Early transcriptional pattern of angiogenesis induced by EGCG treatment in cervical tumour cells

Abstract The major green tea polyphenol (-)-epigallocatechin-3-gallate (EGCG) has been shown to exhibit antitumour activities in several tumour models. One of the possible mechanisms by which EGCG can inhibit cancer progression is through the modulation of angiogenesis signalling cascade. The tumour cells’ ability to tightly adhere to endothelium is a very important process in the metastatic process, because once disseminated into the bloodstream the tumour cells must re-establish adhesive connections to endothelium in order to extravasate into the target tissues. In this study, we investigated the anti-angiogenic effects of EGCG treatment (10 μM) on human cervical tumour cells (HeLa) by evaluating the changes in the expression pattern of 84 genes known to be involved in the angiogenesis process. Transcriptional analysis revealed 11 genes to be differentially expressed and was further validated by measuring the induced biological effects. Our results show that EGCG treatment not only leads to the down-regulation of genes involved in the stimulation of proliferation, adhesion and motility as well as invasion processes, but also to the up-regulation of several genes known to have antagonist effects. We observed reduced proliferation rates, adhesion and spreading ability as well as invasiveness of HeLa tumour cells upon treatment, which suggest that EGCG might be an important anti-angiogenic therapeutic approach in cervical cancers.


Introduction
In 2007, cervical cancer was reported to be the second most common female cancer worldwide, being the third cause of female cancer mortality annually [1]. Although in the developed countries cervical cancer incidence and mortality have decreased, this disease still remains a serious problem with high estimated incidence and mortality [2].
Cervical cancer can be prevented and if detected early it is generally curable [3]. However, the treatments for metastatic or recurrent cervical carcinoma are poorly effective and with serious adverse effects. There is a great need to identify and investigate new systemic acting agents for cervical cancer treatment.
Lately, the research has been focused on natural compounds; several dietary components show prevention and therapeutic proprieties in the pathogenesis of cancer. The use of dietary compounds for cancer prevention and therapy could be of major importance because in addition to the diverse biological effects, they have low toxicity and fewer adverse effects than traditional chemotherapeutic agents.
Green tea has received much attention, particularly its major component (-)-epigallocatechin-3-gallate (EGCG). EGCG has been shown to possess remarkable cancer chemopreventive and therapeutic potentials against various cancers, by modulating cell proliferation and apoptosis both in vitro and in vivo [4][5][6][7][8]. In addition, recent work revealed that EGCG could interfere with cell signalling pathways of angiogenesis, metastasis and migration in prostate, liver and breast cancer cells [9][10][11].
Angiogenesis is a central process in cervical carcinogenesis and progression [12]. Cervical tumour's ability to develop vasculature in order to respond to the metabolic needs is essential to its progression. The antiproliferative effect of EGCG in cervical cancer has been intensively investigated [13][14][15][16][17], but little is known about its anti-angiogenic potential [18].
With these considerations, in this study we report the effect of EGCG on the angiogenic potential of HeLa cervical adenocarcinoma cells, specifically the molecular transcription pattern induced by EGCG treatment. Tumour invasiveness properties including cell morphology change, in vitro invasion, spreading and adhesion to extracellular matrix (ECM) and endothelium were evaluated in order to validate our findings.

EGCG treatment
A total of 5 ϫ 10 5 HeLa cells per well were seeded in 6-well plates in Opti-Mem medium (Gibco, Invitrogen, Carlsbad, CA) and treated with EGCG (Sigma-Aldrich) to a final concentration of 10 M for 24 and 48 hrs. Control cells (UT) were treated with phosphate-buffered saline (PBS) instead of EGCG. All our experiments were conducted under serum starvation conditions in order to synchronize the cells and stimulate growth factor production.

Total RNA isolation
Total RNA was isolated with TriReagent (Sigma-Aldrich), purified with RNAeasy Mini Kit (Qiagen Inc., Valencia, CA, USA) and further analysed for quantity and quality with NanoDrop ND-1000 and Agilent Lab-on-a-chip Bioanalyzer 2100 (Agilent Technology Inc., Santa Clara, CA, USA). All the RNAs presented a RNA Integrity Number (RIN) between 9 and 10.

PCR array
Eighty-four genes involved in angiogenesis modulation were simultaneously evaluated using the Human Angiogenesis RT 2 Profiler™ PCR Array

Attachment assay
Twenty-four and -48 hours after treatment, HeLa cells were harvested with 0.25% trypsin-ethylenediaminetetraacetic acid, followed by centrifugation at 1000 rpm for 5 min. at room temperature (RT). The cell pellets were suspended in serum-free DMEM medium at a density of 2.5 ϫ 10 5 cells/ml. A total of 100 l of cell suspension (25,000 cells/well) were seeded in a 96-well plate which was pre-treated with 15 g/ml type IV collagen (Sigma-Aldrich) or laminin (Sigma-Aldrich) overnight at 4ЊC, rinsed with PBS, followed by blocking with 0.2% bovine serum albumin (BSA) at 37ЊC for 45 min. Cells were allowed to adhere for 20-30 min. in a cell culture incubator and non-attached cells were removed by gently washing with PBS. Attached cells were fixed with 4% formaldehyde for 30 min. at RT, followed by staining in 0.1% (w/v) crystal violet for 30 min. Stained cells were lysed in 1% SDS and the intensity of stain, which is proportional to the number of adherent cells, was quantitated by using a microtitre plate reader at the absorbance of 570 nm.

Spreading assay
HeLa cells were harvested with 0.25% trypsin at 24 and 48 hrs after treatment, centrifuged at 1000 rpm for 5 min., and resuspended at a population density of 5 ϫ 10 4 cells/ml in warm DMEM pre-equilibrated with 5% CO2. A total of 100 l of cell suspension were plated onto a 96-well pre-treated plate coated with 15 g/ml type IV collagen (Sigma-Aldrich) or laminin (Sigma-Aldrich) overnight at 4ЊC, washed with PBS and blocked with 0.2% BSA for 45 min. at 37ЊC. Cells were allowed to spread for 20 hrs at 37ЊC in a cell culture incubator. Cells were washed with PBS and fixed in 3% glutaraldehyde for 30 min. at RT. Cell images were taken by an Axiovert Zeiss (Carl Zeiss Instruments, Bucharest, Romania) microscope with 100ϫ magnification. Spreading cells were defined as cells with extended processes, and non-spreading cells were defined as round cells. After image acquisition, the cells were stained for 30 min. with Giemsa solution, extensively washed and lysed in 1% SDS. The intensity of stain was quantified at 630 nm with a microtitre plate reader.

Adhesion to endothelium assay
Confluent monolayers of HUVEC cells were cultured on plastic cover slips which were pre-treated with 1% gelatine for 30 min. After 24 hrs of culture the medium was removed and 5 ϫ 10 5 HeLa cells treated as previously described were added to each well and incubated for 50 min. at 37ЊC. After incubation the non-adherent HeLa cells were thoroughly washed with PBS-Alb, fixed in 4% paraformaldehyde for 15 min. at RT and stained with Giemsa solution for 5 min. The cover slips were washed with PBS, allowed to air dry and mounted on a microscope slide. Using light microscopy the number of cells adherent to the monolayer were counted in five fields of view per well selected at random.

In vitro invasion assay
The effect of EGCG treatment on the invasive properties of cervical cancer cells was determined using the matrigel transwell migration assay in co-culturing conditions. HeLa and HUVEC cells were trypsinized and resuspended in Opti-Mem and RPMI-1640 medium, respectively. Both cell types were treated with EGCG as described. A total of 1 ϫ 10 5 HeLa cells were plated in the top chamber of the cell culture inserts (6.5 mm diameter insert, 3.0 m pore size; Millipore, Milford, MA, USA) pretreated with 1:10 diluted Matrigel BD (BD Biosciences, Franklin Lakes, NJ, USA) and 3 ϫ 10 5 HUVEC cells were plated onto the bottom well of the chamber also pre-treated with matrigel. A total of 10% FBS was added in the bottom chamber as a chemoattractant. After incubation for 24 and 48 hrs, the cell inserts were removed from the plate and cells that did not migrate were mechanically removed with a cotton swab. Cells on the lower surface of the membrane were fixed in ice cold methanol; the membranes were cut from the housing insert and mounted on microscope slide with 4Ј,6-diamidino-2-phenylindole (DAPI) mounting medium. The migrated cells were examined using inverted phase fluorescence Zeiss Axiovert microscope and counted from six randomly selected fields in a blind way.

Statistical analysis
All experiments were performed in triplicates. All the data are presented as mean Ϯ standard error of mean (S.E.M.). Differences were assessed by parametric and nonparametric tests according to data distribution using the statistical program SPSS 9.0 for Windows. P Ͻ 0.05 was considered to be statistically significant (*P Ͻ 0.05, **P Ͻ 0.01, ***P Ͻ 0.001).

Effect of EGCG on the angiogenic potential of HeLa cells
The total RNA extracted from treated and untreated cells was profiled for 84 genes known to be involved in regulating the angiogenic process. The array includes growth factors and receptors, adhesion molecules, cytokines and chemokines, proteases and matrix proteins. Five housekeeping genes (B2M, HPRT1, RPL13A, GAPDH and ACTB) were used for well to well normalization. Data analysis was done by ⌬⌬Ct method using the Superarray Data Analysis Web Portal. We have considered of interest all the genes with Ϫ1.5 Յ fold regulation Ն 1.5 in order to assess the genes involved in the early anti-angiogenic process. We found significant expression differences (P Ͻ 0.05) for 11 out of the 84 investigated genes. Four genes were up-regulated and seven down-regulated ( Table 1).
Three of these genes were validated at protein level by ELISA method. The results show the same expression pattern for the investigated proteins as their mRNA: we found down-regulated levels of PDGFA and TGF-␤2 proteins and up-regulated levels of IL-1␤ (Fig. 1).

Gene
Fold regulation P value

Effect of EGCG on HeLa cells proliferation
We examined the effect of EGCG on HeLa cells proliferation 24 and 48 hrs after treatment by MTT test. At both time-points, the cell proliferation was found to be significantly inhibited (Fig. 2). EGCG induced a 45% inhibition effect on the proliferation rate measured after 48 hrs of treatment compared to untreated cells; moreover, there was no significant difference between the number of cells evaluated at 24 hrs compared to cells treated for 48 hrs.

Effect of EGCG on cell attachment and spreading
We evaluated the influence of EGCG on cellular adhesion and motility on type IV collagen and laminin coated plates (Figs 3 and 5), two major components of ECM. Our results show (Fig. 3) that after 48 hrs of EGCG treatment the attachment of HeLa cells to type IV collagen was stimulated, but had no effect on the adhesion to laminin. There were no differences in cells attachment observed at 24 hrs after treatment (data not shown). In contrast, if the cells  were allowed to attach for 20 hrs, the adhesion was dramatically inhibited (Fig. 4). Untreated cells presented multiple filopodia and lamellipodia, characteristic for spreading cells, whereas most of the EGCG treated cells maintained their round shape even after 20 hrs, indicating reduced spreading ability (Fig. 5).

Attachment to endothelium assay
In order to quantify the adhesion of the HeLa cells to endothelial HUVEC cells, HeLa cells were treated with EGCG and harvested at 24 and 48 hrs after treatment. The cells were seeded onto the HUVEC cell monolayers, and co-cultured for 50 min. After removing the non-adherent cells, the remaining adherent cells were stained and counted. As shown in Figure 6, the EGCG treatment reduced the adherence of HeLa cells to endothelial cells by 40%.

Inhibition of cell migration by EGCG
Migration towards a chemoattractant is a distinct cellular phenotype of metastatic tumour cells, and it is an essential step for tumour Fig. 3 Adhesion of HeLa cells to collagen and laminin after EGCG treatment. Forty-eight hours after treatment, cells were harvested, resuspended and seeded in a 96-well plate which was pre-treated with 15 g/ml type IV collagen or laminin. Cells were allowed to adhere for 20 min. and non-attached cells were removed by washing. Attached cells were fixed, stained with crystal violet, followed by measuring the absorbance at 570 nm. invasion and metastasis. Because cervical cancer is associated with more aggressive tumour phenotypes, we examined the effect of EGCG on the migration ability of HeLa cells using an in vitro migration assay, which simulates the in vivo metastatic process. As shown in Figure 7, there was a dramatic inhibition of HeLa cells' migration ability after the treatment with EGCG. Compared with the untreated cells, EGCG caused an average of 48% and 68% reduction of migration ability after 24 and 48 hrs, respectively.

Discussions
Several studies describe the beneficial effects of green tea consumption, many of them pointing potential advantages compared to traditional cancer drugs like availability, and low toxicity to healthy cells. These observations led to the investigation of several compounds found in green tea in clinical trials. Currently, there are 51 ongoing clinical trials studying the effects of EGCG on different pathologies, including one on cervical cancer (information available at www.clinicaltrials.gov; accessed 12 March, 2011). To date, the precise molecular mechanism of EGCG anticancer effects remain unclear [8]. In this study, we report the transcriptional mechanism that might modulate the anti-angiogenic and anti-metastatic effects of EGCG on HeLa cervical cells.
By profiling 84 genes involved in angiogenesis modulation we investigated the early molecular transcriptional mechanisms induced by EGCG treatment. The transcriptional analysis showed that EGCG treatment modulates the transcription of several genes involved in the angiogenic process (Table 1). These genes are known to mediate multiple mechanisms like endothelial cell  The anti-proliferative effect of EGCG has been reported by several groups [13][14][15][16][17]. Cell proliferation and growth is promoted through various signalling pathways and EGCG seems to interfere with these pathways and lead to decreased proliferation rates in cervical cancer cells (Fig. 2).
PDGFA is a growth factor that upon binding to their receptors activates a variety of intracellular signalling molecules that control cell growth, proliferation and differentiation, leading to cancer progression. PDGF isoforms act mainly through a paracrine mechanism, their expression promoting the establishment of a well-vascularized and prominent stroma, which in turn provides growth factors and anti-apoptotic signals for tumour cells, support for tumour angiogenesis and facilitate invasion [19,20]. Pietras et al. [21] have recently showed in a mouse model of cervical carcinoma that PDGFA stimulates the stromal components and impairment of the PDGF/PDGFR signalling with imatinib affected the angiogenic phenotype of both premalignant cervical neoplasias and invasive carcinomas; the treated lesions exhibited diminished blood vessel density and reduced pericyte coverage.
A considerable number of solid tumours [22,23] including cervical cancer [24] have been showed to express PDGF receptors which suggest that these molecules also act by an autocrine mechanism leading to increased proliferation rates. We found that EGCG down-regulates the expression of both mRNA and protein levels of PDGFA and this could be one of the mechanisms by which EGCG decreases the proliferation rate as well as the angiogenic potential of cervical cancer cells.
IFN-␤1 antitumour activity has been intensively investigated in human cancers. Studies have shown that its expression leads to immune stimulation [25,26], tumour regression [27,28] and angiogenesis inhibition by down-regulation of proangiogenic factors like bFGF in vivo [29][30][31] and blocking endothelial cell migration in vitro. Vannucchi et al. [32] reported that IFN-␤1 induces a direct anti-proliferative effect on cervical tumour cells by inhibiting the cell cycle progression. This report seems to be in concordance with our study. We observed that the number of cells after 24 and 48 hrs of treatment remained the same and we assume that EGCG treatment decreases the proliferation rate by stopping the cell cycle progression through induced IFN-␤1 expression.
ID genes are regulators of transcription being responsible for changes in gene expression that lead to increased growth, invasion and metastasis in several human carcinomas [33]. The precise role of ID1 in cervical cancer is little investigated; however, several studies correlated increased expression of this protein with higher tumour cell aggressiveness, poor prognosis and survival in early stages [34,35]. Lyden et al. showed that ID proteins are required for the proliferative and invasive phenotype of endothelial cells during tumour-associated angiogenesis [33]; however, there was no association between ID1 expression and angiogenesis in cervical cancer, which suggest that ID1 expression on prognosis cannot be attributed to its proangiogenic effect alone [34]. Another study reported that ID genes are strongly responsive to signalling pathways coupled to growth factors [36]. ID proteins enhance cell cycle progression and their overexpression induces apoptosis in serum-deprived fibroblasts [37]. The exact mechanism of action of EGCG through ID1 gene up-regulation in HeLa cervical cells is still to be determined, ID1 overexpression might be induced as a resistance mechanism against EGCG Fig. 7 Inhibition of HeLa cells migration by EGCG. HeLa cells were trypsinized and plated in the top chamber of the transwell. HUVEC cells cultured in RPMI-1640 medium with 10% FBS was added in the lower chamber as a chemoattractant. After incubation for 24 and 48 hrs, cells that did not migrate through the pores were mechanically removed and cells on the lower surface of the membrane were fixed and stained. The number of migrated cells was counted from six randomly selected fields per well in a blind way. treatment in order to progress the cell cycle or the serum withdrawal coupled with ID1 expression might lead to cell apoptosis as reported for fibroblasts.
IL-1␤ seems to engage different cellular signalling pathways depending on the cell type in a dose-dependent manner, leading to genotoxic damage, cell apoptosis or cell growth [38]. Breast cancer cells under genotoxic stress secreting moderate levels of IL-1␤ stimulated clonal expansion, whereas high levels of IL-1␤ engaged the apoptotic pathway. In angiogenesis, TNF-␣ and IL-1␤ induce increased TNFAIP2 mRNA levels [39,40] during capillary tube formation in vitro [39]. In small cell lung cancer IL-1␤ up-regulates CXCL6 production and increases tumour cell proliferation [41]. CXCL6 overexpression has been associated with increased metastatic phenotype in small cell lung [41] and prostate [42] cancers. In prostate cancer, CXCL6 overexpression was also associated with increased cell adhesion to ECM and endothelium [42]. Roomi et al. reported recently that EGCG and IL-1␤ stimulation on HeLa cells leads to decreased matrix metalloproteinase 2 (MMP-2) secretion [43], a crucial matrix metaloprotease in cell invasion. Our results show that EGCG treatment seems to stimulate the transcription of IL-1␤ and decrease TNFAIP2 and CXCL6 levels.
In addition to decreased proliferation rate, EGCG treatment seems to also modulate the adhesion capability of HeLa cells to ECM and endothelium. The attachment test to ECM showed no effect on the rapid adhesion of HeLa cells to laminin at either timepoints, but showed increased adhesion to type IV collagen after 48 hrs of treatment (Fig. 3). Cell motility is the key step in organ invasion by tumour cells and increased adherence to the ECM is the first process that indicates increased spreading ability and subsequently metastatic potential of a tumour cell. However, if the time of incubation was extended, the cells lost both their adherence as well as spreading potential (Fig. 4). We observed significant morphology change of the cervical cancer cells after the treatment with EGCG, indicating reduced motility (Fig. 5). These observations might be explained by the fact that even if cells establish initial focal complexes with the ECM, they fail to mature into stable focal adhesions, a process that seems to be impaired by the EGCG treatment. Moreover, we further investigated HeLa cells adherence properties to an endothelial cell monolayer. Re-establishment of adhesive connections to endothelium after entering the bloodstream is a key step in tumour metastasis [44]. Our results show that EGCG treatment reduces the adherence of HeLa cells to endothelial cells (Fig. 6).
Adherence of tumour cells to ECM and endothelium is mainly mediated through integrin pathways. THBS-1 has been shown to be involved in modulating these pathways and affect tumour cell adhesion [45], migration and invasion [46]. THBS-1 has been proposed to have both pro-and anti-metastatic properties [46,47] by regulating epithelial cell growth, motility [46], stromal/epithelial interactions and angiogenesis. THBS-1 has been shown to be a potent inhibitor of angiogenesis by multiple mechanisms, including direct interaction with vascular endothelial growth factor (VEGF) and inhibition of MMP-9 activation [47,48]. However, recent studies reported different biological action in tumour progression and metastasis [46,47]. Overall, the effects of THBS-1 on any given tumour's metastatic potential probably depends on genetic and/or epigenetic changes in the tumour cells themselves, as well as its anti-angiogenic effects on endothelial cells [49].
Although generally known for its anti-proliferative effects, TGF-␤ is a well-characterized inducer of epithelial-mesenchymal transformation in tumour cells, resulting in enhanced cell migration and invasion. TGF-␤ controls the transcription of numerous integrins and although the down-regulation of integrin expression has been reported, in most cases TGF-␤ stimulates integrin expression in several cell types and tissues, as well as in various human cancers (rev. in [59]). We assume that EGCG treatment down-regulates the THBS-1 expression (fold regulation -3.2), which in turn downregulates its downstream target, TGF-␤2 (fold regulation -1.67) and subsequently integrin secretion leading to decreased tumour cell adhesion, motility and invasion (Fig. 7).
It is well known that VEGF can regulate the proliferation and migration of both endothelial and tumour cells. EGCG has been shown to induce VEGF down-regulation in different cancers [60][61][62][63], including in HeLa cervical cancer cells [18]. Although we did not find decreased VEGF mRNA levels upon EGCG treatment (probably due to the different experimental conditions) we found decreased levels of EFNA1. EFNA1 is a prototypic ligand which bounds EphA2 receptor tyrosine kinase and triggers downstream signals that regulate cell growth [64,65] and migration [66,67]. EFNA1 and its receptor are known to be involved in the VEGF pathway by up-regulating VEGF expression in tumour cells and subsequently activate endothelial cells [68]. It seems that EGCG treatment down-regulates EFNA1 expression which might lead to VEGF inhibition as reported by Zhang et al. [18]. Concomitant increased expression of VEGF with EFNA1 was observed in patients with cervical cancer; following radiation therapy [69] and blocking EFNA1 signalling cascade decreased tumour-associated angiogenesis and consequently, tumour progression [70,71].
Recently, Galaup et al. [72] showed that ANGPTL4 has an important role in the metastatic process; increased levels of ANGPTL4 inhibited both vascular activity and tumour cell motility and invasiveness. ANGPTL4's role in angiogenesis is controversial suggesting a tissue specific activity, being induced in both endothelial cells [73,74] and tumour cells [75]. Depending on the context, ANGPTL4 has been reported to exert complex biological function, acting as a proangiogenic factor [74,76] by mediating endothelial cell survival [77] or having anti-angiogenic properties [78] by modulating endothelial cell adhesion [79]. It seems that in HeLa cells the ANGPTL4 up-regulation by EGCG treatment induce an anti-angiogenic and anti-metastatic effect.
EGCG treatment seems to mediate the angiogenic phenotype in HeLa cervical cells also through CCL2 inhibition. In human cervical cancer, CCL2 is abundantly expressed in the stroma and associated with macrophage infiltration [80]. CCL2 induces the migration and activation of monocytes, and it has been implicated in the recruitment of tumour-infiltrating macrophages (TIMs) in several tumour types [81,82]. Previous studies showed that TIMs are a major source of MMP-9 [83] as well as a broad range of tumorigenic factors [84] promoting angiogenesis, tumour progression and metastasis. Zijlmans et al. [85] showed that cervical tumours that do not express CCL2 are characterized by a decreased number of TIMs, smaller tumour size and reduced vascular invasion as well as increased overall survival.

Conclusions
In this study, we explored the early transcriptional pattern of angiogenesis induced by EGCG treatment in cervical cancer cells. Our data show that EGCG treatment modulates the angiogenic pathways by interfering with the transcription pattern of several genes that act both on endothelial cells and tumour cells. EGCG treatment decreases the tumour cell proliferation rate, inhibits cell adhesion, motility and invasion potential of HeLa cells. Our results show that EGCG could be used as an anti-angiogenic agent in cervical adenocarcinomas treatment [86].