1

Identification and functional analysis of fusion gene in breast cancer through

comprehensive analysis of genomic structure

by Pang Qing Yuan

Supervisor: Koichiro Inaki Genome Institute of Singapore

Cancer Biology and Pharmacology 2

Submitted to the School of Biological Sciences in partial fulfillment of the requirements for Final Year Project

NANYANG TECHNOLOGICAL UNIVERSITY

April 2010

2

DECLARATION I declare that, in accordance with School requirements this thesis is under 6000 words in length; all presented work was performed within the official project time frame as stated below; all presented work was performed by me, unless otherwise specified; all relevant work experience gained before the Final Year Project is stated below; the input by my supervisor, or delegated supervisor, into this thesis was limited to reviewing of up to two hard copy drafts; this thesis is my own work, unless otherwise referenced, as defined by the NTU policy on plagiarism and I have read the NTU Honour Code and Pledge; the included Abstract, Introduction, Results, Discussion and Conclusions sections will be submitted to SafeAssignment no later than 24 hours after hardcopy submission. Final Year Project start date: 11th January 2010 Final Year Project submission date: 23rd April 2009 Total number of weeks: 16 weeks Pre-Final Year Project experience May-Dec 2008, Research attachment at Genome Institute of Singapore, Cancer Biology and Pharmacology 2 under Dr Koichiro Inaki’s supervision. May-July 2009, Research attachment at Genome Institute of Singapore, Cancer Biology and Pharmacology 2 under Dr Koichiro Inaki’s supervision. Student’s signature: Date:

3

TABLE OF CONTENTS

ACKNOWLEDGEMENTS - 5

ABRREVIATIONS - 6

ABSTRACT - 7

INTRODUCTION - 8

MATERIAL AND METHODS - 10

Functional prediction of fusion gene protein products - 10

Plasmid constructs - 10

Cell culture - 10

Transfection and transient colony formation assay - 10

Scattering assay - 11

Generation of over-expressing stable cell lines - 11

Western Blot - 12

Co-immunoprecipitation - 13

RESULTS - 14

Functional prediction of RPS6KB1:TMEM49 fusion gene

products

- 14

Transient over-expression of HA tagged S6K fusion constructs - 15

S6K fusion constructs when cultured in RPMI-1640

supplemented with 3% FBS showed significant increased colony

size but not when cultured in RPMI-1640 supplemented with

10% FBS

- 16

S6K fusion construct E4-E12 displayed scattered colonies when

cultured in RPMI-1640 supplemented with 10% FBS

- 18

Over-expression of proposed p70-binding partners fused with

FLAG tags in T47D cell lines followed by immuno-precipitation

- 19

DISCUSSION - 22

Functional possibilities of S6K fusion constructs - 22

Over-expression of S6K fusion constructs could confer a cell

growth advantage when cells are cultured in low serum

conditions.

- 23

Scattered colonies observed for over-expressing E4:E12 T47D

cells suggest a migratory phenotype

- 23

4

Higher expression profile of E4-E12 was observed in cells co-

transfected with p85 and p70 as compared to other FLAG

constructs

- 24

Future work 24

CONCLUSIONS 26

REFERENCES 27

5

ACKNOWLEDGEMENTS

I would like to express my gratitude towards Dr Koichiro Inaki for his guidance

and tutelage throughout my duration on this project. The experience in the lab

has allowed me to pick up many techniques and also the process thought

which are essential in being a good scientist. I would also like to extend my

gratitude to Dr Leena Ukil, Edward Chee Yu Wing, Shakeela Banu and

Tareen Ho Yuet Fern for their help and guidance when I am in doubt. Lastly, I

would like to thank Jarius Ng Jia Jin and Natalie See and other lab members

for assisting me in one way or another.

6

ABBREVIATIONS

AKT v-akt murine thymoma viral oncogene homolog 3, protein

kinase B γ

CMV Cytomegalovirus

DMEM Dulbecco Modified Eagle Medium

EDTA Ethylenediaminetetraacetic acid

EGTA Ethylene glycol tetraacetic acid

ETS Homo sapiens ETS variant 1

FISH Fluorescent in situ hybridization

FLAG FLAG octapeptide

G418 Geneticin

HA Hemagglutanin

IRS-1 Insulin receptor substrate 1

mAb Monoclonal antibody

MCF7 Human breast adenocarcinoma cell line, Michigan Cancer

Foundation-7

mTOR Mammalian target of rapamycin

PET Paired end ditag

PPP2CA Protein phosphotase 2 catalytic subunit

qPCR Quantitative polymerase chain reaction

RACE Rapid amplification of cDNA ends

RPMI Roswell Park Memorial Institute

RPS6 Ribosomal protein S6

RPS6KB1 Ribosomal protein S6 Kinase

S6K S6 Kinase

SMART Simple modular architecture research tool

T47D Human ductal breast epithelial tumour cell line

TMEM49 Transmembrane protein 49

TMPRSS2 Transmembrane protease, serine 2

VMP1 Vacuolar membrane protein 1

PI3K Phosphatidylinositol 3-kinase

RPTOR regulatory associated protein of mTOR

7

ABSTRACT

Identification and functional analysis of fusion gene in breast cancer through comprehensive analysis of genomic structure

by

Pang Qing Yuan

Chromosomal abnormalities are common hallmark features in cancer

and they give rise to fusion genes through various mechanical process. In

order to identify these novel fusion genes, PET technology was employed in

this study. From gPET analysis in the frequently amplified locus of 17q23,

multiple gene fusions between RPS6K1 and TMEM49 were identified. This

particular gene fusion was found to be expressed in MCF7 cell lines as well

as 22 out of 70 clinical samples. In our study, we have focused on 6 fusion

structures, which were identified previously by Inaki et al, unpublished data.

We show that by over-expressing these 6 fusion constructs, E4:E12 in

particular displayed significant increase in colony size in low serum conditions,

which suggests that it could confer a cell growth advantage. In addition, we

show that this particular S6K fusion construct displayed scattered colonies

suggesting a migratory phenotype associated with increased p70-S6K activity.

Taken together, these findings suggest that these S6K fusion products could

affect counterparts within the PI3K/Akt/mTOR pathway and contribute towards

oncogenic phenotype. This study also suggests several possibilities into

understanding the functional role of S6K fusion gene and its contribution in

breast cancer.

8

INTRODUCTION

Breast cancer is one of the most common cancers that afflict women

around the world. An average of 1000 women are being diagnosed with

breast cancer annually in Singapore (Singapore Cancer Society). Therefore in

recent years, it has been of increasing importance to discover more

biomarkers as well as anti-cancer drugs that will allow early intervention and

better prognosis.

Chromosomal abnormalities are one of the common features in cancer

cells. These chromosomal abnormalities result in fusion genes that could

eventually cascade towards the progression of cancer. As normal somatic

chromosomes are stable in nature, fusion gene identification serves as an

ideal target in the diagnosis of cancer as well as development of anti-cancer

drugs. Gene fusions could result from translocations, deletions and tandem

duplications of chromosomal structure (Prensner and Chinnaiyan 2009;

Rabbitts et al., 2009). If certain functional gene fusions expressions were to

be up regulated by strong promoters or enhancers, it could result in increased

proliferation of cells and affect the progression of cancer. A classic example

would be Bcr-Abl fusion gene which is formed due to t(9;22)(q34.1;q11.2) and

is observed in 90% of chronic myeloid leukemia.

Tomlins et al (2005) have reported the presence of genomic

rearrangements that resulted in the fusion of TMPRSS2 to ETS family

members. The strong promoter elements of TMPRSS2 mediates the over-

expression of ETS family members which are implicated in 23 out of 29 cases

of prostate cancer. Other studies (Soda et al., 2007; Tomlins et al., 2007)

have reported various gene fusions in different carcinomas. Therefore, by

investigating the transcriptional and functional consequence of these gene

fusions, it would allow us to gain a better understanding of these structural

abnormalities contributing to effective therapy intervention.

Comprehensive analysis of the breast cancer genome structure using

PET (paired end ditag) (Ng et al., 2006; Ruan et al., 2007; Fullwood et al.,

2009) technology developed by Genome Institute of Singapore allowed us to

identify abnormal genetic structures, which produced novel gene fusions

(Inaki et al., unpublished data). From data screening using gPET (genomic

PET) browser, 2687 genetic abnormalities were discovered. Out of the 2687

9

genetic abnormalities, 160 were predicted to produce fusion genes. RT-PCR

validation was conducted for the predicted genes, and 77 out of the 160

(48.1%) were expressed (Sinclair et al., 2003). Among the fusion gene

transcripts, RPS6KB1-TMEM49 fusion gene expressed in MCF-7 cell line was

further analyzed (Inaki et al., unpublished data). This fusion gene is caused by

tandem duplication event within 17q23, which is a high frequently amplified

locus in breast cancer (Sinclair, Rowley et al. 2003). Thus 5’ part of RPS6KB1

is fused with 3’ part of TMEM49. RPS6KB1 is a kinase involved in

PI3K/Akt/mTOR pathway, which is highly activated in the cancers where over-

expression of this gene is known to confer cell growth advantage. Over-

expression of this gene is also reported to promote epithelial to mesenchymal

phenotype in ovarian cell lines. (Mahalingam and Templeton 1996; Dennis et

al., 1998; Pullen et al., 1998) TMEM49/VMP1 is a cell membrane associated

protein reported to be involved in organelle trafficking and triggering

autophagy. Its reduced expression has also been reported to decrease

invasive capability in tumour cells. (Ropolo et al., 2007; Calvo-Garrido et al.,

2008; Sauermann et al., 2008)

In order to identify cancer cell lines and clinical specimens that express

these RPS6KB1-TMEM49 fusion transcripts, RT-PCR was conducted.

Although the expression of the fusion gene was only found in MCF-7 among

the 8 cell lines tested, it showed recurrent expression (22 out of 70) in clinical

samples. In the expressed tumors, 10 multiple fusion points were determined.

After validation of the expression of these fusion transcripts, investigations of

the functional implications and their putative role in the contribution to the

oncogenic phenotype were followed.

10

MATERIALS AND METHODS

Functional prediction of fusion gene protein products

mRNA sequences of functional RPS6KB1 and TMEM49 were obtained from

NCBI. The various exons combinations of the RPS6KB1 and TMEM49 were

created and predicted using Simple Modular Architecture Research Tool

(http://smart.embl-heidelberg.de/).

Plasmid constructs

RPS6, PPP2CA, S6K-p70 and S6K-p85 were constructed by RT-PCR using

primers targeting full-length coding sequence PCR enzyme used was

Phusion® Flash High-Fidelity PCR Master Mix (Finnzymes, Finland). To

construct IRS-1 and mTOR full-length cDNA were first amplified by PCR and

ligated into pSC-B-amp/kan (Stratagene, Agilent Technologies), followed by

PCR targeting full-length coding sequence. As the IRS-1 and mTOR are of

length 3.8-kb and 7.7-kb respectively, sequencing primers (1st Base,

Singapore) were designed in order to validate the sequence. All PCR

constructs were subjected to NotI/XbaI restriction endonuclease digestion

(New England Biolabs, Canada) before ligation into p3XFLAG-CMV

expression vector (Sigma-Aldrich).

Cell culture

T47D cells were propagated at 37°C and 5% CO2 in humidified atmosphere in

RPMI 1640 (BSF, A*STAR) supplemented with 10% heat-inactivated FBS

(Invitrogen). MCF7 cells were propagated with D-MEM (BSF, A*STAR)

supplemented with 10% heat-inactivated FBS.

Transfection and colony formation assay

T47D cells were plated to 90% confluency on a 15 cm2 tissue culture plate

before trypsinization using 0.25% Trypsin-EDTA (Gibco®, Invitrogen). 2x105

cells/well were then plated in triplicate sets into 6 well plates and left into the

incubator for 2 days. Transfection was carried out once the T47D cells

reaches 80% confluency. T47D were transfected with modified pcDNA3.1(-)

(Invitrogen) vectors where inserted constructs are fused with HA tag at C-

11

terminus namely Mock (empty vector), E1:E7, E1:E8, E4:E12, E1:E11,

E2:E11 and E4:E11. 250μL of Opti-MEM® I Reduced Serum Media

(Invitrogen) were added to respective labeled 4 ml polycarbonate tubes.

Following which, 4 μg (equivalent of Mock) of various fusion constructs were

added respectively. In the meantime, 10μL of Lipofectamine™ 2000

(Invitrogen) was added into another set of labeled tubes with 250μL of Opti-

MEM® I Reduced Serum Media (Invitrogen) followed by incubation at room

temperature for 5 minutes before combining the fractions together and adding

it drop-wise into the respective wells of the 6 well plates after 30 minutes

incubation. Transfected cells were then left in the incubator for 24 hours. The

transfected cells were then trypsinized and re-plated into 1x105 cells/well in

triplicate samples. Cells were then left in the incubator for 24 hours. After 24

hours, RPMI-1640 (BSF, A*STAR) supplemented with 3% or 10% FBS was

added to the respective plates. 400 μg/mL of G418 (Gibco®) was added to the

media as means of selection. Fresh selection media was changed every week

and growth was observed. After 3-5 weeks, cells were fixed with methanol

and stained using crystal violet.

Scattering Assay

The scattering assay was performed as described in both Pon et al and

Zhou et al (Zhou and Wong 2006; Pon et al. 2008), colony morphologies were

defined as normal or scattered. After the colony formation assay as

mentioned above, pictures were taken under 10X objective lenses using

image capture (Nikon, Japan). Scattered colonies were judged by a typical

change in morphology, characterized by cell-cell dissociation and the

acquisition of a migratory fibroblast-like phenotype. Photos were randomized

and tagged before they are shown to participants. Scattering activity was

measured in a total number of scattered colonies from 50 colonies for 2 wells

in a single batch for each S6K construct. Participants were recruited to judge

the scattering activity and their scores were recorded and tabulated.

Generation of over-expressing stable cell lines

T47D cells were plated to 90% confluency on a 15 cm2 tissue culture

plate before trypsinization using 0.25% Trypsin-EDTA (Gibco®, Invitrogen).

12

1x106 cells/well were then plated in triplicate sets into 10 cm2 dishes and left

into the incubator for 2 days. Transfection was carried out once the T47D cells

reaches 80% confluency. T47D were transfected with modified pcDNA3.1(-)

(Invitrogen) vectors where inserted constructs are fused with HA tag at C-

terminus namely Mock (empty vector) and E4:E12. 1.25mL of Opti-MEM® I

Reduced Serum Media (Invitrogen) were added to respective labeled 14 ml

polycarbonate tubes. Following which, 20 μg (equivalent of Mock) of fusion

construct was added respectively. In the meantime, 50μL of Lipofectamine™

2000 (Invitrogen) was added into another set of labeled tubes with 1.25mL of

Opti-MEM® I Reduced Serum Media (Invitrogen) followed by incubation at

room temperature for 5 minutes before combining the fractions together and

adding it drop-wise into the respective wells of the 10 cm2 dishes after 30

minutes incubation. Transfected cells were then left in the incubator for 24

hours before RPMI-1640 (BSF, A*STAR) supplemented with 10% FBS and

400 μg/mL of G418 (Gibco®) as mean of selection. Fresh selection media was

changed every week and growth was observed over a period of 2 weeks.

Stable over-expressing cells were harvested after 2 weeks to determine the

level of protein expression.

Western Blot

Transfected cells were lysed in 1X cell lysis buffer (20mM Tris HCl, pH

7.5, 150mM NaCl, 1mM Na2EDTA, 1mM EGTA, 1% Triton, 2.5mM Sodium

pyrophosphate, 1mM β-glycerophosphate, 1mM Na2VO4, 1μg/ml Leupeptin)

(Cell Signaling Technology®) with 1X Complete Proteinase inhibitor (Roche

Applied Science). 15 μL of total cell lysate from each sample was separated

by 15% Tris-Glycine SDS-PAGE and transferred to 0.2 μm nitrocellulose

membrane. Membranes were blocked at room temperature for an hour in 5%

Milk-TBST (Tris-Buffered Saline Tween-20). The membranes were then

incubated overnight with following antibodies overnight: anti HA High Affinity

Rat IgG (Roche Applied Science), Phospho-AKT Thr308 Rabbit mAb (Cell

Signaling Technology®, 9275), AKT Rabbit mAb (Cell Signaling Technology®)

Phospho-S6 Ribosomal Protein Rabbit mAb (Cell Signaling Technology®), S6

Ribosomal Protein (5G10) Mouse mAb (Cell Signaling Technology®, 2217).

Horse-radish peroxidase (HRP) conjugated secondary antibodies – goat anti-

13

rat, goat anti-rabbit, goat anti-mouse (BioRad) - were then used to detect the

primary antibodies. Protein of interest was then visualized using Enhanced

Chemiluminescence (ECL) detection system (Amersham Pharmacia Biotech,

GE Healthcare).

Co-immunoprecipitation

Pre-clearing of lysate was conducted using 100μL of diluted lysate

added with 50μL of Rec. Protein A-Sepharose 4B beads (Zymed Laboratories,

Invitrogen) for 1 hour at 4°C with rotary shaking. Pre-cleared lysate was then

incubated with anti-FLAG®-M2 mouse mAb (Sigma-Aldrich) for 2 hours at 4°C

with rotary shaking. After which, Rec Protein A-Sepharose 4B beads (Zymed

Laboratories, Invitrogen) was added and incubated for an hour at 4°C with

rotary shaking. The mixture was precipitated and wash 5 times before elution

in 1X SDS sample buffer with β-mercaptanol. The eluted mixture was then

subjected to western blot.

14

RESULTS

Functional prediction of RPS6KB1:TMEM49 fusion gene products

Previously, work has been done to validate the presence of fusion

transcripts in MCF7 cell line and clinical samples. In order to predict whether

these fusion transcripts could give rise functional novel protein products, the

sequences of the validated transcripts were subjected to analysis using

SMART. All fusion gene constructs were found to contain no functional kinase

domain. E1:E7 (Fig 1A) and E1:E8 (Fig 1B) were predicted to contain 4 trans-

membrane domains, while E2:E11 (Fig 1C) and E4:E12 (Fig 1D) contained 1

trans-membrane domain, On the other hand, E1:E11 (Fig 1E) and E4:E11

(Fig 1F) were predicted to be truncated protein products due to a frame shift

in the fusion point.

However, all constructs still retain the N terminal region of the

RPS6KB1 protein which is reported to bind to C terminal region. This

interaction results in the kinase activity inhibition (Dennis et al. 1998). The N

terminus is also reported to contain a conserved TOS motif, which is

important in the binding with mTOR/RPTOR to facilitate the phosphorylation

of the T389 residue required in p70 S6K activation. (Schalm and Blenis, 2002;

Nojima et al., 2003)

Figure 1: Various RPS6KB1:TMEM49 fusion gene construct protein domains predictions. (A) E1:E7 and (B) E1:E8 predicted with 4 transmembrane domains, (C) E2:E11 and (D) E4:E12 predicted with 1 transmembrane domain, (E) E1:E11 and (F) E4:E11 were predicted to be truncated proteins.

15

Transient over-expression of HA tagged S6K fusion constructs

6 fusion constructs with various fusion points fused with HA tags were

previously constructed during my attachment. These fusions constructs

consisted of various combinations of fusions between exons of RPS6KB1 and

TMEM49, which were eventually termed like E1:E11 where the former refers

to the exon of RPS6KB1 and the latter being the exon of TMEM49. As both

E1:E11 and E4:E11 were predicted to be truncated products due to out of

frame fusion, the constructs were designed to be in coding frame so as to

express the following truncated products. In order to understand the

expression profile of the various constructs, it was transfected and transiently

over-expressed in T47D (Figure 2A) as well as MCF7 cell lines. (Figure 2B).

Figure 2: Transient over-expression of S6K fusion constructs in T47D and MCF7 cell lines. (A) Over-expression of HA-tagged S6K fusion constructs in T47D cell lines. (B) Over-expression of HA-tagged S6K fusion constructs in MCF7 cell lines.

From the results, expression profiles of the S6K constructs were

identical in both cell lines. Low expression profiles were observed for E1:E7

16

(24.5 kDa), E1:E8 (29.3 kDa) and E1:E11 (9.95 kDa) in both cell lines.

However, E4:E11 (18.7 kDa) was not expressed at all even after it was

designed to be in coding frame. Strong expression profiles were seen for

E2:E11 (16.7 kDa) and E4:E12 (19.7 kDa) constructs. Expected protein

weight was calculated based on the expression sequence of the various

sequences. Some of the over-expressed proteins were detected at higher

molecular weights, which could be a result of post-translational modifications.

We next sought to determine whether these fusion constructs could confer a

cell growth advantage in low S6K expressing T47D cell lines.

S6K fusion constructs when cultured in RPMI-1640 supplemented with

3% FBS showed significant increased colony size but not when

cultured in RPMI-1640 supplemented with 10% FBS

In order to understand if the S6K constructs could confer cell growth

advantage, colony formation assay was performed to find out if there was any

significant colony size phenotype in cancer cell growth by over-expression of

the constructs in T47D cell line which are low RPS6KB1 expressing cells.

Transfected cells with the respective various constructs were observed for

changes in cell morphology, significant colony morphological changes in

comparison to negative control (Mock). When the cells were cultured in 10%

FBS serum, there was significant colony size reduction for 4 of 6 the

constructs only in the 3rd batch (Figure 3C). However, when it was cultured in

3% FBS, there was marked significant increase in colony size observed

(Figure 4). However, this was only observed in the first batch but not in the

subsequent batches. Therefore there is a need to repeat the experiment in

triplicates in order to observe any significant colony size differences.

17

Figure 3: Colony formation assay cultured in RPMI-1640 supplemented with 10% FBS over a period of 5 weeks. Biological replicates were done and fixed after 5 weeks. (A) Batch 1 was conducted in duplicate sets (B) Batch 2 was conducted in duplicate sets (C) Batch 3 was conducted in triplicate sets, Student’s T-test P<0.05 (D) Western Blot of the over-expressing S6K constructs.

18

Figure 4: Colony formation assay cultured in RPMI-1640 supplemented with 3% FBS over a period of 5 weeks. Experiment was conducted in duplicate sets, Student’s T-test P<0.05, P<0.005, P<0.0005 vs control.

We next sought to determine whether there are other phenotypes

associated with the over-expression of these S6K fusion constructs.

S6K fusion construct E4-E12 displayed scattered colonies when

cultured in RPMI-1640 supplemented with 10% FBS

As reported in Pon et al. and Zhou et al. (Zhou and Wong 2006; Pon,

Zhou et al. 2008), over-expression of constitutive active p70 S6K in ovarian

cancer cell lines promotes epithelial to mesenchymal transition (EMT), which

is seen as scattered colonies when cells were cultured in media.

In order to understand if S6K fusion constructs could enhance the over-

expression of p70-S6K, a scattering assay was conducted. Scattered colonies

was observed for E4:E12 (53.5%) when it was cultured in RPMI-1640

supplemented with 10% FBS as scored by participants (Figure 5A). Even

when it was cultured in RPMi-1640 supplemented with 3% FBS (Figure 6), the

scattered colonies was even more evident for E4:E12 construct (77.5%).

19

Figure 5: Scattered colonies cultured in RPMI-1640 supplemented with 10% FBS were scored. (A) Percentage of scattered colonies. Columns, means of three experiments done in duplicates; bars, standard deviation. Student’s T-Test P<0.0005 vs control (B) Representative photos of over-expressing Mock and E4:E12 transfected T47D cells.

Figure 6: Scattered colonies cultured in RPMI-1640 supplemented with 3% FBS were scored. (A) Percentage of scattered colonies. Columns, means of single experiment done in duplicates; bars, standard deviation. Student’s T-Test P<0.05 vs control (B) Representative photos of over-expressing Mock and E4:E12 transfected T47D cells.

Next, we sought to understand how E4:E12 construct could contribute

to such a phenotype observed. Therefore, we proposed that E4:E12 could

possibly bind to various known p70-S6K binding partners to give rise to the

phenotype observed. We would also like to study whether stable over-

expressing E4:E12 cells could possess migratory or invasive properties.

Over-expression of proposed p70-binding partners fused with FLAG

tags in T47D cell lines followed by immuno-precipitation

Proposed partners in the PI3K/Akt/mTOR pathway that may interact

with S6K fusion construct E4:E12 are the 2 isofoms of RPS6KB1 namely p85-

20

S6K and p70-S6K. Other proposed partners included RPS6, PPP2CA, mTOR

and IRS-1 (Ferrari et al. 1993; Nojima et al. 2003; Zhang et al. 2008).

However as mTOR and IRS require more time to generate since they are

rather large fragments. Thus we were able to generate only p85-S6K, p70-

S6K, RPS6 and PPP2CA, which are fused with FLAG tags. We sought to

determine the binding interactions between these proposed partners and

E4:E12 via co-transfection and transient over-expresion (Figure 7).

Figure 7: Expression profile of co-transfected T47D cells. Left: co-transfected Mock-HA cells with proposed binding FLAG tag partners. Right: co-transfected S6K E4:E12-HA with proposed binding FLAG tag partners

Proposed FLAG tagged partners p85-S6K (85 kDa), p70-S6K (70 kDa),

PPP2CA (35.6 kDa) and RPS6 (28.7 kDa) displayed high expression profiles

while E4:E12 displayed a low expression profile across the board. However,

T47D cells that are co-transfected with p85-S6K and E4:E12 displayed slightly

higher expression level as compared to other FLAG tagged constructs.

21

Figure 8: Co-immunoprecipitation using anti-FLAG-m2 mouse IgG and detection using anti-HA rat mAb with anti-rat goat antibody.

Immuno-precipitation was conducted to determine if the protein

products could interact with each other. However, as the antibody used for the

pull-down was anti-FLAG mouse mAb and detected was using anti-HA rat as

primary antibody while anti-rat IgG goat antibody as secondary antibody

resulted in species cross-reactivity and no clear result could be detected

(Figure 8). Anti-HA rabbit antibody will be used in the subsequent experiments

in order to detect the presence of HA tagged E4:E12 in the immuno-

precipitated samples.

22

DISCUSSION

Functional possibilities of S6K fusion constructs

The N terminal region of p85/p70-S6K has been extensively studied

and reported to be involved in the regulation of activity by binding to the C

terminus (Mahalingam and Templeton 1996; Dennis et al. 1998; Pullen et al.

1998). Since all fusion constructs still retained their N terminus despite lacking

a functional kinase domain, they could possibly bind to functional p85/p70-

S6K C terminus in order to relieve the inhibition and bypass the first step of

activation. Thus the fusion constructs would exert a dominant negative effect

and eventually enhance the activity of the functional p85-S6K or p70-S6K.

Recent studies have also identified a conserved TOS motif in both N

terminus and C terminus region, which is required for binding to

mTOR/RPTOR in order to facilitate the phosphorylation of T389 (Schalm et al.

2002; Nojima et al, 2003). As described in Dennis and Pullen et al, the

phosphorylation of this residue allows the S6K protein to adopt a open

confirmation thereby allowing PDK1 to bind and activate it through

phosphorylation of the T229 in the kinase domain. This also suggests that a

fusion construct could possibly bind to mTOR/RPTOR to enhance functional

p85-S6K/ p70-S6K activation.

Functional p70-S6K is also reported to be involved in the regulation of

IRS-1 via negative feedback mechanisms (Zhang et al., 2008). A dominant

negative mutant S6K as described in the same paper prevented the

degradation of the IRS-1 when stimulated by tumour necrosis factor alpha

(TNF-α). This dominant negative mutant described (Addgene, Plasmid 8985:

pRK7-HA-S6K1-KR) only possessed a functional N terminus that is encoded

by exons 1 to 5 and its protein profile is similar to that of our fusion constructs.

In essence, these studies suggest that all fusion constructs could

possibly sustain the activation of PI3K/Akt/mTOR pathway and eventually

result in the constitutive activation of all functional counterparts within this

pathway. Thus in theory, phenotypes that are associated with the over-

expression of constitutive active S6K mutants should manifest with the over-

expression of these fusion constructs should the hypothesis be true.

23

Over-expression of S6K fusion constructs could confer a cell growth

advantage when cells are cultured in low serum conditions.

In low serum conditions (3% FBS), over-expression of S6K fusion

constructs E1:E8, E4:E12, E1:E11 and E2:E11 showed a marked significant

increase in colony size as compared to control. However, this was not the

same case when the cells were cultured in ample serum conditions (10%

FBS). Surprisingly, we observed a larger number of colonies of about two-fold

difference when cells were cultured in ample serum conditions as compared

to low serum conditions. This could be possibly relate to the fact that these

cells aggregate together in low serum conditions in order to survive better

thus resulting in lower numbers of colonies but with increased size. However,

differences in colony size in relation to serum conditions were only significant

(Student’s T-test, p<0.05) for control (Mock) and fusion constructs E1:E7 and

E2:E11. This phenotype observed was also only documented between

samples from a single biological set. It also coincided with preliminary

screening data conducted by other previous students during their stint (not

presented). Nonetheless, biological replicates are still required in order to

reach a definitive conclusion. If this phenotype was to be validated true, then it

would suggest that the fusion constructs have a part to play in the over-

activation of the functional counterparts within the PI3K/Akt/mTOR pathway.

Scattered colonies observed for over-expressing E4:E12 T47D cells

suggest a migratory phenotype.

Over-expression of E4:E12 in T47D resulted in an average of 53.5%

scattered colonies as scored by participants. This scattered colony phenotype

suggests that over-expression of this construct may be involved in intrinsic

pathways that contributed to such an observation.

In order to be able to investigate further, we have to look at pathways

and effectors implicated in the actin cell remodeling. Activating missense

mutations of PI3K reported in breast and ovarian cancers resulted in the

oncogenic phenotype, which promotes cell survival and increased cell

migration (Qian et al., 2004; Levine et al., 2005; Meng et al., 2006). All 3

effectors PI3K, mTOR and p70-S6K are required in the induction of actin

filament remodeling as reported in Qian et al. (2004). Since cell migration is

24

closely related to invasive capability of a tumour cell. It would be interesting to

understand the biological mechanisms in which fusion construct E4:E12 could

give rise to this phenotype observed.

Higher expression profile of E4-E12 was observed in cells co-

transfected with p85 and p70 as compared to other FLAG constructs.

Increased expression of E4:E12 was observed from the western blot

(refer to Figure 7) for T47D cells which were transiently transfected with

functional p85 and p70 S6K. The result suggest 2 possibilities where the first

would be over-expression of p85/p70 S6K may give rise to increased protein

synthesis thus explaining the increased protein expression profile of fusion

construct E4:E12. While the second possibility would be that with over-

expression of the fusion construct E4-E12, IRS-1 is prevented from

degradation thus sustaining the activation of endogenous S6K eventually

resulting in a constant loop activation of increased protein synthesis.

In order to understand whether this over-expression could be due to

which possibility mentioned. Another western blot has to be performed using

antibodies targeted at IRS-1, PI3K phospho-RPS6, RPS6, Akt and phospho-

Akt. If protein expression levels are high across the board, then the second

possibility as mentioned would make a more convincing conclusion. However,

if protein levels are high only for phospho-RPS6 and RPS6 then the first

possibility mentioned would be considered as a plausible conclusion.

Future work

Currently, IRS-1 and mTOR FLAG tagged constructs are still in the

process of being cloned and generated. These constructs upon completion

will be used for transient co-transfection and subjected to immuno-

precipitation in order to find out if E4:E12 fusion construct could interact with

these 2 proposed binding partners. This will allow us to obtain a better

understanding of the mechanisms in which this dominant negative S6K

protein could contribute to the cancer cell progression.

To gain a better conclusive finding on whether over-expression of

E4:E12 could contribute to cell proliferation and migration phenotype similar to

that of PI3K active mutants as reported. Functional PI3K constructs could be

25

generated and over-expressed in T47D cells. If these 2 populations of cells

displayed a similar phenotype, then it would suggest that fusion construct

E4:E12 work via prevention of IRS-1 degradation and eventually leading to

constant loop activation of the PI3K/Akt/mTOR pathway as described in the

first section. Having said so, an over-expressing IRS-1 cell line could also be

generated in the meantime to understand if both IRS-1 over-expression and

PI3K over-expression would give rise to the same phenotype observed.

Other studies such as scratch or wound assay have to be conducted in

order to understand if the cell migratory phenotype is observed in stable over-

expressing E4:E12 cells. Invasion assay will proceed once the migratory

phenotype is validated.

26

CONCLUSIONS

In this study, we have identified that over-expression of fusion

construct E4:E12 could confer a cell growth advantage. We have also

identified a scattering phenotype, which suggests that the over-expression

confers a migratory phenotype. These results suggests a possibility that the

fusion construct may act similarly as a dominant negative mutant of S6K and

could possibly give rise to the phenotypes observed through enhanced activity

of functional p85/p70-S6K within the PI3K/Akt/mTOR pathway. Extensive

studies still has to be conducted in order to seek more insight into the exact

molecular mechanisms of this fusion construct and its functional role in the

oncogenic phenotype.

27

REFERENCES Calvo-Garrido, J., S. Carilla-Latorre, et al. (2008). "Vacuole membrane protein 1 is an endoplasmic reticulum protein required for organelle biogenesis, protein secretion, and development." Mol Biol Cell 19(8): 3442-3453. Dennis, P. B., N. Pullen, et al. (1998). "Phosphorylation sites in the autoinhibitory domain participate in p70(s6k) activation loop phosphorylation." J Biol Chem 273(24): 14845-14852. Ferrari, S., R. B. Pearson, et al. (1993). "The immunosuppressant rapamycin induces inactivation of p70s6k through dephosphorylation of a novel set of sites." J Biol Chem 268(22): 16091-16094. Fullwood, M. J., C. L. Wei, et al. (2009). "Next-generation DNA sequencing of paired-end tags (PET) for transcriptome and genome analyses." Genome Res 19(4): 521-532. Levine DA, et al. (2005). “Frequent mutation of the PIK3CA gene in ovarian and breast cancers.” Clin Cancer Res 11:2875–2878 Mahalingam, M. and D. J. Templeton (1996). "Constitutive activation of S6 kinase by deletion of amino-terminal autoinhibitory and rapamycin sensitivity domains." Mol Cell Biol 16(1): 405-413. Meng, Q., Xia, C., et al. (2006). “Role of PI3K and AKT specific isoforms in ovarian cancer cell migration, invasion and proliferation through the p70S6K1 pathway.” Cell. Signal. 18, 2262–2271 Ng, P., J. J. S. Tan, et al. (2006). "Multiplex sequencing of paired-end ditags (MS-PET): a strategy for the ultra-high-throughput analysis of transcriptomes and genomes." Nucl. Acids Res. 34(12): e84-. Nojima, H., C. Tokunaga, et al. (2003). "The mammalian target of rapamycin (mTOR) partner, raptor, binds the mTOR substrates p70 S6 kinase and 4E-BP1 through their TOR signaling (TOS) motif." J Biol Chem 278(18): 15461-15464. Pon, Y. L., H. Y. Zhou, et al. (2008). "p70 S6 kinase promotes epithelial to mesenchymal transition through snail induction in ovarian cancer cells." Cancer Res 68(16): 6524-6532. Prensner, J. R. and A. M. Chinnaiyan (2009). "Oncogenic gene fusions in epithelial carcinomas." Current Opinion in Genetics & Development 19(1): 82-91. Pullen, N., P. B. Dennis, et al. (1998). "Phosphorylation and activation of p70s6k by PDK1." Science 279(5351): 707-710.

28

Rabbitts, T. H. (2009). "Commonality but Diversity in Cancer Gene Fusions." Cell 137(3): 391-395. Ropolo, A., D. Grasso, et al. (2007). "The pancreatitis-induced vacuole membrane protein 1 triggers autophagy in mammalian cells." J Biol Chem 282(51): 37124-37133. Ruan, Y., H. S. Ooi, et al. (2007). "Fusion transcripts and transcribed retrotransposed loci discovered through comprehensive transcriptome analysis using Paired-End diTags (PETs)." Genome Res 17(6): 828-838. Sauermann, M., O. Sahin, et al. (2008). "Reduced expression of vacuole membrane protein 1 affects the invasion capacity of tumor cells." Oncogene 27(9): 1320-1326. Schalm, S. S. and J. Blenis (2002). "Identification of a conserved motif required for mTOR signaling." Curr Biol 12(8): 632-639. Sinclair, C. S., M. Rowley, et al. (2003). "The 17q23 amplicon and breast cancer." Breast Cancer Res Treat 78(3): 313-322. Singapore Cancer Society. (Retrieved on 09 April 2010) http://www.singaporecancersociety.org.sg/lac-fcb-how-common-is-breast-cancer.shtml Soda, M., Y. L. Choi, et al. (2007). "Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer." Nature 448(7153): 561-566. Tomlins, S. A., B. Laxman, et al. (2007). "Distinct classes of chromosomal rearrangements create oncogenic ETS gene fusions in prostate cancer." Nature 448(7153): 595-599. Qian, Y., Corum L. et al. (2004). “PI3K induced actin filament remodeling through Akt and p70S6K1: implication of essential role in cell migration.” Am. J. Physiol. Cell Physiol. 286: C153–C163 Zhang, J., Z. Gao, et al. (2008). "S6K directly phosphorylates IRS-1 on Ser-270 to promote insulin resistance in response to TNF-(alpha) signaling through IKK2." J Biol Chem 283(51): 35375-35382. Zhou, H. Y. and A. S. Wong (2006). "Activation of p70S6K induces expression of matrix metalloproteinase 9 associated with hepatocyte growth factor-mediated invasion in human ovarian cancer cells." Endocrinology 147(5): 2557-2566.