G-protein-coupled receptor S1P₁ acts within endothelial cells to regulate vascular maturation

Embryonic blood vessel development occurs via vasculogenesis, angiogenesis, and maturation.^1-3  In vasculogenesis, endothelial cells (ECs) differentiate de novo and form the primary vascular plexus. In angiogenesis, the pre-existing network is remodeled by splitting and sprouting, producing a complex vascular tree of variably sized vessels. Finally, during maturation vascular smooth muscle cells (VSMCs) differentiate and are recruited to the channel wall to stabilize the new vessel, to protect it against rupture, and to provide hemostatic control. This final maturation process requires cell-to-cell communication and interactions between ECs and VSMCs.^4,5 

Although many of the signaling pathways that become engaged during blood vessel formation have been defined,^1-3,6-14  the pathways regulating the interactions between ECs and VSMCs during vascular maturation, especially in larger vessels, remain unclear. The sphingosine-1-phosphate receptor (S1P₁) (formally known as Edg1) was shown to be essential for vascular maturation during embryonic development.¹⁵  S1P₁ is a widely distributed G-protein-coupled receptor for sphingosine-1-phosphate, a blood-borne bioactive lipid. Stimulation of the S1P₁ receptor triggers a Gi-linked pathway, leading to growth, survival, migration, and morphogenesis.^16-18  Disruption of the S1P₁ gene in mice caused embryonic lethality because of massive hemorrhage at embryonic day (E) 12.5 to E14.5.¹⁵  Although in the S1P₁^-/- embryos, VSMCs were in the vicinity of aortae and other vessels, they failed to completely and productively surround the nascent endothelial tubes. Because the S1P₁ receptor is expressed in both ECs and VSMCs, the cell type in which the receptor regulated vascular maturation was uncertain. It was not known whether the S1P₁ receptor controlled VSMC coverage of vessels in a cell-autonomous fashion by functioning directly in VSMCs or indirectly through its activity in ECs.

To address this issue we have established mice with the S1P₁ gene deleted specifically in ECs. Our results demonstrate that vessel coverage by VSMCs is directed by the activity of the S1P₁ receptor in ECs.

The structure of the targeting vector is shown in Figure 1. Gene targeting in TC1 ES cells and generation of chimeric and heterozygous mice were done as described.¹⁵  Mice were genotyped by PCR analyses using DNA from yolk sacs or tail biopsies.

S1P₁^loxP/loxP mice were bred with S1P₁^WT/Ko mice expressing Cre recombinase under the control of the EC-specific promoter Tie2 to obtain S1P₁-conditional knock-out (S1P₁^loxP/KoTie2-Cre) and control littermates (S1P₁^loxP/Ko, S1P₁^loxP/WTTie2-Cre, or S1P₁^loxP/WT). To detect the Cre allele, primers Cre1 (5′GCCTGCATTACCGGTCGATGC) and Cre2 (5′CAGGGTGTTATAAGCAATCCC) were used. To detect the S1P₁^Ko allele, primers P4, 5′CCATCCTCTACTGCAGGATCT; P5, 5′TGCTGCGGCTAAATTCCATG; and P6, 5′TCGCCTTCTTGACGAGTTCTTCTGAG were used, giving a 250-bp fragment for the S1P₁^Ko allele and a 500-bp band for the WT band. Tie2-Cre mice were bred with ROSA26 Cre reporter mice²⁰  to determine Cre-expression specificity. To detect the LacZ gene, primers LacZ-F (5′GATCCGCGCTGGCTACCGGC) and LacZ-R (5′GGATACTGACGAAACGCCTGCC) were used. Embryos were obtained from timed-pregnant females with the morning of vaginal plug considered E0.5. Immunohistochemistry and LacZ expression analysis were done as described previously.¹⁵  The S1P₁ antibody was obtained from Suzanne Mandala (Merck, Whitehouse Station, NJ).

The S1P₁ gene consists of 2 exons and 1 intron; the second exon contains the entire coding region²¹  (Figure 1A). To produce an S1P₁ allele that could be conditionally disrupted, a gene-targeting vector was constructed that contained loxP sequences flanking exon 2. One loxP site was placed within the intron, and 2 others, associated with the neomycin selection cassette, were placed 3′ to the end of the gene (Figure 1A). This targeting vector was used to introduce these changes into the endogenous S1P₁ gene in embryonic stem cells by homologous recombination. A clone that contained the S1P₁^loxP allele was injected into blastocytes to establish chimeric and, subsequently, heterozygous mice. Interbreeding the heterozygous mice yielded viable, homozygous S1P₁^loxP/loxP mice at the expected Mendelian frequency. The homozygous S1P₁^loxP/loxP mice appeared to be without abnormalities, indicating that the changes introduced into the S1P₁ gene did not disrupt the locus. A null S1P₁ allele (S1P₁^Ko) produced embryonic lethality when in homozygous form.¹⁵ 

To delete the S1P₁ gene specifically in ECs we used a previously characterized transgenic mouse line that expresses the Cre recombinase under the control of the promoter and enhancer regions of the mouse Tie2 gene.²²  The specificity of Cre expression in ECs was corroborated by using the ROSA26 reporter line, which contains a LacZ gene that is activated by Cre-mediated recombination.²⁰  Embryos containing the Tie2-Cre and ROSA26 reporter genes showed EC-specific Cre expression. Sections of X-gal-stained embryos revealed that LacZ expression was confined to ECs in vessels in brain (Figure 2A) and in the dorsal aorta (not shown).

The specificity of the Cre-mediated recombination of the S1P₁^loxP gene was demonstrated by examining tissues from embryos. In E12.5 embryos carrying the Tie2-Cre gene, substantially more recombination of the S1P₁^loxP allele to produce the S1P₁^ΔEx2 allele (Figure 1B) was found in yolk sac and aorta (tissues highly enriched in ECs) as compared with embryonic neural tissue (Figure 2B). Tie2-Cre mediated deletion of S1P₁ receptor was confirmed by immunohistochemistry, using an S1P₁-specific antibody. On sections of E12.5 embryos, S1P₁ expression in ECs was substantially reduced in S1P₁^loxP/loxP embryos carrying Tie2-Cre compared with control embryos. By contrast, in brain, epithelium S1P₁ expression was similar in S1P₁^loxP/loxPTie2-Cre embryos and in control embryos (Figure 2C).

When S1P₁^loxP/loxP mice were bred with S1P₁^WT/Ko mice carrying the Tie2-Cre transgene, no S1P₁^loxP/KoTie2-Cre mice were found among 74 offspring, suggesting embryonic lethality. To determine the time of lethality in utero, embryos were analyzed at different days after coitum. At E12.5, the S1P₁-conditional mutant embryos (S1P₁^loxP/KoTie2-Cre) (n = 34) appeared phenotypically similar to the global S1P₁ knock-out embryos (S1P₁^Ko/Ko).¹⁵  The S1P₁-conditional mutant embryos showed an enlarged pericardial cavity, undeveloped and rounded limbs, and spots of bleeding along the body. Yolk sacs from S1P₁-conditional mutant embryos were edematous and with less blood in vessels than in control embryos. At E13.5, the appearance of the S1P₁-conditional embryos worsened with massive bleeding (n = 9) (Figure 2D). All S1P₁-conditional mutant embryos examined were without a heartbeat at E14.5 (n = 13). Thus, these conditional mutant embryos died by E14.5 similar to the S1P₁^Ko/Ko embryos,¹⁵  which died between E12.5 and E14.5.

Maturation of vessels in S1P₁-conditional mutant embryos was studied using smooth muscle (SM) α-actin as a marker for VSMCs¹⁵  (Figure 2E). In transverse sections of aortae from control embryos, VSMCs were found to completely surround the vessels. In contrast, the aortae in S1P₁-conditional knock-out embryos were incompletely covered by VSMCs. Cells positive for SM α-actin were found clustered on the ventral side of the vessel. The endothelial cell layer on the uncovered dorsal side of aortae appeared discontinuous, possibly because of the lack of VSMC support.²³  Similarly, vessels in the brains of the S1P₁-conditional mutant embryos were incompletely surrounded by VSMCs. These vessel abnormalities in the S1P₁-conditional mutants resembled those described for the global S1P₁ knock-out embryos.¹⁵ 

Because the S1P₁ receptor is expressed in both ECs and VSMCs, as well as in other cells types, it had been uncertain whether S1P₁ functioned in ECs during vascular maturation. The results shown here indicate that vessel coverage by VSMCs is indeed directed by the activity of the S1P₁ receptor in ECs. In addition, we have found that S1P₁^loxP/loxP mice carrying a Cre transgene under the control of the smooth muscle myosin heavy chain (SM-MHC) promoter²⁴  were viable as adults with an S1P₁ gene deletion in smooth muscle. These smooth muscle-specific S1P₁ KO mice did not show any evidence of hemorrhage or vascular maturation defects (M.L.A., G.K. Owens, R.L.P., unpublished results, 2003). These data suggest that S1P₁ receptor expression in VSMCs may not be essential for the process of vessel coverage during development. However, because the expression of Cre recombinase driven by the SM-MHC promoter begins at about the time that VSMCs start to migrate around the vessels, we cannot be certain that the S1P₁ deletion was timely enough to affect significantly the vascular maturation process. In any case, the evidence indicates that the expression of S1P₁ receptor within ECs has an essential function to regulate the coverage of vessels by VSMCs. At this time an additional function of S1P₁ within VSMCs during this process cannot be totally ruled out.

The promotion of functional EC-VSMC interactions by the S1P₁ receptor may be incorporated into strategies for therapeutic angiogenesis: the formation of new blood vessels into ischemic tissues.^25,26  Endothelial growth factors such as vascular endothelial growth factor (VEGF) promote nascent vessel formation; however, these vessels are leaky and will regress unless stabilized by VSMC interactions.²⁷  Highly effective angiogeneic therapies may ultimately require stimulation of multiple signaling pathways, including the S1P₁ receptor to produce mature, stabile vessels.

We thank Cuiling Li and Chu-Xia Deng for establishing the chimeric mice, Masashi Yanagisawa for the Tie2-Cre mice, and Suzanne Mandala for the antibody to S1P₁.