed by BrdU immunostaining and FACS analysis of dissociated Tg(fli1:EGFP)^y1 embryos. Cell migration was assessed by scratch assay in HUVEC and mouse aortic rings. Lmo2 interactions with Sphk1 promoter were assessed by ChIP-PCR. Lmo2 or Sphk1 KD reduced number and length of intersegmental vessels (ISVs). There was no reduction in the numbers of GFP⁺ endothelial cells following Lmo2 KD. However, reduced numbers of BrdU⁺GFP⁺ nuclei were observed along the dysmorphic ISVs, accumulating instead at the sprouting origin of the ISVs. This anomaly was likely due to impaired EC migration, which was confirmed in migration assays using Lmo2 KD HUVECs and mouse aortic rings. Both in vivo and in vitro, Lmo2 KD reduced Sphk1 gene expression, associated with less Lmo2 binding to the Sphk1 promoter as assessed by ChIP-PCR. Sphk1 mRNA rescued the Lmo2 KD phenotype.

Conclusions

Our data showed that Lmo2 is necessary for Sphk1 gene expression in endothelial cells. Lmo2 KD reduced Lmo2-Sphk1 gene interaction, impaired ISVs formation and reduced cell migration. We identified for the first time Sphk1 as downstream effector of Lmo2.

Introduction

During embryogenesis, blood vessel formation requires vasculogenesis and angiogenesis ¹. Vasculogenesis is the process by which primitive mesodermal cells home to various tissues and differentiate into endothelial cells (ECs) to form the primary capillary network. Angiogenesis is the process by which ECs sprout from pre-existing vessels, proliferate, migrate and assemble into new blood vessels. Migration of ECs is a key process during angiogenesis, in which the directionality of the process is regulated by chemotactic, haptotactic, and mechanotactic stimuli that can activate several signaling pathways converging on cytoskeletal remodeling ².

LIM-domain-only (Lmo)2 (previously known as RBTN2 or TTG2) is a transcription factor essential in both hematopoietic ³ and endothelial pathways ⁴. Lmo2 controls hematopoietic cell fate via protein–protein interactions with key DNA replication proteins MCM6, PRIM1 and POLD1 ^5,6, as well as chromatin-modifying enzymes BAZ1A, SETD8, MYST2, and UHRF1 ^7–9. Lmo2 acts as a bridging molecule through its LIM-domain zinc-finger-like structures ¹⁰, and its canonical function is to assemble a DNA-binding complex which includes the TAL1, E47, GATA1 and Ldb1 proteins in erythroid cells ¹¹. In ECs, a similar multi-complex up-regulates VE-cadherin gene expression through direct binding to the VE-cadherin promoter ¹². Lmo2 is not required for de novo vessel formation during early embryogenesis but is needed for remodeling of vascular networks ⁴ and is normally expressed in vascular endothelium ¹³. However, the role of Lmo2 as well as Lmo2-regulated genes in EC development and migration has not been fully characterized.

In this paper, we provide evidence that during vascular development, Lmo2 plays a critical role in the generation of the intersegmental vessels (ISVs). We find that this critical role of Lmo2 is due to its regulation of EC migration, an effect mediated in part by sphingosine-1 phosphate.

Materials and Methods

Materials and Methods are available in the online-only Data Supplement.

Results

Lmo2 modulates zebrafish vascular development

In our experiments, we used the Tg(fli1:EGFP)^y1 zebrafish line to follow vascular development in vivo. Lmo2-targeted morpholino injection did not affect the gross morphology of the zebrafish, as observed at 24 and 48 hours post fertilization (hpf) (fig. 1A and suppl. fig. I). However, Lmo2 KD significantly impaired the formation of intersegmental vessels (ISVs, fig. 1B & D), but not the larger conduits, compared to the mismatch morpholino (mismatch-Mo) control group. Lmo2 mRNA co-injected with Lmo2 morpholino rescued the phenotype observed in Lmo2 KD embryos, confirming that the impairment of ISV generation and patterning was specifically caused by Lmo2 KD. Western blotting confirmed the Lmo2 KD in morpholino-injected embryos (fig. 1C) whereas co-injection of morpholino with Lmo2 mRNA rescued Lmo2 protein to the level of the control (mismatch-Mo).

Figure 1. Lmo2 KD in tg(fli1:EGFP)^y1 zebrafish embryos.

Zebrafish embryos at 1–2 cell stage were injected with Lmo2 targeted morpholino (Lmo2-Mo), a mismatch morpholino (Mismatch-Mo) as control or co-injected with Lmo2 morpholino and Lmo2 mRNA (mmRNA) for rescue experiments. A. Lmo2 KD did not affect the gross morphology of the whole embryo, here shown at 48 hpf. However, Lmo2 KD reduced significantly ISVs during vascular development compared to mismatch-Mo or rescue experiments [Image in the yellow box for each group is magnified in A′]. B. Western blotting showing an effective Lmo2 KD following Lmo2 morpholino injection and rescue with co-injection of Lmo2 mmRNA. C. Graphical representation showing the significant reduction of ISVs following Lmo2 KD. Data are represented as mean ± S.E.M, n=3, one-way ANOVA test followed by Bonferroni’s post-hoc test; ***=p<0.001, ns= non significant.

The vascular phenotype in Lmo2 KD is not due to impaired EC proliferation

Tg(fli1:EGFP)^y1 embryos were enzymatically digested and cell populations analyzed by FACS. There was no significant difference amongst the Lmo2 KD embryos (8.4% GFP+ cells); those embryos treated with the control mismatch-Mo (9.2% GFP+ cells); and non-injected embryos (9.6% GFP+ cells), with respect to Fli1-expressing cells (fig. 2A; wild-type (Wik) zebrafish line were used to correct for background signal). Because the Fli1-expressing GFP+ cells are primarily of endothelial lineage, these findings suggest that Lmo2 does not regulate EC proliferation. To confirm that Lmo2 was not involved in endothelial cell proliferation and apoptosis, BrdU and TUNEL assay were respectively used. FACS analysis showed that the percentage of BrdU⁺ GFP⁺ (0.81%, fig. 2B) and TUNEL⁺ GFP⁺ (0.34%, fig. 2C) in Lmo2 KD embryos were not different compared to their respective controls.

Figure 2. FACS analysis of proliferation and apoptosis of tg(fli1:EGFP)^y1 zebrafish cells.

Cells were isolated by enzymatic digestion from embryos previously injected with Lmo2 targeting morpholino or controls non injected (NI) and mismatch morpholino. Wik zebrafish line was used as negative control. (A) Analysis of live cells isolated from embryos 48 hpf show that Lmo2 knockdown did not reduce the percentage of GFP⁺ cells in tg(fli1:EGFP)^y1 embryos compared to controls. (B) Analysis of cells isolated from embryos 30 hpf incubated in BrdU 10 mM at 24 hpf and then allowed to develop for 5 hours before enzymatic digestion, fixation and immunostaining. (C) Analysis of cells isolated from embryos 30 hpf following Tunel assay. Data are represented as mean ± S.E.M, n=3, one-way ANOVA test followed by Bonferroni’s post-hoc test, ns= non significant.

More evidence consistent with this notion was derived from BrdU immunostaining in whole embryos. These studies showed that the numbers of proliferating Fli1-expressing cells (GFP⁺ BrdU⁺ nuclei), is similar in control and Lmo2 KD embryos (7.5±0.9 vs 6.8±1.1 proliferating cells × 100 μm of body trunk length (excluding head and tail)). In both groups of embryos, GFP+ BrdU+ nuclei are more evident in larger vessels (fig. 3A–B, shown in dorsal aorta, cardinal vein and dorsal longitudinal anastomotic vessels).

Figure 3. BrdU immunostaining in zebrafish embryos and in vitro migration assay.

A. BrdU immunostaining was used to assess the proliferation of GFP⁺ endothelial cell in vivo and their ability to migrate along the newly forming intersegmental vessels (ISVs). B. There was no difference in the number of double BrdU⁺GFP⁺ cells in whole embryos between control mismatch-MO and Lmo2 KD groups. C. Bar graph showing that in Lmo2 KD, GFP⁺ cell proliferate but fail to migrate as they accumulate where the ISVs branch from the major vessels, resulting in impaired ISVs formation. In control group, BrdU⁺GFP⁺ cells migrate and give rise to the ISVs. D. Migration assay in HUVECs by scratch procedure showed that Lmo2 KD reduces cell migration, as shown by a larger residual gap between the edges of the wounded monolayer, quantified in E. Data are represented as mean ± S.E.M, n=3, Student t-test; ***=p<0.001, ns= non significant.

The vascular phenotype in Lmo2 KD is associated with impaired EC migration

ISVs form by migration of ECs from the larger conduit. Intriguingly, the number of GFP⁺ BrdU+ nuclei found along the ISVs was reduced in Lmo2 KD compared to control embryos (9.2±0.9 vs 3.8±0.6, p≤0.05, measured in all ISVs), fig. 3C. By contrast, the GFP⁺ BrdU⁺ nuclei seem to accumulate at the branch point of the ISVs in Lmo2 KD animals. These data suggest that knockdown of Lmo2 impaired cell migration.

To further delineate if Lmo2 may be required for EC migration, we created a stable line of Lmo2 KD HUVECs using lentiviral shRNA technology followed by puromycin selection (suppl. fig. II). A control (Ctrl) HUVEC line was generated using lentiviral scramble shRNA. Compared with control cells, the Lmo2 KD cells exhibited a 70% decrease in Lmo2 gene expression detected by RT-PCR and an 80% reduction in protein level by western blot (suppl. fig. II A–B). We used a scratch assay to assess EC migration ¹⁴. We observed that Lmo2 KD significantly impaired migration of HUVECs, with an average remaining gap that was larger in Lmo2 KD (590±15 μm) compared to control cells (230±20 μm, p≤0.05; fig. 3D–E).

Role of Sphk1 in zebrafish

To understand what angiogenic genes might be regulated by Lmo2, we analysed mouse Lmo2 ChIP-Seq ¹⁵. We observed Lmo2 binding peaks in the genomic region around VegfA, as well as Flt1 and Kdr ((gene names of VegfR1 and VegfR2, respectively), suggesting these genes as direct targets and downstream effectors of Lmo2 (suppl. fig. III A–C), We did not observe Lmo2 binding peaks in the genomic regions of VegfB, VegfC and Flt4 (gene name of VegfR3) (suppl. fig. III E–F). Real time PCR in Lmo2 KD HUVECs showed a significant reduction in VegfA gene expression but not Kdr (suppl. fig. IV A–B), whereas in Lmo2 KD embryos there was a significant reduction in Kdr gene expression but not VegfA (suppl. fig. IV C–D). These studies indicates that Lmo2 regulates a well-characterized angiogenic pathway.

In addition, our ChIP-Seq studies revealed 2 clear Lmo2 binding peaks in genomic region around Sphk1, but not Sphk2, suggesting Sphk1 is also a direct target gene regulated by Lmo2 (suppl. fig. V). Because of the greater novelty of this observation, we further investigated the possible role of Sphk1 in the action of Lmo2.

Sphk1 phosphorylates sphingosine to sphingosine-1-phosphate (S1P), a mediator of several cellular processes, including cell migration. Therefore, we hypothesized that Lmo2 could regulate endothelial cell migration by modulating Sphk1 (fig. 4A). There is only one group that has examined Sphk in zebrafish, finding that Sphk2, but not Sphk1, ^16,17 mediates cardiac development. Accordingly, it was necessary for us to clone and sequence Sphk1 cDNA before embarking upon Sphk1 loss-of-function studies in vivo. Zebrafish Sphk1 RNA showed 100% coverage with the predicted sequence (suppl. fig. VI). In addition, the percent of identity of zebrafish Sphk1, at the protein level, with human or mouse Sphk1 was 48% or 49%, respectively (suppl. fig. VI A–B).

Figure 4. Lmo2 promotes migration by its direct binding to Sphk1 promoter.

A. Diagram of the hypothesis. Lmo2 activates Sphk1, an enzyme that catalyses the formation of sphingosine-1-phosphate (S1P). In turn, S1P binds to its receptor to promote endothelial cell migration. B. Lmo2 and Sphk1 gene expression patterns during zebrafish development. Real time PCR analysis was performed on total RNA extracted from whole embryos from 15 min post-fertilization up to 120 hpf. C-D-E. Real time PCR analysis in zebrafish embryos (C) and in HUVECs (D) showed that Lmo2 KD or Lmo2 overexpression reduced and increased, respectively, Sphk1 gene expression. This was confirmed by western blot detection (E) of Sphk1 in Lmo2 KD and OE cells. B-actin gene expression was used to normalize data. Data are represented as mean ± S.E.M, n=3. F–G. ChIP-PCR analysis in HUVECs showing Lmo2 binding complex to Sphk1 promoter compared to normal rabbit IgG, used as a negative control. Data are represented as mean ± S.E.M, n=3, one-way ANOVA test followed by Bonferroni’s post-hoc test; ***=p<0.001.

Lmo2 and Sphk1 interactions in ECs and vascular development

The expression of Lmo2 and Sphk1 as assessed by real time PCR is very low between 0.15 and 6 hours post-fertilization (fig. 4B). Both genes increased significantly and with very similar patterns from 6 to 72 hpf. Lmo2 KD significantly suppressed Sphk1 gene expression in zebrafish embryos 48 hpf, whereas Lmo2 overexpression by Lmo2 mRNA upregulated Sphk1 expression (fig. 4C), as shown by real time PCR.

To determine if Lmo2 regulated Sphk1 in a similar fashion in mammalian cells we employed the Lmo2 KD HUVECs described above. In addition, we generated a stable line of Lmo2 overexpressing (OE) HUVECs by transfection of lentiviral particles encoding Lmo2 ORF or virus without the ORF, and selected with puromycin. Compared with ctrl cells, the Lmo2 OE cells exhibited a 9-fold increase in Lmo2 gene expression detected by RT-PCR and a 40-fold increase in protein level by western blot (suppl. fig. IV C–D). We observed that, by comparison to sham-transfected cells, the Lmo2 KD cells expressed less Sphk1; by contrast, the Lmo2 OE HUVECs manifested an increased expression of Sphk1 (fig. 4D–E).

Bioinformatic analyses predicted that Lmo2 could bind to the sequence ACCGATAAGG, located at +3986 position of the Sphk1 gene promoter (Gene ID: 8877). We found that Sphk1 gene expression was significantly reduced in Lmo2 KD ECs (fig. 4F). Furthermore, a ChIP-PCR assay in HUVECs indicated that Lmo2 KD significantly decreases the binding of the Lmo2 associated complex to the Sphk1 gene. This experiment suggests that Lmo2 binds to the Sphk1 promoter to positively regulate Lmo2 expression (fig. 4G).

Similar to the effects of Lmo2 KD, Sphk1 KD did not affect the gross morphology of the zebrafish but significantly impaired the formation of ISVs (figure 5A–B). Again, similar to the Lmo2 KD, the Sphk1 KD HUVECs also manifested a significant impairment in cell migration compared to control (a remaining gap of 595±30 μm vs 260±55 μm, p≤0.05) (fig. 5C–D). Finally, the impaired ISV formation in both Sphk1 KD zebrafish, and in Lmo2 KD zebrafish, could each be rescued by mRNA encoding Sphk1.

Figure 5. Effects of Sphk1 KD in tg(fli1:EGFP)^y1 zebrafish embryos.

Zebrafish embryos at 1–2 cell stage were injected with Sphk1 targeted morpholino (Sphk1-Mo), or a mismatch morpholino (Mismatch-Mo) as control or co-injected with Sphk1 morpholino and Sphk1 mmRNA for rescue experiments. A. Sphk1 KD did not affect significantly the morphology of the whole embryo, here shown at 48 hpf. However, Sphk1 KD reduced ISVs during vascular development (see images in the yellow box that is a magnified view in A′). Sphk1 mRNA rescued the Lmo2 KD phenotype when co-injected with Lmo2 -Mo. B. Graphical representation showing the significant reduction of ISVs following Sphk1 KD. Data are represented as mean ± S.E.M, n=3, one-way ANOVA test followed by Bonferroni’s post-hoc test; ***=p<0.001, ns= non significant.

We also confirmed the relation between Lmo2 and sphingosine-1-phosphate (S1P), of which Sphk1 is an important mediator, by studying cell migration in mice using an additional model, the aortic ring assay (fig. 6). Treatment with Lmo2 SiRNA reduced cell migration in aortic rings, whereas the addition of S1P partially rescued cell migration (fig. 6A–B). Sk1-I, a selective inhibitor of Sphk1, also impaired cell migration, whereas co-incubation with S1P rescued cell migration to a level similar to the control (fig. 6A–B). These data strongly support a model whereby Lmo2 upregulation of Sphk1 plays a critical role in EC migration and ISV generation.

Figure 6. S1P rescues the defect endothelial migration by Lmo2 knockdown or Sphk1 inhibition.

A. Representative image of aortic ring transfected with siRNA ctrl 40nM, Lmo2 siRNA 40nM, or Lmo2 siRNA 40nM plus S1P 100nM, n=6. B. Quantification of the number of junctions per aortic ring showed in A. C. Representative image of aortic ring treated with vehicle control, SK1-I (Sphk1 inhibitor) 10μM, and SK1-I 10μM plus S1P 100nM, n=6. D. Quantification of the number of junctions per aortic ring in C. Data are represented as mean ± SEM. *, p<0.05.

Discussion

Salient findings

The major new findings of our study are that: 1) Lmo2 regulates the generation of intersegmental vessels in zebrafish, a phenomenon that is known to be dependent upon angiogenesis; 2) The effect of Lmo2 on ISV generation appears to be mediated by its promotion of EC migration; 3) The effect of Lmo2 to enhance EC migration and ISV formation is mediated at least in part by its upregulation of Sphk1.

Role of Lmo2 in adult angiogenesis

Angiogenesis is required in many physiological and pathological conditions, including embryonic development, wound healing, tissue regeneration, and tumor growth ¹⁸. The process is characterized by a highly regulated balance between pro- and antiangiogenic agents and involves a cascade of events in which the migration of capillary endothelial cells is an essential component ². Indeed, Lmo2 has been proposed as a marker of vascular differentiation in a variety of native and neoplastic tissues ¹³. In a previous report we showed that Lmo2 regulates proliferation of adult ECs by the control of the endothelial G1/S transition and of the transcriptional regulation of cyclin-dependent kinase (CDK) 2, CDK4, Cyclin D1 and Cyclin A1 ¹⁹. We also showed that Lmo2 affected EC proliferation and tailfin regeneration in the adult zebrafish.

New insights regarding the role of Lmo2 in developmental angiogenesis

In the present study, we investigated the role of Lmo2 during zebrafish vascular development. Uniquely, the zebrafish embryo survives and develops even in complete absence of blood circulation until several days old ^20,21, gaining sufficient oxygenation via diffusion from the environment. This facilitates phenotypic analysis of animals with even severe circulation defects ²². Indeed, Lmo2 KD embryos were viable, and there was no significant alteration in gross morphology, despite a severe vascular anomaly. Specifically, Lmo2 KD substantially reduced the number and length of intersegmental vessels (ISVs). This finding is consistent with a previous observation in murine embryogenesis that the primitive vascular network formed by vasculogenesis does not undergo maturation into functional vascular structures in the absence of Lmo2 ⁴.

It is known that ISVs form by angiogenesis, a process that is known to involve proliferation, survival, migration and differentiation of endothelial cells. However, as opposed to our studies in adult angiogenesis, Lmo2 KD did not have a significant effect on EC proliferation. Specifically, Lmo2 KD did not substantially reduce GFP+ cells in the tg(fli1:EGFP)^y1 embryos. This did not involve proliferation nor apoptosis, as the percentages of BrdU⁺GFP⁺.and TUNEL⁺GFP⁺ cells were similar in Lmo2 KD and control embryos. Furthermore, BrdU immunostaining in the whole embryo also showed that the total number of BrdU+ GFP+ was not different in Lmo2 KD embryos, further supporting the notion that Lmo2 was not critical for EC proliferation in the embryo

In Lmo2 KD embryos the majority of BrdU+ GFP+ cells seemed to accumulate at the branchpoint of ISVs, suggesting that they failed to migrate to form ISVs ²³. To confirm our hypothesis, we used a well-established EC migration assay in vitro. These studies revealed that Lmo2 KD impaired the ability of HUVECs to migrate. The data suggest that Lmo2 governs EC migration, and that during developmental angiogenesis the effect of Lmo2 on EC migration (rather than proliferation) plays a predominant role in ISV formation.

Sphk1 mediates the effect of Lmo2 on developmental angiogenesis

We were interested to understand the gene(s) downstream of Lmo2 that played a role in EC migration ^11,24. Bioinformatic analysis of mouse Lmo2 ChIP-Seq data showed that Lmo2 complex binds to both the VegfA, Flt1 and Kdr as well as the Sphk1 genes, each of which are good candidates for the vascular effects of Lmo2. In addition, in our previous studies Sphk1 was the most downregulated genes in a PCR array performed in Lmo2 KD HUVECs ¹⁹. Because there are no prior studies of the role of Sphk1 in zebrafish vasculogenesis, our subsequent studies focused on this candidate.

Sphingosine kinases phosphorylate sphingosine to form sphingosine-1-phosphate (S1P). S1P can act as an intracellular second messenger ²⁵ or can be released by S1P transporters, including SPNS2 ²⁶. S1P is recognized by G protein-coupled S1P receptors, activating numerous downstream signaling pathways ²⁷. S1P regulates a plethora of cellular events, including cell survival, growth, differentiation, motility and cellular architecture ²⁸ as well as angiogenesis and vascular integrity ^29–31. S1P pathways have been implicated in a broad range of human disease, including endothelial disorders ³², cancer ³³ and inflammatory disorders ³⁴.

Sphks are highly conserved ³⁵ and widely expressed ³⁶ enzymes. Knockout of either Sphk1 or Sphk2 generates phenotypically normal mice or zebrafish ^16,17, suggesting that Sphk1 and Sphk2 have redundant functions that can compensate for each other. However recent studies revealed that overexpression of Sphk2 suppressed cell growth and enhanced apoptosis in vitro ³⁰, in contrast to Sphk1, which generally promoted cell survival and growth ^16,17. The contrasting effects of their overexpression may be due to their different cellular localization, as Sphk1 in human is mostly localized in the cytosol whereas Sphk2 localizes to the mitochondria, endoplasmic reticulum and nucleus ²⁷. Sphk1 promotes proliferation and survival by catalyzing a reaction that produces S1P and depletes levels of ceramide and sphingosine, which are proapoptotic precursors of S1P. Signaling via S1P receptors in vascular endothelial cells results in integrin activation and focal contact assembly, events that are required for cell migration ³⁷.

Our ChIP-PCR data in HUVECs confirmed that Lmo2 associates with Sphk1 gene. Lmo2 KD reduced Sphk1 gene expression, whereas Lmo2 mRNA treatment increased the expression of Sphk1, validating our in vitro data in HUVECs and providing further evidence of the relationship between Lmo2 and Sphk1. In addition, the mouse aortic ring assay showed that sphingosine-1-phosphate rescues the effects of Lmo2 KD or Sphk1 pharmacological inhibition on cell migration. Taken together, these data provide evidence that Lmo2 modulates Sphk1. Consistent with this notion was our observation that in the zebrafish embryo, Lmo2 and Sphk1 shared a similar pattern of gene expression up to 96 hpf, with Lmo2 expression beginning somewhat earlier. To confirm the relationship between these two genes, we needed to design a proper knockdown strategy. We first sequenced the putative Sphk1 mRNA, using primers based on predictions from the Pubmed and Ensembl databases. We confirmed that this predicted sequence was indeed Sphk1.

Sphk1 morpholino KD produced very few effects on the global phenotype, as observed in previous studies ^16,17. However, Sphk1 KD impaired ISV formation, with a phenotype similar, although less severe, compared to that observed following Lmo2 KD. One possible explanation for the milder phenotype is that Sphk1 and Sphk2 have redundant functions that can compensate for each other. In addition, as previously mentioned, the effects of Lmo2 are also likely to be mediated in part through the VEGF pathway. Certainly, other genes downstream of Lmo2 other than Sphk1 may be involved in the effects of Lmo2 on vascular development. It is known that Lmo2 directly interacts with transcription factors mostly of the bHLH family, SCL/TAL1, TAL2, and LYL1 or GATA proteins ^11,24. These interactions have not been reported to be involved in endothelial cell migration, and are worthy of further examination in Lmo2-mediated EC migration. That being said, the ISV defect in Lmo2 KD zebrafish could be almost entirely rescued by Sphk1 mRNA, providing further evidence that Sphk1 mediates the effect of Lmo2 on EC migration and ISV generation. Consistent with this observation is that Sphk1 KD HUVECs manifested an impairment of EC migration, similar to that of Lmo2 KD HUVECs.

Conclusion

To conclude, in this study, we showed that Lmo2 is a determinant of vascular development in the zebrafish, due to an effect on embryonic angiogenesis. As opposed to adult angiogenesis, the predominant effect of Lmo2 during embryonic angiogenesis appears to be on EC migration, rather than proliferation. These effects of Lmo2 are mediated in part by the sphingosine-1-phosphate axis. It is possible that modulation of Lmo2 or Sphk1 expression or activity may provide a novel therapeutic approach either for syndromes of inadequate or pathological angiogenesis.

Highlights.

• Lmo2 or Sphk1 KD impaired intersegmental vessels formation and endothelial cells migration.

• Lmo2 modulated Sphk1 gene expression.

• Lmo2 KD reduced Lmo2 binding to the Sphk1 gene.

• Sphk1-mRNA rescued zebrafish vascular phenotype caused by Lmo2 knockdown.

• The effects of Lmo2 are partially mediated by sphingosine-1-phosphate axis.

Nonstandard Abbreviations and Acronyms

ECs

Endothelial cells

Lmo2

LIM-domain-only 2

Sphk1

Sphingosine kinase 1

S1P

Sphingosine-1-phosphate

ISVs

Intersegmental vessels

Mo

Morpholino

HUVEC

Human umbilical vein endothelial cells

CDK

Cyclin-dependent kinase

KD

Knockdown

Ctrl

Control

OE

Overexpression

BrdU

Bromodeoxyuridine

Hpf

Hours post-fertilization

TUNEL

Terminal deoxynucleotidyl transferase dUTP nick end labeling