DNA methylation regulates expression of VEGF-R2 (KDR) and VEGF-R3 (FLT4)

Background Vascular Endothelial Growth Factors (VEGFs) and their receptors (VEGF-Rs) are important regulators for angiogenesis and lymphangiogenesis. VEGFs and VEGF-Rs are not only expressed on endothelial cells but also on various subtypes of solid tumors and leukemias contributing to the growth of the malignant cells. This study was performed to examine whether VEGF-R2 (KDR) and VEGF-R3 (FLT4) are regulated by DNA methylation. Methods Real-time (RT) PCR analysis was performed to quantify KDR and FLT4 expression in some ninety leukemia/lymphoma cell lines, human umbilical vein endothelial cells (HUVECs) and dermal microvascular endothelial cells (HDMECs). Western blot analyses and flow cytometric analyses confirmed results at the protein level. After bisulfite conversion of DNA we determined the methylation status of KDR and FLT4 by DNA sequencing and by methylation specific PCR (MSP). Western blot analyses were performed to examine the effect of VEGF-C on p42/44 MAPK activation. Results Expression of KDR and FLT4 was observed in cell lines from various leukemic entities, but not in lymphoma cell lines: 16% (10/62) of the leukemia cell lines expressed KDR, 42% (27/65) were FLT4 positive. None of thirty cell lines representing six lymphoma subtypes showed more than marginal expression of KDR or FLT4. Western blot analyses confirmed KDR and FLT4 protein expression in HDMECs, HUVECs and in cell lines with high VEGF-R mRNA levels. Mature VEGF-C induced p42/44 MAPK activation in the KDR- /FLT4+ cell line OCI-AML1 verifying the model character of this cell line for VEGF-C signal transduction studies. Bisulfite sequencing and MSP revealed that GpG islands in the promoter regions of KDR and FLT4 were unmethylated in HUVECs, HDMECs and KDR+ and FLT4+ cell lines, whereas methylated cell lines did not express these genes. In hypermethylated cell lines, KDR and FLT4 were re-inducible by treatment with the DNA demethylating agent 5-Aza-2'deoxycytidine, confirming epigenetic regulation of both genes. Conclusions Our data show that VEGF-Rs KDR and FLT4 are silenced by DNA methylation. However, if the promoters are unmethylated, other factors (e.g. transactivation factors) determine the extent of KDR and FLT4 expression.


Background
Vascular endothelial growth factors (VEGFs) and their corresponding receptors (VEGF-Rs) are important regulators of angiogenesis and lymphangiogensis. VEGF-A binds VEGF-R1 (FLT1) and VEGF-R2 (KDR). Both tyrosine kinase receptors are expressed on blood vessel endothelial cells. VEGF-C and VEGF-D bind to VEGF-R3 (FLT4) and the fully processed, mature forms also to KDR. FLT4 is primarily expressed on cells of the lymphatic endothelium [1]. VEGFs and VEGF-Rs are important for vessel formation in healthy individuals, but also for tumor angiogenesis [2]. Moreover, the VEGF-Rs are not only expressed on endothelia, but also on different types of solid tumor cells and on leukemic cells [3][4][5][6][7][8][9][10][11]. The interaction of receptors with their ligands mediates survival and can lead to proliferation of the malignant cells [2,12].
Even twenty years after their discovery, little is known about the regulation of the three VEGF-Rs. On the transcriptional level, NF-B and the NF-B target Prox1 have been described as activators of FLT4 in lymphatic endothelial cells [13]. Epigenetic mechanisms contribute to the regulation of FLT1 and KDR but this is not investigated in great detail [14,15].
We set out to test whether DNA methylation is also responsible for the silencing of FLT4. We determined the methylation status of KDR and FLT4 in human umbilical vein endothelial cells (HUVECs), dermal microvascular endothelial cells (HDMECs) and in a large panel of leukemia and lymphoma cell lines. Confirming that expression of KDR and FLT4 is epigenetically regulated, we observed an inverse correlation between promoter methylation and receptor expression. Furthermore, the demethylating agent 5-Aza-2'deoxycytidine (5-Aza-dC) induced expression of KDR and FLT4 in methylated, but not in unmethylated cell lines.

Cell lines and primary cell cultures
The cell lines in this study were taken from the stock of the cell bank (DSMZ-German Collection of Microorganisms and Cell Cultures; http://www.dsmz.de). Detailed references and cultivation protocols have been described previously [16]. Primary HDMECs were purchased from Clonetics/Lonza (Verviers, Belgium). Primary HUVECS (pooled) were purchased from PromoCell (Heidelberg, Germany). HDMECs and HUVECs were cultured in endothelial cell growth medium MV (Promo Cell).

Bisulfite sequencing
To confirm the methylation status of the FLT4 and KDR promoters, genomic DNA was bisulfite converted according to the manufacturer's instructions (Qiagen). Subsequently, amplification of the promoter regions (FLT4: 337 bp; KDR: 612 bp) was performed using BSP primers, specifically binding bisulfite converted DNA (for primer sequence and PCR conditions see MSP section). Resulting PCR products were purified, cloned into the pGEM-TEasy vector (Promega, Madison, WI, USA) and sequenced. Sequences were evaluated using BiQ Analyzer (http://biq-analyzer.bioinf.mpi-sb.mpg.de) and had to conform to at least 90% bisulfite conversion rate. In addition, identical clones were excluded from the analysis.

Gene expression analyses
Quantitative PCR was performed on a 7500 Applied Biosystems (Darmstadt, Germany) real-time PCR system using the manufacturer's protocol. RNA was prepared using the RNeasy Mini kit (Qiagen). This kit includes a DNase digestion step to avoid false positives resulting from contaminating genomic DNA. For mRNA quantification, reverse transcription was performed using the SuperScript II reverse transcriptase kit (Invitrogen, Karlsruhe, Germany). TaqMan probes (Applied Biosystems) were used to quantify human FLT4 (Hs 01047677 m1) and KDR (Hs 00911700 m1) expression levels with TATA box binding protein (TBP) as endogenous control. For interferon gamma inducible protein 10 (IP-10) and tumor necrosis factor alpha (TNFa), SYTO-82 (Molecular Probes, Leiden, Netherlands) was used as fluorescent dye, ImmoMix (Biline, Luckenwalde, Germany) as PCR master mix, and ribosomal protein S9 (RPS9) as endogenous control. The following primers were used: TNFa  Western blot analysis, antibodies, reagents Samples were prepared as described previously [17]. Anti FLT4, ERK and pERK antibodies were purchased from Santa Cruz (Heidelberg, Germany). Anti IB, pIB, KDR, p38 MAPK and pp38 MAPK antisera were obtained from Cell Signalling (New England Biolabs, Frankfurt, Germany). The anti GAPDH monoclonal antibody (mAb) was purchased from Abcam (Cambridge, UK). Specific bands on nitrocellulose membranes were visualized with the biotin/streptavidin-horseradish peroxidase system (Amersham, Freiburg, Germany) in combination with the "Renaissance Western Blot Chemoluminescence Reagent" protocol (Perkin Elmer, Waltham, MA, USA). Synthetic macrophage activating factor of 2 kDa molecular mass (MALP-2) was a gift from P. Mühlradt to H. Weich. The preparation was free of endotoxin.

Analysis of FLT4 and KDR protein expression by flow cytometry
For detection of CD31 (Becton Dickinson Biosciences, Heidelberg, Germany), FLT4 (R&D Systems, Wiesbaden, Germany), KDR (Reliatech, Wolfenbüttel, Germany) and podoplanin (Reliatech) on the cell surface, cells were washed and incubated with the mouse mAb or with the isotope-matched control mouse immunoglobulin (BD Biosciences) for 30 min at 4°C. Subsequently, cells were treated with FITC conjugated anti-mouse secondary Ab (Biozol, Eching, Germany) and propidium iodide. Labeled cells were analyzed on a FACSCalibur (BD Biosciences) using CellQuest Pro software.
Treatment with DNA demethylating agent 5-Aza-2'deoxycytidine (5-Aza-dC) 5-Aza-dC (Sigma Aldrich, Taufkirchen, Germany) dissolved in DMSO was used to verify the effect of methylation on expression of FLT4 and KDR. Cells were seeded at a cell density of 5 × 10 5 cells/ml, 5-Aza-dC was added at a final concentration of 5 μM. Control cells were treated with 0.05% DMSO. After 2 d, half of the medium was replenished with medium with/without 5-Aza-dC (5 μM). After 3 d, cells were harvested to prepare RNA and protein.

Results and Discussion
Expression of KDR and FLT4 in leukemia and lymphoma cell lines The VEGF-Rs KDR (VEGF-R2) and FLT4 (VEGF-R3) are not only expressed on blood endothelial and lymphendothelial cells, but also on solid tumors and leukemias. Leukemia-derived VEGFs may induce the growth of leukemic cells in an autocrine or paracrine fashion [7,10,18,19]. The promoters of VEGF-Rs and their ligands contain CpG islands, regulatory regions that are typically methylated in epigenetically silenced genes [14]. Recent reports show that expression of FLT1 and KDR are controlled by promoter methylation [14,15]. However, only a limited number of leukemia and lymphoma cell lines have been tested for VEGF-R expression and promoter methylation hitherto.

OCI-AML1: a model system for VEGF-C induced cell signaling
Cytokine-dependent cell lines have often and successfully been used as model systems for signal transduction studies. In contrast to primary cells, no contaminating cell fraction effects "false" signals in cell lines, and in contrast to cytokine-independently growing cell lines, cytokine starvation silences the relevant enzymes in cytokinedependent cell lines. We chose cell line OCI-AML1 as this was the only cytokine dependent, FLT4 positive cell line tested ( Table 2). The cytokine response profile of this cell line has been published previously [20]. Cell line OCI-AML1 did not show a proliferative response on VEGF-C (data not shown). However, short-term (5 min) stimulation with VEGF-C induced phosphorylation of ERK1/2 ( Figure 2). Preincubation with the FLT4 inhibitor MAZ51 inhibited this effect, confirming the specificity of the VEGF-C induced EKR1/2 activation (Figure 2). ERK1/2 phosphorylation was tested because the p42/44 MAPK pathway is a known FLT4 target [21,22]. The results of cell signaling experiments shown in Figure 2 confirm that cell line OCI-AML1 is a model system for FLT4 signaling, expecially as KDR, the second receptor for VEGF-C is not expressed in this cell line (Table 2).

KDR: promoter methylation and gene expression
To test whether KDR is epigenetically regulated, we performed bisulfite sequencing of KDR negative and positive cell lines and of primary endothelial cells. The KDR negative cell line DOHH-2 had a highly methylated KDR promoter, the KDR positive cell line HEL was nearly unmethylated (Figure 3). Largely unmethylated were also HDMECs and HUVECs, both expressing KDR (Figure 3). To assess the KDR methylation status for a larger number of cell lines, we performed methylationspecific PCR (MSP), a technique less costly and laborious than bisulfite sequencing. The majority of KDR negative cell lines were methylated, KDR positive HUVECs were unmethylated (Figure 4). However, even HDMECs were U-and M-PCR positive although they expressed high KDR levels and although only a small minority of clones were methylated according to sequencing analysis (Figures 3 and 4). Apparently, a low proportion of methylated CpGs was sufficient to yield signals in the M-PCR. The same was true for U-PCR: the KDR negative cell line DOHH-2 -highly methylated according to the results of bisulfite sequencing-showed signals in M-and in U-PCR ( Figure 3, Table 2).
In spite of the high sensitivity-a certain drawback of the PCR-based MSP technique-the accuracy of KDR M-PCR was 88% supporting the notion that KDR expression is regulated by DNA methylation (Table 2).

FLT4: promoter methylation and gene expression
Bisulfite sequencing and BSP analysis were also performed to analyze the methylation status of FLT4 in cell lines, HUVECs and HDMECs. Results of bisulfite sequencing showed that FLT4 was largely methylated in the FLT4 negative cell line EM-2 and unmethylated in the FLT4 positive cell line SUP-B15 as it was in HUVECs and HDMECs ( Figure 5). MSP analysis confirmed that FLT4 exhibited the inverse correlation   between promoter methylation and gene expression that is indicative for epigenetic regulation (Figure 4, Table 2). However, the accuracy of FLT4 M-PCR (80%) was lower than for KDR M-PCR (88%). Of note was also that TF-1 cells did not express FLT4 although the promoter was unmethylated ( Table 2). These data suggested that regulatory mechanisms other than DNA methylation are also important for the regulation of FLT4.

Effect of DNA demethylating agent 5-Aza-dC on expression of KDR and FLT4
To test whether KDR and FLT4 were silenced by promoter methylation, we treated methylated and unmethylated cell lines with the DNA demethylating agent 5-Aza-dC. In 4/5 KDR-negative cell lines, expression of KDR was induced by DNA demethylation (Table 3). FLT4 expression was upregulated in 4/4 negative cell lines (Table 3). KDR and FLT4 expression in positive cell lines were not affected (Table 3). Although these results confirmed that promoter methylation plays a role for the regulation of these VEGF-Rs, we also noted substantial differences in the levels of 5-Aza-dC-triggered gene induction between different cell lines (Table  3). Furthermore, even in the most sensitive cell lines (HL-60 for KDR induction, EM-2 for FLT4 induction), demethylation did not induce mRNA expression that would translate into protein levels detectable by Western blot analysis (data not shown). These results suggest that other mechanisms than DNA methylation are also involved in the regulation of KDR and FLT4.
Besides DNA methylation, also histone modifications are epigenetic mechanisms that affect the expression of individual genes. Just to mention two examples, acetylated histone H3 (at lysine 9 and 14) is a marker for gene activation [23], tri-methylation of histone H3 lysine 27 stands for gene suppression [24]. Furthermore, epigenetic modifications can influence each other: methylated CpGs in a promoter region can be targeted by proteins that interact with histone deacetylases. The consequence is an inactive chromatin status and transcriptional repression [25,26].  However, besides epigenetic mechanisms, also the presence or absence of trans-acting factors may govern the expression of KDR and FLT4. Thus, it has been shown that transcription factor binding sites (Sp1, AP-2 and NFB) are essential for the base-line activity of the KDR promoter [27]. Here, we set out to find whether NFB also plays a role for the expression of FLT4.

Influence of transactivating factors
During inflammation, new lymphatic vessels are formed. NF-B is a key mediator of inflammatory processes and has recently been identified as inducer of FLT4 on lymphatic endothelial cells [13]. To test whether NF-B contributes to FLT4 expression in leukemic cells, we stimulated the FLT4 negative cell line EM-2 and the FLT4 positive cell line OCI-AML1 with synthetic MALP-2. MALP-2 binds to toll-like receptors-2 and -6 [28]. MALP-2 triggers the NF-B pathway [29] which leads to the expression of NF-B targets like TNFa [30].
Accordingly, MALP-2 (100 ng/ml, 7 min) induced phosphorylation and degradation of the NF-B inhibitor IB and stimulated phosphorylation of p38 in cell lines EM-2 and OCI-AML1 ( Figure 6). MALP-2 (100 ng/ml, 1 h) triggered expression of the NK-B targets TNFa (80× in EM-2, 1000× in OCI-AML1) and IP-10 (600× in EM-2, > 1000× in OCI-AML1) in both cell lines. However, the expression of FLT4 was not affected, neither in the FLT4 positive cell line OCI-AML1 nor in the FLT4 negative (methylated) cell line EM-2. MALP-2 did also not increase the FLT4 stimulating effect of 5-Aza-dC on EM-2 cells (data not shown). Thus, our results do not support the view that NF-B is a transactivator of FLT4. We observed a 10-fold increase in KDR in cell lines OCI-AML1 and EM-2 (the latter pretreated with 5-Aza-dC). However, as the level reached after stimulation was still extremely low, it appears unlikey that NF-B is an important regulator for KDR either.   Table 2. NTC, non template control.

KDR and FLT4 in HDMECs and HUVECs
HDMECs, HUVECs and KDR positive leukemia cell lines exhibited demethylated KDR promoters ( Figure 3, Table 2). However, mRNA and protein levels were distinctly higher in the primary cells than in the leukemia cell lines (Figure 1). Results of Western blot analysis were confirmed by flow cytometry (Figure 7). HDMECs and HUVECs expressed the pan-endothelial marker CD31 (Figure 7). HDMECs, primarily consisting of lymphatic endothelial cells, were also positive for the  Induction of KDR and FLT4 by 5-Aza-dC (5 μM, 3 d) when compared to untreated control cells: -< 2.5-fold; (+) > 2.5-fold; + > 10-fold; ++ > 40-fold; + ++ > 160-fold. Expression levels were determined by qRT-PCR. TBP expression was used as endogenous control, DMSO (0.05%) treated cells were used for normalization Figure 6 Degradation of IB and phosphorylation of p38MAPK in MALP-2 responsive cell lines. Cell lines EM-2 (FLT4 negative) and OCI-AML1 (FLT4 positive) were stimulated with MALP-2 (100 ng/ml). The NF-B inhibitor IB was phosphorylated (7 min) and degraded (7 min, 21 min) in both cell lines. Likewise, rapid phosphorylation of p38 MAPK was observed in both cell lines by Western blot analysis. GAPDH is shown as loading control.
lymphatic vessel marker podoplanin, HUVECs were podoplanin negative (Figure 7). Both types of primary cells expressed much higher levels of KDR than KDRpositive cell lines (Figure 1). Also FLT4 expression levels varied greatly from one cell type to the other: FLT4 expression of HUVECs was comparable to those of FLT4 positive cell lines, while HDMECs showed much higher FLT4 expression levels (Figures 1 and 7). These results are in line with our data of MSP analyses and 5-Aza-dC experiments suggesting that DNA methylation Figure 7 Expression of KDR and FLT4 on HDMECs and HUVECs. Flow cytometry analysis for KDR, FLT4, the lymphatic vessel marker podoplanin and the panendothelial marker CD31. Note that HDMECs and HUVECs show comparable KDR expression levels, while FLT4 is stronger in lymphatic vessel cells (HDMECs) than in blood vessel cells (HUCECs). The non-filled peak shows the corresponding isotype control. is not the only mechanism that controls KDR and FLT4 gene expression.

Conclusions
Our data obtained from primary endothelial cells and from leukemia/lymphoma cell lines show that KDR and FLT4 are epigenetically regulated genes. Both genes can be silenced by methylation. However, if the promoters are unmethylated, other factors are responsible for the extent of KDR and FLT4 expression. Furthermore, we show that the KDR negative/FLT4 positive cell line OCI-AML1 is a model system for FLT4 signal transduction studies.