RNAi Mediated Tiam1 Gene Knockdown Inhibits Invasion of Retinoblastoma

T lymphoma invasion and metastasis protein (Tiam1) is up-regulated in variety of cancers and its expression level is related to metastatic potential of the type of cancer. Earlier, Tiam1 was shown to be overexpressed in retinoblastoma (RB) and we hypothesized that it was involved in invasiveness of RB. This was tested by silencing Tiam1 in RB cell lines (Y79 and Weri-Rb1) using siRNA pool, targeting different regions of Tiam1 mRNA. The cDNA microarray of Tiam1 silenced cells showed gene regulations altered by Tiam1 were predominantly on the actin cytoskeleton interacting proteins, apoptotic initiators and tumorogenic potential targets. The silenced phenotype resulted in decreased growth and increased apoptosis with non-invasive characteristics. Transfection of full length and N-terminal truncated construct (C1199) clearly revealed membrane localization of Tiam1 and not in the case of C580 construct. F-actin staining showed the interaction of Tiam1 with actin in the membrane edges that leads to ruffling, and also imparts varying invasive potential to the cell. The results obtained from our study show for the first time that Tiam1 modulates the cell invasion, mediated by actin cytoskeleton remodeling in RB.


Introduction
Retinoblastoma (RB) is an intraocular tumor of childhood. As per International Retinoblastoma Staging Working Group (IRSWG) most of the RB tumors are found to have massive choroidal, optic nerve, and anterior segment invasion [1]. The risk factors include choroidal invasion .3 mm (CI.3 mm), post laminar and surgical end of optic nerve invasion. Understanding the molecular regulation of tumor cell invasion and apoptosis helps in identifying new therapeutic targets.
T lymphoma invasion and metastasis protein (Tiam1) was first identified as an invasion and metastasis inducing gene using T lymphoma cells by proviral tagging and in vitro selection for invasiveness [2,3]. Tiam1 is a guanine nucleotide exchange factor (GEF) that mediates the specific activation of Rac1 [4,5,6]. Small guanine triphosphate (GTP) binding proteins belonging to Ras superfamily act as molecular switches for activation of cellular activities such as signal transduction, actin cytoskeleton remodeling, microtubule stabilization, centrosome reorganization and intracellular trafficking [7,8,9,10]. Aberrations or mutations of these proteins lead to malignancy of the cell. In response to extracellular signals, GEFs play a major role by catalyzing the activation of GTP-binding proteins by dissociation of guanosine diphosphate bound to it. RhoA, Rac1 and Cdc42 are key proteins of Rho family that depends on GEFs for their acitivation [11,12].
Tiam1 has been linked with cancer progression and having growth promoting functions based on the tumor type. Overexpression of N-terminus truncated Tiam1 is found to impart oncogenic activity in NIH 3T3 cells [13,14]. Similarly, mutations in Tiam1 gene are able to transform NIH 3T3 cells [15]. Oncogenic potential of Tiam1 was found to be present in various tumors with respect to the tumor grade and stage. The over expression of Tiam1 in breast carcinoma, nasopharyngeal carcinoma, hepatocellular carcinoma, renal cell carcinoma, retinoblastoma, colorectal carcinoma, lung and prostate cancer has been previously reported [16,17,18,19,20,21,22]. Tiam1 is negatively correlated in case of renal carcinoma, where it inhibits invasion by promoting E-cadherin mediated adhesion [15].
Tiam1 contains consensus myristoylation sequence at the amino terminus, two PEST sequences, a Ras binding domain (RBD), PSD-95/DlgA/ZO-1 domain (PDZ), two pleckstrin homology (PH) domains and DH domain [23,24,25] ( Figure 1A). The PH domain present in carboxy terminal next to DH domain is similar in all other GEFs. The DHR domain is important for the proteinprotein interaction [26]. Presence of Tiam1 on the membrane surface is accomplished through the N-terminal PH domain, and not by the c-terminal PH domain or DHR domain. Using truncated constructs of Tiam1, localization of Tiam1 in the membrane is shown to be necessary for the membrane ruffling [24,27,28].
Earlier, we showed the expression of Rho, Rac, Cdc42 and Tiam1 in RB. Additional studies on E-Cadherin and N-Cadherin has helped us to correlate the expression levels of these antigens with respect to the function they mediate in the RB tumor progression. Especially the expression of Tiam1 was positively correlated with the invasive potential of the tumors [18]. The functional relevance behind this overexpression of Tiam1 in tumorogenesis and invasion of RB is not yet elucidated. In the current study, we analyzed the effect of Tiam1 on cell proliferation and invasion using RB cell lines, Y79 and Weri-Rb1. Additionally, we addressed the effects of the truncated constructs with respect to subcellular localization and F-actin based interaction and invasiveness attributed to the RB cell lines.

Cell culture
Human RB cell lines Y79 and Weri-Rb1 were obtained from the cell bank, RIKEN BioResource Center (Ibaraki, Japan).

Sample collection and Ethics statement
Tumors samples were collected from enucleated eyeballs of RB patients as a part of treatment with written consent from patient/ guardian and utilized for research purpose anonymously. Retina samples were collected from normal cadaveric donor eyeballs from C U shah eye bank, Medical Research Foundation, Sankara Nethralaya, India. This work was approved by Vision Research Foundation ethics committee and done at Vision Research Foundation, Sankara Nethralaya (Ethical clearance no: 136-2008-P).

Transfection and RNA interference of Tiam1
Full-length and N-terminal truncated C1199 and C580 Tiam1 containing a hemagglutinin tag at the 39 end, cloned in the eukaryotic expression vector pUTSV1 (Eurogentec, Belgium) were procured from John G Collard, Netherlands Cancer Institute [13].

Total RNA isolation and Quantitative PCR
The total mRNA was isolated from transfected cells and fresh RB tumors by RNAeasy kit (Qiagen, Valencia, USA) and Trizol respectively. 1 mg of total RNA was transcribed into cDNA using oligo-dT and random hexamers. The quantitative PCR was performed in Applied Biosystem 7500 by using Sybr-green (Thermoscientific, Mumbai, India). The primer sequences for Tiam1, 18S and b2M were as follows, FP: 59AAGACGTACT-CAGGCCATGTCC-39and RP: 59-GACCCAAATGTCG-CAGTCAG-39 for Tiam1, FP: 59-AACCCGTTGAACCC-CATT-39 and RP: 59-CCATCCAATCGGTAGTAG-39 for 18S, and FP: 59-TATCCAGCGTACTCCAAAGA-39 and RP: 59-GACAAGTCTGAATGCTCCAC -39 for b2M. Comparative quantification (normal retina vs primary RB tumors or untreated cells vs Tiam1 siRNA treated cells) was determined using the formula 2 2DDCt and relative expression values normalized to the 18S or b2M endogenous control were used for plotting. Experiments were performed in triplicate for the same samples.

Western Blotting
Transfected cells and RB tumors were collected and washed with 1X PBS followed by cell lysis in RIPA buffer (containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM PMSF, 1 mM EDTA, 1%Triton X-100, 1% Sodium deoxycholate, 0.1% SDS and Protease inhibitor cocktail, Sigma). 50-100 mg of protein lysate was resolved on 8% SDS Polyacrylamide gel and then electroblotted onto nitrocellulose membrane at 100V for 90 min. The membrane was blocked in 5% skimmed milk, further incubated with primary anti-Tiam1 polyclonal antibody (raised in rabbit, C-16, Santacruz) in 1:100 dilutions and anti-b actin monoclonal antibody (raised in mouse, Sigma) in 1:2000 at 4uC for overnight. The membrane was washed thrice with TBST (1 M Tris-HCl pH-7.6 containing 0.1% Tween-20 and 0.8% NaCl) and incubated with HRP-conjugated secondary antibody at 1:2500 dilution for 2 hrs. The membrane was detected by supersignal west femto maximum sensitivity substrate (Thermoscientific, Rockford, USA). The blots are representative of three experiments.   Immunofluorescence For immunofluorescence, cells were seeded on Poly-L-Lysine coated cover slips in 24-well plate. After 48 hrs of transfection, cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 and blocked with 5% BSA in PBS for 30 mins. The cells were incubated with polyclonal anti-Tiam1 antibody (1:25) for overnight at 4uC. After PBS wash cells were incubated with FITC conjugated anti-rabbit secondary antibody (1:500) for

Flow cytometry analysis of apoptosis
For the apoptosis assay, Annexin V kit from BD biosciences(-San Diego, CA) was used. Briefly, cells were washed with ice cold 1X PBS twice and resuspended in 1X binding buffer and incubated with annexin V-FITC and Propidium iodide for 15 mins, followed by flow cytometry. Experiments were performed thrice individually. Cell proliferation assay 3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay was performed to evaluate the percentage viability of silenced cells. Transfected cells were incubated with 100 ml of media containing 10 ml of MTT (5 mg/ml) and incubated for 4 hrs at 37uC. Then media with MTT was removed and 100 ml of DMSO was added to each well the absorbance was measured at 570 nm. All experiments were performed in triplicate.

cDNA microarray
Whole genome microarray was performed in Tiam1 siRNA treated Y79 cells along with untransfected Y79 cells. The experiment was performed in triplicates. In brief, 500 ng of total RNA was used for cDNA synthesis, followed by amplification/ labeling using TotalPrep RNA Amplification kit (Ambion Inc., Austin, TX) to synthesize biotin-labeled cRNA. The concentration of cRNA was measured by spectrophotometer (Nanodrop, ND-1000, Thermo scientific). Labeled, amplified cRNA (750 ng per array) was hybridized to a ver. 3 of the Illumina Human-Ht-12 BeadChip (48 K) according to the Manufacturer's instructions (Illumina, Inc., San Diego, CA). The whole 48803 probes on the Human-Ht12 beadChip ver. 3 were used and the arrays were scanned with an Illumina Bead array Reader confocal scanner (BeadStation 500GXDW; Illumina, Inc.). Sample Gene Profile option of Illumina BeadStudio software was used to export the gene expression data.

Wound healing assay
The cells were grown on PLL coated culture plate until confluent in 1% serum medium. A wound was made by scratching a straight line using a 200 ml pipette tip. The cells were then washed twice, transfected and incubated further for 48 hrs with 10% FBS containing media. Images were taken under Phase contrast microscope at 0 hr and 48 hr using AxioObserver fluorescent microscope.

Matrigel invasion assay
Y79 and Weri-Rb1 cells were counted 24 hr post transfection of the siTiam1 and scramble siRNA and 5610 4 cells in serum free media were seeeded respectively in the rehydraded matrigel invasion chamber with 10% FBS containing media added as chemoattractant. Cells were allowed to incubate for 48 hr. Chambers were removed, washed twice with 16 PBS, cleaned the matrigel using cotton plug, fixed cells in methanol, stained with cystal violet. The membranes were cut, removed and mounted with DPX mountant and viewed under 206 objective of Axio-Observer microscope.

Statistical analysis
Statistical analysis was performed by unpaired students t-test using Graphpad software on the triplicate data generated from three individual experiments. The two-tailed P values less than 0.05 were considered as significant.

Differential expression of Tiam1 in RB tumors
In our earlier study, we reported the expression of Tiam1 protein in RB tumors by immunohistochemistry. In this study, we analyzed both the mRNA and protein levels of Tiam1 through qPCR and western blotting. The differential expression of Tiam1 in RB tumors (tumors with choroid invasion (CI) ,3 mm and CI.3 mm), Y79 and Weri-Rb1 cell lines were compared to normal cadaveric human retina (n = 2). On average of 7.5661.8, 3.360.3 and 5.360.5 fold were expressed higher in RB tumors, Y79 and Weri-Rb1 respectively ( Figure 1B) using b-2 microglobulin for normalization in relative quantitative PCR. Tiam1 protein levels were higher in expression compared to the normal retina on Western blot. Tiam1 protein bands were observed at ,180 KDa, a loading control b-actin was detected at ,42 KDa in Western blot ( Figure 1C). The protein bands intensity were represented as graph in Figure 1D.

RNA interference of Tiam1 in RB cell lines
To study the cellular events mediated by Tiam1 in RB, Tiam1 knockdown is performed transiently in Y79 cells with three different siRNA sequences (T1, T5 and T6) as mentioned in methods. After 48 h of transfection, the silencing efficiency of each siRNA duplexes was determined by Real-time PCR (Figure 2A). Individual siRNA duplexes did not show significant downregulation of Tiam1 in Y79. As per Parsons et al. (2009) the silencing efficiency was increased when all the three siRNA sequences were pooled and transfected [29]. 200 nM of siRNA pools showed -2.75 and 22.0 fold down regulation in Y79 and Weri-Rb1 respectively ( Figure 2B). This was further confirmed at protein level by Western blotting ( Figure 2C). The intensity of protein bands were measured using ImageJ software and percentage of Tiam1 expression upon siRNA treatment was calculated and represented as graph. In general, Tiam1 is localized in both plasma membrane and cytoplasm of untransfected cells but the expression level on plasma membrane was drastically reduced upon Tiam1 silencing in Y79 and Weri-Rb1 cells ( Figure 2D).

Tiam1 regulates apoptosis and viability in RB cells
Further to understand the functional relevance behind the expression of Tiam1 in cellular activities such as apoptosis and viability, studies were carried out in RB cell lines, Y79 and Weri-Rb1 in the presence and absence of Tiam1. To elucidate apoptotic effects upon silencing of Tiam1 Annexin-V assay was performed. Results showed 15% and 50% of late apoptotic cells in Y79 and Weri-Rb1 cells respectively upon siTiam1 ( Figure 3A). Upon silencing with Tiam1, the cellular metabolic activity of Y79 and Weri Rb1 decreased by 25% in MTT assay ( Figure 3B). siRNA mediated Tiam1 silencing results in de-regulation of gene expression in Y79 cells cDNA microarray was performed to analyze the genes regulated upon Tiam1 knock down using Y79 RB cell line. Raw data files were normalized using GeneSpring GX v 12.0. A total of 790 transcripts were observed to be differentially expressed in Tiam1 silenced cells above 1.0 fold (p,0.05) of which 302 genes were up-regulated and 488 genes were down-regulated. Hierarchical clustering of differentially regulated genes was done using Pearson uncentered distance matrix and average linkage rule to establish gene clusters that differentiate the two groups of samples ( Figure 4A). Biological analysis of differentially expressed genes was done for Gene Ontology and Pathways using DAVID tool (http://david.abcc.ncifcrf.gov/). Statistically significant ontologies and pathways were filtered based on p-Value ,0.05 (Obtained using Fischer Exact Test) with Benjamini Hocheberg FDR correction ( Figure S1). The data obtained from microarray analysis was deposited in NCBI's Gene Expression Omnibus (GEO45130) as per MIAME guidelines. The significantly deregulated genes in response to Tiam1 silencing are listed in Table S1.
Validation of de-regulated genes by qPCR in Tiam1 silenced RB cell lines and primary RB tumors From the de-regulated genes list, a panel of genes namely PAK2, MAP1B, RABL3, RAP2A, GAB1, MLCK2, MYOT, CALPAIN7, BAX, PDCD, TRAF6, DFFB were selected for the confirmation of microarray analysis by quantitative PCR in Tiam1 silenced Y79 and Weri-Rb1 cells ( Figure 4B). The list of primer sequences used for the SYBR-green based qPCR is given in Table 1. Genes involved in actin cytoskeleton were downregulated where as apoptotic genes were up-regulated in post Tiam1 silenced Y79 and Weri-Rb1 cells. The mRNA expressions of these genes in both RB cell lines were consistent with microarray analysis. Similarly qRT-PCR was done to evaluate the correlation of the validated genes expression in primary RB tumors ( Figure 4C). mRNA expression of these selected genes were negatively correlate with primary RB tumors. The average fold expression of 12 genes in 12 tumors normalized to two normal retina were, PAK2 (5.  Table 2. Tumors with CI,3 mm showed lesser extent of changes in gene expression compared to tumors with CI.3 mm (tumor 2, 3, 5, 6, 7). qPCR results were shown in Figure 4 & CI status were shown in Table 2. The expression levels of Tiam1 is also directly correlating to the changes in gene expression of validated targets in RB tumor.

Tiam1 regulates actin polymerization, cell migration and invasion in RB cell lines
To investigate the involvement of Tiam1 in actin cytoskeleton regulation, F-actin staining was performed using phalloidin in Tiam1 silenced Y79 and Weri-Rb1 cells ( Figure 5A). As mentioned above, Tiam1 was co-localized with actin at cell junctions. In case of Tiam1 knockdown, the cells exhibited lesser extent of actin polymerization at the cellular junctions and actin co-localization. Thus the results show that Tiam1 is essential for actin re-organisation in RB cells. Additionally, the function of Tiam1 in cell migration was assessed by wound healing assay ( Figure 5B). Tiam1 silenced RB cell lines showed impairment of cell migration, compared to untransfected and scrambled siRNA transfected RB cells. Silencing of Tiam1 in Weri-Rb1 had resulted in lesser number of cells invaded across the matrigel coated invasion chamber. This indicated the decrease in invasion potential depends on Tiam1 expression as it is not observed in the scramble siRNA transfected cells ( Figure 6).

Plasma membrane localization of Tiam1 is mediated by N-terminal PH domain
Since Tiam1 localizes along with F-actin and controls the actin cytoskeleton, we investigated which domain of the protein regulates the localization Tiam1 on plasma membrane in RB. RB cell lines (Y79 and Weri-Rb1) transfected with full length and C1199 Tiam1 showed the membrane localization and induced the membrane ruffling (Figure 7). In contrast, C580 Tiam1 was localized to the nucleus and failed to induce membrane ruffling.

N-terminal PH domain regulates retinoblastoma cell motility
We observed that N-terminal PH domain, but not C-terminal PH domain modulates the localization of Tiam1 and induces membrane ruffling in RB cells. Further we determined the association of membrane localization of Tiam1 and cell migration in Y79, Weri-Rb1 cells. To elucidate this, wound healing assay was performed in Y79 and Weri-Rb1 cells. The cells transfected with full length Tiam1 and C1199 Tiam1 were showing more cell migration towards the wound, whereas C580 Tiam1 transfected cells showed delayed cell migration into the wound in both cell lines (Figure 8).

Discussion
Though Tiam1 has been shown to play a crucial role in actin cytoskeleton, cell migration and invasion, various intracellular pathways are involved in activation of upstream and downstream signaling of Tiam1 ( Figure 9) [9,10,13,27]. In particular, Tiam1 is required for activation of Rac1 and cdc42 to produce specific actin rich structures like membrane ruffling, lamellipodia, filopodia [30,31,32,33,34,35] and for neurite outgrowth [36]. Addition to that, Tiam1 interacts with CADM1, Ephrin and Arp2/3 to initiate Rac1 mediated actin cytoskeleton remodeling [37,38,39]. Hence in the current study, we analysed the importance of Tiam1 in RB cells using short interfering RNA (siRNA) mediated knockdown studies.
Silencing of Tiam1 in RB cell lines followed by cDNA microarray showed various pathways and genes altered, predominantly genes related to MAPK pathway, small GTPase, apoptosis and cell migration. The impairment of cell migration in Tiam1 silenced Y79 and Weri-Rb1 cells was proved by Wound healing and matrigel invasion assay. This correlates with the microarray analysis where the actin cytoskeleton genes were down-regulated in Tiam1 silenced Y79 cells resulting in lesser cellular migration potential ( Figure 5). One of the down-regulated actin cytoskeleton gene in our microarray analysis is MAP1B (Microtubule-associated protein 1B) which is needed for axonal development, reported to be involved in neurite growth, neuron migration and metastasis [40,41,42]. MAP1B is found to interact with Tiam1 thereby activating Rac1 and cdc42 and further inhibiting RhoA activity which leads to actin polymerization and axonal elongation [43,44]. MAP1B deficient cells exhibit a decreased cell migration and axonal development [45,46]. The other mechanism might be PAK mediated activation of MyosinII, a protein involved in stress fiber formation and contraction [47]. The phosphorylation of myosinII light chain (MLC) by myosinII light chain kinase (MLCK) regulates actin-myosin II interaction [48]. We observed that PAK2 and MLCK were down-regulated in Tiam1 silenced retinoblastoma Y79 cells. Moreover, small GTPase subfamily Rab-like 3 (Rabl3) and actin-binding protein Myotilin, promote motility, tumor cell survival [49]. The down-regulation of Rabl3 and myotilin might as well attribute to the suppression of cell motility in Tiam1 silenced cells. The expression level of these genes when validated in primary RB tumors showed differential expression correlating with their CI status. CI represents the invasion potential of the given tumor during the enucleation, which may or may not have undergone chemotherapeutic treatment. The tumors with CI.3 mm compared to CI,3 mm cases showed increased expression of pro-survival genes and decreased expression of apoptotic genes.
From our Annexin-V assay results, a remarkable increase in apoptosis was observed. The mechanism of apoptotic induction might be due the up-regulation of pro-apoptotic gene BAX (Bcl-2 associated X), PDCD7 (Programmed cell death protein), TRAF6 (Tumor necrosis factor receptor-associated factor 6) and DFFB (DNA fragmentation factor subunit beta) upon Tiam1 knockdown [50,51,52,53,54]. Similar to our results, apoptotic cell death has been observed earlier upon Tiam1 silencing [55].
Since endogenous Tiam1 is localized in both plasma membrane and cytoplasm, we were interested to find out which domain of the Tiam1 protein regulates the localization intracellularly in RB cells. We elucidated that N-terminal PH domain of Tiam1 mediates the membrane localization and invasion but not the C-terminal PH domain and targeting the N-terminal PH domain of Tiam1 affects the cell migration and invasion. Earlier studies showed that localization of Tiam1 to plasma membrane requires N-terminal PH domain. Also the membrane localization of Tiam1 is required for the activation of the c-Jun NH 2 -terminal kinase (JNK) thus leading to successful Rac mediated signaling [13]. N-terminal PH domain of Tiam1 has affinity towards phosphoinositides and its interaction with transmembrane domain accelerates the membrane localization. For the first time our data has additionally elucidated the importance of the N-terminal region with respect to the migration as it is directly correlating with actin cytoskeleton modulation in RB. From our findings, we suggest that Tiam1 can be utilized as a potential target for RB therapy. In future, docking studies to find potent inhibitors to RBD binding or alternatively aptamers targeting the N-terminal PH domain of Tiam1 need to be isolated, which have better application in vivo for drug delivery purposes. Figure S1 Pathways and gene ontology altered by Tiam1. A. Pie chart represents significantly de-regulated pathways in response to Tiam1 silencing. B. Graphical representation of significantly de-regulated gene ontology. (TIF)