1. Systems Genetics Analysis Reveals Higher-Order Transcriptional Regulatory Elements of Breast Cancer Metastasis
Farhoud Faraji , Ying Hu , Gang Wu , Jinghui Zhang , Kent W. Hunter
1
2
3
3
1
Laboratory of Cancer Biology and Genetics and 2Laboratory of Population Genetics, National Cancer Institute. Bethesda, MD.
3
Department of Computational Biology, Saint Jude Children’s Research Hospital. Memphis, TN.
1
Introduction
Prognostic gene signatures that predict patient risk for metastatic disease are in clinical trials. These
gene sets, however, provide little insight into mechanisms of metastasis. Here we exploit the principle
that metastatic propensity is modified by the genetic background to link prognostic gene signatures with
molecular mechanisms driving metastasis. In doing so we transcend single gene functional analysis and
unveil the higher-order transcriptional architecture of heritable predisposition to breast cancer
metastasis.
Key Points
A.Network analysis of global expression profiles from genetically defined AKXD recombinant inbred
panel primary tumors identified a network of co-expressed genes centered on Cnot2 that predicts
metastasis free survival in human breast cancer.
 Cnot2 over-expression regulates expression of network genes.
 Modulating Cnot2 expression inversely impacts tumor cell metastatic potential in vivo.
 Cnot2 binds the metastasis driver genes Brd4-SF, Rrp1b, and Sipa1
A.Small RNA sequencing of the same tumor panel revealed miRNA-3470b as a potential upstream
regulator of the Cnot2 network.
 miR-3470b down-regulates Cnot2 network hub gene expression.
 miR-3470b expression down-regulated anti-metastatic genes and upregulated pro-metastatic
genes
 miR-3470b promotes metastasis.
Conclusion
Our systems genetics strategy provides a higher-order view of metastatic susceptibility. We identify and
validate a co-expressed module of transcripts that is post-transcriptionally regulated by miR-3470b and
whose central node, Cnot2, functionally regulates metastasis. The physical interaction of CNOT2 with
previously identified metastasis modifier proteins BRD4-SF, SIPA1, and RRP1B implicates CNOT2 in a
larger nuclear complex that regulates metastatic potential, further demonstrating the value of
undertaking higher-order analyses to interrogate mechanisms of metastasis.
The Cnot2
3 A module of co-expressed Network: by Cnot2
genes regulated
1
2
Employing meiotic genetics to understand
heritable predisposition to metastasis
AKR/J
Systems genetics strategy to unveil the
higher-order transcriptional structure of metastasis
AKXD x PyMT F1 Tumors
DBA/2J
Microarray based
mRNA profiling
x
P
Metastasis-prone
F1
x
F2
x
x
F3
x
+
FVB PyMT
]
+
4
Prone
x
Network Hub
Kaplan-Meier
Analysis
Intermediate
Resistant
5
51
4
Signal
Absorption
51
4
nm
*
p=0.05
nm
C-terminal YFP
Fragment
CNOT2
miR-3470b down-regulates Cnot2 network hubs
Proteins in close proximity
No Signal
N-terminal YFP
Fragment
Biological
Validation
miR-3470b is a candidate post-transcriptional
regulator of the Cnot2 network
Bi-molecular fluorescent complementation (BiFC)
Proteins not in close proximity
Identify miRNAs
Highly Represented
In Networks
Biological
Validation
CNOT2 binds known metastasis modifier proteins
Absorption
Negatively Correlated
miRNA-mRNA
Filter for miR-mRNA
Target Pairs
AKXDnPyMT+ F1
Metastasis
Susceptibility
miRNA Sequencing
Network
Generation
x
x
[
F>20
Metastasis-resistant
x
Contact: farajif@mail.nih.gov
***
*
**
m
0n
53
Reconstituted
YFP
*
*
**
Protein B
CNOT2
Protein B
**
**
**
Cnot2 module expression predicts distant metastasis free survival in lymph node negative non-adjuvant-treated breast cancer
CNOT2 complements YFP fluorescence in conjunction with
BRD4 short isoform (BRD4-SF), SIPA1, and RRP1B
GSE11121
1.0
Protein B = BRD4-SF
0.8
Protein B = SIPA1
Protein B = RRP1B
miR-3470b modulates metastasis driver expression to drive a pro-metastatic transcriptional profile
0.6
**
**
0.2
**
––– Low risk
----- High risk
––– Low risk
----- High risk
p = 0.004
50
100
150
0
50
Months
BiFC
p = 0.004
100
150
Transfection Control
BiFC
Transfection Control
BiFC
Transfection Control
200
Months
In vivo metastasis assays validate Cnot2 as a metastasis suppressor
Primary Tumor Burden
Primary Tumor Burden
**
0.5
DAPI
Merge
6DT1
Control
6DT1
Control
80
Surface Metastasis Count
0.5
6DT1
Cnot2
Pulmonary Metastases
Pulmonary Metastases
*
**
WB
6DT1
shControl
6DT1
sh62
6DT1
sh64
Control
WB
40
anti-MYC
anti-MYC
6DT1
shControl
6DT1
sh62
6DT1
sh64
1% Input
IP: anti-MYC
1.5
1.0
0.5
0.0
1% Input
IP: anti-MYC
6DT1
miR-3470a
Primary Tumor Burden
Primary Tumor Burden
p=0.11
*
2.0
anti-HA
anti-FLAG
60
20
6DT1
Control
2.5
0
0.0
miR-3470a/b RNA
Merge
20
PrimaryTumor Burden
Primary Tumor Burden
1.0
DAPI
CNOT2 co-precipitates BRD4-SF and RRP1B
40
6DT1
Cnot2
1.5
Merge
60
0
0.0
DAPI
80
Surface Metastasis Count
1.0
6DT1 sh64
Tumor Mass (g)
1.5
6DT1 sh62
Ne
ga
ti v
e
Cn
ot
2M
YC
Rr
p1
bHA
Cn
ot
2
+
Rr
Ne
p1
ga
b
ti v
e
Cn
ot
2M
YC
Rr
p1
bHA
Cn
ot
2
+
Rr
p1
b
Surface Metastasis Count
80
6DT1 shControl
Ne
ga
tiv
e
Cn
ot
2M
YC
Br
d4
-S
FFL
Cn
AG
ot
2
+
Br
Cn
d4
ot
-S
2F
M
YC
Ne
ga
tiv
e
Br
d4
-S
FFL
Cn
AG
ot
2
+
Br
d4
-S
F
*
2.0
Tumor Mass (g)
Pulmonary Metastases
Pulmonary Metastases
miR-3470a and miR-3470b expression promote metastasis in vivo
6DT1
miR-3470b
Pulmonary Metastases
Pulmonary Metastases
**
**
60
40
20
0
6DT1
Control
6DT1
miR-3470a
6DT1
miR-3470b
miR-3470a
0.0
0
Tumor Mass (g)
**
0.4
miR-3470b
Distant Metastasis Free Survival
GSE2034
Underlined:
Predicted miRNA
Recognition Element
6DT1
Control
6DT1
miR-3470a
6DT1
miR-3470b