.. 3 g ,_

 

THESB

/)_

l
A IA/‘L

This is to certify that the
thesis entitled

MOLECULAR MECHANISM FOR CD82 SUPPRESSION OF
HGF-INDUCED MET ACTIVATION AND INVASION

presented by

SUSAN M SPOTTS

has been accepted towards fulfillment
of the requirements for the

MS. degree in Biochemistry and Molecular
Biology

 

 

C; Mr; /C /W

/ Major Professor’s Signature
57,] fill/2,1910

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

 

LIBRARY
Michigan State
University

 

 

 

V_!._._.—._--.----‘_‘-—;-‘—4—-—n-.-n-g-n-u-

_A_.—.—---o-l--n-o-I-I-IQv-a--.‘_A4

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5108 K:IProj/Acc&Pres/ClRC/DateDue.indd

MOLECULAR MECHANISM FOR C082 SUPPRESSION OF HGF-INDUCED
MET ACTIVATION AND INVASION

By

Susan M. Spotts

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Biochemistry and Molecular Biology

2010

ABSTRACT

MOLECULAR MECHANISM FOR C082 SUPPRESSION OF HGF-INDUCED
MET ACTIVATION AND INVASION

By
Susan M. Spotts

Met is up-regulated in metastatic prostate cancer which can lead to
enhanced migration and invasion of tumor cells [1]. 0082 is a member of the
tetraspanin family and has been identified as a metastasis suppressor in prostate
cancer [2, 3]. CD82 can suppress ligand-induced Met activation and invasion in
prostate tumor cells [4]. The objective of this project was to determine the
mechanism by which 0082 suppresses ligand-dependent activation of Met and
invasion.

We’ve determined that 009, but not CD151, integrin 03 or 06, is required
for 0082 suppression of HGF-induced Met activation. 009 was not required for
0082 suppression of invasion. We have determined that C0151 and 03 integrin
contribute to Met activation and invasion in P03 cells. lntegrin 0681 and C051
were found to be associated with Met in the absence of 0082. CD151and
integrin 03 were also shown to associate in the absence of C082. In the
presence of C082, CD151 and integrin (13 associations are reduced and 0082 is
now able to associate with integrin 06 and 00151. Our current model is that
without 0082; integrins (1381 and 0681 are able to interact with Met via
interactions with 00151 to promote Met activation and invasion. In the presence
of 0082, 009 and CD151interact with 0082 drawing both CD151 and integrin

0681 into the TERM resulting in suppression of Met activation and invasion.

ACKNOWLEDGEMENTS

There have been many people who have contributed to and supported me
through the process of earning my master's degree. First, I would like to thank
my mentor, 0r. Cindy Miranti, whose continued support, and guidance has made
this possible. Thank you for encouraging me and giving me an opportunity to
grow as a scientist.

I want to thank my committee members, Dr. John Wang and Dr. Kathy
Gallo, for their advice and guidance throughout this process. I am grateful for
your support and assistance.

To all past and current Miranti lab members, I am indebted to you all for
the support and friendship you have shown me over the past three years. I will
never forget your kind words and camaraderie throughout my time here. I will
never be able to put into words all that you have meant to me and my
appreciation.

I would also like to thank my family for their continuous love and support
throughout my life. To my mother and father, thank you for believing in me even
when I couldn’t. I would have never made it without you. To my aunt and uncle,
thank you for opening your home to me and providing love and support when
times were difficult.

It’s with your support that I am encouraged to continue my education to

become a better scientist.

iii

TABLE OF CONTENTS

LIST OF FIGURES ............................................................................... v
INTRODUCTION ................................................................................... 1
Hypothesis .................................................................................. 5
CHAPTER 1: IDENTIFICATION OF REQUIRED TETRASPANIN ENRICHED
MICRODOMAINCONSTITUENTS ........................................................... 10
Rationale ................................................................................... 10
Results ......................... 11
lntegrin 03 is not required for CD82 suppression of HGF-induced
Met activation but contributes to Met activation ......................... 11
lntegrin 03 is not required for CD82 suppression of invasion but
contributes to P03 cell invasion ............................................. 15
lntegrin 06 is not required for Met activation nor for CD82
suppression of Met .............................................................. 19
lntegrin 06 is not required for 0082 suppression of invasion but
contributes to P03 cell invasion ............................................. 22
009 is required for 0082 suppression of HGF-induced Met
activation .......................................................................... 25
009 is not required for 0082 suppression of invasion ................ 28
00151 is not required for 0082 suppression of HGF-induced Met
activation but contributes to Met activation in PC3 cells .............. 31
00151 is not required for 0082 suppression of invasion but
contributes to P03 cell invasion ............................................. 32
Discussion ........................................................................ 37
CHAPTER 2: PROTEIN-PROTEIN INTERACTIONS ................................... 39
Rationale ................................................................................... 39
Results ..................................................................................... 40
Met associates with integrin 0681, but not with integrin 0381 ....... 40
In the presence of CD82, integrin 06 is drawn into the tetraspanin-
enriched microdomain and associates with 0082 and CD15 ........ 45
In the presence of 0082, there is a reduction in the number of CD9
and 00151 associations ...................................................... 49
Discussion ................................................................................. 49
CHAPTER 3: CONCLUSIONS AND PERSPECTIVES ................................ 55
Future Work ............................................................................... 60
APPENDIX A: Materials and Methods ...................................................... 62
REFERENCES ..................................................................................... 66

iv

INTRODUCTION

Prostate cancer is the second leading cause of cancer deaths in men in
the United States [5]. While primary prostate tumors are often curable,
metastatic prostate tumors are not and ultimately fatal. Understanding the
mechanisms of metastasis development is crucial for developing novel
therapeutic targets for this incurable disease.

Met is a receptor tyrosine kinase that has been shown to be up regulated
in metastatic prostate cancer which can lead to proliferation and invasion of
tumor cells [1]. According to a study in 2002, 51% of locally invasive prostate
carcinomas were Met positive and 100% of prostate metastases were Met
positive. Met can be activated through its ligand, hepatocyte growth factor
(HGF), or through a ligand-independent mechanism mediated by cell adhesion
via integrins.

0082/KAI1 was first identified in a rat prostate cancer model as a
metastasis suppressor gene [2]. 0082 expression is often lost in metastatic
prostate tumors and correlates with poor prognosis. Re-expression of 0082 in
human tumor cell xenografts suppresses the development of metastatic tumors
with little to no effect on primary tumor growth [3]. In vitro studies have shown
that functional 0082 can regulate cell aggregation, adhesion, motility, invasion,
apoptosis, and morphogenic processes [3]. Normal prostate cells express 0082
which inhibits cross-talk between Met and integrins, but the loss of 0082 in

prostate tumor cells allows for enhanced integrin- and ligand-mediated activation

of Met, which may promote metastasis [4]. These studies provide evidence that
0082 functions as an inhibitor of metastatic potential.

0082, a member of the tetraspanin super family, has classical properties
shared by other tetraspanin family members [3]. Tetraspanins possess short
carboxy— and amino-terminal tails, four transmembrane domains, two
extracellular loops, and a short intracellular loop. The larger extracellular domain
(E02) of the tetraspanin 0081 has been crystallized and molecular modeling
studies have provided significant insight into the predicted structure of
tetraspanins. Within the family, tetraspanins share very little sequence
homology; however, there are a few key residues that all tetraspanins share
which are important for maintaining their 30 structure. The E02 domain is
divided into a constant region which allows for dimerization and a variable region
[6]. The variable region is thought to confer specificity via interactions with non-
tetraspanin protein binding partners. The E02 domain contains either four or six
conserved cysteine residues which have been proposed to form 2 or 3 disulfide
bridges giving tetraspanins their characteristic tertiary structure. A short region
within the E02 of the tetraspanin 00151 has been shown to be important for
interacting with the extracellular domain of the 03 integrin subunit [7].

Tetraspanins lack catalytic activity and function as ‘molecular facilitators’
to organize surrounding membrane proteins into a complex known as the
tetraspanin enriched microdomain (TERM) [3]. Tetraspanins associate laterally
within the membrane to cluster various protein binding partners in tetraspanin-

enriched microdomains to regulate downstream signaling pathways [8]. The

tetraspanin-enriched microdomain, TERM, is distinct from lipid rafts [9]. Various
reports have found tetraspanin-enriched microdomains to be located on the
plasma membrane, endosomes, and exosomes.

Various tetraspanins, integrins, receptor tyrosine kinases, and other
membrane proteins have been shown to exist within tetraspanin-enriched
microdomains. Our lab has previously shown that 0082 suppresses HGF-
induced Met activation through an unknown mechanism. Supporting papers
have since been published providing additional evidence that 0082 inhibits Met
activation and its cross-talk with integrins [10, 11]; however, the mechanism is
still elusive. We hypothesize that the suppressive function of CD82 is regulated
by specific constituents of the tetraspanin-enriched microdomains.

Some of the tetraspanin-enriched microdomain associated proteins
include 00151, 009, and integrins. 00151 has been shown to up regulated in
metastatic prostate cancer. 00151 is highly expressed in epithelial and
endothelial cells where it localizes to cell-cell junctions to modulate cell migration
and invasion [9]. Studies have shown that CD151 forms a stable, specific, and
stoichiometric association with integrin 0381 via the extracellular E02 domain [7].
The same group has also determined that CD151 associates indirectly with
integrin 06 [12].

009 has been implicated in cell adhesion, motility, proliferation, and
differentiation with a functional role as a tumor suppressor [9]. 009 has been
shown to be up or down regulated in metastatic tumors depending on the study.

In 2000, Shi et al. reported that CD9 associates with TGF-d and regulates TGF-

induced EGFR activation [25]. It has been reported that 009 and 00151
associate with 0082 through conserved polar residues within the
transmembrane domains, which are conserved throughout the tetraspanin super
family [13]. This group also reports that 009 and 00151 are not required for
0082 suppression of invasion.

Tetraspanins are palmitoylated and palmitoylation is important for the
assembly of the tetraspanin-enriched microdomains [9]. Multiple palmitoylation
sites have been indentified on 009, 00151, and 0082. Palmitoylation partly
stabilizes tetraspanin—tetraspanin interactions within the tetraspanin-enriched
microdomain [14]. Mutations of CD9 palmitoylation sites impaired its association
with 0081 and 0053 [14]. A previous paper showed that palmitoylation-deficient
CD151 weakens its association with the 0381 integrin within the tetraspanin web
[15]. A palmitoylation-deficient 0082 mutant largely reverses the suppression
function of CD82 on cell migration and invasion of prostate tumor cells [16]. In
2008, Shanna et al. identified DHH02 as the enzyme responsible for
palmitoylating 009, 00151, and 0082. DHH02-dependent palmitoylation aided
009 and 00151 associations as well as associations with integrin 03 [17].
Induction of DHH02 resulted in increased cell-cell contact [17]. There is also
evidence to show that integrins may be palmitoylated which may aid in their
ability to associate with tetraspanins.

Metastasis is a disease of abnormal cell adhesion in which a small
population of tumor cells detaches from the primary tumor, migrate, and invade

other tissues in the body [18]. These processes are linked to cell adhesion.

Normal epithelial cells balance cell-cell interactions, generally known to
negatively regulate cell growth and migration, with cell-extracellular matrix (ECM)
interactions that positively regulate cell growth, migration, invasion, and survival
[18]. During metastasis development there is a shift in the balance towards
increased cell-ECM interactions and decreased cell-cell interactions. lntegrins
are heterodimeric transmembrane proteins that are able to activate signaling
transduction pathways by interacting with extracellular matrix proteins [19].
lntegrins consist of an alpha subunit and a beta subunit that specifies which
particular extracellular matrix protein the integrin will bind. Changes in integrin
and matrix protein expression as well as changes in integrin affinity for their ECM
substrate are observed in most cancers and this can lead to an increase in
invasion and metastasis [19]. These particular integrins are specific for binding
to the laminin extracellular matrix protein. Multiple groups have reported that
integrin 03 and integrin 06 associate with 00151 via the E02 domain [16]. In
2005, Zhang et al. reported that integrin 06 physically interacts with 0082 [16].
How 0082 association with integrins affects integrin function and whether this

complex influences other signaling transduction pathways remains controversial.

Hypothesis

0082 has been shown to suppress HGF-induced c-Met activation [4];
however, the molecular mechanism of this process is still elusive. 0082 lacks
catalytic activity and has been suggested to function as a ‘molecular facilitator’ to

organize tetraspanins, integrins, and other membrane proteins into a tetraspanin-

enriched microdomain [3]. If 0082 functions through its involvement in the
TERM, we would predict that the specific constituents and specific protein
interactions within the microdomain are important for the ability of 0082 to
suppress Met activation.

Our lab has previously shown that in the absence of HGF, there are basal
levels of Met activation when the cells are plated on matrix. This basal level of
activation is through an integrin-mediated mechanism. In the absence of HGF
and when the cells are in suspension, there is no Met activation. Therefore, our
hypothesis is that 0082 suppresses Met activation by associating with integrins
to displace them from a complex with Met. 0082 may regulate the spatial and
temporal positions of integrin and Met associations.

Preliminary data generated in the lab has shown through co-
immunoprecipitation studies that 0082, 009, and 00151 complex together when
extracted with a mild detergent. Within these tetraspanin complexes, the 0381
integrin, but not the 0681 integrin, was found to be associated with these
tetraspanins complexes as well.

My hypothesis is that in the absence of 0082, integrin 0381 interacts with
Met resulting in activation of downstream signaling pathways for invasion and
metastasis (Figure 1A). In the presence of 0082, 0082 interacts with integrin
0381 and draws them into the tetraspanin-enriched microdomain (Figure 1B).
Withdrawal of the integrins from the Met complex disrupts the ability of Met to

aggregate and to signal downstream. The specific composition of this

microdomain is essential for the ability of 0082 to suppress HGF-induced Met

activation.

Figure 1. Proposed molecular mechanism for 0082 suppression of Met
activation. (A) In 0082 non-expressing cells, 0381 integrin associates with Met
leading to Met aggregation and downstream signaling of cell invasion and
metastasis. (B) In 0082-expressing cells, CD82 associates with the integrin
0381 which becomes displaced into the tetraspanin-enriched microdomain

(TERM). Therefore, Met is unable to aggregate and signal downstream.

Figure 1. Proposed molecular mechanism for CD82 suppression of Met
activation.

  
      
        
 

        

          
 

      
  

[I'll I’ll-iii]

ltlr III!!!

    
  

             

 

'l'l’l’l’l’llljljlj

; ‘f'hzf‘h .1" " :I': ‘ I q»
. 3‘1","!.:"’l.'.'._'.'.’.l."-'.'.'.'.".’, ., , .
ft‘ thITIfitththhtfktthhtttifig ‘ _f? K» t;
. .l i1)? L?” 3‘ 2;:-

 

‘IIIIIII

   
   

   

  

'1'!"".'I

rill-Hill. I ,(r j l '_

    

 

Illllllli

  

 

Invasion 8.
Metastasis

 

 

 

 

          

   

 

 

 

IIIII'

fittfifitI

                          
     

     

2' . “I: .34! i will} till II'lg'Ia'll'lflg-III‘IIIIIII'IIII‘I‘II.[III’I'I

i'Ill .

  

         

Ilj'll'lfi'l'l-I'I'I’l'i

I liltttlr

 

      
   

 

 

rrlv it If! it I II"

IIII'I'I'III .I lift ill I It" i It". 5

 

          

        
 
     

         
 

                          
      

 

Invasion 8
Metastasis

 

 

CHAPTER 1: IDENTIFICATION OF REQUIRED TETRASPANIN-ENRICHED

MICRODOMAIN CONSTITUTUENTS

Rationale

My hypothesis states that in the presence of 0082, Met activation and
signaling are disrupted by CD82 displacing integrin 0381 from the Met complex
and drawing it into the tetraspanin-enriched microdomain. We also hypothesize
that CD82 functions through a definite microdomain which requires the presence
of specific protein-binding partners such as integrins and other tetraspanins.

To determine these particular protein-binding partners that are responsible
for 0082 suppression, PC3 cells expressing either empty vector or 0082 were
transiently transfected with siRNA specific to the protein of interest. The cells
were stimulated with increasing concentrations of HGF, immunoprecipitated for
Met, and then immunoblotted
for tyrosine phosphorylation and total Met protein. We expect that in 0082-
expressing cells, a larger concentration of HGF will be required to reach an
equivalent amount of Met activation in a non-expressing 0082 cell. We then
predicted that P03 cells expressing 0082 and have a reduction in a protein that
is required for CD82 suppression function will have adequate Met activation at

lower concentrations of HGF similar to non-0082 expressing cells.

10

Results

lntegrin 03 is not required for 0082 suppression of HGF-induced Met
activation but contributes to Met activation.

Based upon the model presented in Figure 1, it would be predicted that 03
integrin is necessary for dimerization of Met to stimulate its activity. To
determine whether integrin 03 is required for Met activation, PC3 cells stably
expressing CD82 or empty vector were transiently transfected with either a pool
of integrin 03-specific siRNAs or a control siRNA. The control siRNA is a
scrambled sequence that has no known target based on a BLAST search. The
same control siRNA sequence will be used throughout all assays requiring
siRNA. After transfection, cells were stimulated with increasing concentrations of
hepatocyte growth factor (HGF). Met protein was immunoprecipitated from cell
lysates and analyzed by immunoblotting for tyrosine phosphorylation and total
Met protein. My data shows that approximately 85% of integrin 03 protein
expression was knocked down in P03 cells expressing empty vector resulting in
reduced Met activation compared to cells transfected with control siRNA (Figure
2A,B). Knockdown of integrin 03 by approximately 90% had no effect on Met
activation in 0082-expressing PC3 cells.

This data suggests that integrin 0381 may be required for Met to reach maximal
activation. However, this particular assay lacks an output to determine what level
of Met activation results in increased migration, proliferation, or invasion. An

experiment that includes a physiological output such as proliferation or migration

is required to firmly state that 0082 suppression of Met activation results in a

specific physiological response.

12

Figure 2. lntegrin 03 is not required for 0082 suppression of HGF-induced
Met activation but contributes to Met activation. (A) P03 cells stably
expressing 0082 (cl57) or empty vector (cl31-) were transiently transfected with
either a pool of 4 integrin 03 (ITGA3) specific siRNAs or scramble siRNA
(control) and starved for 48 hours. The cells were then stimulated with O, 10, 25,
50, and 100 ng/mL of HGF for 10 minutes. Met phosphorylation was measured
by immunoblotting immunoprecipitates with a phospho-specific antibody. Total
levels of Met protein in immunoprecipitates were measured by immunoblotting
stripped blots with Met antibody. (B) ITGA3 and CD82 expression were
monitored by immunoblotting of whole cell lysates. The numbers below the
figures indicate the percent knockdown values of ITGA3, as determined by

densitometry. Susan Spotts, unpublished data

Figure 2. lntegrin 03 is not required for 0082 suppression of
HGF-induced Met activation but contributes to Met activation.

 

 

 

 

 

 

 

 

 

 

 

 

A. PC3.Vector.cl31-
Control sIRNA ITGA3 siRNA
Met IP: .. 0 10 25 50 100 0 10 25 50 100 ng/mL HGF
P-Tyr _ ..
Blot: I "" H . . ’-"“'“‘== '7“. ~ ~|- Met
aroc|~ H BIIfl‘HE‘IMet
PC3.0082.cI57
Control siRNA ITGA3 siRNA
Met IP: 7 0 10 25 50 100 0 10 25 50 100nglmL HGF
P-TYr ..., ’
Blot: it” Q g .
Met muflfi‘W'i no...

Blot: uufluu—~--—n Met

 

 

 

  

 

 

 

 

 

 

 

B. PC3.Vector.cl31-
Control siRNA ITGA3 siRNA
0‘ 1O 25 50100 0 10 25 50 100nglmLHGF
"GA? - ITGA3
wcl. 7
C08? ~0082
wcl.
0 0 0 0 0 82 85 85 85 87 % Knockdown
PC3.0082.cI57
Control siRNA ITGA3 siRNA
10 25 50100 0 10 25 50 100ngImL HGF
ITGA3

 

 

“MD...- -
r'Trt r . . I

wcl:

I- ITGA3

 

LLA LELEL A LMhAJ

0 O O 94 91 90 85 90 % Knockdown

 

lntegrin 03 is not required for 0082 suppression of invasion but
contributes to P03 cell invasion.

Our lab has previously demonstrated that Met is required for P03 cell
invasion of Matrigel [4]. Thus, if 03 integrin is required for Met activation, then
blocking 03 integrin should also suppress invasion. To determine if integrin 03 is
required for PC3 cell invasion, PC3 cells stably expressing either empty vector or
CD82 were transiently transfected with a pool of integrin 03-specific siRNAs.

The cells were then placed in a Boyden chamber containing Matrigel and laminin
as a chemo-attractant and allowed to invade for 72 hours. Knockdown of integrin
03 in PC3 cells expressing empty vector reduced the number of cells that were
able to invade the Matrigel, and appear in the lower chamber, compared to their
control counterparts (Figure 3A). Knockdown of 03 integrin was quantified using
Quantity One software to determine the density of the bands in Figure 3C. The
bands present in the integrin 03 siRNA lanes were compared to the bands in the
control siRNA lanes for each cell type. Knockdown of 03 integrin in empty-vector
PC3 cells by 76% resulted in 1.8—fold decrease in invasion as shown in Figure
3B,C. PC3 cells expressing 0082 had only a 40% reduction in integrin 03
protein expression (Figure 30). While the presence of 0082 suppressed
invasion 4-fold, loss of 03 integrin did not restore the invasive capacity of P03
parental cells, but rather resulted in a further reduction of invasive capabilities by
an additional 65-fold (Figure 3A,B). These data suggest that integrin 03 partially
contributes to P03 invasion, and may be reflective of its action on Met based on

the effect of 03 integrin on Met activation (see Figure 2). However, the presence

IS

of 03 integrin with 0082 is not required for 0082 to suppress invasion. The
further reduction in invasion in CD82 cells may reflect other 03-regulated events

that are not controlled by CD82.

16

Figure 3. lntegrin 03 is not required for 0082 suppression of invasion but
contributes to P03 cell invasion. (A) PC3 cells stably expressing CD82 (cl57)
or empty vector (cl31-) were transiently transfected with either integrin 03
(ITGA3) specific siRNA or scramble siRNA (ctrl) and starved for 48 hours. Cells
were placed in a Boyden chamber and allowed to invade for 72 hours. Cells
which passed through the matrigel and appeared on the lower membrane were
stained with crystal violet. (B) Cells were counted in 5 different fields and
averaged. Three separate experiments were then averaged together. The error
bar represents one standard deviation above the mean. (0) 72 hours after cells
were placed in the Boyden chamber, parallel cultures of cells were lysed and
immunoblotted for ITGA3 and CD82 expression. The numbers below the figure
indicate the percent knockdown values of ITGA3, as determined by densitometry.

Susan Spotts, unpublished data

Figure 3. lntegrin a3 is not required for CD82 suppression of
invasion but contributes to P03 cell invasion.

A.

 

 

 

 

 

 
 

 

 

 

 

 

 

   

 

 

PC3.Vector.cl3 PC3.CD82.cI5
5"".13‘5.‘ ' _' \ a.- I '
- . 5' 27.. . + Scramble
(f 1‘. \“u '2' .1; m. siRNA
. j k" . ’1
z," - . Q
5.1 " } - '-
~ + ITGA3 siRNA
.'- 1,.
160 - IScramble siRNA
95’ 14° ‘ ITGA3 siRNA
0
0 120 q
’8
13 100 ~
3
5 80 ~
‘5 so . l
is
E 40
5 ’°’ I
o .l
PC3.Vector PC3.0082
Vector.cl31- 0082.cl57
: siRNA

Ctrl

 

   

ITGA3 blot:

ITGA3 Ctrl ITGA3

     
 

 

 

76

18

 

: - ITGA3

- CD82

lntegrin 06 is not required for Met activation nor for 0082 suppression of
Met.

Loss of integrin 03 did not fully suppress invasion or completely block Met
activation, and 0082 still maintained its capacity to suppress Met activation.
However, loss of 03 integrin strengthened 0082’s inhibition of invasion which
suggests that 0082 may also suppress invasion by an additional mechanism.
lntegrin 06 is another laminin-specific integrin endogenously expressed on
prostate cells [4]. Preliminary data generated in the lab had also shown that
integrin 06 was present in 0082 complexes, albeit at much lower levels. Thus,
0082 might be controlling Met and invasion by regulating both 03 integrin and 06
integrin. To determine if 06 integrin can also control Met activation, PC3 cells
stably expressing 0082 or empty vector were transiently transfected with either a
pool of integrin 06-specific siRNAs or a control siRNA. The cells were stimulated
with increasing concentrations of HGF prior to immunoprecipitation of Met and
immunoblot analysis of Met tyrosine phosphorylation. Knockdown of integrin 06
protein expression by 75 to 92% in P03 cells expressing empty-vector seems to
have had no effect on Met activation. P03 cells expressing 0082 with a 35-72%
reduction in 06 integrin protein did not have altered Met activation (Figure 4A,B).
These data suggest that integrin 06 does not contribute to the ability of 0082 to
suppress Met activation nor does it play a role in stimulating Met activation.
However, the knockdown efficiency in control P03 cells was not optimal to make
a final conclusion (Figure 4B). Knockdown efficiency was determined by

densitometry using

19

Figure 4. lntegrin 06 is not required for Met activation nor for 0082
suppression Met. (A) P03 cells stably expressing 0082 (cl57) or empty vector
(cl31-) were transiently transfected with either integrin 06 (ITGA6) specific siRNA
or scramble siRNA (ctrl) and starved for 48 hours. The cells were then
stimulated with 0, 10, 25, 50, and 100 ng/mL of HGF for 10 minutes. Met
phosphorylation was measured by immunoblotting immunoprecipitates with a
phospho-specific antibody. Total levels of Met protein in immunoprecipitates
were measured by immunoblotting stripped blots with Met antibody. (B) ITGA6
and 0082 expression were monitored by immunoblotting of whole cell lysates.
The numbers below the figures indicate the percent knockdown values of ITGA6,

as determined by densitometry. Susan Spotts, unpublished data

 

 

Figure 4. lntegrin 06 is not required for Met activation nor for
0082 suppression of Met.

A. PC3.Vector.cl31-
Control siRNA ITGA6 siRNA
Met IP: 0 10 25 50 100 0 10 25 50 100ngImL HGF

P-TYI’ 1 .. s? ..=
| a... mu-UI-Met

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Blot:
Met *‘"“"“"‘"t--=-—- Hun-o '.
Blot~“~~ a-na‘al'mt
PC3.0082.cI57
Control siRNA ITGA6 siRNA
Met IP: _ 0 10 25 50 100 0 10 25 50 100nglmL HGF
P-Tyr . . _
BIOt! w . - if?“ a - u ‘ MGt
Met “HM '"'°*'**
Blot: -‘~‘---~- Met
3- PC3.Vector.cl31-
0Control siRNA ITGA6 siRNA
10 25 50 100 0 10 25 50 100ngImL HGF
wcL . ~ .- - a . all-I-
0082 """“‘* " * " - c032
wc': ~ ~ a New 7" 4,, .. ._., «w

 

 

 

PC3.0082.cl57
Control siRNA ITGA6 siRNA

10 25 50100 0 10 25 50 100091mLHGF

ITGA6
“.9- I'm

0082
wcl:

   

 

 

 

 

21

Quantity One software and comparing control cells with HGF treatments to

integrin 06 siRNA treated cells stimulated with equal concentrations of HGF.

lntegrin a6 is not required for C082 suppression of invasion but
contributes to PC3 cell invasion.

lntegrin 0684 has previously been shown to be essential for HGF-induced
invasion in a highly tumorigenic human gastric cancer cell line [20]. To
determine if integrin 06 is required for PC3 cell invasion, PCB cells stably
expressing either empty vector or 0082 were transiently transfected with a pool
of four integrin aS-specific siRNAs. The cells were then placed in a Boyden
chamber containing Matrigel and laminin as a chemo-attractant and allowed to
invade for 72 hours. A 39% reduction of integrin 06 in PC3 cells expressing
empty vector and integrin 06 siRNA displayed a 2.8-fold reduction in
invasiveness compared to their control counterparts (Fig 5A,C). The difference
in invasiveness may have been more significant had I been able to achieve a
greater reduction in (16 integrin protein. Loss of (:6 integrin in 0082 expressing
cells had no effect on the ability of CD82 to inhibit invasion (Figure 5A,B). These
data suggest that integrin (16, like 03, is required for invasion of PCB cells, but its

interaction with C082 is not required for C082 to suppress invasion.

22

Figure 5. lntegrin 06 is not required for C082 suppression of invasion but
contributes to PC3 cell invasion. (A) P03 cells stably expressing C082 (cl57)
or empty vector (cl31-) were transiently transfected with either integrin 06
(ITGA6) specific siRNA or scramble siRNA (ctrl) and starved for 48 hours. Cells
were placed in a Boyden chamber and allowed to invade for 72 hours. Cells
which passed through the matrigel and appeared on the lower membrane were
stained with crystal violet. (B) Cells were counted in 5 different fields and
averaged. Three separate experiments were then averaged together. (C) 72
hours after cells were placed in the Boyden chamber, parallel cultures of cells
were lysed and immunoblotted for ITGA6 and 0082 expression. The numbers
below the figure indicate the percent knockdown values of ITGA6, as determined

by densitometry. Susan Spotts, unpublished data

23

Figure 5. lntegrin a6 is not required for C082 suppression of
invasion but contributes to PC3 cell invasion.

A. PC3.Vector.cl31 - PC3.C082.c|57

 

 

9". o , +Scramble siRNA

 

 

 

\ ’h +|TGA6 siRNA

 

 

l Scramble siRNA
14° 1 , ITGA6 siRNA

 

* .

PC3.Vector PC3£082
C. Vector.cl31- C082.c|57
Ctrl ITGA6 Ctrl ITGA6 :siRNA

ITGA6 blot:|‘_i .. “ m .... I- ITGA6

 

 

 

 

C082 blot 1 - C082

 

 

C09 is required for C082 suppression of HGF-induced Met activation.

If 0082 functions through the tetraspanin-enriched microdomain, we
predict that the specific constituents within the microdomain are important for the
ability of CD82 to suppress Met activation. Preliminary co-immunoprecipitation
data generated in the lab had shown that C09 protein in tetraspanin-enriched
microdomains was increased in the presence of 0082. To determine if 009 is
required for 0082 suppression of Met activity, PCB cells stably expressing CD82
or empty vector were transiently transfected with either 009-specific siRNA or a
control siRNA. These experiments were repeated multiple times using two
unique siRNA sequences that are specific to 009 based upon a BLAST search.
The cells were stimulated with increasing concentrations of HGF, Met was
immunoprecipitated from cell lysates, and analyzed by immunoblotting for
tyrosine phosphorylation. An approximately 75% knockdown of 009 resulted in
increased Met activation in PCB cells expressing CD82 compared to CD82
expressing PCB cells transfected with control siRNA (Figure 6A,B). Knockdown
of C09 in PCB cells expressing empty vector appears to have had little to no
effect on Met activation compared to cells transfected with control siRNA.
However, the knockdown of 009 in empty vector—expressing PCB cells was not
very efficient with only a 13-70% reduction in 009 protein levels. Greater
knockdown of 009 may be required to see a change in Met activation. These
data suggest that C082 requires the presence of 009 to suppress HGF-induced

Met activation.

25

Figure 6. C09 is required for CD82 suppression of HGF-induced Met
activation. (A) PCB cells stably expressing C082 (cl57) or empty vector (clB1-)
were transiently transfected with either C09 specific siRNA or control siRNA (ctrl)
and starved for 48 hours. The cells were then stimulated with 0, 10, 25, 50, and
100 ng/mL of HGF for 10 minutes. Met phosphorylation was measured by
immunoblotting immunoprecipitates with a phospho-specific antibody. Total
levels of Met protein in immunoprecipitates were measured by immunoblotting
stripped blots with Met antibody. (B) C09 and CD82 expression were monitored
by immunoblotting of whole cell lysates. The numbers below the figure indicate
the percent knockdown values of C09, as determined by densitometry. Susan

Spotts, unpublished data.

26

Figure 6. CD9 is required for C082 suppression of HGF-induced Met acti-

 

 

 

vation.
A. PC3.Vector.cl31-
Control siRNA 009 siRNA
Meth:_ 0 10 25 50 100 0 10 25 50 100 ngImL HGF

 

Blot: m m-uflmnflflm .Met

 

 

 

       
 

 

 

P03.0082.cl57
Control siRNA CD9 siRNA
Met leg 0 10 25 50 100 0 10 25 50 100nglmLHGF
P-Tyrl ’_ .. :fl
Blot: m in an d - Met
Met ‘
Blot:

 

 

 

 

   

PCB.Vector.cl31-
Control siRNA C09 siRNA
_ 0 1O 25 50100 O 10 25 50 100 nglmL HGF

 

 

 

 

 

 

 

 

 

     
      

C082 -C082
wcl:
0 O 0 O O 13 32 46 66 69% Knockdown
PC3.C082.cl57
Control siRNA CD9 siRNA
7 0 10 25 50100 0 10 25 50100nglmLHGF
009 .. .
wc|: - - _ . mm g .~ . * “ ”all _CDg
r“ r "FT 1 r r l v
323.2 r i a . -coaz

   

Lt i t a t t L

0 0 0 0 43 68 75 80 80% Knockdown

27

C09 is not required for CDBZ suppression of invasion.

Since 009 is required for CD82 suppression of HGF-induced Met
activation and Met is required for PC3 cell invasion, we predicted that knockdown
of 009 in PC3 cells expressing C082 would result in increased invasion. To
determine if CDQ is required for CD82 suppression of invasion, PC3 cells stably
expressing either empty vector or CD82 were transiently transfected with either
CDQ-specific siRNA or control siRNA. The cells were then placed in a Boyden
chamber containing Matrigel and laminin as a chemo-attractant and allowed to
invade for 72 hours. A 50% reduction of CD9 protein in PCB cells expressing
0082 resulted in a minimal increase in invasion that was not statistically
significant compared to control siRNA cells (Figure 7A,B,C). PC3 cells
expressing empty vector had a 54% reduction in CDQ protein and displayed no
change in invasion compared to their control counterparts. These data suggest
that C09 is not required for CD82 to suppress invasion. Due to CDBZ’s
requirement of C09 for HGF-induced Met activation but not for invasion, it
appears that 0082 may function to suppress Met activation and invasion through
two separate mechanisms. Met activation leads to multiple pathways being
activated to regulate cell proliferation, migration, and invasion. An assay that
measures both Met activation via tyrosine phosphorylation and a physiological
output such as growth is needed to understand the disconnection in the

requirement for 009.

28

Figure 7. C09 is not required for C082 suppression of invasion. (A) PC3
cells stably expressing C082 (cl57) or empty vector (cl31-) were transiently
transfected with either CD9 specific siRNA or scramble siRNA (ctrl) and starved
for 48 hours. Cells were placed in a Boyden chamber and allowed to invade for
72 hours. Cells which passed through the matrigel and appeared on the lower
membrane were stained with crystal violet. (B) Cells were counted in 5 different
fields and averaged. (n=6 for all samples except PC3.Vecter.c|31- + Scramble
siRNA where n=3) (C) 72 hours after cells were placed in the Boyden chamber,
parallel cultures of cells were lysed and immunoblotted for C09 and C082
expression. The numbers below the figure indicate the percent knockdown

values of CD9, as determined by densitometry. Susan Spotts, unpublished data.

29

Figure 7. CD9 is not required for CD82 suppression of invasion.

A. PC3.Vector.cl31- PC3.CD82.c|57

+Scramble siRNA

 

 

+ C09 siRNA

 

 

 

 

 

 

 

 

 

 

B.
1.2 - IScramble siRNA
3*? C09 siRNA
1 ..
8» 0.8 -
c
m
5 0.6 « -
E , ..
L2 0.4 . . . _
0.2 . .
0 _ ,
PC3.Vector PC3.0082
C. PC3.Vector.cl31-PCB.CD82.cI57
Ctrl CD9 Ctrl C09 : siRNA
l _ ,. .
C09 b °t I” w a. m - cog
CD82 blot

.. - C082

 

 

 

 

 

30

CD151 is not required for CD82 suppression of HGF-induced Met
activation, but contributes to Met activation in PC3 cells.

Once we had determined that CDQ was required for C082 suppression of
Met activation, I assessed the requirement of other tetraspanins for 0082
inhibition of Met activation. CD151 is often up—regulated in a variety of cancers
and CD151 ablation results in decreased cell migration, invasion, spreading, and
signaling [21]. C0151 has been shown to directly interact with integrin 03 and
indirectly with integrin 06. Klosek et al. reported that CD151 associates with the
(:3 integrin and Met, providing a functional link for integrins to associate with Met
in human salivary gland tumors [12]. Knowing that integrin 03 can regulate Met,
we hypothesized that C082 may suppress Met activation by drawing integrin 03
and/or integrin 06 and CD151 into tetraspanin enriched microdomains. To
determine whether CD151 is required for 0082 to suppress Met activation, PC3
cells stably expressing C082 or empty vector were transiently transfected with
either a pool of CD151-specific siRNA or a control siRNA. These cells were then
stimulated with increasing concentrations of HGF. Met protein was
immunoprecipitated from cell lysates and analyzed by immunoblotting for
tyrosine phosphorylation and total Met protein. The results show that a 70—90%
knockdown of 00151 had no effect on CD82-expressing PC3 cells’ ability to
suppress HGF-induced Met activation (Figure 8A,B). However, a reduction of
CD151 protein by approximately 90% in PC3 cells expressing empty vector

reduced Met activation compared to cells transfected with control siRNA (Figure

31

8A,B). These data suggest that C0151 is required for Met to reach maximal

activation, but that CD151 is not required for CD82 suppression of Met.

CD151 is not required for CD82 suppression of invasion, but contributes to
PC3 cell invasion.

Since CD151 ablation has been shown to reduce invasion, we expected to
see a reduction in the number of cells that were able to invade in the PCB cells
expressing empty-vector. However, we also sought to determine if CD151 is
required for CD82 to suppress invasion. PC3 cells stably expressing either
empty vector or CD82 were transiently transfected with either a pool of four
CD151-specific siRNAs or control siRNA. The cells were then placed in a
Boyden chamber containing Matrigel and laminin as a chemo-attractant and
allowed to invade for 72 hours. A 88% knockdown of CD151 in vector-
expressing PC3 cells reduced invasiveness by 1.7-fold compared to their control
counterparts (Figure 9A,B,C). Loss of CD151 in C082 expressing cells did not
restore the invasive capacity of control PC3 cells, but resulted in an additional
3.5-fold reduction in invasive capabilities (Figure 9A,B). The CD151 knockdown
invasion assay results mimic the (13 integrin knockdown invasion assay data.
These data suggest that CD151 and integrin 03 may be present in the same
pathway regulating invasion. Overall, these results suggest that CD151 is not

required for CDBZ to suppress invasion, but contributes to invasion of PC3 cells.

32

Figure 8. CD151 is not required for CD82 suppression of HGF-induced Met
activation, but contributes to Met activation in PC3 cells. (A) PC3 cells
stably expressing C082 (cl57) or empty vector (cl31-) were transiently
transfected with either CD151 specific siRNA or control siRNA (ctrl) and starved
for 48 hours. The cells were then stimulated with O, 10, 25, 50, and 100 ng/mL of
HGF for 10 minutes. Met phosphorylation was measured by immunoblotting
immunoprecipitates with a phospho-specific antibody. Total levels of Met protein
in immunoprecipitates were measured by immunoblotting stripped blots with Met
antibody. (B) CD151 and C082 expression were monitored by immunoblotting of
whole cell lysates. The numbers below the figure indicate the percent
knockdown values of CD151, as determined by densitometry. Susan Spotts,

unpublished data

33

Figure 8: CD151 is not required for C082 suppression of HGF-induced Met
activation but contributes to Met activation in PC3 cells.

 

 

 

 

     

 

 

A. PCS.Vector.cl31-
Control siRNA CD151 siRNA
Met lP _0 10 25 50 100 0 10 25 50100ng/mLHGF
P-Tyr “5
Blot: W" if'flffi m “I Met
Blot: ---C~- -~

 

 

PC3.0082.c|57
Control siRNA CD151 siRNA

Methzg 0 10 25 50100 0 10 25 50 100 nglmLHGF

P-Tyr

Blot: *" "'~‘ “l. - I. ' 4““

 

 

 

ggtl“"”m““”-"~ -Met

 

B. PC3.Vector.cl31-
0Control siRNA CD151 siRNA
10 25 50100 0 10 25 50 100nglmL HGF

 

 

 

 

 

+
6233+... : I-
c323 - CD82

 

 

 

0 0 0 0 0* 84 93 92 93 87% Knockdown

PC3.CD82.cl57
Control siRNA CD151 siRNA

- 0 10 25 50100 0 10 25 50 100nglmLHGF
co151[a---.' -c0151

wcl:

0 0 0 0 70 83 80 86 90 % Knockdown
34

 

 

C082

wcl: - C082

   

 

 

 

Figure 9. CD151 is not required for CD82 suppression of invasion, but
contributes to PC3 cell invasion. (A) PC3 cells stably expressing C082 (cl57)
or empty vector (cl31-) were transiently transfected with either CD151 specific
siRNA or scramble siRNA (ctrl) and starved for 48 hours. Cells were placed in a
Boyden chamber and allowed to invade for 72 hours. Cells which passed
through the matrigel and appeared on the lower membrane were stained with
crystal violet. (B) Cells were counted in 5 different fields and averaged. Four
separate experiments were then averaged together. (C) 72 hours after cells
were placed in the Boyden chamber, parallel cultures of cells were lysed and
immunoblotted for CD151 and C082 expression. The numbers below the figure
indicate the percent knockdown values of CD151, as determined by

densitometry. Susan Spotts, unpublished data

35

Figure 9: CD151 is not required for CD82 suppression of invasion but con-
tributes to PC3 cell invasion.

 

 

 

 

 

 

 

 

 

   

A. PC3.vector.cl31- PC3.CD82.cl57
b ‘ " ‘2 1" ' .'
.\' .'2,Z.7No\ «'3
if: i. ' . . ’ 1"
L;- .5 i' ‘ . ..‘_ +Scramble siRNA
,jj.‘,')fg, ,- ' _ ..
v. ' . 3‘
o . ' +CD151siRNA
‘., . . .. / .
:3. .f f
1"... ‘ O
B' 180
m 160 - IScramble siRNA
E CD151siRNA
0 140 ~
3 120 ~
'0
g 100 -
C
E 80 ~
0
g 60 -
E 40 ~
2 20 « .
o .l
PCB.Vector PC3.C082
C. PC3.Vector.cl31- PC3.CD8Z.CIS7

 

Ctrl CD151 Ctrl CD151 :siRNA

CD151blotzl -‘ ' .

 

I- CD151

 

 

 

C082 blot:

 

 

 

36

Discussion

Our lab had previously established that CD82 is capable of suppressing
both integrin-mediated and ligand-induced Met activation and that Met requires
the cross-talk with integrin to achieve maximal activation [4]. Multiple labs within
the tetraspanin field have provided information regarding the requirement for the
correct composition of tetraspanins and their protein-binding partners within the
tetraspanin-enriched microdomain to achieve functional tetraspanins [22].

l was interested in determining which constituents within the tetraspanin-
enriched microdomain were required for C082 to suppress HGF-induced Met
activation. Based upon these results, we have determined that CD82 requires
the presence of CD9 for suppression of HGF-induced Met activation and its
function is independent of the presence of CD151, 03 integrin, and 06 integrin.
On the other hand, CD151, and integrins 03 and 06 do not appear to be required
for CD82 suppression of Met activation.

However, the experiment chosen-to determine the requirement for other
tetraspanins and integrins on C082 suppression of HGF-induced Met activation
contains a major flaw. This experiment does not establish the threshold of Met
activation required to result in a physiological change within the cell. To provide
a more concrete understanding of which proteins are required for CD82 function,
Met activation studies should be conducted simultaneously with an assay that
has a physiological output such as a wound healing assay or scratch test that

measures HGF-dependent migration. Multiple tests may need to be completed

37

to determine whether C082 only suppresses migration, proliferation, or invasion,
or whether it’s a combination of pathways it affects.

Met contains multiple tyrosine residues within the cytoplasmic domain.
The phosphorylation states of these residues in specific patterns results in
regulation of different signaling transduction pathways. Identifying the specific
tyrosine residues which are regulated by CD82 and result in decreased cell
invasion or proliferation will provide valuable insight into CD82’s molecular
mechanism. Birchmeier et al. have shown that mutations of the tyrosine residues
in the C-terminal tail result in reduced ligand-induced cell motility and branching
morphogenesis [1]. Preliminary studies to identify the tyrosine residues being
modulated by CD82 has been completed in our lab and the data suggest that
CD82 regulates tyrosine residue 1365. Studies involving knockdown of various
tetraspanins and integrins should be completed in which the phosphorylation
state of specific tyrosine residues on Met would be monitored.

C082 has also been shown to suppress invasion to the same extent as
loss of Met has on reduced invasiveness [4]. Therefore, we sought to
understand whether the same proteins required for C082 suppression of Met
activation was also required for CD82 inhibition of invasion. The results suggest
that CDQ, 00151, integrin 03, and integrin 06 are not required for C082
suppression. However, 03 and 06 integrins and CD151 contribute to invasion of
PC3 cells. interesting, the loss of CD151 and integrin 03 in addition to the

presence of C082 resulted in further reduced invasion. These data seem to

38

further strengthen the evidence that C0151 and integrin d3 associate to regulate

Met as well as other receptor tyrosine kinases.

CHAPTER 2: IDENTIFICATION OF PROTEIN-PROTIEN INTERACTIONS

Rationale

The literature has suggested that 0082 functions as a ‘molecular
facilitator’ to organize tetraspanins, integrins, and other membrane proteins into a
tetraspanin-enriched microdomain [3]. If C082 functions through its involvement
in the TERM, we would predict that specific protein interactions within the
microdomain are important for the ability of CD82 to suppress c-Met activation.

Met can be activated through an integrin-mediated mechanism and 0082
was shown to co-immunoprecipitate with integrins. In addition, I have
demonstrated that (:3 integrin is required for efficient Met activation (see Figure
2). Thus our hypothesis is that C082 suppresses Met activation by drawing the
integrins out of the Met complex and into a tetraspanin-enriched microdomain.

Protein-protein interactions within the tetraspanin-enriched microdomain
have been divided into three types based on the strength of the detergent
required to disrupt their interactions [6]. Type 1 interactions are direct physical
interactions which include tetraspanin homodimers and some heterodimer
interactions such as CD151 and integrin 03 interactions [6]. Type 2 interactions
are stable under mild lysis conditions and consist of tetraspanin-tetraspanin

heterodimer interactions and tetraspanin-integrin interactions which are stabilized

39

mainly by palmitoylation [6]. Type 3 interactions are very weak associations
between tetraspanins and receptor protein-binding partners which are stabilized
also by palmitoylation [6].

To identify the protein-protein interactions which exist in the presence and
absence of C082, we chose to conduct co-immunoprecipitation experiments.
While co-immunoprecipitation assays allow for the identification of protein-protein
associations, it does not provide any information regarding direct physical
interactions. PC3 cells expressing either 0082 or empty vector were stimulated
with 0 or 25 ng/mL of HGF for 10 minutes prior to lysing. Since we are interested
in determining hetero-tetraspanin interactions and tetraspanin-integrin
associations which are Type 2 interactions, we chose to use the mild detergent
1% CHAPS in our lysis buffer. l immunoprecipitated the protein of interest, then
immunoblotted for the corresponding proteins of interest before immunoblotting
for total protein. We expected to see a decrease in Met-integrin associations in
the presence of CD82 and an increase in associations with HGF stimulation. We
also expected to see an increase in tetraspanin-integrin interactions in the

presence of C082 with little change occurring with HGF treatment.

Results:
Met associates with integrin G681 but not with integrin 0301.

Met can be activated through an integrin-mediated mechanism and CDBZ
was previously shown to co-immunoprecipitate with integrins. I predict that

integrin 03 will be the predominant integrin a subunit to associate with Met since

40

we have data suggesting that integrin 03 contributes to both HGF-induced Met
activation and invasion of PC3 cells (see Figure 2). To determine which integrins
are present in the Met complex and whether the presence of C082 has an effect
on Met/integrin complexes, PCB cells expressing either empty vector or C082
were left unstimulated or stimulated with 25 ng/mL of HGF for 10 minutes and
lysed under mild conditions using 1% CHAPS lysis buffer. Met, integrin 03,
integrin (16, and integrin 81 protein was then immunoprecipitated and
immunoblotted for various integrin subunits and tetraspanins. I chose the HGF
concentration of 25 ng/mL based on the fact that this concentration is where I
saw the largest differential effect on HGF-induced Met activation between vector
and CD82 expressing cells. I was unable to detect integrin (13 or 06 subunits and
the integrin B1 subunit within the Met immunoprecipitates (Figure 10A). While it
appears that small amounts of integrin a3 and 06 might be associating with Met,
the integrin B1 subunit, which is the other half of the integrin dimmer, was not
detected. lntegrin [31 serves as a positive control throughout these co-
immunoprecipitations since integrins function as heterodimers consisting of an
alpha and a beta subunit. However, I was also able to detect Met and 81 within
integrin 06 immunoprecipitates (Figure 10C). Upon HGF treatment, there
appears to be a slight increase in the amount of Met associating with (16 integrin
in both C082-expressing cells and non-CD82 expressing cells. I was able to
detect B1 integrin, but not Met within the integrin 03 immunoprecipitates (Figure
108). Overall these data suggest that in PC3 cells, Met predominantly

associates with integrin 06131. This result conflicts with our original hypothesis of

41

Met associating predominantly with integrin a3 subunit since integrin (13

contributes to HGF-induced Met activation and invasion.

42

Figure 10. Met associates with integrin (1681 but not with integrin (1381.
PC3 cells stably transfected with either empty vector (V) or CD82 (CD) were
serum starved for 48 hours and then stimulated with 25 ng/mL of HGF for 10
minutes. The cells were lysed in 1% CHAPS lysis buffer for 30 minutes. (A)
Lysates were immunoprecipitated with anti-Met antibody, then immunoblotted
with anti-integrin (13 subunit (ITGA3), anti-integrin 06 subunit (ITGA6), anti-
integrin B1 subunit (ITGB1), and anti-Met antibodies. (B) Lysates were
immunoprecipitated with anti-integrin (13 (ITGA3) antibodies, then immunoblotted
with anti-Met, anti-integrin B1 (ITGB1), and anti-integrin 03 (ITGA3) antibodies.
(C) Lysates were immunoprecipitated with anti-integrin 06 (ITGA6) antibodies,
then immunoblotted with anti-Met, anti-integrin B1 (ITGB1), and anti-integrin r16
(ITGA6) antibodies. Whole cell lysates (WCL) served as a positive control.
Respective whole lgG molecules served as negative controls. Susan Spotts,

unpublished data

43

Figure 10: Met associates with integrin a681 but not with integ-
rin (1331.

A. W: Met 3 _v_v__cl

 

 

 

. ... I u
. I—ll
...._.'
. . . . u -
' ' " ' ’ . . .’.:‘.. ' ...-L.T. .
' n u. -d i. I. - fl
. ... .. . ... ...-. . "

- + + + - - :25nglmL HGF
V V CD CD V CD V CD cell line

   

 

 

 

B. IP: ITGA3 I69 WCL
Met
blot:I _ I Me

 

 

”331: [El H H H .I- ITGB1

 

 

 

 

 

 

 

 

 

 

 

ITGA3 :4 --.-_
blot: fl “H ‘t H -I- ITGA3
"' - :25ngImL HGF
V V CD CD V CD V CD cell line
c. IP: ITGA6 lgG w_c__l_
Met 717T v: ‘ -
blot: T” ,h.”
ITGB1
blot: I- - H H
ITGA6

blot:

 

 

 

- :25nglmLHGF
V V CD CD V CD V CD cell line

44

In the presence of CD82, integrin 06 is drawn into the tetraspanin-enriched
microdomain and associates with CD82 and CD151.

Once we had determined which integrin associates with Met, we sought to
identify which tetraspanins associate with Met, and the 03 and (:6 integrin
subunits. I was able to detect CD82 within 06 integrin immunoprecipitates
(Figure 11A), but I was unable to detect C082 within (:3 integrin or Met
immunoprecipitates (Figure 11B,C). I was also able to detect an association
between C09 and both integrin 0 subunits. However, there was a reduction in
the amount of C09 that was able to associate with the 0 integrin subunits in the
presence of C082 (Figure 11A,B). However, in the reverse co-
immunoprecipitations, in which CD9 was pulled down in immunoprecipitates and
then immunoblotted for integrin a3 and integrin 06, I was unable to detect Met,
(:3 integrin, or 06 integrin within the C09 immunoprecipitates (Figure 110). I also
evaluated the associations between CD151 and (1 integrin subunits. I predicted
that there would be an association between CD151 and both 03 integrin and 06
integrin. Both integrin <13 and 06 were found to interact with CD151; however in
the presence of 0082, there was a reduction in the amount of CD151 interacting
with integrin (13 (Figure 118). In the reverse co-immunoprecipitation, CD151
associated with both integrin 0 subunits with some decrease in the amount of
association due to the presence of C082 (Figure 11E). l was also able to detect
a weak association with Met within the CD151 immunoprecipitates; however,
because of the low signal, i am unable to determine if there is a change in Met-

CD151 interactions due to 0082 or HGF stimulation.

45

Figure 11. In the presence of CD82, integrin (16 is drawn into the
tetraspanin-enriched microdomain and associates with C082 and CD151.
PC3 cells stably transfected with either empty vector (V) or CD82 (CD) were
serum starved for 48 hours and then stimulated with 25 ng/mL of HGF for 10
minutes. The cells were lysed in 1% CHAPS lysis buffer for 30 minutes. (A)
Lysates were immunoprecipitated with anti-integrin 06 (ITGA6) antibody, then
immunoblotted with anti-CD82, anti-CD9, and anti-CD151 antibodies. (B)
Lysates were immunoprecipitated with anti-integrin Cl3 (ITGA3) antibodies, then
immunoblotted with anti—CD82, anti—C09, and anti-CD151 antibodies. (C)
Lysates were immunoprecipitated with anti-Met antibodies, then immunoblotted
with anti-C082, anti-C09, and anti-CD151 antibodies. (D) Lysates were
immunoprecipitated with anti-C09 antibody, then immunoblotted with anti-Met,
anti-integrin 03 subunit (ITGA3), and anti-integrin B1 subunit (ITGB1) antibodies.
(E) Lysates were immunoprecipitated with anti-CD151 antibody, then
immunoblotted with anti-Met, anti-integrin 03 subunit (ITGA3), and anti-integrin
B1 subunit (ITGB1) antibodies. Whole cell lysates (WCL) served as a positive
control. Respective whole lgG molecules served as negative controls. Susan

Spotts, unpublished data

46

Figure 11: In the presence of CD82, integrin a6 is drawn into the
tetraspanin-enriched microdomain and associates with CD82,

 

 

and CD151
A. IP: ITGA6 lgG WCL
CD82
blot:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

B. iP: ITGA3

C082

blot:

CD9 .. ”I“, II...

blot: '

CD151 ”m - CD151
blot: ,

+ - + - - 2:5nglmL HGF
V V CD CD V CD V CD cell Ilne

C. IP: Met lgG WCL

CD82 €7.

blot: CDB2
CD9

blotzv -cog
CD151

blot: -CD151

 

 

- + - - :25nglmL HGF
V V CD CD V CD V CD :cellline

47

Figure 11 can’t: In the presence of CDBZ, integrin a6 is drawn
into the tetraspanin-enriched microdomain and associates with
CD82, and CD151

 

 

 

 

 

- ITGB1

    

 

 

 

 

 

 

 

 

 

 

 

ITGB1
blot: . ..
i - + . + - - - - :25nglmL HGF

V V D CD V CD V CD :cellline

E. IP: c0151 I96
' 1' J ' -ITGA3
a -ITGA6
“ll-”Gm

. 5. f .‘ l-Met

 

 

- + - 7- - - :25nglmL HGF
V V C CD V CD V CD :cell line

48

In the presence of CD82, there is a reduction in the number of CD9 and
CD151 associations.

To further understand the protein-protein interactions that are specific for
C082 suppression of HGF-induced Met activation, I sought to identify the
tetraspanins that were interacting in the presence and absence of CDBZ. I was
able to detect C082 and CD151 in CDQ immunoprecipitates (Figure 12A);
however, less CD151 was pulled down in the presence of CD82. I was also able
to detect CD9 and CDBZ in CD151 immunoprecipitates (Figure 128). While
CD151 appears to only associate with C082 when the cells are stimulated with
HGF, CD151 may also be able to associate with 0082 in the absence of HGF. I
may not have immunoprecipitated enough CD151 in the control, non-HGF
treated sample to determine if the interaction is dependent on cell stimulation

with HGF.

Discussion

These data suggest that in the absence of CDBZ, CD151 interacts with
integrin (:3, integrin 06, and Met within the tetraspanin enriched microdomain.
However, while the co-immunoprecipitation data did not show an association
between Met and CD9, I did detect a decrease in CDQ-CD151 interactions in the
presence of C082. Therefore, I am assuming that CD9 may also exist within the
Met complex in association with CD151. It is also possible that CBS) and CD151
form a complex that is separate from the Met and integrin complex. These data

further suggest that when 0082 is present within the cell; C082, CD151, and

49

integrin 06 associate within a complex and that CD9, CD151, and integrin (13

may form a separate complex.

50

Figure 12. In the presence of 0082, there is a reduction in the number of
C09 and CD151 associations. PC3 cells stably transfected with either empty
vector (cl31) or CD82 (cl57) were serum starved for 48 hours and then stimulated
with 25 ng/mL of HGF for 10 minutes. The cells were lysed in 1% CHAPS lysis
buffer for 30 minutes. (A) Lysates were immunoprecipitated with anti-C09
antibody, then immunoblotted with anti-0082, anti-CDQ, and anti-CD151
antibodies. (B) Lysates were immunoprecipitated with anti-00151 antibodies,
then immunoblotted with anti-CD82, anti-C09, and anti-CD151 antibodies.
Whole cell lysates (WCL) served as a positive control. Respective whole lgG

molecules served as negative controls. Susan Spotts, unpublished data

51

Figure 12: In the presence of CD82, there is a reduction in the
number of CD9 and CD151 associations.

A. C09 I69 _W___CL

c082I " "
I_}

CD9 a... '

 

 

 

 

 

 

CD151
blot:[ 1 J -CD151
" - + + - :25nglmL HGF
V V CD CD V CD V CD: cell line
8. IP: CD151 lgG WCL

 

 

 

blot:| m: In} C082

cos ..
W: _

blotzl "'" “I cos

cm 51 if ”an- - c0151
blot: m W
F - + - + - - 2:5nglmL HGF

v v co co v co v co cell line

 

 

 

52

However, the data still fails to show certain interactions that are needed to
completely validate my working hypothesis. While integrin 06 did show an
association with Met, I would have expected that in the presence of C082 there
would be a reduction in the amount of integrin 06 able to associate with Met. I
would also assume that integrin 03 should be detected within the Met complex
independent of the absence or presence of CD82. Some of these interactions
may not have been captured because these microdomains are transient and
dilute within the membrane making them hard to capture in a co-
immunoprecipitation. The antibodies used for these experiments may also
recognize the site in which Met and the integrins interact and therefore would not
be able to bind or disrupt the binding.

Yamada et al. conducted a monoclonal antibody study to further
understand the interaction between CD151 and integrin 03B1 [23]. They
reported that the antibodies had distinct reactivities towards this particular
tetraspanin-integrin complex and the antibodies were divided into three
categories [23]. Group 1 antibodies were devoid of sufficient protein when
integrin 03 was co-immunoprecipitated, but had increased affinity when unbound
integrin 03 was immunoprecipitated [23]. This suggests that the epitope of
bound integrin 0361 may be blocked by the association with CD151 [23]. Based
on this observation, I may not have been able to detect an association between
integrin 03B1 and Met, because the epitope may have been blocked by

association with Met.

Alternatively, my initial hypothesis may have been incorrect and based
upon the data I have I would propose the following model that integrins 06B1 and
0361, CD151, and Met form a complex in the absence of 0082. However in
0082-expressing cells, there is a partial shift of integrin 06131 and CD151 from
the Met complex into the tetraspanin-enriched microdomain with 0082.

Upon stimulation with HGF, l was able to detect a small increase in Met
associating with integrin 06 in both the absence and presence of 0082.
However, all other observable interactions seemed to be independent of HGF
treatment. I expected to see an increase in integrin-Met associations with HGF
stimulation because HGF aids in clustering Met for increased auto-
phosphorylation of the receptor. While HGF increases Met clustering, it may also
draw more integrins into the complex to further increase Met activation.

However, for the most part I did not detect this.

I was careful to also include reverse co-immunoprecipitations to help
validate the interactions; but in some cases, the associations were not detected
in both immunoprecipitations. If an association can only be seen in one direction,
I may have to conclude that the interaction is not real but an artifact of the co-
immunoprecipitation conditions I used. However, the interaction may only be
observed in one direction due to antibody epitopes being blocked by a particular
association. Time courses may also be required to understand the kinetics of

these interactions and how this regulates CDBZ’s function.

54

CHAPTER 3: CONCLUSIONS AND PERSCPECTIVES

The overall objective of the project was to determine the molecular
mechanism by which C082 suppresses tumor metastasis. Based upon the
literature, we had initially hypothesized that 0082 suppresses both HGF-induced
Met activation and invasion by disrupting integrin 0381-Met associations and
drawing the integrins into the tetraspanin-enriched microdomain. The decrease
in Met-integrin interactions would result in reduced Met activation which would
lead to a decrease in downstream invasion. We also hypothesized that the
specific constituents within the tetraspanin-enriched microdomain would be
critical for CD82 function.

To determine which integrins and tetraspanins were required for CD82
suppression, we conducted Met activation studies on PC3 cells expressing 0082
or empty vector that were also transiently expressing either a specific siRNA or
control siRNA. The results suggest that only CD9 is required for C082
suppression of HGF-induced Met activation and that C082 function is
independent of integrins (13 and (16, and CD151. However, these data also
suggest that integrin a3 and CD151 may contribute to Met activation in PC3 cells.

Once we had completed the Met activation studies, we wanted to assess
the requirement of these proteins for CDB2 suppression of invasion.

Surprisingly, the invasion assay data was not entirely consistent with the Met
activation studies. The data suggest that CD151 and integrin a3 and 06 are not

required for 0082 suppression of invasion but may contribute to invasion of PC3

55

cells. We also observed that knockdown of either CD151 or integrin (13 in 0082-
expresing cells had a further similar reduction in the number of cells able to
invade the Matrigel. These data suggest that CD151 and integrin (13 may
regulate invasion through the same pathway. Overall, the results seem to
suggest that CD82 suppression of HGF-induced Met activation is an independent
process of CDB2 suppression of invasion.

We sought to determine which integrins are present in the Met complex
and how the presence of C082 changes these interactions. These data suggest
that in the absence of CD82, CD151 interacts with integrin (:3, integrin 06, and
Met within the tetraspanin enriched microdomain. However, while the co-
immunoprecipitation data did not show an association between Met and 009, I
did detect a decrease in CDQ-CD151 interactions in the presence of 0082.
Therefore, I am assuming that CDQ may also exist within the Met complex in
association with CD151; however, it may exist in a complex with CD151 that is
separate from Met-integrin complexes. These data further suggest that when
CDB2 is present within the cell; 0082, CD151, and integrin 06 associate within a
complex and that CDQ, CD151, and integrin 03 may form a separate complex.

Our current hypotheses are that in the absence of CD82, integrins (1331
and 0681 are able to interact with Met via interactions with CD151 (Figure 13A).
In the presence of CD82, 0082 interacts with 009 and CD151 which draws both
CD151 and integrin (1661 into the tetraspanin-enriched microdomain (Figure
138). With fewer integrins within the Met complex, Met is unable to reach full

activation and signal downstream for invasion and metastasis.

56

0082 has been also been shown to inhibit EGF receptor signaling by
directly associating with CDB2 and co-intemalizing via the 0082’s internalization
motif and its “sorting motif” [24]. The increased rate of EGF R internalization
results in decreased EGF R signaling and migration [24]. Yang et al. has shown
that CD151 ablation in immortalized and metastatic breast Cancer cell lines
results in decreased migration, invasion, and signaling and disruption of integrin
dBB4-EGF R associations [21]. Although this paper is taken in the context of
breast cancer, this system could be applied to prostate cancer in which Met up-

regulation results in increased metastatic potential.

57

Figure 13: Working model of molecular mechanism for C082 suppression
of Met activation. (A) In the absence of CD82, integrin 0661 associates with
Met and integrin 0361 is able to interact with Met via interactions with CD151.
(B) In the presence of C082, CD82 interacts with CD9 and CD151 which draws
both CD151 and integrin 0681 into the tetraspanin-enriched microdomain. With
fewer integrins within the Met complex, Met is unable to reach full activation and

signal downstream for invasion and metastasis.

Figure 13. Working model of molecular mechanism for CD82 suppression
of Met activation.

    

 

        

      

Met
Extracellular Matri '
"“~.;:MW:"'::I:‘)::;:i:‘é!.$ 31.: 35".”: :{15' .
I II I IIII ‘- I.I 55- :l 1‘ II II ‘ I I III II IIII I II II
II I I III II III I , ' I I I ‘II I II I * I I III II I
1I‘IIIIIA‘I’I It‘l‘l‘l’fl,‘ I ail: L ‘J’z‘i‘t‘ll It} I ’I’ m . z._ I V 1'; t I,” ‘I‘I‘I’I II“: ‘I‘I‘I‘I gigwgfli t I I’t’ I “II ‘Ilt, I‘l‘tllll‘tllél‘I'I‘I‘t'IlI‘I‘l‘I'
fix I: it “ "' ‘

 

 

D9 D151 “3‘“ cm \1 l/ “55‘ 0015 CD
I/

Invasion &
Metastasis

      
      
    
 

. ’I'. \v
\r-.,. .. I . ’8'
-: :1; 0"“..‘.‘

II: ’I II: II I'. f: a. I .I‘I II'III'IIIIIII HI'IIII II ‘ iI ”‘5 a” 1 ‘-
IM III“ W” I‘II W W a m IIIfII'I I' "WI III" I IIII'I‘III’II ”W”- I “III I I m 1%
' ' ‘22.: i;- " I

‘ I I I IIIIIII IIIIIII

IIII :I II l l-IIWIIII I: IIII:

II I I I IIIIIIIII IIIII

IIIII 'I
. ' = ' is,
'53

C08 CD (1631 D15 CD82 CD' D151 03$“
TERM

Invasion 8.
Metastasis

59

Future Work

The Miranti lab has in vivo data, demonstrating that C082 suppresses
metastases but not primary tumors in SCID mice expressing human HGF. To
determine if human HGF (huHGF) is required for the development of metastasis,
DU145 cells were injected into the prostates of 20 SCID mice and 20 huHGF
SCID mice. Of these mice, 100% produced primary tumors but only the SCID
mice expressing human HGF developed metastases. Once we had determined
that human HGF is required for the development of metastases, 20 huHGF SCID
mice were injected with DU145 cells expressing 0082 and another 20 huHGF
SCID mice were injected with control DU145 cells. Of these mice, 100%
developed primary tumors and only the mice injected with control DU145 cells
developed metastases. These experiments provide in vivo evidence that 0082
suppresses HGF-dependent processes.

My thesis work has all been completed in cell culture and therefore is not
a very good “in vivo” model. These experiments need to be repeated in a mouse
model to provide further evidence for our model of the mechanism for CD82
suppression. PC3 cells stably expressing either CDBZ or empty vector should be
transfected with short hairpin RNAs that are specific for CD9, CD151, integrin a3,
integrin (16, or a control sequence. PCB clones which stably express both 0082
or empty vector and a specific shRNA should be placed into SCID mice
expressing human HGF and monitored for the development of primary tumors
and metastases. Based upon my results, I would expect that injection of control

PCB cells expressing shRNAs of CD151, integrin (13, or integrin 06 would result

60

in reduced size or number of primary tumors and metastases compared to mice
injected with control PC3 cells expressing control shRNA. If this experiment
were repeated with PCS cells expressing 0082, I would expect there would be a
reduction in primary tumors and metastases compared to PC3 cells expressing

CD82 and control shRNA.

61

APPENDIX A: MATERIALS AND METHODS

Antibodies

The antibody against Met for immunoprecipitation (Met-D1) was obtained from
Dr. Brian Cao at the monoclonal antibody core facility at the Van Andel Institute.
The C082 and CDQ antibodies used for immunoprecipitation and immunoblotting
were purchased from Diaclone. Antibodies against CD151 and Met used for
immunoblotting were purchased from Santa Cruz. The anti-phosphotyrosine
monoclonal antibody 4G10 was purchased from Upstate. The CD151 antibody
utilized in co-immunoprecipitations was purchased from Abcam. The integrin 03
used for immunoprecipitations and from immunoblotting was a generous gift from
Dr. Christopher Stipp (University of Iowa, Iowa City, IA, USA). lntegrin 06
antibody used for immunoblotting was a generous gift from Dr. Anne Cress

(University of Arizona, Tucson, AZ, USA).

Cell Culture and Transfection

PC3 cells were obtained from the American Type Culture Collection (ATCC).
PC3 cells were maintained in F12K medium (lnvitrogen) supplemented with 10%
fetal bovine serum (Gibco), 2mM glutamine, 50U of penicillin, and 50ug of
streptomycin/ml. PC3 cells were transfected with a plasmid construct
pcDNA3.1.CD82. PC3 cells at 2x106 per 10-cm diameter plate were transfected
with 5 pg of DNA using Lipofectamine 2000 lipid as described by the

manufacture (lnvitrogen). At 6 hours following transfection, the culture media

62

was changed to regular F12K medium with supplements and replaced 48 hours
later with the same culture medium containing puromycin (2pg/ml). Puromycin-
resistant clones were picked 10 days later and screened for CD82 expression by
immunoblotting with T882b antibody. 0082-expressing clones were maintained

in growth medium containing puromycin and used in assays described below.

lmmunoprecipitation and Immunoblotting

PC3 cells expressing either vector or C082 were serum starved for 48 hours and
then treated for 10 minutes with increasing concentrations of hepatocyte growth
factor (HGF) at 37°C (Calbiochem and a generous gift from George Vande
Woude). Cells were lysed in either 1X MAPK lysis buffer (50mM Tris pH 7.5,
100mM EDTA, 1% TritonX-100, 50mM NaF, 50mM B-glycerophosphate, 5mM
sodium pyrophosphate, 1mM Na3VO4, 1mM PMSF, 100U/mL aprotinin, 10
ug/mL pepstatin, and 10 ug/mL Ieupeptin) or 1% CHAPS lysis buffer. Protein
concentration of the samples was determined using either the bicinchoninic acid
assay (Pierce) or Bradford assay (BioRad). lmmunoprecipitation mixtures
containing 500-2000pg of protein were incubated with the appropriate antibodies
for 3h at 4°C with protein A-conjugated agarose beads (Pierce) to capture
complexes. All immunoprecipitated complexes were washed three times with
their respective lysis buffer. lmmunoprecipitated samples were resuspended in
2X SDS sample buffer, boiled for 5-10 minutes. For immunoblotting of CD82 in
whole cell lysates, 50pg of protein was resuspended in 2X SDS sample buffer

under non-reducing conditions. All resuspended samples were then subjected to

63

SDS poly—acrylamide gel electrophoresis, transferred to a polyvinylidene
difluoride membrane (PVDF). The PVDF membranes were blocked with 5%
bovine serum albumin in Tris-buffered saline containing 0.1% Tween20 (TBST)
for 2 hours, followed by a 2 hour incubation with the appropriate primary
antibodies. After several washes, blots were incubated with a horseradish
peroxidase-conjugated secondary antibody (BioRad) for 1 hour and visualized
with a chemiluminescence reagent. Blots were stripped in low-pH 2% SDS at
65°C for 60 minutes, rinsed, and reprobed for total levels of protein in the

immunoprecipitates or cell lysates.

Invasion Assay

For invasion assays, the underside of the membrane of the upper chamber of
matrigel invasion chambers (BD Biosciences) was coated with 10pg/ml of laminin
(lnvitrogen). Approximately 62,500 cells were serum starved for 48 hours and
then added to the upper chamber. The cells were incubated at 37°C for 72 hours
prior to the cells being washed, fixed, and stained with 0.09% crystal violet stain
(Chemicon) for 20 minutes. Cells remaining on the upper side of the membrane
were removed with a cotton swab. Stained cells that had invaded through to the
underside were visualized on a Nikon TE300 microscope and pictures taken
using a CCD camera (Hammamatsu) attached to an Apple Macintosh G5.
Images were compiled using OpenLab software (lmprovision). Pictures were
collected from five different fields of each chamber and counted. Excess dye

was removed from the exterior sides of each chamber with 70% ethanol. Crystal

64

violet stain was eluted from chambers using extraction buffer [25% Acetate
Buffer pH4.5 (.122M Acetic Acid, .078M Sodium Acetate), 50% Reagent Agent,

25% DI water] and then read the CD at 570nm.

Small Interfering RNA Transfections

Two unique small interfering RNA (siRNA) sequences targeting CDQ and a non-
targeting control sequence were designed and purchased from IDT. (CD9
sequence 1: GAGCAUCUUCGAGCAAGAA; C09 sequence 2:
UGCUGUUCGGAUUUAACUU; Scramble sequence:
ACUACCGUUGUUAUAGGUGTT) A pool of four siRNA against CD151, integrin
03, and integrin (16 were purchased from Dharmacon. Approximately 7.5x105
cells were transfected with either 10nM (lntegrin o3, lntegrin 06, and CD151),
25nM (CD9 Sequence 1), or 50nM (CD9 Sequence 2) siRNA in F12K serum free
media using siLentFect lipid reagent (Bio-Rad) and Opti-MEM (lnvitrogen) media
following manufacturer's directions. The media was changed 24 hours after

transfection.

65

10.

11.

12.

REFERENCES

Birchmeier, C., et al., Met, metastasis, motility and more. Nat Rev Mol Cell
Biol, 2003. 4(12): p. 915-925.

Dong, J.T., et al., KAI 1, a metastasis suppressor gene for prostate cancer
on human chromosome 11p11.2. Science, 1995. 268(5212): p. 884-886.

Tonoli, H. and J.C. Barrett, CDBZ metastasis suppressor gene: a potential
target for new therapeutics? Trends in Molecular Medicine, 2005. 11(12):
p. 563-570.

Sridhar, SC. and CK. Miranti, Tetraspanin KAI1//C082 suppresses
invasion by inhibiting integrin-dependent crosstalk with c-Met receptor and
Src kinases. Oncogene, 2005. 25(16): p. 2367-2378.

Cancer Facts & Figures 2006. Atlanta (American Cancer Society), 2006.

Zoller, M., Tetraspanins: push and pull in suppressing and promoting
metastasis. Nat Rev Cancer, 2009. 9(1): p. 16.

Yauch, R.L., et al., Direct extracellular contact between integrin alpha 3
beta 1 and TM4SF protein CD151. J. Biol. Chem., 2000. 275(13): p. 9230-
9238.

Levy, S. and T. Shoham, Protein-protein interactions in the tetraspanin
web. Physiology. 2005. 20(4): p. 218-224.

Lazo, P.A., Functional implications of tetraspanin proteins in cancer
biology. Cancer Science, 2007. 98(11): p. 1666-1677.

Todeschini, A.R., et al., Gang/ioside GM2-tetraspanin 0082 complex
inhibits Met and its cross-talk with integn‘ns, providing a basis for control of
cell motility through glycosynapse. J. Biol. Chem., 2007. 282(11): p. 8123-
81 33.

Miho Takahashi, T.S.M.A.K.I.K.S., Regulation of c-Met signaling by the
tetraspanin KAI-1/CD82 affects cancer cell migration. lntemational Journal
of Cancer, 2007. 121(9): p. 1919-1929.

Klosek, S.K., et al., CD151 regulates HGF-stimulated morphogenesis of

human breast cancer cells. Biochemical and Biophysical Research
Communications, 2009. 379(4): p. 1097—1100.

66

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

Bari, R., et al., Transmembrane interactions are needed for KAI 1/CDBZ-
mediated suppression of cancer invasion and metastasis. Am J Pathol,
2009. 174(2): p. 647-660.

Yang, X., et al., Palmitoylation supports assembly and function of integrin-
tetraspanin complexes. J. Cell Biol., 2004. 167(6): p. 1231 -1 240.

Berditchevski, F ., et al., Expression of the palmitoylation-deficient CD151
weakens the association of alpha 3 beta 1 integrin with the tetraspanin-
enn’ched microdomains and affects integrin-dependent signaling. J. Biol.
Chem., 2002. 277(40): p. 36991-37000.

He, B., et al., Tetraspanin CD82 attenuates cellular morphogenesis
through Ddown-regu/ating integrin alpha6-mediated cell adhesion. J. Biol.
Chem., 2005. 280(5): p. 3346-3354.

Sharma, C., X.H. Yang, and ME. Hemler, DHH02 affects palmitoylation,
stability, and functions of tetraspanins C09 and CD151. Mol. Biol. Cell,
2008. 19(8): p. 3415-3425.

Kinch, M.S.a.K.C.-K., Overexpression and functional alterations of the
EphA2 tyrosine kinase in cancer. Clinical and Experimental Metastasis,
2003. 20(1): p. 59-68.

Cheresh, H.J.D.a.D.A., Role of integrins in cell invasion and migration.
Nature Reviews Cancer, 2002. 2(2): p. 91-100.

Klosek, S.K., et al., CD151 forms a functional complex with c-Met in
human salivary gland cancer cells. Biochemical and Biophysical Research
Communications, 2005. 336(2): p. 408-416.

Yang, X.H., et al., CD151 accelerates breast cancer by regulating alpha6
integrin function, signaling, and molecular organization. Cancer Res,
2008. 68(9): p. 3204-3213.

Hemler, M.E., Targeting of tetraspanin proteins -] potential benefits and
strategies. Nat Rev Drug Discov, 2008. 7(9): p. 747-758.

Yamada M, T.Y., Sanzen N, Sato-Nishiuchi R, Hasegawa H, Ashman LK,
Rubinstein E, Yanez-Mo M, Sanchez-Madrid F, Sekiguchi K, Probing the
interaction of tetraspanin CD151 with the integrin alpha 3 beta 1 using a
panel of monoclonal antibodies with distinct reactivities toward CD151-
integrin alpha 3 beta 1 complex. Biochem J., 2008. 415(3): p. 417-27.

67

24.

25.

Odintsova, E., T. Sugiura, and F. Berditchevski, Attenuation of EGF
receptor signaling by a metastasis suppressor, the tetraspanin CD82/KAI-
1. Current Biology, 2000. 10(16): p. 1009-1012.

Shi, W., H. Fan, L. Shum, and R. Derynck, The tetraspanin CDQ
associates with transmembrane TGF-alpha and regulates TGF-alpha-
induced EGF receptor activation and cell proliferation. J Cell Biol, 2000.
148(3): p. 591-602.

68

 

"711111111111fillllll'llillllllllllllS