The Transcription Factor SOX18 Regulates the Expression of Matrix Metalloproteinase 7 and Guidance Molecules in Human Endothelial Cells

Abstract

Background

Mutations in the transcription factor SOX18 are responsible for specific cardiovascular defects in humans and mice. In order to gain insight into the molecular basis of its action, we identified target genes of SOX18 and analyzed one, MMP7, in detail.

Methodology/Principal Findings

SOX18 was expressed in HUVEC using a recombinant adenoviral vector and the altered gene expression profile was analyzed using microarrays. Expression of several regulated candidate SOX18 target genes was verified by real-time PCR. Knock-down of SOX18 using RNA interference was then used to confirm the effect of the transcription factor on selected genes that included the guidance molecules ephrin B2 and semaphorin 3G. One gene, MMP7, was chosen for further analysis, including detailed promoter studies using reporter gene assays, electrophoretic mobility shift analysis and chromatin-immunoprecipitation, revealing that it responds directly to SOX18. Immunohistochemical analysis demonstrated the co-expression of SOX18 and MMP7 in blood vessels of human skin.

Conclusions/Significance

The identification of MMP7 as a direct SOX18 target gene as well as other potential candidates including guidance molecules provides a molecular basis for the proposed function of this transcription factor in the regulation of vessel formation.

About the Authors:

Martina Hoeth

Current address: Baxter Innovations GmbH, Vienna, Austria

Affiliation: Department of Vascular Biology and Thrombosis Research, Medical University of Vienna, Vienna, Austria

Heide Niederleithner

Affiliation: Department of Dermatology, Medical University of Vienna, Vienna, Austria

Renate Hofer-Warbinek

Affiliation: Department of Vascular Biology and Thrombosis Research, Medical University of Vienna, Vienna, Austria

Martin Bilban

Affiliation: Clinical Institute of Medical and Chemical Laboratory Diagnostics, Medical University of Vienna, Vienna, Austria

Herbert Mayer

Current address: Austrian Science Fund, Vienna, Austria

Affiliation: Department of Vascular Biology and Thrombosis Research, Medical University of Vienna, Vienna, Austria

Ulrike Resch

Affiliation: Department of Vascular Biology and Thrombosis Research, Medical University of Vienna, Vienna, Austria

Christof Lemberger

Affiliation: Department of Vascular Biology and Thrombosis Research, Medical University of Vienna, Vienna, Austria

Oswald Wagner

Affiliation: Clinical Institute of Medical and Chemical Laboratory Diagnostics, Medical University of Vienna, Vienna, Austria

Erhard Hofer

Affiliation: Department of Vascular Biology and Thrombosis Research, Medical University of Vienna, Vienna, Austria

Peter Petzelbauer

Affiliation: Department of Dermatology, Medical University of Vienna, Vienna, Austria

Rainer de Martin

* E-mail: [email protected]

Affiliation: Department of Vascular Biology and Thrombosis Research, Medical University of Vienna, Vienna, Austria

Introduction

The SOX family of High Mobility Group (HMG) box transcription factors plays important roles in embryonic development. SOX18, together with SOX7 and SOX17, constitutes the subgroup F within this family. Mutations in SOX18 are the underlying cause of hypotrichosis-lymphedema-telangiectasia (HLT) in humans, a disease that is characterized by sparse hair, bleeding and lymphedema [1]. A similar phenotype (ragged) exists in mice and is characterized not only by the name-giving coat defects but also by cardiovascular abnormalities, e.g., edema, cyanosis, dilation, distention and rupture of peripheral blood vessels, including lymphatic defects [2]–[5]. The structure of SOX18 comprises an N-terminal DNA binding domain, a central transactivation domain, and a C-terminal region that contains a recently discovered additional transactivation motif [6]. Several of the cognate mutations in murine Sox18 are located in a short region at the end of the central transactivation and the beginning of the C-terminal part, leading to premature stop codons, and thereby generating dominant-negative (dn) proteins. In contrast to the ragged alleles, Sox18^−/− mice show only a mild phenotype [7] which, though it may in part be explained by strain differences, suggests that other transcription factors, possibly members of the Sox subgroup F family, may have redundant function [8]. Indeed, the generation of Sox18/Sox17 double mutant mice has at least partially supported this view [9], [10].

In the developing mouse embryo, Sox18 is expressed in the allantois and yolk sack blood islands, in the heart, the paired dorsal aortae, in expanding intersomitic and peripheral vessels, in the pancreas, and in nascent vibrissae follicles [5], [11]. In chickens, expression has been described in feather follicles [12], and in humans in fetal brain [13]. In Xenopus laevis, sox18 together with sox7 appears to be necessary for cardiogenesis [14]. The combined knock-down of SOX18 and SOX7 in the zebrafish resulted in defective arteriovenous specification [15], [16].

In the adult organism, SOX18 expression occurs in ventricles and the inter-ventricular septum of the heart [17]. In the vasculature, it is transiently expressed in capillaries within granulation tissue of skin wounds [18]. In atherosclerotic lesions, SOX18 was localized to endothelial cells of the vasa vasorum and intimal neovessels, and also to vascular smooth muscle cells (SMC) in the intima [19]. Expression was also detected in human umbilical vein endothelial cells (HUVEC) and SMC in culture, and found to be necessary for SMC growth in an in vitro injury model [19].

Together, the phenotypes observed in human as well as in different experimental model organisms suggest a predominant role of SOX18 in the vasculature, both during development and in the adult. The finding of its expression in a number of tumor cell lines [19], the observation that ragged mice show reduced growth of vascularized tumors [20], and the successful inhibition of tumor angiogenesis using cell-permeable dn SOX18 mutants [21] support the view that SOX18 could be a valuable target for interfering with (tumor) angiogenesis. However, despite these important aspects, very little is known about the molecular mechanisms underlying the function(s) of SOX18, i.e. which genes are regulated by the transcription factor. The most prominent one in the context of lymphatic vasculature development is Prox1, however, it requires the venous endothelial-specific nuclear hormone receptor Coup-TFII [2], [22] Two other target genes, VCAM-1 and the μ-opioid receptor, have been described previously [23], [24], however, these can only partially explain the observed phenotypes. Another one, Claudin-5, was reported more recently [25]. Since claudins constitute tight junction strands and maintain cellular barriers, impaired expression could account for the observed edema formation in individuals affected by SOX18 mutations. However, claudin-5 knock-out mice did not show a corresponding phenotype [26]. Last not least, ROBO4 was found to be transcriptionally regulated by Sox18 in the zebrafish, suggesting a role for Sox18 in vessel guidance [27].

We present here the results of a more global approach aiming at the identification of SOX18 target genes in endothelial cells. Using ectopic SOX18 expression in primary human endothelial cells followed by microarray-based gene expression analysis, we have obtained a comprehensive list of potential target genes. Selected genes were confirmed by real-time PCR and by knock-down experiments. In addition, and to verify the validity of the approach, we have characterized the SOX18-dependent regulation of one of them in more detail, namely matrix metalloproteinase (MMP7). The multiple biological effects of this protein during vessel formation may well contribute to the phenotype caused by Sox18 deficiency.

Materials and Methods

Plasmids

Human SOX7, SOX17, and SOX18 cDNAs were isolated by RT-PCR from HUVEC and cloned into the vector pCMV-myc (Clontech). Promoter fragments for MMP7 (345 and 196 bps) were isolated by PCR (High Fidelity, Roche), and inserted into the luciferase reporter vector pUBT-Luc [28]. Mutation of the potential SOX18 binding site in the MMP7 promoter was done using the QuikChange Mutagenesis Kit (Stratagene). The sequence of the primers used for construction is given in Table S1. All constructs were verified by sequencing.

Cell culture and transfection

HEK293 cells were obtained from ATCC. HUVEC were isolated from human umbilical cords derived from human subjects and propagated as described previously [29]. The use of human umbilical cords for the isolation of HUVEC and the use of skin samples has been approved by the Ethics Commission of the Medical University of Vienna. Written informed consent was obtained from all patients (in the case of umbilical cords, written informed consent was obtained from the parents). Umbilical cords and skin samples were obtained from the Department of Obstetrics and Gynecology, and the Department of Dermatology, respectively, both at the Vienna General Hospital, Medical University of Vienna. HEK293 cells were transfected with the different reporter constructs using the calcium-phosphate method. All transfections were done in triplicates. Luciferase levels were normalized for ß-Gal expression.

RNA interference

HUVEC were grown in 6-well plates and transfected with siRNAs using polyethylenimine (PEI). The final concentrations in the transfection mix were 250 mM siRNAs and 0.003% PEI. siRNAs directed against either human SOX18, SOX17 or control were purchased from Ambion. Primer sequences are given in Table S1.

Recombinant adenoviral vectors

SOX18 was inserted into pIRES-EGFP (Clontech), and then the entire SOX18-IRES-EGFP casette was subcloned into the adenoviral transfer vector pACCMVpLpASR+ [30]. The control vector contained only the IRES-EGFP part. Recombinant adenovirus (Adv) was generated by recombination in HEK293 cells as described previously [29]. For transduction, HUVEC were washed twice with serum-free medium and incubated with the adenovirus at a final conc. of 10⁸ pfu/ml for 30 min., followed by replacement with complete medium.

Microarray analysis

RNA was isolated from non-transduced, Adv-SOX18 and control Adv transduced HUVEC using the RNeasy Mini Kit (Qiagen). Probe labeling and hybridization to HG-U133A 2.0 arrays (Affymetrix) interrogating the expression of 18.400 transcripts, as well as data analysis, was done as described previously [31]. Data conforming to MIAME standards were deposited in the GEO database under the accession nr. GSE8952.

Real-time PCR

Expression of selected genes was analyzed using a LightCycler with Fast Start SYBR Green I Kit (Roche). Values were normalized for ß2 microglobulin expression. Primers were designed using the program Primer3. Results were confirmed by different sources of RNA. Primer sequences are given in Table S1.

Electrophoretic mobility shift assay

The Lightshift chemiluminescence EMSA Kit (Pierce) was used with biotinylated oligonucleotides as given in the supplemental materials. Competition experiments were carried out using a 100-fold excess of unlabelled oligonucleotides with either the same, a mutated, or an unrelated sequence. Primer sequences are given in Table S1.

Chromatin Immunoprecipitation

We used a protocol with two-step cross-linking as decribed [32]. Briefly, 10⁷ cells were pelleted and incubated with 2 mM disuccinylglutarate in PBS for 45 min at room temp, washed with PBS, and crosslinked with 1% paraformaldehyde for 15 min. at RT. The reaction was stopped with 0.125 mM glycine in PBS, cells were washed and suspended in cold lysis buffer containing protease inhibitors (Roche). After 10 min on ice, DNA was sonicated 4× for 25 sec. at 25% power (Branson Sonic Power Company B-12 sonifyer). Debris was removed by centrifugation and the samples stored at −70°C. After pre-clearing with protein A sepharose, 100 µl aliquots were incubated with 2 µg of different anti-SOX18 antibodies (Santa Cruz sc-20100, Thermo Scientific PA1-24474, and Chemicon AB-3207) or rabbit IgG over-night, precipitated with blocked protein A sepharose, and washed with buffer containing increasing salt concentrations. The antibody was eluted from the DNA-protein complex in 1% SDS/0.1 M sodium bicarbonate, the agarose removed by centrifugation, and cross-linking reversed by addition of NaCl to a final conc. of 0.2 M and incubation at 65°C for 4 hours. DNA was purified over Qiagen columns, and analyzed by PCR using the primers given in Table S1. The anti-Sox18 antibody recognized Sox17 10–20 times less efficiently, no detectable cross-reactivity with Sox7 was observed (Figure S2).

Immunohistochemistry

Serial sections from paraffin-embedded normal human trunk skin were stained with rabbit antibodies against SOX18 (sc-20100; 1∶2500, Santa Cruz), MMP7 (ab4044, Abcam, 1∶500), LYVE-1 (102PA50, Research Diagnostics, Inc., 1∶100), or von Willebrand factor (A0082, Dako, 1∶5000), followed by incubation with biotinylated secondary antibody (BA-1000, anti rabbit IgG, Vector Laboratories, 1∶200). Bound antibodies were visualized by incubation with a streptavidin-HRP conjugate (Dako) and 3-amino-9-ethylcarbazole resulting in a red reaction product. Sections were counterstained with haematoxylin.

Results

Identification of SOX18 target genes

Our strategy to identify SOX18 target genes was to ectopically express the transcription factor in primary endothelial cells followed by microarray analysis of the resulting gene repertoire. In order to achieve high transfection efficiency we generated a SOX18 adenoviral vector containing an IRES-EGFP cassette. The control virus contained the IRES-EGFP cassette alone. As judged by EGFP expression, transduction efficiency was close to 100%. As assayed by real-time PCR, expression of Sox18 was approx. 70-fold higher as compared to controls (Figure S1), and did not alter Sox7 and -17 levels. Expression of selected candidate genes was confirmed by real-time PCR. Knock-down of SOX18 by RNA interference was then used as a complementary approach to verify the dependence of selected genes on the transcription factor.

In order to obtain primary target genes and avoid the isolation of genes that are up-regulated secondary to the induction of others, we harvested the SOX18 expressing HUVEC at an early time point after transduction. SOX18 protein could be detected as early as 16 hours (not shown). RNA from SOX18-, EGFP-, and non-transduced HUVEC was subjected to microarray analysis. Using a threshold of 3-fold changes, 30 genes were found to be upregulated, and 13 down-regulated (0.23% of the analyzed transcripts, Table 1). Selected genes were further analyzed using real-time PCR, and induction was confirmed for MMP7, ephrin-B2 (EFNB2), EPHA7, semaphorin 3G (SEMA3G), interleukin 7 receptor (IL-7R), cannabinoid receptor 1 (CNR1) (Figure 1), and others (Table 1). Whereas the published expression of VCAM-1 could not be confirmed, both the μ-opioid receptor and claudin-5 were found to be 2-fold increased in the array (not shown). The entire data set has been deposited in the GEO data base (NCBI) under the accession nr. GSE8952.

[Figure omitted. See PDF.]

Figure 1. Real-time PCR analysis of SOX18 target genes.

RNA was isolated from non-transduced (mock), control virus (AdvEGFP) and SOX18 adenovirus (AdvSOX18) transduced HUVEC and analyzed for the expression of matrix metalloproteinase 7 (MMP7), ephrinB2 (EFNB2), EPHA7, semaphorin 3G (SEMA3G), interleukin 7 receptor (IL-7R), and cannabinoid receptor 1 (CNR1). Values were normalized for ß2 microglobulin expression, and are expressed as fold induction compared to control cells.

https://doi.org/10.1371/journal.pone.0030982.g001

[Figure omitted. See PDF.]

Table 1. Top SOX18 regulated genes.

https://doi.org/10.1371/journal.pone.0030982.t001

In a complementary approach we then knocked down SOX18 in HUVEC (and also SOX17 which has been reported to have partially redundant function) using siRNA, and analyzed a subset of genes by real-time PCR. Of these transcripts, MMP7, EFNB2, and SEMA3G were reduced by SOX18 knock-down (Figure 2), the last two also responded to SOX17 knock-down. Two genes (CNR1 and EPHA7) responded weakly, whereas IL-7R expression was enhanced (data not shown).

[Figure omitted. See PDF.]

Figure 2. Knock-down of SOX18 diminishes the expression of target genes.

HUVEC were transfected with siRNAs directed against SOX17 or SOX18, or with a scrambled control (Scr), and expression of target genes analyzed by real-time PCR. Values were normalized for ß2 microglobulin levels, and are expressed as fold induction compared to control cells.

https://doi.org/10.1371/journal.pone.0030982.g002

MMP7 is a direct target gene of SOX18

Based on its multiple functions in angiogenesis [33]–[37] and clear response to SOX18 in the overexpression and knock-down experiments, we selected MMP7 for further analysis. A set of experiments was performed to investigate whether it is a primary target or expressed indirectly, i.e., secondary to other regulators that may have been induced by SOX18. A direct target gene is supposed to contain one or more functional SOX18 binding site(s) in the promoter. We therefore analyzed the MMP7 promoter by reporter gene analysis including mutation of the potential SOX18 binding site, binding of SOX18 in vitro using electrophoretic mobility shift assay (EMSA), and in vivo using chromatin immunoprecipitation (ChIP).

First, bioinformatic analysis using the program TOUCAN [38] revealed the presence of two potential SOX18 binding sites. We isolated two promoter fragments that contained either both or only the proximal site (Figure 3A), and cloned them into the luciferase reporter vector UBT-Luc [28]. Cotransfection of these constructs together with SOX18 resulted in a strong dose-dependent induction of both MMP7 promoter constructs. In contrast, cotransfected SOX7 or SOX17, the other members of the subgroup F, had only negligible effects (Figure 3B). Since cotransfection experiments showed that the proximal SOX18 binding site is sufficient to induce a strong response only this site was chosen for mutation experiments. Its mutation (Figure 3C) resulted in strongly diminished MMP7 induction, demonstrating that this site is functional (Figure 3D). Some activity was observed with high amounts of co-transfected Sox18, possibly reflecting non-physiological effects. Furthermore, the short MMP7 promoter construct did not respond to cotransfected SOX7 or SOX17, the latter being in line with the lack of response to knock-down of SOX17 (Figure 3D).

[Figure omitted. See PDF.]

Figure 3. Promoter analysis of MMP7.

A, Schematic representation of the two MMP7 promoter-luciferase reporter constructs containing either only the proximal or both potential SOX18 binding sites (black diamonds). B, the wild type (wt) constructs were cotransfected together with increasing amounts of SOX18, SOX7 or SOX17 expression vectors into HEK293 cells as indicated. Luciferase values were normalized for ß-Gal expression. C, in the shorter construct, the single potential SOX18 binding site was mutated. D, The short MMP7 promoter construct containing the mutated SOX18 binding site was analyzed as in (B).

https://doi.org/10.1371/journal.pone.0030982.g003

Next, we determined whether SOX18 binds to the proximal site in vitro and in vivo. EMSA revealed that nuclear extracts from HUVEC contained a protein binding to the SOX18 binding site, and competition experiments using unlabelled wild type (wt) and mutated oligonucleotides demonstrated the specificity of binding (Figure 4A). However, probably due to technical reasons we were not able to perform a conclusive supershift to directly demonstrate the presence of Sox18 in the complex. HUVEC also express Sox-7 and -17, so in order to assess whether this site was indeed occupied by SOX18 in vivo, we performed ChIP. Using primers flanking the proximal SOX18 binding site in the MMP7 promoter, a specific band could be detected by PCR after precipitation with three different SOX18 antibodies, but not with a control antibody (Figure 4B). Since the Sox18 antibody from SantaCruz recognizes a region with some homology to the other SoxF group members we tested its specificity and found a weak cross-reactivity (1/10–1/20) with Sox17, and none with Sox7 (Figure S2). The Sox18 antibody from Chemicon recognizes an N-terminal 17 aa peptide sequence specific for Sox18. The correct identity of the amplified bands was confirmed by sequencing. Therefore, and in accordance with the reporter and EMSA data, binding of SOX18 to the proximal MMP7 SOX18 binding site could be demonstrated. We conclude that, using the criteria of in vitro and in vivo binding and of reporter gene analysis, MMP7 contains at least one functional SOX18 binding site and represents a direct SOX18 target gene.

[Figure omitted. See PDF.]

Figure 4. SOX18 binds to the proximal site in the MMP7 promoter in vitro and in vivo.

A, Electrophoretic mobility shift assay (EMSA). Binding of SOX18 to its proximal site in the MMP7 promoter was analyzed using nuclear extracts from HUVEC. Specificity of binding was demonstrated by competition experiments using the same (wt), a mutated (mut) or an unrelated (unrel) oligonucleotide as indicated, the specific complex is indicated by an arrow. B, The same site was analyzed by chromatin immunoprecipitation in HUVEC. Lane 1–3: different α-Sox18 antibodies (SantaCruz, Thermo, Chemicon, respectively) as described in Materials and Methods, Ig: IgG control antibody; inp: input, M: 100 bp marker. The 120 bp PCR fragments obtained after amplification ofDNA precipitated by the anti-SOX18 antibodies and input DNA are marked by an arrow.

https://doi.org/10.1371/journal.pone.0030982.g004

Co-expression of SOX18 and MMP7 in skin vessels

SOX18 has previously been found to be expressed in endothelial and smooth muscle cells [19]. To correlate SOX18 expression with its target gene we performed immunohistochemical analysis of serial sections of normal adult human skin samples (Figure 5). We found SOX18 expressed in arteries, veins and some small capillaries, the latter all being LYVE1 negative and thus belonging to the blood vessel tree. SOX18 positive vessels co-expressed MMP7. Moreover, SOX18 stained positive in the muscularis of arteries and veins, and these also expressed MMP7. The lack of a complete correlation of SOX18 and MMP7 expression might be due to the fact that MMP7 is also expressed in other cells and in response to stimuli that do not utilize SOX18 (e.g., FGF regulates MMP7 via AP1) [39], or due to diffusion of MMP7 as a secreted protein. Sections were stained with von Willebrand factor to identify all vessels, and LYVE1 for lymphatics. Of note, all lymphatics were negative for SOX18. Furthermore, we have analyzed tissue arrays containing 300 skin samples from basal cell carcinomas of a total of 100 different patients. Similar to the situation in normal skin, all LYVE1 positive lymphatics were negative for SOX18 (data not shown). Although these findings are in contrast with the notion that mutations in SOX18 affect the lymphatic vasculature [2], it may well be that these alterations occur during development and persist during adulthood without the necessity for continued expression of SOX18.

[Figure omitted. See PDF.]

Figure 5. SOX18 and MMP7 are co-expressed in arteries and veins, but not in lymphatic vessels.

Serial sections from human skin were analyzed by immunohistochemical staining with the indicated antibodies. Endothelial cells were revealed by von Willebrand factor (vWF) staining, lymphatic vessels by LYVE-1.

https://doi.org/10.1371/journal.pone.0030982.g005

Discussion

The notion that SOX18 plays an important role in angiogenesis is well supported by the occurrence of mutations in humans and mice that result in vascular (including lymphatic) defects; however, mechanistic insight how this transcription factor exerts its function is still very limited. As a first step to answer this question we have identified SOX18 target genes; one of them, MMP7, was analyzed in more detail, and we could show that it represents a direct SOX18 target gene. A recent study described lower levels of Mmp7 and Il-7r in ragged mice [40]; while the former is supported by our findings, our data were controversial about the latter: Although expression of IL-7R was upregulated upon SOX18 overexpression, it was not reduced, for reasons unknown, upon knock-down in our experimental system.

In vascular cells, multiple functions have been described for MMP7. In general, MMPs degrade a broad variety of proteins, including extracellular matrix proteins. MMP7 substrates include proteoglycans, fibronectin, elastin and casein. SOX18 dependent expression of MMP7 during angiogenesis could therefore function to promote the sprouting process. However, MMP7 appears to have more functions than just clearing barriers: It has been implicated in ectodomain shedding of several cell surface molecules, including heparin-binding epidermal growth factor precursor, membrane-bound Fas ligand, E-cadherin, and TNFα precursor, all of which are cleaved to release the respective soluble forms [35]. Moreover, VEGF-165, which is bound by connective tissue growth factor, can be released from this inactive complex by MMP7 [34]. Similarly, KC, a member of the CXC family of small chemokines can be liberated from its complex with syndecan-1, thereby generating a KC gradient capable of directing neutrophil migration [33]. These biological activities would suggest a pro-angiogenic function.

However, also anti-angiogenic properties have been attributed to MMP7. First, it can cleave plasminogen and collagen XVIII to generate angiostatin, as well as a 28 kD endostatin-spanning fragment, respectively [41]. Moreover, it cleaves decorin, a proteoglycan that can serve as a reservoir for TGF-ß, a potent anti-angiogenic factor in the extracellular matrix [36]. In a mouse model of corneal wound healing, MMP7 served as an inhibitor of corneal avascularity [37]. Therefore, the net outcome of SOX18 dependent MMP7 expression cannot be predicted, and may depend on the specific site and context of the angiogenic process.

Beside MMP7, we have identified several other potential target genes of SOX18. Combining SOX18 overexpression with the more stringent criterion of SOX18 knock-down, EFNB2 and SEMA3G turned out to be reliable candidates. It should be noted that MMP7 responded only to SOX18 knock-down, whereas the other two were dependent (however, to a lesser degree) on SOX17 levels. Both EFNB2 and SEMA3G serve as guidance molecules in endothelial cells. EFNB2 and its counter-receptor EPHB4 are expressed on developing arteries and veins respectively, and repulsive signaling between the two molecules has been suggested to maintain the boundaries between the two types of vessels during network formation [42]. The identification of EFNB2 as a SOX18 target fits well to the recent observations that SOX18 knock-down in the zebrafish results in impaired arteriovenous specification [15], [16]. Less information exists for SEMA3G, a member of larger family of phylogenetically conserved secreted and membrane-bound proteins. SEMA3G is a secreted molecule that binds to Neuropilin-2, but not Neuropilin-1, receptors that are also bound by certain VEGF forms. In neuronal development, SEMA3G induced the repulsion of sympathetic, but not dorsal root ganglion axons, indicating that it can selectivity repel specific types of axons [43]. Semaphorin class 3 family members have been suggested as modulators of angiogenesis [44]. Together, and also considering the additional upregulation of EPHA7 (which, however, did not clearly respond to SOX18 knock-down), suggests that one function of SOX18 may be the control of directed vessel formation. For reasons unknown, we did not identify VCAM-1, a published Sox18 target gene; one explanation could be that our experimental setup of Sox18 overexpression as a primary screen would allow solely the identification of genes for which Sox18 is alone sufficient, but not those where it is necessary (in combination with other factors). We also did not find Prox1 in HUVEC, however Sox18 could induce its expression in endothelial progenitor cells (not shown). This is consistent with the finding that Coup-TFII has been identified as a cofactor for Prox1 initial expression and maintainance [22].

Since a single transcription factor usually participates in the regulation of several genes, mutations that impair its function can be expected to affect a broad spectrum of genes and result in a complex phenotype. Diminished MMP7 expression as a result of a mutated SOX18 may already affect several aspects of angiogenesis, both in the developing as well as the adult organism. In addition, the identification of EFNB2 and SEMA3G suggests a role for SOX18 in vessel guidance.

Taken together, the present work represents a functional way to identify potential SOX18 target genes. With the identification of MMP7 and performed functional analysis, the direct regulatory relationship between SOX18 and MMP7 was demonstrated. Future studies will focus on other potential target genes identified by this approach, to further investigate the role of SOX18 and its target genes in vessel formation. With the more detailed characterization of additional direct Sox18 target genes a better understanding of the molecular mechanisms of SOX18 in the ragged phenotype in mice and the HLT syndrome in humans will be gained.

Supporting Information

[Figure omitted. See PDF.]

Figure S1.

Overexpression of Sox18 by a recombinant adenovirus. HUVEC were either mock transfected or transfected with a recombinant adenovirus for expression of GFP or of Sox18 as indicated, and mRNA isolated 16 hours later and analyzed by real-time PCR for expression of Sox7, -17 and -18. Values were normalized to ß2-microglobulin.

https://doi.org/10.1371/journal.pone.0030982.s001

(TIF)

Figure S2.

Crossreactivity of Sox18 antibody. A, HEK293 cells were transfected with myc-tagged expression vectors for Sox7, -17, and -18, or empty vector (EV), and protein extracts analyzed by Western blotting using anti-myc and anti-Sox18 (SantaCruz). B, repetition of the experiment using Sox17 and -18 and different times of exposure.

https://doi.org/10.1371/journal.pone.0030982.s002

(TIF)

Table S1.

Primers used for constructs, real-time PCR, EMSA, ChIP, and knock-down experiments.

https://doi.org/10.1371/journal.pone.0030982.s003

(DOC)

Acknowledgments

The authors wish to thank the members of the Austrian Angiogenesis Network for continuous support and discussion, and Y. Schichl for proof-reading of the manuscript.

Author Contributions

Conceived and designed the experiments: OW PP RdM EH. Performed the experiments: MH HN RHW MB UR CL. Analyzed the data: HM RdM. Wrote the paper: RdM.

Citation: Hoeth M, Niederleithner H, Hofer-Warbinek R, Bilban M, Mayer H, Resch U, et al. (2012) The Transcription Factor SOX18 Regulates the Expression of Matrix Metalloproteinase 7 and Guidance Molecules in Human Endothelial Cells. PLoS ONE 7(1): e30982. https://doi.org/10.1371/journal.pone.0030982

References

1. Irrthum A, Devriendt K, Chitayat D, Matthijs G, Glade C, et al. (2003) Mutations in the transcription factor gene SOX18 underlie recessive and dominant forms of hypotrichosis-lymphedema-telangiectasia. Am J Hum Genet 72(6): 1470–1478.A. IrrthumK. DevriendtD. ChitayatG. MatthijsC. Glade2003Mutations in the transcription factor gene SOX18 underlie recessive and dominant forms of hypotrichosis-lymphedema-telangiectasia.Am J Hum Genet72614701478

2. Francois M, Caprini A, Hosking B, Orsenigo F, Wilhelm D, et al. (2008) Sox18 induces development of the lymphatic vasculature in mice. Nature 456(7222): 643–647.M. FrancoisA. CapriniB. HoskingF. OrsenigoD. Wilhelm2008Sox18 induces development of the lymphatic vasculature in mice.Nature4567222643647

3. Hosking BM, Wang SC, Chen SL, Penning S, Koopman P, et al. (2001) SOX18 directly interacts with MEF2C in endothelial cells. Biochem Biophys Res Commun 287(2): 493–500.BM HoskingSC WangSL ChenS. PenningP. Koopman2001SOX18 directly interacts with MEF2C in endothelial cells.Biochem Biophys Res Commun2872493500

4. James K, Hosking B, Gardner J, Muscat GE, Koopman P (2003) Sox18 mutations in the ragged mouse alleles ragged-like and opossum. Genesis 36(1): 1–6.K. JamesB. HoskingJ. GardnerGE MuscatP. Koopman2003Sox18 mutations in the ragged mouse alleles ragged-like and opossum.Genesis36116

5. Pennisi D, Gardner J, Chambers D, Hosking B, Peters J, et al. (2000) Mutations in Sox18 underlie cardiovascular and hair follicle defects in ragged mice. Nat Genet 24(4): 434–437.D. PennisiJ. GardnerD. ChambersB. HoskingJ. Peters2000Mutations in Sox18 underlie cardiovascular and hair follicle defects in ragged mice.Nat Genet244434437

6. Sandholzer J, Hoeth M, Piskacek M, Mayer H, de Martin R (2007) A novel 9-amino-acid transactivation domain in the C-terminal part of Sox18. Biochem Biophys Res Commun 360(2): 370–374.J. SandholzerM. HoethM. PiskacekH. MayerR. de Martin2007A novel 9-amino-acid transactivation domain in the C-terminal part of Sox18.Biochem Biophys Res Commun3602370374

7. Pennisi D, Bowles J, Nagy A, Muscat G, Koopman P (2000) Mice null for sox18 are viable and display a mild coat defect. Mol Cell Biol 20(24): 9331–9336.D. PennisiJ. BowlesA. NagyG. MuscatP. Koopman2000Mice null for sox18 are viable and display a mild coat defect.Mol Cell Biol202493319336

8. Hosking B, Francois M, Wilhelm D, Orsenigo F, Caprini A, et al. (2009) Sox7 and Sox17 are strain-specific modifiers of the lymphangiogenic defects caused by Sox18 dysfunction in mice. Development 136(14): 2385–2391.B. HoskingM. FrancoisD. WilhelmF. OrsenigoA. Caprini2009Sox7 and Sox17 are strain-specific modifiers of the lymphangiogenic defects caused by Sox18 dysfunction in mice.Development1361423852391

9. Matsui T, Kanai-Azuma M, Hara K, Matoba S, Hiramatsu R, et al. (2006) Redundant roles of Sox17 and Sox18 in postnatal angiogenesis in mice. J Cell Sci 119(Pt 17): 3513–3526.T. MatsuiM. Kanai-AzumaK. HaraS. MatobaR. Hiramatsu2006Redundant roles of Sox17 and Sox18 in postnatal angiogenesis in mice.J Cell Sci119Pt 1735133526

10. Sakamoto Y, Hara K, Kanai-Azuma M, Matsui T, Miura Y, et al. (2007) Redundant roles of Sox17 and Sox18 in early cardiovascular development of mouse embryos. Biochem Biophys Res Commun 360(3): 539–544.Y. SakamotoK. HaraM. Kanai-AzumaT. MatsuiY. Miura2007Redundant roles of Sox17 and Sox18 in early cardiovascular development of mouse embryos.Biochem Biophys Res Commun3603539544

11. Lioubinski O, Muller M, Wegner M, Sander M (2003) Expression of Sox transcription factors in the developing mouse pancreas. Dev Dyn 227(3): 402–408.O. LioubinskiM. MullerM. WegnerM. Sander2003Expression of Sox transcription factors in the developing mouse pancreas.Dev Dyn2273402408

12. Olsson JE, Kamachi Y, Penning S, Muscat GE, Kondoh H, et al. (2001) Sox18 expression in blood vessels and feather buds during chicken embryogenesis. Gene 271(2): 151–158.JE OlssonY. KamachiS. PenningGE MuscatH. Kondoh2001Sox18 expression in blood vessels and feather buds during chicken embryogenesis.Gene2712151158

13. Stanojcic S, Stevanovic M (2000) The human SOX18 gene: cDNA cloning and high resolution mapping. Biochim Biophys Acta 1492(1): 237–241.S. StanojcicM. Stevanovic2000The human SOX18 gene: cDNA cloning and high resolution mapping.Biochim Biophys Acta14921237241

14. Zhang C, Basta T, Klymkowsky MW (2005) SOX7 and SOX18 are essential for cardiogenesis in Xenopus. Dev Dyn 234(4): 878–891.C. ZhangT. BastaMW Klymkowsky2005SOX7 and SOX18 are essential for cardiogenesis in Xenopus.Dev Dyn2344878891

15. Cermenati S, Moleri S, Cimbro S, Corti P, Del Giacco L, et al. (2008) Sox18 and Sox7 play redundant roles in vascular development. Blood 111(5): 2657–2666.S. CermenatiS. MoleriS. CimbroP. CortiL. Del Giacco2008Sox18 and Sox7 play redundant roles in vascular development.Blood111526572666

16. Herpers R, van de Kamp E, Duckers HJ, Schulte-Merker S (2008) Redundant Roles for Sox7 and Sox18 in Arteriovenous Specification in Zebrafish. Circ Res 102(1): 12–15.R. HerpersE. van de KampHJ DuckersS. Schulte-Merker2008Redundant Roles for Sox7 and Sox18 in Arteriovenous Specification in Zebrafish.Circ Res10211215

17. Saitoh T, Katoh M (2002) Expression of human SOX18 in normal tissues and tumors. Int J Mol Med 10(3): 339–344.T. SaitohM. Katoh2002Expression of human SOX18 in normal tissues and tumors.Int J Mol Med103339344

18. Darby IA, Bisucci T, Raghoenath S, Olsson J, Muscat GE, et al. (2001) Sox18 is transiently expressed during angiogenesis in granulation tissue of skin wounds with an identical expression pattern to Flk-1 mRNA. Lab Invest 81(7): 937–943.IA DarbyT. BisucciS. RaghoenathJ. OlssonGE Muscat2001Sox18 is transiently expressed during angiogenesis in granulation tissue of skin wounds with an identical expression pattern to Flk-1 mRNA.Lab Invest817937943

19. Garcia-Ramirez M, Martinez-Gonzalez J, Juan-Babot JO, Rodriguez C, Badimon L (2005) Transcription factor SOX18 is expressed in human coronary atherosclerotic lesions and regulates DNA synthesis and vascular cell growth. Arterioscler Thromb Vasc Biol 25(11): 2398–2403.M. Garcia-RamirezJ. Martinez-GonzalezJO Juan-BabotC. RodriguezL. Badimon2005Transcription factor SOX18 is expressed in human coronary atherosclerotic lesions and regulates DNA synthesis and vascular cell growth.Arterioscler Thromb Vasc Biol251123982403

20. Young N, Hahn CN, Poh A, Dong C, Wilhelm D, et al. (2006) Effect of disrupted SOX18 transcription factor function on tumor growth, vascularization, and endothelial development. J Natl Cancer Inst 98(15): 1060–1067.N. YoungCN HahnA. PohC. DongD. Wilhelm2006Effect of disrupted SOX18 transcription factor function on tumor growth, vascularization, and endothelial development.J Natl Cancer Inst981510601067

21. Luo M, Guo XT, Yang W, Liu LQ, Li LW, et al. (2008) Inhibition of tumor angiogenesis by cell-permeable dominant negative SOX18 mutants. Med Hypotheses 70(4): 880–882.M. LuoXT GuoW. YangLQ LiuLW Li2008Inhibition of tumor angiogenesis by cell-permeable dominant negative SOX18 mutants.Med Hypotheses704880882

22. Srinivasan RS, Geng X, Yang Y, Wang Y, Mukatira S, et al. (2010) The nuclear hormone receptor Coup-TFII is required for the initiation and early maintenance of Prox1 expression in lymphatic endothelial cells. Genes Dev 24(7): 696–707.RS SrinivasanX. GengY. YangY. WangS. Mukatira2010The nuclear hormone receptor Coup-TFII is required for the initiation and early maintenance of Prox1 expression in lymphatic endothelial cells.Genes Dev247696707

23. Hosking BM, Wang SC, Downes M, Koopman P, Muscat GE (2004) The VCAM-1 gene that encodes the vascular cell adhesion molecule is a target of the Sry-related high mobility group box gene, Sox18. J Biol Chem 279(7): 5314–5322.BM HoskingSC WangM. DownesP. KoopmanGE Muscat2004The VCAM-1 gene that encodes the vascular cell adhesion molecule is a target of the Sry-related high mobility group box gene, Sox18.J Biol Chem279753145322

24. Im HJ, Smirnov D, Yuhi T, Raghavan S, Olsson JE, et al. (2001) Transcriptional modulation of mouse mu-opioid receptor distal promoter activity by Sox18. Mol Pharmacol 59(6): 1486–1496.HJ ImD. SmirnovT. YuhiS. RaghavanJE Olsson2001Transcriptional modulation of mouse mu-opioid receptor distal promoter activity by Sox18.Mol Pharmacol59614861496

25. Fontijn RD, Volger OL, Fledderus JO, Reijerkerk A, de Vries HE, et al. (2008) SOX-18 controls endothelial-specific Claudin-5 gene expression and barrier function. Am J Physiol Heart Circ Physiol 294(2): H891–900.RD FontijnOL VolgerJO FledderusA. ReijerkerkHE de Vries2008SOX-18 controls endothelial-specific Claudin-5 gene expression and barrier function.Am J Physiol Heart Circ Physiol2942H891900

26. Nitta T, Hata M, Gotoh S, Seo Y, Sasaki H, et al. (2003) Size-selective loosening of the blood-brain barrier in claudin-5-deficient mice. J Cell Biol 161(3): 653–660.T. NittaM. HataS. GotohY. SeoH. Sasaki2003Size-selective loosening of the blood-brain barrier in claudin-5-deficient mice.J Cell Biol1613653660

27. Samant GV, Schupp MO, Francois M, Moleri S, Kothinti RK, et al. (2011) Sox factors transcriptionally regulate ROBO4 gene expression in developing vasculature in zebrafish. J Biol Chem 286(35): 30740–30747.GV SamantMO SchuppM. FrancoisS. MoleriRK Kothinti2011Sox factors transcriptionally regulate ROBO4 gene expression in developing vasculature in zebrafish.J Biol Chem286353074030747

28. de Martin R, Strasswimmer J, Philipson L (1993) A new luciferase promoter insertion vector for the analysis of weak transcriptional activities. Gene 124(1): 137–138.R. de MartinJ. StrasswimmerL. Philipson1993A new luciferase promoter insertion vector for the analysis of weak transcriptional activities.Gene1241137138

29. Wrighton CJ, Hofer-Warbinek R, Moll T, Eytner R, Bach FH, et al. (1996) Inhibition of endothelial cell activation by adenovirus-mediated expression of I kappa B alpha, an inhibitor of the transcription factor NF-kappa B. J Exp Med 183(3): 1013–1022.CJ WrightonR. Hofer-WarbinekT. MollR. EytnerFH Bach1996Inhibition of endothelial cell activation by adenovirus-mediated expression of I kappa B alpha, an inhibitor of the transcription factor NF-kappa B.J Exp Med183310131022

30. McGrory WJ, Bautista DS, Graham FL (1988) A simple technique for the rescue of early region I mutations into infectious human adenovirus type 5. Virology 163(2): 614–617.WJ McGroryDS BautistaFL Graham1988A simple technique for the rescue of early region I mutations into infectious human adenovirus type 5.Virology1632614617

31. Mayer H, Bilban M, Kurtev V, Gruber F, Wagner O, et al. (2004) Deciphering regulatory patterns of inflammatory gene expression from interleukin-1-stimulated human endothelial cells. Arterioscler Thromb Vasc Biol 24(7): 1192–1198.H. MayerM. BilbanV. KurtevF. GruberO. Wagner2004Deciphering regulatory patterns of inflammatory gene expression from interleukin-1-stimulated human endothelial cells.Arterioscler Thromb Vasc Biol24711921198

32. Nowak DE, Tian B, Brasier AR (2005) Two-step cross-linking method for identification of NF-kappaB gene network by chromatin immunoprecipitation. Biotechniques 39(5): 715–725.DE NowakB. TianAR Brasier2005Two-step cross-linking method for identification of NF-kappaB gene network by chromatin immunoprecipitation.Biotechniques395715725

33. Ding K, Lopez-Burks M, Sanchez-Duran JA, Korc M, Lander AD (2005) Growth factor-induced shedding of syndecan-1 confers glypican-1 dependence on mitogenic responses of cancer cells. J Cell Biol 171(4): 729–738.K. DingM. Lopez-BurksJA Sanchez-DuranM. KorcAD Lander2005Growth factor-induced shedding of syndecan-1 confers glypican-1 dependence on mitogenic responses of cancer cells.J Cell Biol1714729738

34. Hashimoto G, Inoki I, Fujii Y, Aoki T, Ikeda E, et al. (2002) Matrix metalloproteinases cleave connective tissue growth factor and reactivate angiogenic activity of vascular endothelial growth factor 165. J Biol Chem 277(39): 36288–36295.G. HashimotoI. InokiY. FujiiT. AokiE. Ikeda2002Matrix metalloproteinases cleave connective tissue growth factor and reactivate angiogenic activity of vascular endothelial growth factor 165.J Biol Chem277393628836295

35. Ii M, Yamamoto H, Adachi Y, Maruyama Y, Shinomura Y (2006) Role of matrix metalloproteinase-7 (matrilysin) in human cancer invasion, apoptosis, growth, and angiogenesis. Exp Biol Med (Maywood) 231(1): 20–27.M. IiH. YamamotoY. AdachiY. MaruyamaY. Shinomura2006Role of matrix metalloproteinase-7 (matrilysin) in human cancer invasion, apoptosis, growth, and angiogenesis.Exp Biol Med (Maywood)23112027

36. Imai K, Hiramatsu A, Fukushima D, Pierschbacher MD, Okada Y (1997) Degradation of decorin by matrix metalloproteinases: identification of the cleavage sites, kinetic analyses and transforming growth factor-beta1 release. Biochem J 322(Pt 3): 809–814.K. ImaiA. HiramatsuD. FukushimaMD PierschbacherY. Okada1997Degradation of decorin by matrix metalloproteinases: identification of the cleavage sites, kinetic analyses and transforming growth factor-beta1 release.Biochem J322Pt 3809814

37. Kure T, Chang JH, Kato T, Hernandez-Quintela E, Ye H, et al. (2003) Corneal neovascularization after excimer keratectomy wounds in matrilysin-deficient mice. Invest Ophthalmol Vis Sci 44(1): 137–144.T. KureJH ChangT. KatoE. Hernandez-QuintelaH. Ye2003Corneal neovascularization after excimer keratectomy wounds in matrilysin-deficient mice.Invest Ophthalmol Vis Sci441137144

38. Aerts S, Van Loo P, Thijs G, Mayer H, de Martin R, et al. (2005) TOUCAN 2: the all-inclusive open source workbench for regulatory sequence analysis. Nucleic Acids Res 33: W393–396.S. AertsP. Van LooG. ThijsH. MayerR. de Martin2005TOUCAN 2: the all-inclusive open source workbench for regulatory sequence analysis.Nucleic Acids Res33W393396

39. Holnthoner W, Kerenyi M, Groger M, Kratochvill F, Petzelbauer P (2006) Regulation of matrilysin expression in endothelium by fibroblast growth factor-2. Biochem Biophys Res Commun 342(3): 725–733.W. HolnthonerM. KerenyiM. GrogerF. KratochvillP. Petzelbauer2006Regulation of matrilysin expression in endothelium by fibroblast growth factor-2.Biochem Biophys Res Commun3423725733

40. Downes M, Francois M, Ferguson C, Parton RG, Koopman P (2009) Vascular defects in a mouse model of hypotrichosis-lymphedema-telangiectasia syndrome indicate a role for SOX18 in blood vessel maturation. Hum Mol Genet 18(15): 2839–2850.M. DownesM. FrancoisC. FergusonRG PartonP. Koopman2009Vascular defects in a mouse model of hypotrichosis-lymphedema-telangiectasia syndrome indicate a role for SOX18 in blood vessel maturation.Hum Mol Genet181528392850

41. Patterson BC, Sang QA (1997) Angiostatin-converting enzyme activities of human matrilysin (MMP-7) and gelatinase B/type IV collagenase (MMP-9). J Biol Chem 272(46): 28823–28825.BC PattersonQA Sang1997Angiostatin-converting enzyme activities of human matrilysin (MMP-7) and gelatinase B/type IV collagenase (MMP-9).J Biol Chem272462882328825

42. Fuller T, Korff T, Kilian A, Dandekar G, Augustin HG (2003) Forward EphB4 signaling in endothelial cells controls cellular repulsion and segregation from ephrinB2 positive cells. J Cell Sci 116(Pt 12): 2461–2470.T. FullerT. KorffA. KilianG. DandekarHG Augustin2003Forward EphB4 signaling in endothelial cells controls cellular repulsion and segregation from ephrinB2 positive cells.J Cell Sci116Pt 1224612470

43. Taniguchi M, Masuda T, Fukaya M, Kataoka H, Mishina M, et al. (2005) Identification and characterization of a novel member of murine semaphorin family. Genes Cells 10(8): 785–792.M. TaniguchiT. MasudaM. FukayaH. KataokaM. Mishina2005Identification and characterization of a novel member of murine semaphorin family.Genes Cells108785792

44. Neufeld G, Lange T, Varshavsky A, Kessler O (2007) Semaphorin signaling in vascular and tumor biology. Adv Exp Med Biol 600: 118–131.G. NeufeldT. LangeA. VarshavskyO. Kessler2007Semaphorin signaling in vascular and tumor biology.Adv Exp Med Biol600118131