Metastasis

Malignant tumors usually have some characteristics that are life-threatening, such as uncontrolled proliferation and immortality. Progression toward metastasis is one of the key factors that affects patient recovery after surgery or chemotherapy. It represents all of the end products of a multistep cell–biological process termed the invasion-metastasis cascade. In general, the invasion-metastasis cascade can be distributed into seven steps, (1) local invasion, (2) intravasation, (3) forming circulating tumor cell, (4) arrest at distant organ sites, (5) extravasation, (6) forming micrometastases, (7) metastatic colonization (Friedl & Wolf, 2003). Each steps is important for cancer metastasis.

During local invasion, the malignant cancer cells may secret some proteases, metalloprotease and/or cathepsin, which can digest the basement membranes and ECM surrounding epithelial cells and make cancer cells touching stroma cells, which secrete

11

many growth factors or stimulate tumor-associated macrophages activity. Therefore, they establish a potentially self-amplifying positive feedback loop. Most types of carcinoma can invade as cohesive multicellular units, jointly called a “collective invasion.” However, individual tumor cells may invade through two ways: (1) mesenchymal invasion or (2) amoeboid invasion (Friedl & Wolf, 2003). Cancer cells invade their neighbor cells via “epithelial-mesenchymal transition”, which decreases the E-cadherin expression and increase a set of pleiotropically acting transcription factors, including Snail, Twist and ZEB1, making them more stem-like (Thiery et al, 2009).

Even though there are many drugs that inhibit the protease activity, cancer cells also can invade, via amoeboid invasion, a mode of invasion that depends on diffusion of non-clustered integrins(Wang et al, 2004).

All the factors in the ECM can make cancer cells more malignant and may even lead to poor diagnosis. After cancer cells intravasate into blood vessels, they can form relatively large emboli via interaction with blood platelets through L-/P-selectins (Joyce

& Pollard, 2009). They shield themselves from shear forces and evade immune detection. When the CTCs disseminate to appropriate sites where they may make contact through ligand-receptor interactions or physical trapping, they can extravasate into the tissue. During the cancer cells extravasation, the primary tumors are capable of secreting some factors to perturb microvessel permeability, such as MMP-1, MMP-2,

12

and COX-2 to make them pass easily (Gupta et al, 2007; Padua et al, 2008). In order to form micrometastasis, the cancer cells change their gene expression and receive stimulation by some growth factors from stroma cells to maintain their life. They change their state via “mesenchymal-epithelial transition”. Finally, the primary cancer cells may now thrive at the secondary site and become more malignant.

1.5 The nineteen membrane proteins in the present study

Two different A549 cell lines, A549-mock and A549-Fut IV, were derived by Dr. Yu of the Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine. The A549-FutIV cell line is transfected with gene fut4 for over-expression of FucT IV while A549-mock is transfected with empty plasmid as control. The only difference between these two cell lines is that the former can attach more fucose residues on the membrane proteins and affect cancer biology.

In a cite, comparative glycoproteomic approached were used to study the differences in the fucosylated membrane proteins when comparing the two A549 cell lines.

Nineteen membrane proteins were identified (Table1) that had high fucosylation. While the functions of some of these proteins remains unknown, for example, TM9SF3, TMEM206 and GOLM1, some of them seem to participate in cell adhesion, such as CAMD4, DSG2 and CNTN1. In normal cells, PLXNB2 is a cell surface receptor for

13

SEMA4C, SEMA4 and SEMA4G and plays an important role in cell-cell signaling.

Moreover, PLXNB2 affects neuron migration and is associated with RhoA activation (Conrotto et al, 2004; Perrot et al, 2002). EphA2 is a receptor tyrosine kinase which binds to ephrin-A family ligands on adjacent cells (Wykosky & Debinski, 2008).

Moreover, it can regulate cell migration by ephrin-A1/EFNA1 or promote cell adhesion by DSG1/desmoglein-1 (Lin et al, 2010; Ogawa et al, 2000; Wykosky & Debinski, 2008). It also may affect cancer metastasis through RhoA GTPase activation and make tumors become more malignant.

14

Gene entry name Protein name Sequence of identified

peptide

F/M

IMPAD1

Q9NX62

Inositol monophosphatase 3 QVALQTFGN²⁵⁹QTTIIPAG 1.683 GAGYK

CNTN1 Q12860 Isoform1 of Contactin-1 GTEWLVN⁴⁵⁷SSR 16.084 AN⁴⁹⁴STGTLVITDPTR 16.392 GKAN⁴⁹⁴STGTLVITDPTR 7.356 TM9SF3

Transmembrane protein206 IN¹⁵⁵YTDPFSN¹⁶²QTVK 2.186

GOLM1

Q8NBJ4

Isoform 2 of Golgi membrane protein 1

AVLVNN¹⁰⁹ITTGER 6.27 Table 1. 19 membrane proteins identified from A549-mock and A549-FutIV

cell lines.

15

DSG2 Q14126 Desmoglein-2 YVQN⁴⁶²GTYTVK 11.169

PTPRJ

Isoform 1 of Major prion protein

GEN¹⁹⁷FTETDVK 2.685

EPHA2 P29317 Ephrin type-A receptor 2 TASVSIN⁴³⁵QTEPPK 5.79 GLG1

Q92896

Isoform 2 of Golgi apparatus protein 1

N⁶²MTFDLPSDATVVLN⁷⁶R 11.398 (S)

CD44 P16070 Isoform 12 of CD44 antigen AFN⁵⁷STLPTMAQMEK N/A

ODZ3 Q9P273 Teneurin-3 IGPFAN²¹²⁴TTK 4.213

CPD O75976 Carboxypeptidase D FANEYPN⁵²²ITR 9999 CADM4 Q8NFZ8 Cell adhesion molecule 4 QTLFFN⁶⁷GTR 9999

16

1.6 Target genes: SLC3A2, ALCAM, CD44

1.6.1 SLC3A2

The 4F2 cell-surface antigen heavy chain is the protein encoded by the SLC3A2 gene. It is a multifunctional protein involved in cell transformation, integrin signaling, cell fusion and amino acid transport. 4F2hc is a type II glycoprotein, which forms heterodimers with different light subunits. When it associate with different light chains, it can transport different amino acids involved in gastric acid secretion (Drummond et al, 2010; Kirchhoff et al, 2006; Pfeiffer et al, 1999; Verrey et al, 2009). For example, SLC3A2 associates with SLC7A6 or SLC7A7, acting as an arginine/glutamine exchanger that follows an antiport mechanism for amino acid transport(Nel et al, 2012;

Verrey et al, 2009). Another function of the SLC3A2 is to make cell exert force on the matrix to interact with integrins to support downstream signals that lead to activation of RhoA small GTPase (Feral et al, 2007).

1.6.2 CD166

The CD166 antigen is a type I transmembrane glycoprotein, encoded by the ALCAM gene. It mediates both heterophilic (ALCAM-CD6) and homophilic

(ALCAM-ALCAM) cell-cell interactions. When it binds to CD6, recruited the antigen-induced dendritic cell (DC) to the T cell contact zone and sustains DC-induced T-cell proliferation after initial contact (Skonier et al, 1996b; Zimmerman et al, 2006a).

17

Although CD166 attracts lymphocytes through CD6 interaction, it also can help cell migration and make cancer cell more malignant. In colorectal carcinoma, ALCAM is frequently up-regulated and is a new independent prognostic marker (Weichert et al, 2004). Truncation of CD166 was found to diminish primary tumor growth and enhance melanoma metastasis (van Kempen et al, 2004). ALCAM in breast cancer is a tumor suppressor because lower ALCAM expression makes tumors more aggressive (Jezierska et al, 2006). It may have different effects in different tumors.

1.6.3 CD44

CD44 is a multi-structural and multi-functional cell surface molecule. The variety of CD44 isoforms can be divided into three groups: (1) CD44s the standard isoform, (2) CD44v which contains variable exons, and (3) CD44E which includes exons v8-10 (Goodison et al, 1999). CD44 is a receptor for hyaluronic acid (Aruffo et al, 1990). It also mediates cell-cell and cell-matrix interaction through its affinity for HA, collagens, osteopontin and matrix metalloproteinases (MMPs). CD44 may provide a platform for tumor invasion because CD44 forms hyaluronan-induced aggregates that make MMP9 work easily (Yu & Stamenkovic, 1999). In addition, CD44 can be cleaved by membrane-type 1 matrix metalloproteinase (MT1-MMP) and promotes cell migration (Kajita et al, 2001).

18

2. Materials and Methods

2.1 Cell lines and cell culture

The human lung adenocarcinoma cell lines, A549-Mock and fucosyltranferase IV-transfected A549 (A549-Fut4) were obtained from the Department of Internal Medicine, College of Medicine, National Taiwan University. Both of these cell lines were grown in RPMI1640 (Hyclone, Thermo Scientific,) supplemented with 10%(v/v) Fetal bovine serum (Biological industries), 100 unit/mL penicillin, 0.1 mg/mL streptomycin, 0.25 μg/mL amphotericin(Biological industries), 0.15 %(w/v) sodium bicarbonate (Sigma), 10mM HEPES (Sigma), 0.25% (w/v) glucose, and 1mM sodium pyruvate(Caissonlabs, North Logan, US). They were incubated in a humidified atmosphere with 5% CO2 at 37℃. The cells were subcultured every 2-3 times per week.

2.2 siRNA knockdown

The siRNA of target genes (SLC3A2, CD166, CD44) were ordered from Thermo Scientific Dharmacon. All siRNA of these genes were dissolved in RNase-free water to produce 20 μM as stock, then stored in a -80℃ refrigerator. Before the siRNA transfection, both cell lines (A549-Mock and A549-FutIV) were seeding in 24-well plate and 5*10⁴/mL/well. The culture medium was replaced with antibiotic-free complete medium and incubated overnight at 37℃. Then, 50μL of the siRNA in opti-medium (invitrogen) was prepared by adding 2.5μL of 5μM siRNA to 47.5μL of

19

opti-medium in an 1.5mL eppendorf and gently mixed by pipetting carefully up and down, then incubated for 5 minutes at room temperature. This preparation was labeled as tube 1.

DharmaFECT transfection reagent in opti-medium was prepared by adding 1μL DharmaFECT reagent and 49μL opti-medium gently mixed and incubated for 5 minutes at room temperature, then labeled as tube 2. The content of tube 1 was then added to tube 2, mixed by pipetting carefully up and down and incubated for 20 minutes at room temperature. Then, 400 μL of antibiotic-free complete medium was added to get 500μL transfection medium with a final concentration of 25nM siRNA. This tranfection medium was prepared fresh before use. The antibiotic-free medium in the 24-well cultured plate was removed and 500 μL of fresh transfection medium added to each well.

The cell were incubated for 24 hours.

2.3 RNA preparation, reverse transcription polymerase chain reaction (RT-PCR) and

quantitative PCR(qPCR)

2.3.1 RNA preparation

The total RNA of A549-Mock and A549-Fut4 in the 24-well plate were extracted using a TRIzol Reagent (Invitrogen) following the protocol provided by the manufacturer. Briefly, the medium was removed, the cells were washed by PBS twice and then 1 mL TRIzol reagent was added to each well. The cell lysate was incubated 5

20

minutes to homogenize the RNA and then transferred to a new tube. Two hundred microliters of chloroform was added into 1 mL of homogenized mixture, which was shaken vigorously, and then centrifuged at 12000g for 15 minutes at 4℃ to separate the RNA into an aqueous phase, DNA into an internal phase and protein into an organic phase. For RNA precipitation, the aqueous phase was transferred into a new tube and the tube was inverted several times after adding isopropyl alcohol. Then, the aqueous-isopropyl alcohol mixture was centrifuged at 12000g for 10 minutes at 4℃, producing an RNA pellet at the bottom of tube. The supernatant was discarded and 75%

ethyl alcohol added twice to wash the pellet. Finally, the RNA pellet was dried out and dissolved in 50 μL RNase-free water..

2.3.2 Reverse transcription polymerase chain reaction (RT-PCR)

To obtain 1μg cDNA, the total RNA concentration obtained by the previous procedures was measured using NanoDrop(Thermo scientific) and appropriate volume of total RNA was used as template in a sterile PCR tube. One hundred pmol of oligodeoxythymidine primer (oligo(dT)18) was mixed with template RNA and filled with RNase-free water to 12.5 μL. The mixture would chill on ice after incubating at 65℃ for 5 minutes. Meanwhile, pre-mix was prepared, which contained 4μL of 5X reaction buffer, 0.5μL of RNaseOUT, 2μL of 10mM dNTP mix and 1μL of RevertAid reverse transcriptase (Fermentas). These two mixtures were mixed to a total volume of

21

20μL and the mixture incubated at 42℃ for 60 minutes. Finally, the reaction was

terminated by heating at 70℃ for 10 minutes and used in further experiments.

2.3.3 Quantitative PCR(qPCR)

The qPCR probe for these genes (Fut4, SLC3A2, CD166, CD44) were designed by LightCycler Probe Design Software (Roche) and the sequence of these genes were obtained from NCBI(accession number: NM_002033.3, AB018010.1, NM_001627.3, and NM_000610.3, respectively). The replicon of qPCR length was generally about 160-180 base pairs and the melting temperature was about 60℃. To avoid primer dimer, the probes used did not have ΔG less than -2000, a threshold value estimated by the software. The probes used in this study are shown in the following table.

2.4 Migration/Invasion assay

2.4.1 Wound healing assay

A549-mock and A549-FutIV cell lines were seeded at a density of 2 × 10⁵cells primer name F primer 5'->3' R primer 5'->3'

Fut4 CCCAgACCgTgCCAACTA ggAggTgATgTggACAgC SLC3A2 CCAgAAggATgATgTCgCT CAACCTgAgTggAgAACC

ALCAM ggCAgTggAAgCgTCATA AgCAgAgACATTCAAggAgT

CD44 TCAACAgTggCAATggAgC gCAggTTCCTTgTCTCATCA

22

per well in a 24-well plate and grown overnight to confluence in complete medium. The monolayer was scratched with a pipette tip and washed with PBS to remove floating cells. The scrape was monitored and photographed after 4, 8, 12, 24 h incubation.

2.4.2 Invasion assay

A cell invasion assay was performed using a Boyden chamber (Millpore, Co., USA) with 8 μm pore polycarbonate filters that were coated with 40μg geltrex (Invitrogen)

which was diluted by serum-free RPMI-1640. After 72 hrs of transfection, 2*10⁴ cells were resuspended by 200μL serum-free medium and seeded to the upper compartment

of the Boyden chamber. The lower chamber was filled with 1 mL of RPMI-1640 complete medium. The cells were allowed to migrate for 48 hrs, incubated at 37℃, 5%

CO2. The cells were then removed from the upper chamber using a cotton swab. The cells on the lower surface of the chamber were stained with Liu’s stain and counted.

Data representing the average number of cells per pixel in five fields were compared between the siRNA groups, NC groups, and blank control groups.

2.5 Adhesion assay

Ninety-six-well plates were coated with Collagen (Cohesion; Vitrogen), Fibronectin

(Sigma-Aldrich), Gelatin (Sigma-Aldrich) and BSA for 12 hours at 4°C. Each coating protein was dissolved in PBS (pH: 7.4) to yield a final concentration of 60 μg/mL, and a volume of 100 μL was added to each individual wells. The plates were then blocked

23

with 10 mg/mL Bovine serum albumin (Sigma-Aldrich) which was heated at 85℃ for 10 minutes in PBS for 1 hour at 37°C and washed with PBS twice.

Targeted knockdown cells were isolated by trypsinization and washed once in RPMI1640 with 10% FBS to stop trypsin activity. Then the cells were resolved in serum-free RPMI1640 to remove serum components. Suspensions of 10⁴ cells/mL viable targeted knockdown cells were then added to each well and allowed to attach for 1 hour at 37°C, 5% CO2. To determine the cell adhesion ability, plates were then carefully washed three times with PBS and fixed by 5% paraformaldehyde for 15 minutes. And then they were washed by ddH2O three times and stained by crystal violet for 30 minutes. Then, they were washed by ddH2O three times and resolved crystal violet by 10 % acetic acid. Finally, the optical density was measure at 570 nm.

2.6 Protein extraction

The organic phase from the RNA extraction which contained protein was added to

300 μL of 100% ethanol per 1 mL of TRIzol reagent and then mixed by inverting the

sample. The mixture was stored for 3 minutes at room temperature and centrifuged for 2000 g for 5 minutes at 4℃. The supernatant was transferred to a new eppendorf and then 1.5 mL of isopropanol was added for homogenization. The mixture was stored at room temperature for 15 minutes and centrifuged at 12000 g for 10 minutes at 4℃. The supernatant was removed and the protein pellet was washed by a solution containing

24

0.3M guanindine hydrochloride in 95% ethanol three times. During each wash step, the protein pellet was stored in wash solution for 20 minutes at room temperature and centrifuged at 7500 g for 5 minutes at 4℃. After the final washing, the protein pellet was vortexed in 2 ml absolute ethanol and allowed to sit for 20 minutes at room temperature. In the final step, it was centrifuged at 7500g for 5 minutes at 4℃ and the supernatant was discarded. The protein pellet was resolved by the solution with 9.5 M Urea and 2% CHAPS and the sample was stored at -20 ℃ for further use.

2.7 Western blot

Protein samples (15 μg) were separated by SDS-PAGE and transferred to a PVDF membrane The membrane was blocked with 5% (w/v) BSA (sigma) in PBST for 1hr at room temperature, incubated with (1) 200 ng/mL rabbit polyclonal IgG to SLC3A2 (santa cruz, CA. U.S.A.), (2) 200 ng/mL mouse monoclonal IgG to CD166 (santa cruz, CA. U.S.A.), or (3) 400 ng/mL rabbit polyclonal IgG to CD44 (santa cruz, CA. U.S.A.)

in PBST involving 5% (w/v) BSA overnight at 4 ℃. The membrane was washed with

PBST for 30 mins twice and incubated with (1) 80 ng/mL Goat polyclonal secondary antibody to rabbit IgG HRP (Abcam, Cambridge, UK), or (2) 200 ng/mL rabbit polyclonal secondary antibody to Mouse IgG HRP (Abcam, Cambridge, UK) in PBST involving 5% BSA (v/v) for 1hr. The membrane was washed with PBST for 30 minutes twice. The immunoreactive bands were visualized by film exposure (GE Healthcare,

25

Uppsala, Sweden) through the ECL enhanced chemiluminescence detection system.

26

3. Results

3.1 Characteristics of A549-Mock and A549-FutIV

Under microscopic observation, the morphology of A549-Mock and A549-FutIV appeared different. The former had a thinner spindle shape and the latter had a round shape (Fig.1). To confirm the FUT IV expression, real-time PCR and western blot had been employed. The total RNA extractions from the two cell lines were extracted and reversely transcribed into cDNA. The FUT IV primers had been designed and it had about seventy-four folds change in real-time PCR examination (Fig.2). The FUT IV protein expression and fucosylation also had a significant change in A549-FutIV (Fig.3-4). In order to identify which cell lines were more malignant, we analyzed invasion ability to investigate their metastasis in vitro. The result showed that A549-FutIV had a higher degree of cell invasion to another transwell site than A549-Mock (Fig.5).

3.2 Functional network analysis of the nineteen genes

The 19 membrane proteins that were more or less saturated with fucose were analyzed by Ingenuity Pathways Analysis. There were 12 membrane proteins that could be linked to form a network which were associated with cell morphology, cell movement and cell assembly (Fig.6). Six of them were relative to tumorigenesis, metastasis and cell migration, namely PRNP, ITGA3, CD44, EPHA2, DSG2 and

27

CD166. There was one protein associated with amino acid transport, SLC3A2 (Fig.7).

3.3 Analysis of the siRNA knockdown effect in A549-Mock and A549-FutIV

siRNA had been employed to decrease the target genes expression level for SLC3A2, CD166 and CD44. Under the rules of siRNA knockdown, the knockdown efficiency of target genes must be over 90%. In other words, the target gene expression must lower than 10% compared to the control. From the quantitative polymerase chain reaction examination, all the genes’ expressions decreased below 0.06 compared to the control after 24h transfection (Fig.8). Transfected-cell lines had no cytotoxicity from siRNA or transfection reagent and their viabilities also still had over 80% during the experiments (Fig.9). After siRNA transfection 3 days, the target proteins decreased(Fig.10-10.3).

3.4 Functional analysis of the target genes after genes knockdown

3.4.1 The effect of target proteins invasion ability

Since membrane proteins could affect cancer metastasis, we investigated whether the membrane proteins influenced it and made cancer more malignant. In an in vitro study, cell invasion was assessed using Boyden chambers. While cancer cells invaded another side of the transwell, they would be stained by Liu’s stain, causing them to displayed a purple color(Fig.11). Knockdown for all the three genes showed different invasion ability compared to the control (Fig.12) and had different fold changes in cell counting (Table2). From the results, CD166 had the greatest ability to inhibit cancer cell

28

invasion, especially in the more malignant cell line, A549-FutIV. The second strongest invasion inhibitor was CD44 and the weakest was SLC3A2. All transfected A549-mock cell lines had no significant difference regarding cancer cells invasion inhibition. When all the transfected-cell lines were compared to A549-Mock, the invasion ability of A549-FutIV that was transfected with CD166 was closest in similarity to A549-mock (Fig.13 Table3).

3.4.2 The effect of target proteins adhesion ability

In order to investigate how these fucosylated membrane proteins after knockdown could affect cell adhesion, the cell adhesion ability was studied with different ECMs.

We employed fibronectin, collagen, gelatin and BSA as negative controls. With different ECM coatings, the transfected cell lines showed different adhesion abilities to different ECMs (Fig.14). CD166 knockdown of the A549-FutIV cell line had significantly decreased its adhesion ability to collagen, gelatin and fibronectin. SLC3A2 had slightly reduced its binding ability to fibronectin and CD44 had slightly reduced its binding ability to collagen.

3.4.3 The effect of target proteins migration ability

To investigate how these proteins after knockdown could affect cell migration. We

used would healing assay to study their migration ability and found that when SLC3A2 was knockdown, the migration ability of the cells increased comparing to control after