Stable Knock-Down of the Sphingosine 1-Phosphate Receptor S1P1 Influences Multiple Functions of Human Endothelial Cells
http://www.100md.com
动脉硬化血栓血管生物学 2005年第3期
From the Institute for Prevention of Cardiovascular Diseases (V.K.-K., T.R., W.A., C.V., W.S.), Ludwig Maximilian University, Munich, Germany; the Department of Physiology (S.Y., N.M., G.T.), University of Tennessee Health Science Center, Memphis; the Institute of Medical Microbiology, Immunology and Hygiene (J.M.), Technical University, Munich, Germany; and the Institute of Pathology (C.J.D.), Johannes Gutenberg University, Mainz, Germany.
Correspondence to Dr Vera Krump-Konvalinkova, Institute for Prevention of Cardiovascular Diseases, Universit?t München, Pettenkoferstr 9, D 80336 München, Germany. E-mail vera.krump-konvalinkova@klp.med.uni-muenchen.de
Abstract
Objectives— Sphingosine 1-phosphate (S1P) is a bioactive phospholipid acting both as a ligand for the G protein–coupled receptors S1P1-5 and as a second messenger. Because S1P1 knockout is lethal in the transgenic mouse, an alternative approach to study the function of S1P1 in endothelial cells is needed.
Methods and Results— All human endothelial cells analyzed expressed abundant S1P1 transcripts. We permanently silenced (by RNA interference) the expression of S1P1 in the human endothelial cell lines AS-M.5 and ISO-HAS.1. The S1P1 knock-down cells manifested a distinct morphology and showed neither actin ruffles in response to S1P nor an angiogenic reaction. In addition, these cells were more sensitive to oxidant stress–mediated injury. New S1P1-dependent gene targets were identified in human endothelial cells. S1P1 silencing decreased the expression of platelet–endothelial cell adhesion molecule-1 and VE-cadherin and abolished the induction of E-selectin after cell stimulation with lipopolysaccharide or tumor necrosis factor-. Microarray analysis revealed downregulation of further endothelial specific transcripts after S1P1 silencing.
Conclusions— Long-term silencing of S1P1 enabled us for the first time to demonstrate the involvement of S1P1 in key functions of endothelial cells and to identify new S1P1-dependent gene targets.
Stable silencing by RNA interference of S1P1 expression in human endothelial cell lines demonstrates the involvement of S1P1 in key functions of human endothelial cells and identifies new S1P1-dependent gene targets.
Key Words: S1P1 ? functional analysis ? siRNA ? permanent inhibition ? endothelial cells
Introduction
The bioactive sphingolipid metabolite, sphingosine 1-phosphate (S1P), is an important component of serum that is released primarily from activated platelets. S1P is a multifunctional physiologic mediator implicated in the regulation of a broad spectrum of biologic processes, including proliferation, survival, regulation of cytoskeletal reorganization, motility, and differentiation in many cell types.1–5
The response of cells to S1P has been shown to be mediated predominantly by G protein–coupled receptors. Five receptors encoded by the endothelial differentiation gene family that bind S1P with high affinity have been described.6–10 The receptors couple to multiple G proteins that activate different intracellular signaling pathways. Several of these receptors are simultaneously coexpressed on the same cell.
In addition to acting as a ligand of cell surface receptors, S1P can function as a second messenger.3,4,11 A variety of external stimuli, particularly growth factors and chemoattractants, as well as lysophosphatidic acid and S1P, have been reported to strongly stimulate S1P kinase (SphK) to generate intracellular S1P, which can mobilize calcium from internal stores,12 regulate cell survival by activating the transcription factor nuclear factor-B,13 and control cell proliferation by mediating Ras and extracellular signal–regulated kinase 1/2 activation in cells stimulated with vascular endothelial growth factor (VEGF).14
Stimulation of the S1P1 receptor (previously known as endothelial differentiation gene-1, edg-1) activates a Gi-linked pathway leading to cell growth, survival, and migration.15,16 The signaling pathways mediating these responses include activation of the Ras and Rho GTPases, which direct mitogenesis and cytoskeletal remodeling, respectively. S1P1 receptor–knockout mice are not viable and die in utero because of defects in vascular maturation.17 Coordinate endothelial expression of S1P1, S1P2, and S1P3 is essential for embryonic development.18,19
The cytoskeletal and morphogenic response of human umbilical vein endothelial cells (HUVECs) to S1P has been evaluated with the use of antisense oligonucleotides directed against S1P1 and S1P3.20–22 Receptor expression inhibition studies indicated that S1P1 couples to Rac, regulating cortical actin assembly, whereas S1P3 couples to Rho, involved in stress fiber formation.20 However, measurements of Rac and Rho activity by biochemical assays indicated that both receptors are capable of activating both Rac and Rho pathways involved in migration and morphogenic differentiation of endothelial cells into capillary-like structures.21 Both receptors mediate the activation of membrane type 1-matrix metalloproteinase by S1P.22
To analyze the function of S1P1 in human endothelial cells, we generated human endothelial cell lines with a permanently reduced expression of S1P1 by introducing plasmids encoding small interfering RNA (siRNA) targeted to S1P1 into the cell lines AS-M.523 and ISO-HAS.1.24,25 The use of established human endothelial cell lines enabled us to study the long-term functional consequences of S1P1 silencing.
Methods
Please see the expanded online Methods section at http://atvb.ahajournals.org.
Cell Culture
HUVECs were isolated as described26 and propagated in ECG medium (PromoCell). The human cell lines derived from human pulmonary microvascular endothelial cells (HPMEC-ST1.6R)27 and from patients with angiosarcoma (AS-M.523 and ISO-HAS.124,25) were cultured in ECG medium-MV (PromoCell).
Cloning of siS1P1 Sequences
The S1P1-specific siRNA expression vectors were generated by cloning the sequences encoding the hairpin siRNA targeted to S1P1 into the siRNA expression vector pSilencer 1.0-U6 (Ambion, Austin, Tex). The targeted sequences were as follows: #1, 5'-GGAGATGCGTCGGGCCTTC-3'; #2, 5'-CTGCATCAGTGCGCTGTCC-3'; and #3, TGATCGATCATCTATAGCA-3', located at bp 1196 to 1214, 797 to 816, and 1940 to 1958, respectively (NCBI accession number BC018650). The control plasmid was pSilencer negative control (Silco) that encodes siRNA having no significant sequence similarity to human gene sequences (Ambion). The secondary structure was determined by using the Vienna RNA secondary structure prediction program.28
Generation of S1P1–Knock-Down Endothelial Cell Lines
Plasmids encoding siRNA targeted to the S1P1 transcript were introduced into endothelial cells, and stably transfected cells were selected as described at www.ahajournals.org.
Results
Expression Profile of S1P Receptors in Cultured HUVECs
Transcripts encoding S1P1 were abundant in all cultures of endothelial cells tested. Transcripts encoding S1P3 were detected by reverse transcription–polymerase chain reaction (RT-PCR) in all but ISO-HAS.1 cells (Figure 1). By quantitative RT-PCR, the S1P3 transcripts were less abundant than S1P1 transcripts, representing 4.3%, 3.5%, and 0.75% of S1P1 expression levels in HUVEC, AS-M.5, and ISO-HAS.1, respectively. A very weak expression of S1P4 was observed in HUVECs, ISO-HAS, and a subclone of HPMEC-ST1.6R, designated HPMEC-ST1.6R.D. Transcripts encoding either S1P2 or S1P5 were not detected (Figure 1).
Figure 1. Expression of S1P receptors in human endothelial cells. Total RNA from the endothelial cell lines ISO-HAS.1,24,25 HPMEC.ST1.6R,27 subclones D and S, AS-M.5,23 and HUVECs was analyzed for expression of S1P1, S1P2, S1P3, S1P4, S1P5, and ?-actin. RT-PCR products were separated on agarose gel and stained with ethidium bromide. Abbreviations are as defined in text.
siRNA Efficiently Reduces the Concentration of S1P1 Transcripts
To gain insight into the biologic function of S1P1 in human endothelial cells, we introduced into the cell lines AS-M.5 and ISO-HAS.1 plasmids encoding siRNA targeted to 3 different sequences in the S1P1 transcript (Sil#1, Sil#2, Sil#3) and isolated stably transfected cells. We measured by quantitative RT-PCR the levels of S1P1 transcripts in the transfected cells and compared these with unmanipulated cells and with cells stably transfected with a control plasmid (Silco; Figure 2). The levels of S1P1 transcripts differed substantially between the different cells. In the AS-M.5 cells transfected with Sil#1 (AS-M.5-Sil#1), S1P1 transcripts were repeatedly reduced to 20% of levels detected in either AS-M.5 or AS-M.5-Silco cells (Figure 2). We isolated single cell–derived clones of AS-M.5-Sil#1 and analyzed them for expression of S1P1. Individual clones (cl.) differed in the levels of S1P1 expression, reaching average values of 15% to 70% of S1P1 levels detected in the control pool of cells (Figure 2). The low S1P1 expression level remained stable for at least 6 months in a selected clone, AS-M.5-Sil#1, T2, cl.3. In contrast, expression of S1P1 in the cells containing the construct Sil#2 and Sil#3 was reduced to only 60% of control or not affected, respectively (Figure 2).
Figure 2. Knockdown of S1P1 by RNA interference. AS-M.5 (gray) and ISO-HAS.1 (dark gray) cells were transfected with pSilencer negative control vector (Silco) and pSilencer encoding siRNA targeted to sequence #1 and #2 in the S1P1 transcript. Two independent transfections (T1, T2) were performed. Pools of transfected cells (Silco, Sil#1, and Sil#2) as well as single cell–derived clones (cl.) isolated from the pools of transfected cells were analyzed for S1P1 expression by real-time RT-PCR. Values (% of control) are mean±SEM. All other abbreviations are as defined in text.
The 3 pSil-derived plasmids encoding siRNA sequences targeted to S1P1 transcript also were stably introduced into the ISO-HAS.1 cell line. In this cell line, only the pSil#1 construct was efficient in reducing the expression of S1P1 (Figure 2, dark gray bars). S1P1 expression in single cell–derived clones of these cells was stably reduced by 50% to 80%.
The low expression of S1P3 transcripts, as measured by quantitative RT-PCR, was unchanged in the clones of S1P1-knockdown AS-M.5 clones (T2, Sil#1, cl.2, 3, and 6) and also not affected by transfection of ISO-HAS.1 with Sil#1 plasmid (not shown), suggesting that S1P1 downregulation had no effect on expression of S1P3.
Cells With Reduced S1P1 Expression Display Altered Morphology and Have Altered Proliferation Characteristics
The S1P1-knockdown AS-M.5 cells exhibited a distinctive morphology (Figure Ia, A and B, available online at http://atvb.ahajournals.org) compared with either the nonmanipulated cells (Figure Ia, D) or the cells transfected with Silco (Figure Ia, C). S1P1-knockdown cells contained vesicles on their surface, a feature that was more prominent in low-density cultures (Figure Ia, B). These profound changes in cellular morphology were observed in cultures of AS-M.5 cells transfected with the construct Sil#1 before the clonal selection in 2 independent experiments. After single-cell cloning, 3 single cell–derived clones with very low S1P1 expression levels (T2, cl.2, 3, and 6) displayed this morphology, suggesting that this morphology may be correlated with the low expression of S1P1.
In low-density cultures, AS-M.5-pSil#1 cells grew considerably slower than did the control AS-M.5 cells (not shown). However, no major differences in growth characteristics were observed in high-density cultures, suggesting that the AS-M-Sil#1 cells may depend on substances released by neighboring cells available in high-density cultures.
S1P1-Knockdown Cells Do Not Show S1P-Induced Reorganization of Actin
S1P-induced remodeling of the actin cytoskeleton is a characteristic response in many cell types. To verify the role of S1P1 in this reaction, we evaluated cytoskeletal responses in serum-starved cells. Confluent monolayers of both AS-M.5 and AS-M.5-Silco cells showed F-actin fibers distributed throughout the cytoplasm (Figure Ib, A and C). Treatment with S1P for 2 minutes caused spectacular reorganization of the actin cytoskeleton: fibers of F-actin were concentrated at cell-to-cell contacts in confluent cultures (Figure Ib, D and F) and into a subcortical network in subconfluent cultures (not shown). In contrast, neither actin fibers nor a response to S1P treatment was observed in AS-M.5-Sil#1 cells (Figure Ib, B and E). These observations demonstrate that the S1P1 receptor is essential for the reorganization of the actin cytoskeleton induced by S1P in human endothelial cells.
S1P1 Is Indispensable for the Angiogenic Response of Endothelial Cells
To evaluate the effect of reduced expression of S1P1 on the angiogenic reaction, we examined endothelial morphogenesis into capillary-like networks. Capillary formation was evaluated after 6 (Figure 3A through 3C) and 24 (Figure 3D through 3F) hours. The AS-M.5 (Figure 3A and 3D) and control AS-M.5-Silco (Figure 3C and 3F) cells showed the characteristic formation of capillary structures. In contrast, no angiogenic reaction was observed in AS-M.5-Sil#1 cultures (Figure 3B and 3E). Therefore, S1P1 is required for the angiogenic reaction of endothelial cells. A similar angiogenic reaction was observed in ISO-HAS.1 cells expressing S1P1 but very little S1P3 (not shown).
Figure 3. S1P1-knockdown endothelial cells show no angiogenic response. AS-M.5 (A, D), AS-M.5 S1P1 knockdown cells (AS-M.5-Sil#1; B, E), and AS-M.5 cells containing the control plasmid (AS-M.5-Silco; C, F) were plated on Matrigel and incubated for 6 (A–C) or 24 (D–F) hours. Phase-contrast micrographs of capillary structures. Bar represents 100 μm. Abbreviations are as defined in text.
Responses to Oxidative Stress in S1P1-Knockdown Cells
The response to oxidative stress induced by 4-hydroxynonenal (4-HNE; 10 to 20 μmol/L, 6 hours, 37°C) was evaluated by quantifying the ratio of necrotic to apoptotic cells. The percentage of both apoptotic and necrotic cells increased dependent on the concentration of 4-HNE and reached 10% in ISO-HAS.1 (not shown) and AS-M.5 cells. Whereas apoptosis was the same in control and AS-M.5-sil#1 cells, the proportion of necrotic cells in AS-M.5-Sil#1 cultures increased dramatically after 4-HNE treatment, attaining 30% (Figure II, available online at http://atvb.ahajournals.org).
Reduced S1P1 Expression Abolishes the Expression of Cell Adhesion Molecules
S1P1 silencing led to a reduction in expression of both VE-cadherin and platelet–endothelial cell adhesion molecule-1 (PECAM-1). The degree of S1P1 knockdown was correlated with the expression levels of these cell surface proteins in both AS-M.5 and ISO-HAS.1 cells (Figure 4a). The typical cell surface expression pattern of VE-cadherin and PECAM-1 was observed in AS-M.5 cells (Figure 4b, A and D) and in HUVECs (Figure 4b, C and F). In contrast, no expression of VE-cadherin or PECAM-1 could be detected at cell-to-cell contacts in monolayers of AS-M.5-Sil#1 cells (Figure 4b, B and E). Gene expression analyses by microarrays showed downregulation of transcripts encoding PECAM-1 and VE-cadherin in 2 different clones of S1P1-silenced AS-M.5 cells (Table I, available online at http://atvb.ahajournals.org). These observations indicate that S1P1 is required for gene transcription and expression of VE-cadherin and PECAM-1 in human endothelial cells.
Figure 4. S1P1 knockdown alters the expression of adhesion molecules on endothelial cells. A, Degree of downregulation of S1P1 transcripts is correlated with reduced expression levels of PECAM-1 and VE-cadherin in human endothelial cells. Cell lines ISO-HAS.1 (I-H) and AS-M.5 (AS) were transfected with plasmids encoding siRNA control sequences (Silco) or siRNA targeted to different portions of S1P1 transcript (Sil#1, Sil#2, Sil#3). Cell surface expression of PECAM-1 (light gray), VE-cadherin (dark gray), and levels of S1P1 transcripts (black) were analyzed by fluorescence-activated cell sorter and quantitative RT-PCR, respectively, in both whole populations of transfected cells (Silco, Sil#2, Sil#3) and single cell–derived clones (I-H, Sil#1,cl.19 and cl.5; AS, Sil#1, cl.3). The values are mean±SEM Abbreviations are as defined in text. B, No expression of VE-cadherin or PECAM-1 in S1P1-knockdown endothelial cells. AS-M.5 (A, D), HUVECs (C, F), and AS-M.5-Sil#1 (B, E) were grown to confluence. VE-cadherin (A–C) and PECAM-1 (D–F) were visualized by immunofluorescence microscopy. Bar represents 100 μm. B, ICAM-1 and E-selectin expression in response to proinflammatory stimuli is reduced in S1P1-knockdown human endothelial cells. AS-M.5-Sil#1 and control AS-M.5-Silco were treated with TNF- (100 ng/mL, 5 hours, 37°C; gray bars) or LPS (1 μg/mL, 5 hours, 37°C; black bars). The expression of ICAM-1 and E-selectin was analyzed by flow cytometry. Specific mean fluorescence intensity was expressed as a percentage of the TNF-–induced stimulation in control (Silco) cells. Values are mean±SEM. Abbreviations are as defined in text.
S1P1 Modulates the Endothelial Cell Response to Proinflammatory Stimuli
Similar to freshly isolated HUVECs that respond to proinflammatory stimuli by the transcriptional upregulation of cell adhesion molecules, both AS-M.5 (not shown) and AS-M.5-Silco cells responded to stimulation by tumor necrosis factor- (TNF-) (100 ng/mL, 6 hours) and lipopolysaccharide (LPS; 1 μg/mL, 6 hours) by a transient induction of transcription of E-selectin and intercellular adhesion molecule-1 (ICAM-1) on the cell surface (Figure 4c). Albeit diminished in comparison with the control cells, induction of ICAM-1 expression in response to TNF- and LPS was clearly detectable in the knockdown AS-M.5 cells. However, in contrast to either the AS-M.5 or AS-M.5-Silco cells, no induction of E-selectin production was observed in AS-M.5-Sil#1 (Figure 4c). Analysis of E-selectin transcripts by real-time RT-PCR showed that E-selectin transcripts were highly reduced in the LPS-stimulated, S1P1-knockdown cells compared with the control cells, suggesting that E-selectin expression was regulated at the transcriptional level (not shown). Thus, we conclude that S1P1 is involved in the regulation of the inflammatory response of endothelial cells.
Microarray Analysis of Genes Expressed in S1P1-Silenced Endothelial Cells
Searching for gene targets that are dependent on S1P1 expression, we compared gene expression patterns of 2 clones of S1P1-knockdown AS-M.5 cells (T2, cl.2 and cl.3) with those of AS-M.5-Silco and AS-M.5. Among the 22 238 gene probes that the microarrays contained, we identified 82 genes with increased expression and 220 downregulated transcripts in the S1P1-knockdown cells (2-fold decrease and increase; error probability <5%). Many endothelial specific transcripts were downregulated (Table II, available online at http://atvb.ahajournals.org), but the expression of other genes was not affected in the S1P1-knockdown AS-M.5 cells. The latter genes include angiotensin II receptor (Hs.405348), endothelial lipase (Hs.65370), endothelin-2 (Hs.1407), nitric oxide synthetase-3 (NOS3, Hs.446303), P-selectin (Hs, 73800), the VEGF receptor FLT1 (Hs.347713), and fms-related tyrosine kinase FLT3 (Hs.385), vimentin (Hs.43800), and vitronectin (Hs.2257; not shown).
Discussion
The analysis of expression profiles of S1P receptors showed that all human endothelial cells tested (AS-M.5,23 ISO-HAS.1,24,25 HPMEC-ST1,27 and HUVECs) expressed transcripts encoding S1P1, suggesting that the S1P1 receptor may have a ubiquitous function in endothelial cells. All human endothelial cells tested also contained transcripts encoding S1P3, at a much lower level, however, as measured by quantitative RT-PCR. Expression of both S1P1 and S1P3 characterizes the typical expression pattern of S1P receptors in cultured human endothelial cells.20
To examine the function of S1P1 in human endothelial cells, we permanently silenced by RNA interference the expression of S1P1 in the established endothelial cell lines AS-M.5 and ISO-HAS.1, previously shown to possess many characteristics of freshly isolated endothelial cells.23–25 The siRNA targeted to a large loop in the secondary structure of the S1P1 transcript (Sil#1) was more efficient in silencing than were the less exposed sequences (Sil#2, Sil#3) in the S1P1 transcripts. A similar observation was reported for silencing of the insulin-like growth factor receptor,29 suggesting that the secondary structure of the targeted transcripts may be important for efficient silencing by siRNA. The reduction of S1P1 expression in the cells transfected with plasmids Sil#1 did not affect the already very low level of S1P3 expression and was stable for at least 6 months of continuous culturing.
Silencing of S1P1 transcripts to <20% of control profoundly altered the phenotype of endothelial cells. The S1P1-knockdown endothelial cells (AS-M.5-Sil#1) manifested an altered morphology and proliferated slowly in low-density cultures. In contrast to the control cells showing spectacular reorganization of the actin cytoskeleton in response to S1P, the S1P1-knockdown AS-M.5 cells showed no reorganization of actin after S1P treatment. Actin reorganization is an immediate response triggered by S1P.20,30 It has been reported to be regulated through balance between Rho and Rac activity in HUVECs21,31 and to be correlated with an enhancement of endothelial cell barrier function.32 Our results demonstrate that the S1P1 receptor is required for S1P-induced actin reorganization in human endothelial cells.
The S1P1-knockdown cells exhibited altered angiogenic properties. No formation of typical capillary-like networks of differentiated endothelial cells on Matrigel could be observed with AS-M.5-Sil#1, consistent with the reports on the involvement of S1P1 in angiogenesis.33 Both the lack of angiogenic response of AS-M.5-Sil#1 and the capacity to form capillary-like structures34 of ISO-HAS.1 expressing S1P1 but almost no S1P3 suggest that S1P1 but not S1P3 is essential for endothelial cell morphogenesis. However, we cannot exclude a cooperative effect of S1P1 and S1P3 previously observed in short-term inhibition studies with antisense oligonucleotides.20 In a recent study, downregulation of S1P1 expression effectively suppressed tumor angiogenesis in a murine model in vivo.35
We observed an increased proportion of both apoptotic and necrotic cells in both control and AS-M.5-Sil#1 cells after treatment with 4-HNE, a mediator of oxidative stress36 shown to induce apoptotic death in HUVECs.37 Whereas no major difference between control and AS-M.5-Sil#1 cells was observed in the proportion of apoptotic cells, the proportion of necrotic cells in AS-M.5-Sil#1 increased significantly. S1P has been previously reported to activate SphK, which results in the intracellular production of S1P and has an antiapoptotic effect in endothelial cells.38,39 However, our observations indicate that S1P1 is not involved in the antiapoptotic activity of S1P but may rather protect endothelial cells from necrosis.
Long-term silencing of S1P1 enabled us to identify new S1P1-dependent gene targets in endothelial cells. Downregulation of S1P1 led to a reduced expression of the adhesion molecules PECAM-1, expressed predominantly at the endothelial cell contacts,40 and VE-cadherin, usually expressed at intercellular junctions.41 The degree of S1P1 knockdown was correlated with the extent of suppression of PECAM-1 and VE-cadherin. This suggests that the expression of these surface molecules is controlled by S1P1. VE-cadherin expression was previously reported to be reduced after injection of antisense oligonucleotides directed against S1P1 into HUVECs.20 Because functional VE-cadherin was demonstrated to be required for endothelial cell morphogenesis42 and PECAM-1–null endothelial cells failed to migrate in response to S1P,43 the lack of expression of both VE-cadherin and PECAM-1 in S1P1 knockdown endothelial cells may impede their morphogenetic reaction on Matrigel.
The S1P1-knockdown endothelial cells differed from the control cells also in their response to proinflammatory stimuli. Silencing of S1P1 had no major effect on LPS-induced expression of ICAM-1 but reduced the TNF-–stimulated expression of ICAM-1 by 60%. No expression of E-selectin in response to stimulation with either LPS or TNF- was observed in the S1P1-knockdown endothelial cells, implicating involvement of S1P1 in the regulation of E-selectin expression. Several lines of evidence suggest a possible participation of S1P in the inflammatory reaction of endothelial cells. TNF- was found to induce rapid activation of SphK and subsequent generation of intracellular S1P that resulted in strong stimulation of E-selectin expression in HUVECs.44 The defective control of E-selectin expression in the S1P1 knockdown cells suggests that a functional S1P1 receptor is required to mediate expression of E-selectin in stimulated endothelial cells and implies that the intracellularly produced S1P does not function as a second messenger but rather is released from the cells and then activates S1P1. S1P treatment of confluent HUVEC cultures was reported to induce the expression of E-selectin to levels comparable to those induced with TNF- in 144 but not in another45 study, Also, we found only negligible induction of E-selectin expression after S1P treatment of both HUVEC and AS-M.5 cultures (not shown). The latter observations suggest that a synergistic interaction of the signal transduction pathways induced by TNF- and S1P1 receptor activation is required for robust E-selectin expression observed after treatment of endothelial cells with TNF-.
Further S1P1-dependent gene targets were identified by microarray analysis of gene expression. Many endothelial specific transcripts, such as the VEGF-receptor KDR, were downregulated in S1P1-silenced endothelial cells, but other endothelial specific genes, such as the VEGF-receptor Flt-1 and the fms-related kinase Flt-3, were not affected.
In conclusion, our results demonstrate that long-term silencing of S1P1 profoundly alters gene expression and key functions of cultured human endothelial cells. This observation, if confirmed in vivo, could have important therapeutic consequences for chronic inhibition of the S1P1 receptor with receptor antagonists or low concentrations of agonists leading to S1P1 downregulation. The immunosuppressive drug FTY720, an agonist of S1P-receptors currently used in clinical trials, at therapeutic concentrations has been recently reported to block S1P signaling through downregulation of S1P1,2,5.46 Our results imply that FTY720 could be harmful if it induced downregulation of S1P1 on vascular endothelium in vivo.
Acknowledgments
The study was supported by the August-Lenz-Stiftung and the Deutsche Forschungsgemeinschaft (graduate program Vascular Biology in Medicine GRK 438: T.R. and C.V., SFB 413 and Si 247/9). We thank Dr Reinhard Hoffmann for his help with the microarray analyses and Nicole Wilke for technical assistance.
References
Hla T, Lee MJ, Ancellin N, Paik JH, Kluk MJ. Lysophospholipids–receptor revelations. Science. 2001; 294: 1875–1878.
Kluk MJ, Hla T. Signaling of sphingosine-1-phosphate via the S1P/EDG-family of G-protein-coupled receptors. Biochim Biophys Acta. 2002; 1582: 72–80.
Spiegel S, Milstien S. Sphingosine-1-phosphate: an enigmatic signalling lipid. Nat Rev Mol Cell Biol. 2003; 4: 397–407.
Spiegel S, Milstien S. Sphingosine 1-phosphate, a key cell signaling molecule. J Biol Chem. 2002; 29: 25851–25854.
Watterson K, Sankala H, Milstien S, Spiegel S. Pleiotropic actions of sphingosine-1-phosphate. Prog Lipid Res. 2003; 42: 344–357.
Hla T, Maciag T. An abundant transcript induced in differentiating human endothelial cells encodes a polypeptide with structural similarities to G-protein-coupled receptors. J Biol Chem. 1990; 265: 9308–9313.
Okazaki H, Ishizaka N, Sakurai T, Kurokawa K, Goto K, Kumada M, Takuwa Y. Molecular cloning of a novel putative G protein-coupled receptor expressed in the cardiovascular system. Biochem Biophys Res Commun. 1993; 190: 1104–1109.
Yamaguchi F, Tokuda M, Hatase O, Brenner S. Molecular cloning of the novel human G protein-coupled receptor (GPCR) gene mapped on chromosome 9. Biochem Biophys Res Commun. 1996; 227: 608–614.
Graler MH, Bernhardt G, Lipp M. EDG6, a novel G-protein-coupled receptor related to receptors for bioactive lysophospholipids, is specifically expressed in lymphoid tissue. Genomics. 1998; 53: 164–169.
Im DS, Heise CE, Ancellin N, O’Dowd BF, Shei GJ, Heavens RP, Rigby MR, Hla T, Mandala S, McAllister G, George SR, Lynch KR. Characterization of a novel sphingosine 1-phosphate receptor, Edg-8. J Biol Chem. 2000; 275: 14281–14286.
Olivera A, Spiegel S. Sphingosine-1-phosphate as second messenger in cell proliferation induced by PDGF and FCS mitogens. Nature. 1993; 365: 557–560.
Choi OH, Kim JH, Kinet JP. Calcium mobilization via sphingosine kinase in signalling by the Fc epsilon RI antigen receptor. Nature. 1996; 380: 634–636.
Xia P, Wang L, Moretti PA, Albanese N, Chai F, Pitson SM, D’Andrea RJ, Gamble JR, Vadas MA. Sphingosine kinase interacts with TRAF2 and dissects tumor necrosis factor- signaling. J Biol Chem. 2002; 277: 7996–8003.
Shu X, Wu W, Mosteller RD, Broek D. Sphingosine kinase mediates vascular endothelial growth factor-induced activation of ras and mitogen-activated protein kinases. Mol Cell Biol. 2002; 22: 7758–7768.
Spiegel S, Milstien S. Sphingosine-1-phosphate: signaling inside and out. FEBS Lett. 2000; 476: 55–57.
Hla T. Sphingosine 1-phosphate receptors. Prostaglandins Other Lipid Mediat. 2001; 64: 135–142.
Liu Y, Wada R, Yamashita T, Mi Y, Deng CX, Hobson JP, Rosenfeldt HM, Nava VE, Chae SS, Lee MJ, Liu CH, Hla T, Spiegel S, Proia RL. Edg-1, the G protein-coupled receptor for sphingosine-1-phosphate, is essential for vascular maturation. J Clin Invest. 2000; 106: 951–961.
Kono M, Mi Y, Liu Y, Sasaki T, Allende ML, Wu YP, Yamashita T, Proia RL. S1P1, S1P2 and S1P3 receptors coordinately function during embryonic angiogenesis. J Biol Chem. 2004; 279: 29367–29373.
Allende ML, Yamashita T, Proia RL. G-protein-coupled receptor S1P1 acts within endothelial cells to regulate vascular maturation. Blood. 2003; 102: 3665–3667.
Lee MJ, Thangada S, Claffey KP, Ancellin N, Liu CH, Kluk M, Volpi M, Sha’afi RI, Hla T. Vascular endothelial cell adherens junction assembly and morphogenesis induced by sphingosine-1-phosphate. Cell. 1999; 99: 301–312.
Paik JH, Chae S, Lee MJ, Thangada S, Hla T. Sphingosine 1-phosphate-induced endothelial cell migration requires the expression of EDG-1 and EDG-3 receptors and Rho-dependent activation of v3- and 1-containing integrins. J Biol Chem. 2001; 276: 11830–11837.
Langlois S, Gingras D, Beliveau R. Membrane type 1-matrix metalloproteinase (MT1-MMP) cooperates with sphingosine 1-phosphate to induce endothelial cell migration and morphogenic differentiation. Blood. 2004; 103: 3020–3028.
Krump-Konvalinkova V, Bittinger F, Olert J, Br?uninger W, Brunner J, Kirkpatrick CJ. Establishment and characterization of an angiosarcoma-derived cell line(AS-M). Endothelium. 2003; 10: 319–328.
Masuzawa M, Fujimura T, Hamada Y, Fujita Y, Hara H, Nishiyama S, Katsuoka K, Tamauchi H, Sakurai Y. Establishment of a human hemangiosarcoma cell line (ISO-HAS). Int J Cancer. 1999; 81: 305–308.
Unger RE, Krump-Konvalinkova V, Peters K, Kirkpatrick CJ. In vitro expression of the endothelial phenotype: comparative study of primary isolated cells and cell lines, including the novel cell line HPMEC-ST1.6R. Microvasc Res. 2002; 64: 384–397.
Jaffe EA, Nachman RL, Becker CG, Minick CR. Culture of human endothelial cells derived from umbilical veins: identification by morphologic and immunologic criteria. J Clin Invest. 1973; 52: 2745–2756.
Krump-Konvalinkova V, Bittinger F, Unger RE, Peters K, Lehr HA, Kirkpatrick CJ. Generation of human pulmonary microvascular endothelial cell lines. Lab Invest. 2001; 81: 1717–1727.
Hofacker IL. Vienna RNA secondary structure server. Nucleic Acids Res. 2003; 31: 3429–3431.
Bohula EA, Salisbury AJ, Sohail M, Playford MP, Riedemann J, Southern EM, Macaulay VM. The efficacy of small interfering RNAs targeted to the type 1 insulin-like growth factor receptor (IGF1R) is influenced by secondary structure in the IGF1R transcript. J Biol Chem. 2003; 278: 15991–15997.
Yatomi Y, Ohmori T, Rile G, Kazama F, Okamoto H, Sano T, Satoh K, Kume S, Tigyi G, Igarashi Y, Ozaki Y. Sphingosine 1-phosphate as a major bioactive lysophospholipid that is released from platelets and interacts with endothelial cells. Blood. 2000; 96: 3431–3438.
Essler M, Retzer M, Ilchmann H, Linder S, Weber PC. Sphingosine 1-phosphate dynamically regulates myosin light chain phosphatase activity in human endothelial cells. Cell Signal. 2002; 14: 607–613.
Garcia JG, Liu F, Verin AD, Birukova A, Dechert MA, Gerthoffer WT, Bamberg JR, English D. Sphingosine 1-phosphate promotes endothelial cell barrier integrity by Edg-dependent cytoskeletal rearrangement. J Clin Invest. 2001; 108: 689–701.
English D, Brindley DN, Spiegel S, Garcia JG. Lipid mediators of angiogenesis and the signalling pathways they initiate. Biochim Biophys Acta. 2002; 1582: 228–239.
Krump-Konvalinkova V, Kleideiter E, Friedrich U, Klotz U, Kirkpatrick CJ. Tumorigenic conversion of endothelial cells. Exp Mol Pathol. 2003; 75: 154–159.
Chae SS, Paik JH, Furneaux H, Hla T. Requirement for sphingosine 1-phosphate receptor-1 in tumor angiogenesis demonstrated by in vivo RNA interference. J Clin Invest. 2004; 114: 1082–1089.
Uchida K. 4-Hydroxy-2-nonenal: a product and mediator of oxidative stress. Prog Lipid Res. 2003; 42: 318–343.
Herbst U, Toborek M, Kaiser S, Mattson MP, Hennig B. 4-Hydroxynonenal induces dysfunction and apoptosis of cultured endothelial cells. J Cell Physiol. 1999; 181: 295–303.
Xia P, Wang L, Gamble JR, Vadas MA. Activation of sphingosine kinase by tumor necrosis factor- inhibits apoptosis in human endothelial cells. J Biol Chem. 1999; 274: 34499–34505.
Hisano N, Yatomi Y, Satoh K, Akimoto S, Mitsumata M, Fujino MA, Ozaki Y. Induction and suppression of endothelial cell apoptosis by sphingolipids: a possible in vitro model for cell-cell interactions between platelets and endothelial cells. Blood. 1999; 93: 4293–4299.
Albelda SM, Oliver PD, Romer LH, Buck CA. EndoCAM: a novel endothelial cell-cell adhesion molecule. J Cell Biol. 1990; 110: 1227–1237.
Ayalon O, Sabanai H, Lampugnani MG, Dejana E, Geiger B. Spatial and temporal relationships between cadherins and PECAM-1 in cell-cell junctions of human endothelial cells. J Cell Biol. 1994; 126: 247–258.
Bach TL, Barsigian C, Chalupowicz DG, Busler D, Yaen CH, Grant DS, Martinez J. VE-Cadherin mediates endothelial cell capillary tube formation in fibrin and collagen gels. Exp Cell Res. 1998; 238: 324–334.
Gratzinger D, Canosa S, Engelhardt B, Madri JA. Platelet endothelial cell adhesion molecule-1 modulates endothelial cell motility through the small G-protein Rho. FASEB J. 2003; 17: 1458–1469.
Xia P, Gamble JR, Rye KA, Wang L, Hii CS, Cockerill P, Khew-Goodall Y, Bert AG, Barter PJ, Vadas MA. Tumor necrosis factor- induces adhesion molecule expression through the sphingosine kinase pathway. Proc Natl Acad Sci U S A. 1998; 95: 14196–14201.
Miura Y, Yatomi Y, Ohmori T, Osada M, Ozaki Y. Independence of tumor necrosis factor--induced adhesion molecule expression from sphingosine 1-phosphate signaling in vascular endothelial cells. J Thromb Haemost. 2004; 2: 1019–1021.
Gr?ler MH, Goetzl EJ. The immunosuppressant FTY720 down-regulates sphingosine 1-phosphate G-protein-coupled receptors. FASEB J. 2004; 18: 551–553.(Vera Krump-Konvalinkova; )
Correspondence to Dr Vera Krump-Konvalinkova, Institute for Prevention of Cardiovascular Diseases, Universit?t München, Pettenkoferstr 9, D 80336 München, Germany. E-mail vera.krump-konvalinkova@klp.med.uni-muenchen.de
Abstract
Objectives— Sphingosine 1-phosphate (S1P) is a bioactive phospholipid acting both as a ligand for the G protein–coupled receptors S1P1-5 and as a second messenger. Because S1P1 knockout is lethal in the transgenic mouse, an alternative approach to study the function of S1P1 in endothelial cells is needed.
Methods and Results— All human endothelial cells analyzed expressed abundant S1P1 transcripts. We permanently silenced (by RNA interference) the expression of S1P1 in the human endothelial cell lines AS-M.5 and ISO-HAS.1. The S1P1 knock-down cells manifested a distinct morphology and showed neither actin ruffles in response to S1P nor an angiogenic reaction. In addition, these cells were more sensitive to oxidant stress–mediated injury. New S1P1-dependent gene targets were identified in human endothelial cells. S1P1 silencing decreased the expression of platelet–endothelial cell adhesion molecule-1 and VE-cadherin and abolished the induction of E-selectin after cell stimulation with lipopolysaccharide or tumor necrosis factor-. Microarray analysis revealed downregulation of further endothelial specific transcripts after S1P1 silencing.
Conclusions— Long-term silencing of S1P1 enabled us for the first time to demonstrate the involvement of S1P1 in key functions of endothelial cells and to identify new S1P1-dependent gene targets.
Stable silencing by RNA interference of S1P1 expression in human endothelial cell lines demonstrates the involvement of S1P1 in key functions of human endothelial cells and identifies new S1P1-dependent gene targets.
Key Words: S1P1 ? functional analysis ? siRNA ? permanent inhibition ? endothelial cells
Introduction
The bioactive sphingolipid metabolite, sphingosine 1-phosphate (S1P), is an important component of serum that is released primarily from activated platelets. S1P is a multifunctional physiologic mediator implicated in the regulation of a broad spectrum of biologic processes, including proliferation, survival, regulation of cytoskeletal reorganization, motility, and differentiation in many cell types.1–5
The response of cells to S1P has been shown to be mediated predominantly by G protein–coupled receptors. Five receptors encoded by the endothelial differentiation gene family that bind S1P with high affinity have been described.6–10 The receptors couple to multiple G proteins that activate different intracellular signaling pathways. Several of these receptors are simultaneously coexpressed on the same cell.
In addition to acting as a ligand of cell surface receptors, S1P can function as a second messenger.3,4,11 A variety of external stimuli, particularly growth factors and chemoattractants, as well as lysophosphatidic acid and S1P, have been reported to strongly stimulate S1P kinase (SphK) to generate intracellular S1P, which can mobilize calcium from internal stores,12 regulate cell survival by activating the transcription factor nuclear factor-B,13 and control cell proliferation by mediating Ras and extracellular signal–regulated kinase 1/2 activation in cells stimulated with vascular endothelial growth factor (VEGF).14
Stimulation of the S1P1 receptor (previously known as endothelial differentiation gene-1, edg-1) activates a Gi-linked pathway leading to cell growth, survival, and migration.15,16 The signaling pathways mediating these responses include activation of the Ras and Rho GTPases, which direct mitogenesis and cytoskeletal remodeling, respectively. S1P1 receptor–knockout mice are not viable and die in utero because of defects in vascular maturation.17 Coordinate endothelial expression of S1P1, S1P2, and S1P3 is essential for embryonic development.18,19
The cytoskeletal and morphogenic response of human umbilical vein endothelial cells (HUVECs) to S1P has been evaluated with the use of antisense oligonucleotides directed against S1P1 and S1P3.20–22 Receptor expression inhibition studies indicated that S1P1 couples to Rac, regulating cortical actin assembly, whereas S1P3 couples to Rho, involved in stress fiber formation.20 However, measurements of Rac and Rho activity by biochemical assays indicated that both receptors are capable of activating both Rac and Rho pathways involved in migration and morphogenic differentiation of endothelial cells into capillary-like structures.21 Both receptors mediate the activation of membrane type 1-matrix metalloproteinase by S1P.22
To analyze the function of S1P1 in human endothelial cells, we generated human endothelial cell lines with a permanently reduced expression of S1P1 by introducing plasmids encoding small interfering RNA (siRNA) targeted to S1P1 into the cell lines AS-M.523 and ISO-HAS.1.24,25 The use of established human endothelial cell lines enabled us to study the long-term functional consequences of S1P1 silencing.
Methods
Please see the expanded online Methods section at http://atvb.ahajournals.org.
Cell Culture
HUVECs were isolated as described26 and propagated in ECG medium (PromoCell). The human cell lines derived from human pulmonary microvascular endothelial cells (HPMEC-ST1.6R)27 and from patients with angiosarcoma (AS-M.523 and ISO-HAS.124,25) were cultured in ECG medium-MV (PromoCell).
Cloning of siS1P1 Sequences
The S1P1-specific siRNA expression vectors were generated by cloning the sequences encoding the hairpin siRNA targeted to S1P1 into the siRNA expression vector pSilencer 1.0-U6 (Ambion, Austin, Tex). The targeted sequences were as follows: #1, 5'-GGAGATGCGTCGGGCCTTC-3'; #2, 5'-CTGCATCAGTGCGCTGTCC-3'; and #3, TGATCGATCATCTATAGCA-3', located at bp 1196 to 1214, 797 to 816, and 1940 to 1958, respectively (NCBI accession number BC018650). The control plasmid was pSilencer negative control (Silco) that encodes siRNA having no significant sequence similarity to human gene sequences (Ambion). The secondary structure was determined by using the Vienna RNA secondary structure prediction program.28
Generation of S1P1–Knock-Down Endothelial Cell Lines
Plasmids encoding siRNA targeted to the S1P1 transcript were introduced into endothelial cells, and stably transfected cells were selected as described at www.ahajournals.org.
Results
Expression Profile of S1P Receptors in Cultured HUVECs
Transcripts encoding S1P1 were abundant in all cultures of endothelial cells tested. Transcripts encoding S1P3 were detected by reverse transcription–polymerase chain reaction (RT-PCR) in all but ISO-HAS.1 cells (Figure 1). By quantitative RT-PCR, the S1P3 transcripts were less abundant than S1P1 transcripts, representing 4.3%, 3.5%, and 0.75% of S1P1 expression levels in HUVEC, AS-M.5, and ISO-HAS.1, respectively. A very weak expression of S1P4 was observed in HUVECs, ISO-HAS, and a subclone of HPMEC-ST1.6R, designated HPMEC-ST1.6R.D. Transcripts encoding either S1P2 or S1P5 were not detected (Figure 1).
Figure 1. Expression of S1P receptors in human endothelial cells. Total RNA from the endothelial cell lines ISO-HAS.1,24,25 HPMEC.ST1.6R,27 subclones D and S, AS-M.5,23 and HUVECs was analyzed for expression of S1P1, S1P2, S1P3, S1P4, S1P5, and ?-actin. RT-PCR products were separated on agarose gel and stained with ethidium bromide. Abbreviations are as defined in text.
siRNA Efficiently Reduces the Concentration of S1P1 Transcripts
To gain insight into the biologic function of S1P1 in human endothelial cells, we introduced into the cell lines AS-M.5 and ISO-HAS.1 plasmids encoding siRNA targeted to 3 different sequences in the S1P1 transcript (Sil#1, Sil#2, Sil#3) and isolated stably transfected cells. We measured by quantitative RT-PCR the levels of S1P1 transcripts in the transfected cells and compared these with unmanipulated cells and with cells stably transfected with a control plasmid (Silco; Figure 2). The levels of S1P1 transcripts differed substantially between the different cells. In the AS-M.5 cells transfected with Sil#1 (AS-M.5-Sil#1), S1P1 transcripts were repeatedly reduced to 20% of levels detected in either AS-M.5 or AS-M.5-Silco cells (Figure 2). We isolated single cell–derived clones of AS-M.5-Sil#1 and analyzed them for expression of S1P1. Individual clones (cl.) differed in the levels of S1P1 expression, reaching average values of 15% to 70% of S1P1 levels detected in the control pool of cells (Figure 2). The low S1P1 expression level remained stable for at least 6 months in a selected clone, AS-M.5-Sil#1, T2, cl.3. In contrast, expression of S1P1 in the cells containing the construct Sil#2 and Sil#3 was reduced to only 60% of control or not affected, respectively (Figure 2).
Figure 2. Knockdown of S1P1 by RNA interference. AS-M.5 (gray) and ISO-HAS.1 (dark gray) cells were transfected with pSilencer negative control vector (Silco) and pSilencer encoding siRNA targeted to sequence #1 and #2 in the S1P1 transcript. Two independent transfections (T1, T2) were performed. Pools of transfected cells (Silco, Sil#1, and Sil#2) as well as single cell–derived clones (cl.) isolated from the pools of transfected cells were analyzed for S1P1 expression by real-time RT-PCR. Values (% of control) are mean±SEM. All other abbreviations are as defined in text.
The 3 pSil-derived plasmids encoding siRNA sequences targeted to S1P1 transcript also were stably introduced into the ISO-HAS.1 cell line. In this cell line, only the pSil#1 construct was efficient in reducing the expression of S1P1 (Figure 2, dark gray bars). S1P1 expression in single cell–derived clones of these cells was stably reduced by 50% to 80%.
The low expression of S1P3 transcripts, as measured by quantitative RT-PCR, was unchanged in the clones of S1P1-knockdown AS-M.5 clones (T2, Sil#1, cl.2, 3, and 6) and also not affected by transfection of ISO-HAS.1 with Sil#1 plasmid (not shown), suggesting that S1P1 downregulation had no effect on expression of S1P3.
Cells With Reduced S1P1 Expression Display Altered Morphology and Have Altered Proliferation Characteristics
The S1P1-knockdown AS-M.5 cells exhibited a distinctive morphology (Figure Ia, A and B, available online at http://atvb.ahajournals.org) compared with either the nonmanipulated cells (Figure Ia, D) or the cells transfected with Silco (Figure Ia, C). S1P1-knockdown cells contained vesicles on their surface, a feature that was more prominent in low-density cultures (Figure Ia, B). These profound changes in cellular morphology were observed in cultures of AS-M.5 cells transfected with the construct Sil#1 before the clonal selection in 2 independent experiments. After single-cell cloning, 3 single cell–derived clones with very low S1P1 expression levels (T2, cl.2, 3, and 6) displayed this morphology, suggesting that this morphology may be correlated with the low expression of S1P1.
In low-density cultures, AS-M.5-pSil#1 cells grew considerably slower than did the control AS-M.5 cells (not shown). However, no major differences in growth characteristics were observed in high-density cultures, suggesting that the AS-M-Sil#1 cells may depend on substances released by neighboring cells available in high-density cultures.
S1P1-Knockdown Cells Do Not Show S1P-Induced Reorganization of Actin
S1P-induced remodeling of the actin cytoskeleton is a characteristic response in many cell types. To verify the role of S1P1 in this reaction, we evaluated cytoskeletal responses in serum-starved cells. Confluent monolayers of both AS-M.5 and AS-M.5-Silco cells showed F-actin fibers distributed throughout the cytoplasm (Figure Ib, A and C). Treatment with S1P for 2 minutes caused spectacular reorganization of the actin cytoskeleton: fibers of F-actin were concentrated at cell-to-cell contacts in confluent cultures (Figure Ib, D and F) and into a subcortical network in subconfluent cultures (not shown). In contrast, neither actin fibers nor a response to S1P treatment was observed in AS-M.5-Sil#1 cells (Figure Ib, B and E). These observations demonstrate that the S1P1 receptor is essential for the reorganization of the actin cytoskeleton induced by S1P in human endothelial cells.
S1P1 Is Indispensable for the Angiogenic Response of Endothelial Cells
To evaluate the effect of reduced expression of S1P1 on the angiogenic reaction, we examined endothelial morphogenesis into capillary-like networks. Capillary formation was evaluated after 6 (Figure 3A through 3C) and 24 (Figure 3D through 3F) hours. The AS-M.5 (Figure 3A and 3D) and control AS-M.5-Silco (Figure 3C and 3F) cells showed the characteristic formation of capillary structures. In contrast, no angiogenic reaction was observed in AS-M.5-Sil#1 cultures (Figure 3B and 3E). Therefore, S1P1 is required for the angiogenic reaction of endothelial cells. A similar angiogenic reaction was observed in ISO-HAS.1 cells expressing S1P1 but very little S1P3 (not shown).
Figure 3. S1P1-knockdown endothelial cells show no angiogenic response. AS-M.5 (A, D), AS-M.5 S1P1 knockdown cells (AS-M.5-Sil#1; B, E), and AS-M.5 cells containing the control plasmid (AS-M.5-Silco; C, F) were plated on Matrigel and incubated for 6 (A–C) or 24 (D–F) hours. Phase-contrast micrographs of capillary structures. Bar represents 100 μm. Abbreviations are as defined in text.
Responses to Oxidative Stress in S1P1-Knockdown Cells
The response to oxidative stress induced by 4-hydroxynonenal (4-HNE; 10 to 20 μmol/L, 6 hours, 37°C) was evaluated by quantifying the ratio of necrotic to apoptotic cells. The percentage of both apoptotic and necrotic cells increased dependent on the concentration of 4-HNE and reached 10% in ISO-HAS.1 (not shown) and AS-M.5 cells. Whereas apoptosis was the same in control and AS-M.5-sil#1 cells, the proportion of necrotic cells in AS-M.5-Sil#1 cultures increased dramatically after 4-HNE treatment, attaining 30% (Figure II, available online at http://atvb.ahajournals.org).
Reduced S1P1 Expression Abolishes the Expression of Cell Adhesion Molecules
S1P1 silencing led to a reduction in expression of both VE-cadherin and platelet–endothelial cell adhesion molecule-1 (PECAM-1). The degree of S1P1 knockdown was correlated with the expression levels of these cell surface proteins in both AS-M.5 and ISO-HAS.1 cells (Figure 4a). The typical cell surface expression pattern of VE-cadherin and PECAM-1 was observed in AS-M.5 cells (Figure 4b, A and D) and in HUVECs (Figure 4b, C and F). In contrast, no expression of VE-cadherin or PECAM-1 could be detected at cell-to-cell contacts in monolayers of AS-M.5-Sil#1 cells (Figure 4b, B and E). Gene expression analyses by microarrays showed downregulation of transcripts encoding PECAM-1 and VE-cadherin in 2 different clones of S1P1-silenced AS-M.5 cells (Table I, available online at http://atvb.ahajournals.org). These observations indicate that S1P1 is required for gene transcription and expression of VE-cadherin and PECAM-1 in human endothelial cells.
Figure 4. S1P1 knockdown alters the expression of adhesion molecules on endothelial cells. A, Degree of downregulation of S1P1 transcripts is correlated with reduced expression levels of PECAM-1 and VE-cadherin in human endothelial cells. Cell lines ISO-HAS.1 (I-H) and AS-M.5 (AS) were transfected with plasmids encoding siRNA control sequences (Silco) or siRNA targeted to different portions of S1P1 transcript (Sil#1, Sil#2, Sil#3). Cell surface expression of PECAM-1 (light gray), VE-cadherin (dark gray), and levels of S1P1 transcripts (black) were analyzed by fluorescence-activated cell sorter and quantitative RT-PCR, respectively, in both whole populations of transfected cells (Silco, Sil#2, Sil#3) and single cell–derived clones (I-H, Sil#1,cl.19 and cl.5; AS, Sil#1, cl.3). The values are mean±SEM Abbreviations are as defined in text. B, No expression of VE-cadherin or PECAM-1 in S1P1-knockdown endothelial cells. AS-M.5 (A, D), HUVECs (C, F), and AS-M.5-Sil#1 (B, E) were grown to confluence. VE-cadherin (A–C) and PECAM-1 (D–F) were visualized by immunofluorescence microscopy. Bar represents 100 μm. B, ICAM-1 and E-selectin expression in response to proinflammatory stimuli is reduced in S1P1-knockdown human endothelial cells. AS-M.5-Sil#1 and control AS-M.5-Silco were treated with TNF- (100 ng/mL, 5 hours, 37°C; gray bars) or LPS (1 μg/mL, 5 hours, 37°C; black bars). The expression of ICAM-1 and E-selectin was analyzed by flow cytometry. Specific mean fluorescence intensity was expressed as a percentage of the TNF-–induced stimulation in control (Silco) cells. Values are mean±SEM. Abbreviations are as defined in text.
S1P1 Modulates the Endothelial Cell Response to Proinflammatory Stimuli
Similar to freshly isolated HUVECs that respond to proinflammatory stimuli by the transcriptional upregulation of cell adhesion molecules, both AS-M.5 (not shown) and AS-M.5-Silco cells responded to stimulation by tumor necrosis factor- (TNF-) (100 ng/mL, 6 hours) and lipopolysaccharide (LPS; 1 μg/mL, 6 hours) by a transient induction of transcription of E-selectin and intercellular adhesion molecule-1 (ICAM-1) on the cell surface (Figure 4c). Albeit diminished in comparison with the control cells, induction of ICAM-1 expression in response to TNF- and LPS was clearly detectable in the knockdown AS-M.5 cells. However, in contrast to either the AS-M.5 or AS-M.5-Silco cells, no induction of E-selectin production was observed in AS-M.5-Sil#1 (Figure 4c). Analysis of E-selectin transcripts by real-time RT-PCR showed that E-selectin transcripts were highly reduced in the LPS-stimulated, S1P1-knockdown cells compared with the control cells, suggesting that E-selectin expression was regulated at the transcriptional level (not shown). Thus, we conclude that S1P1 is involved in the regulation of the inflammatory response of endothelial cells.
Microarray Analysis of Genes Expressed in S1P1-Silenced Endothelial Cells
Searching for gene targets that are dependent on S1P1 expression, we compared gene expression patterns of 2 clones of S1P1-knockdown AS-M.5 cells (T2, cl.2 and cl.3) with those of AS-M.5-Silco and AS-M.5. Among the 22 238 gene probes that the microarrays contained, we identified 82 genes with increased expression and 220 downregulated transcripts in the S1P1-knockdown cells (2-fold decrease and increase; error probability <5%). Many endothelial specific transcripts were downregulated (Table II, available online at http://atvb.ahajournals.org), but the expression of other genes was not affected in the S1P1-knockdown AS-M.5 cells. The latter genes include angiotensin II receptor (Hs.405348), endothelial lipase (Hs.65370), endothelin-2 (Hs.1407), nitric oxide synthetase-3 (NOS3, Hs.446303), P-selectin (Hs, 73800), the VEGF receptor FLT1 (Hs.347713), and fms-related tyrosine kinase FLT3 (Hs.385), vimentin (Hs.43800), and vitronectin (Hs.2257; not shown).
Discussion
The analysis of expression profiles of S1P receptors showed that all human endothelial cells tested (AS-M.5,23 ISO-HAS.1,24,25 HPMEC-ST1,27 and HUVECs) expressed transcripts encoding S1P1, suggesting that the S1P1 receptor may have a ubiquitous function in endothelial cells. All human endothelial cells tested also contained transcripts encoding S1P3, at a much lower level, however, as measured by quantitative RT-PCR. Expression of both S1P1 and S1P3 characterizes the typical expression pattern of S1P receptors in cultured human endothelial cells.20
To examine the function of S1P1 in human endothelial cells, we permanently silenced by RNA interference the expression of S1P1 in the established endothelial cell lines AS-M.5 and ISO-HAS.1, previously shown to possess many characteristics of freshly isolated endothelial cells.23–25 The siRNA targeted to a large loop in the secondary structure of the S1P1 transcript (Sil#1) was more efficient in silencing than were the less exposed sequences (Sil#2, Sil#3) in the S1P1 transcripts. A similar observation was reported for silencing of the insulin-like growth factor receptor,29 suggesting that the secondary structure of the targeted transcripts may be important for efficient silencing by siRNA. The reduction of S1P1 expression in the cells transfected with plasmids Sil#1 did not affect the already very low level of S1P3 expression and was stable for at least 6 months of continuous culturing.
Silencing of S1P1 transcripts to <20% of control profoundly altered the phenotype of endothelial cells. The S1P1-knockdown endothelial cells (AS-M.5-Sil#1) manifested an altered morphology and proliferated slowly in low-density cultures. In contrast to the control cells showing spectacular reorganization of the actin cytoskeleton in response to S1P, the S1P1-knockdown AS-M.5 cells showed no reorganization of actin after S1P treatment. Actin reorganization is an immediate response triggered by S1P.20,30 It has been reported to be regulated through balance between Rho and Rac activity in HUVECs21,31 and to be correlated with an enhancement of endothelial cell barrier function.32 Our results demonstrate that the S1P1 receptor is required for S1P-induced actin reorganization in human endothelial cells.
The S1P1-knockdown cells exhibited altered angiogenic properties. No formation of typical capillary-like networks of differentiated endothelial cells on Matrigel could be observed with AS-M.5-Sil#1, consistent with the reports on the involvement of S1P1 in angiogenesis.33 Both the lack of angiogenic response of AS-M.5-Sil#1 and the capacity to form capillary-like structures34 of ISO-HAS.1 expressing S1P1 but almost no S1P3 suggest that S1P1 but not S1P3 is essential for endothelial cell morphogenesis. However, we cannot exclude a cooperative effect of S1P1 and S1P3 previously observed in short-term inhibition studies with antisense oligonucleotides.20 In a recent study, downregulation of S1P1 expression effectively suppressed tumor angiogenesis in a murine model in vivo.35
We observed an increased proportion of both apoptotic and necrotic cells in both control and AS-M.5-Sil#1 cells after treatment with 4-HNE, a mediator of oxidative stress36 shown to induce apoptotic death in HUVECs.37 Whereas no major difference between control and AS-M.5-Sil#1 cells was observed in the proportion of apoptotic cells, the proportion of necrotic cells in AS-M.5-Sil#1 increased significantly. S1P has been previously reported to activate SphK, which results in the intracellular production of S1P and has an antiapoptotic effect in endothelial cells.38,39 However, our observations indicate that S1P1 is not involved in the antiapoptotic activity of S1P but may rather protect endothelial cells from necrosis.
Long-term silencing of S1P1 enabled us to identify new S1P1-dependent gene targets in endothelial cells. Downregulation of S1P1 led to a reduced expression of the adhesion molecules PECAM-1, expressed predominantly at the endothelial cell contacts,40 and VE-cadherin, usually expressed at intercellular junctions.41 The degree of S1P1 knockdown was correlated with the extent of suppression of PECAM-1 and VE-cadherin. This suggests that the expression of these surface molecules is controlled by S1P1. VE-cadherin expression was previously reported to be reduced after injection of antisense oligonucleotides directed against S1P1 into HUVECs.20 Because functional VE-cadherin was demonstrated to be required for endothelial cell morphogenesis42 and PECAM-1–null endothelial cells failed to migrate in response to S1P,43 the lack of expression of both VE-cadherin and PECAM-1 in S1P1 knockdown endothelial cells may impede their morphogenetic reaction on Matrigel.
The S1P1-knockdown endothelial cells differed from the control cells also in their response to proinflammatory stimuli. Silencing of S1P1 had no major effect on LPS-induced expression of ICAM-1 but reduced the TNF-–stimulated expression of ICAM-1 by 60%. No expression of E-selectin in response to stimulation with either LPS or TNF- was observed in the S1P1-knockdown endothelial cells, implicating involvement of S1P1 in the regulation of E-selectin expression. Several lines of evidence suggest a possible participation of S1P in the inflammatory reaction of endothelial cells. TNF- was found to induce rapid activation of SphK and subsequent generation of intracellular S1P that resulted in strong stimulation of E-selectin expression in HUVECs.44 The defective control of E-selectin expression in the S1P1 knockdown cells suggests that a functional S1P1 receptor is required to mediate expression of E-selectin in stimulated endothelial cells and implies that the intracellularly produced S1P does not function as a second messenger but rather is released from the cells and then activates S1P1. S1P treatment of confluent HUVEC cultures was reported to induce the expression of E-selectin to levels comparable to those induced with TNF- in 144 but not in another45 study, Also, we found only negligible induction of E-selectin expression after S1P treatment of both HUVEC and AS-M.5 cultures (not shown). The latter observations suggest that a synergistic interaction of the signal transduction pathways induced by TNF- and S1P1 receptor activation is required for robust E-selectin expression observed after treatment of endothelial cells with TNF-.
Further S1P1-dependent gene targets were identified by microarray analysis of gene expression. Many endothelial specific transcripts, such as the VEGF-receptor KDR, were downregulated in S1P1-silenced endothelial cells, but other endothelial specific genes, such as the VEGF-receptor Flt-1 and the fms-related kinase Flt-3, were not affected.
In conclusion, our results demonstrate that long-term silencing of S1P1 profoundly alters gene expression and key functions of cultured human endothelial cells. This observation, if confirmed in vivo, could have important therapeutic consequences for chronic inhibition of the S1P1 receptor with receptor antagonists or low concentrations of agonists leading to S1P1 downregulation. The immunosuppressive drug FTY720, an agonist of S1P-receptors currently used in clinical trials, at therapeutic concentrations has been recently reported to block S1P signaling through downregulation of S1P1,2,5.46 Our results imply that FTY720 could be harmful if it induced downregulation of S1P1 on vascular endothelium in vivo.
Acknowledgments
The study was supported by the August-Lenz-Stiftung and the Deutsche Forschungsgemeinschaft (graduate program Vascular Biology in Medicine GRK 438: T.R. and C.V., SFB 413 and Si 247/9). We thank Dr Reinhard Hoffmann for his help with the microarray analyses and Nicole Wilke for technical assistance.
References
Hla T, Lee MJ, Ancellin N, Paik JH, Kluk MJ. Lysophospholipids–receptor revelations. Science. 2001; 294: 1875–1878.
Kluk MJ, Hla T. Signaling of sphingosine-1-phosphate via the S1P/EDG-family of G-protein-coupled receptors. Biochim Biophys Acta. 2002; 1582: 72–80.
Spiegel S, Milstien S. Sphingosine-1-phosphate: an enigmatic signalling lipid. Nat Rev Mol Cell Biol. 2003; 4: 397–407.
Spiegel S, Milstien S. Sphingosine 1-phosphate, a key cell signaling molecule. J Biol Chem. 2002; 29: 25851–25854.
Watterson K, Sankala H, Milstien S, Spiegel S. Pleiotropic actions of sphingosine-1-phosphate. Prog Lipid Res. 2003; 42: 344–357.
Hla T, Maciag T. An abundant transcript induced in differentiating human endothelial cells encodes a polypeptide with structural similarities to G-protein-coupled receptors. J Biol Chem. 1990; 265: 9308–9313.
Okazaki H, Ishizaka N, Sakurai T, Kurokawa K, Goto K, Kumada M, Takuwa Y. Molecular cloning of a novel putative G protein-coupled receptor expressed in the cardiovascular system. Biochem Biophys Res Commun. 1993; 190: 1104–1109.
Yamaguchi F, Tokuda M, Hatase O, Brenner S. Molecular cloning of the novel human G protein-coupled receptor (GPCR) gene mapped on chromosome 9. Biochem Biophys Res Commun. 1996; 227: 608–614.
Graler MH, Bernhardt G, Lipp M. EDG6, a novel G-protein-coupled receptor related to receptors for bioactive lysophospholipids, is specifically expressed in lymphoid tissue. Genomics. 1998; 53: 164–169.
Im DS, Heise CE, Ancellin N, O’Dowd BF, Shei GJ, Heavens RP, Rigby MR, Hla T, Mandala S, McAllister G, George SR, Lynch KR. Characterization of a novel sphingosine 1-phosphate receptor, Edg-8. J Biol Chem. 2000; 275: 14281–14286.
Olivera A, Spiegel S. Sphingosine-1-phosphate as second messenger in cell proliferation induced by PDGF and FCS mitogens. Nature. 1993; 365: 557–560.
Choi OH, Kim JH, Kinet JP. Calcium mobilization via sphingosine kinase in signalling by the Fc epsilon RI antigen receptor. Nature. 1996; 380: 634–636.
Xia P, Wang L, Moretti PA, Albanese N, Chai F, Pitson SM, D’Andrea RJ, Gamble JR, Vadas MA. Sphingosine kinase interacts with TRAF2 and dissects tumor necrosis factor- signaling. J Biol Chem. 2002; 277: 7996–8003.
Shu X, Wu W, Mosteller RD, Broek D. Sphingosine kinase mediates vascular endothelial growth factor-induced activation of ras and mitogen-activated protein kinases. Mol Cell Biol. 2002; 22: 7758–7768.
Spiegel S, Milstien S. Sphingosine-1-phosphate: signaling inside and out. FEBS Lett. 2000; 476: 55–57.
Hla T. Sphingosine 1-phosphate receptors. Prostaglandins Other Lipid Mediat. 2001; 64: 135–142.
Liu Y, Wada R, Yamashita T, Mi Y, Deng CX, Hobson JP, Rosenfeldt HM, Nava VE, Chae SS, Lee MJ, Liu CH, Hla T, Spiegel S, Proia RL. Edg-1, the G protein-coupled receptor for sphingosine-1-phosphate, is essential for vascular maturation. J Clin Invest. 2000; 106: 951–961.
Kono M, Mi Y, Liu Y, Sasaki T, Allende ML, Wu YP, Yamashita T, Proia RL. S1P1, S1P2 and S1P3 receptors coordinately function during embryonic angiogenesis. J Biol Chem. 2004; 279: 29367–29373.
Allende ML, Yamashita T, Proia RL. G-protein-coupled receptor S1P1 acts within endothelial cells to regulate vascular maturation. Blood. 2003; 102: 3665–3667.
Lee MJ, Thangada S, Claffey KP, Ancellin N, Liu CH, Kluk M, Volpi M, Sha’afi RI, Hla T. Vascular endothelial cell adherens junction assembly and morphogenesis induced by sphingosine-1-phosphate. Cell. 1999; 99: 301–312.
Paik JH, Chae S, Lee MJ, Thangada S, Hla T. Sphingosine 1-phosphate-induced endothelial cell migration requires the expression of EDG-1 and EDG-3 receptors and Rho-dependent activation of v3- and 1-containing integrins. J Biol Chem. 2001; 276: 11830–11837.
Langlois S, Gingras D, Beliveau R. Membrane type 1-matrix metalloproteinase (MT1-MMP) cooperates with sphingosine 1-phosphate to induce endothelial cell migration and morphogenic differentiation. Blood. 2004; 103: 3020–3028.
Krump-Konvalinkova V, Bittinger F, Olert J, Br?uninger W, Brunner J, Kirkpatrick CJ. Establishment and characterization of an angiosarcoma-derived cell line(AS-M). Endothelium. 2003; 10: 319–328.
Masuzawa M, Fujimura T, Hamada Y, Fujita Y, Hara H, Nishiyama S, Katsuoka K, Tamauchi H, Sakurai Y. Establishment of a human hemangiosarcoma cell line (ISO-HAS). Int J Cancer. 1999; 81: 305–308.
Unger RE, Krump-Konvalinkova V, Peters K, Kirkpatrick CJ. In vitro expression of the endothelial phenotype: comparative study of primary isolated cells and cell lines, including the novel cell line HPMEC-ST1.6R. Microvasc Res. 2002; 64: 384–397.
Jaffe EA, Nachman RL, Becker CG, Minick CR. Culture of human endothelial cells derived from umbilical veins: identification by morphologic and immunologic criteria. J Clin Invest. 1973; 52: 2745–2756.
Krump-Konvalinkova V, Bittinger F, Unger RE, Peters K, Lehr HA, Kirkpatrick CJ. Generation of human pulmonary microvascular endothelial cell lines. Lab Invest. 2001; 81: 1717–1727.
Hofacker IL. Vienna RNA secondary structure server. Nucleic Acids Res. 2003; 31: 3429–3431.
Bohula EA, Salisbury AJ, Sohail M, Playford MP, Riedemann J, Southern EM, Macaulay VM. The efficacy of small interfering RNAs targeted to the type 1 insulin-like growth factor receptor (IGF1R) is influenced by secondary structure in the IGF1R transcript. J Biol Chem. 2003; 278: 15991–15997.
Yatomi Y, Ohmori T, Rile G, Kazama F, Okamoto H, Sano T, Satoh K, Kume S, Tigyi G, Igarashi Y, Ozaki Y. Sphingosine 1-phosphate as a major bioactive lysophospholipid that is released from platelets and interacts with endothelial cells. Blood. 2000; 96: 3431–3438.
Essler M, Retzer M, Ilchmann H, Linder S, Weber PC. Sphingosine 1-phosphate dynamically regulates myosin light chain phosphatase activity in human endothelial cells. Cell Signal. 2002; 14: 607–613.
Garcia JG, Liu F, Verin AD, Birukova A, Dechert MA, Gerthoffer WT, Bamberg JR, English D. Sphingosine 1-phosphate promotes endothelial cell barrier integrity by Edg-dependent cytoskeletal rearrangement. J Clin Invest. 2001; 108: 689–701.
English D, Brindley DN, Spiegel S, Garcia JG. Lipid mediators of angiogenesis and the signalling pathways they initiate. Biochim Biophys Acta. 2002; 1582: 228–239.
Krump-Konvalinkova V, Kleideiter E, Friedrich U, Klotz U, Kirkpatrick CJ. Tumorigenic conversion of endothelial cells. Exp Mol Pathol. 2003; 75: 154–159.
Chae SS, Paik JH, Furneaux H, Hla T. Requirement for sphingosine 1-phosphate receptor-1 in tumor angiogenesis demonstrated by in vivo RNA interference. J Clin Invest. 2004; 114: 1082–1089.
Uchida K. 4-Hydroxy-2-nonenal: a product and mediator of oxidative stress. Prog Lipid Res. 2003; 42: 318–343.
Herbst U, Toborek M, Kaiser S, Mattson MP, Hennig B. 4-Hydroxynonenal induces dysfunction and apoptosis of cultured endothelial cells. J Cell Physiol. 1999; 181: 295–303.
Xia P, Wang L, Gamble JR, Vadas MA. Activation of sphingosine kinase by tumor necrosis factor- inhibits apoptosis in human endothelial cells. J Biol Chem. 1999; 274: 34499–34505.
Hisano N, Yatomi Y, Satoh K, Akimoto S, Mitsumata M, Fujino MA, Ozaki Y. Induction and suppression of endothelial cell apoptosis by sphingolipids: a possible in vitro model for cell-cell interactions between platelets and endothelial cells. Blood. 1999; 93: 4293–4299.
Albelda SM, Oliver PD, Romer LH, Buck CA. EndoCAM: a novel endothelial cell-cell adhesion molecule. J Cell Biol. 1990; 110: 1227–1237.
Ayalon O, Sabanai H, Lampugnani MG, Dejana E, Geiger B. Spatial and temporal relationships between cadherins and PECAM-1 in cell-cell junctions of human endothelial cells. J Cell Biol. 1994; 126: 247–258.
Bach TL, Barsigian C, Chalupowicz DG, Busler D, Yaen CH, Grant DS, Martinez J. VE-Cadherin mediates endothelial cell capillary tube formation in fibrin and collagen gels. Exp Cell Res. 1998; 238: 324–334.
Gratzinger D, Canosa S, Engelhardt B, Madri JA. Platelet endothelial cell adhesion molecule-1 modulates endothelial cell motility through the small G-protein Rho. FASEB J. 2003; 17: 1458–1469.
Xia P, Gamble JR, Rye KA, Wang L, Hii CS, Cockerill P, Khew-Goodall Y, Bert AG, Barter PJ, Vadas MA. Tumor necrosis factor- induces adhesion molecule expression through the sphingosine kinase pathway. Proc Natl Acad Sci U S A. 1998; 95: 14196–14201.
Miura Y, Yatomi Y, Ohmori T, Osada M, Ozaki Y. Independence of tumor necrosis factor--induced adhesion molecule expression from sphingosine 1-phosphate signaling in vascular endothelial cells. J Thromb Haemost. 2004; 2: 1019–1021.
Gr?ler MH, Goetzl EJ. The immunosuppressant FTY720 down-regulates sphingosine 1-phosphate G-protein-coupled receptors. FASEB J. 2004; 18: 551–553.(Vera Krump-Konvalinkova; )