hypothesis that TIC represent a unique sub-population of cells within a tumor possessing properties of self-renewal and the ability to give rise to the characteristic cell types present within a given tumor. Because of their unique abilities, we hypothesize further that TIC are localized and function within a spatially and molecularly-regulated microenvironment (mE) (a.k.a. niche).

militaryzoologistΤεχνίτη Νοημοσύνη και Ρομποτική

1 Δεκ 2013 (πριν από 3 χρόνια και 6 μήνες)

59 εμφανίσεις

A. Specific Aims


The specific aims have not been significantly modified from the original submission, as
stated below.

The only major change we have to this center is Dr. Vittorio Cristini moving from
the University of Texas H
ealth

Science Center

at Houston,

to
T
he University of New Mexico.
Despite this move h
e will maintain a significant partnership on this grant, providing mathematical
modeling and computer simulations of tumor growth.


He will have access to data generated by
BCM and TMHRI thru

email
, ad hoc conference calls, and our

password protected
wiki
website.


In addition, he will attend our annual symposium to provide updates on his progress and assist
us in the design and interpretation of data
. We do not see any major problems with ma
intaining
our partnership with Dr. Cristini at his new institution.


Component 1

is guided by the
hypothesis that TIC represent a unique sub
-
population
of cells within a tumor possessing properties of self
-
renewal and the ability to give rise
to the char
acteristic cell types present within a given tumor. Because of their unique
abilities, we hypothesize further that TIC are localized and function within a spatially and
molecularly
-
regulated microenvironment (mE) (a.k.a. niche).
To identify, localize, and
functionally interrogate TIC in vivo in sufficient detail to allow mathematical modeling of their
behaviors and responses to genetic and pharmacological manipulation in Component 2,

the
Specific Aims of Component 1 are:

Aim 1.1: To identify tumor
-
initiatin
g cells (cancer stem cells) using newly developed
lentiviral fluorescent signaling reporters and to characterize their spatial distribution and
behaviors during tumor growth using in vivo imaging.

Aim 1.2: To identify candidate genes and pathways that may
regulate TIC behaviors (e.g.
self
-
renewal, differentiation, and metastasis)

Aim 1.3: To conduct a “Directed Iterative Functional Genomic Screen” (DIFGS) to
characterize genes functionally that either increase or decrease tumor
-
initiating capacity.

Aim 1.4
: To define the cellular responses of TIC to genetic and pharmacological
manipulation of genes regulating TIC survival or function in vivo.

Component 2

is guided by the
hypothesis

that TIC behavior during tumor development
can be simulated using
a robust,
multiparameter mathematical/computational model of
TIC behavior during breast cancer development. Further, that these models can be built
to reflect not only the molecular, cellular, and tissue
-
level dynamics, but also to allow
prediction of the response o
f TIC to experimental therapeutics.

Thus, the central goal of
Component 2

is to build a multi
-
scale model platform of TIC mE for investigating TIC self
-
renewal, proliferation, localization, and other functions within a spatially and molecularly
-
regulated m
icroenvironment.

Based on the experimental data obtained from Component 1 and
published knowledge of TIC, we will model the TIC tissue microenvironment

(TIC mE
) from the
molecular and cellular level up to the tissue level.

The TIC mE model can further pred
ict and
guide the pathway analysis, the candidate gene selection,

genetic and pharmacological
manipulation

in Component 1.

Accordingly, the Specific Aims of Component 2 are:

Aim 2.1: To model the TIC tissue mE mathematically based on 2D and 3D microscopy
a
nd image analysis

Aim 2.2: To predict the TIC pathways or key genes related to specific cancer subtypes so
to refine the TIC microenvironment model

Aim 2.3: To develop bioimaging informatics models for mapping gene functional
networks within and among TIC
and niche cells from the directed iterative shRNA screen
and further refine

the

TIC mE model

Aim 2.4: To model the response of TIC and their microenvironment to genetic and
pharmacological manipulations of TIC function
in vivo


B. Studies and Results


Component 1 (Experimental)


Aim 1.1

TIC Identification & Purification

In the current funding period, we have:

1.

Constructed and validated three sets of lentiviral fluorescent signaling reporters (with
negative controls) for the Wnt
, Hedgehog, and STAT3
-
mediated signaling pathways
using positive control cell lines in vitro. These reporters have been validated in vitro
using specific agonists and antagonists of each pathway whenever possible. In addition,
we have tested these three re
porters in xenograft
-
derived cells and show Wnt and Stat3
activity. Our
Hedgehog

and Hes1 promoter
-
driven reporter
s

show no activity in any of
the xenografts tested thus

far in vitro, however, we have preliminary indication that the
Hedgehog

reporter is
active in vivo, and that a different Notch signaling reporter
(detecting CBF1 activation downstream of Notch) may be superior to the Hes1p
construct

in vitro since Hes1 upregulation is not a universal readout of notch activation,
but CBF1 is thought to be
the primary transcription factor mediating responses to all four
Notch receptors.
.

Table 1. Expression of lentiviral signaling reporters in xenograft
-
derived cells in vitro.








2.

Report
er viruses have been screened for function in a set of four p53 null mouse models
and
selected

human xenograft models.
p53 tumors

used thus

far show unique patterns
of signaling activation of these
four

pathways.

Table 2. Expression of lentiviral

signaling reporters (% posit
i
ve) in p53 null mouse
tumors in vivo

p53 null tumors

Wnt
Reporter

Notch
Reporter

Hh
Reporter

Stat3
Reporter

T1
-
squamous

6
-
9%

~4%

ND

~10%

T2
-
papillary

1.80%

~4%

ND

ND

T6
-
typical solid

0.70%

~0.6%

ND

~5%

T7
-
solid, usual
ND

ND

ND

ND

Reporter
activity in xenograft lines
(In vitro)
Xenograft
line
Wnt
(TCF)
Stat3
(M67)
Hedgehog
(Gli)
Notch
(Hes1p)
MC1
-
+
-
-
2147
+
-
-
-
2665
+
+
-
ND
3887
+
-
-
-
undifferentiated

Tumor (luminal)

ND

1.1
-
2.5%

ND

ND

3.

Demonstrated that Wnt
-
responsive cells in p53
-
null mouse tumors are typically an
average of 20um from the nearest blood vessel. Using the Wnt

signaling reporter, which
enriches for TIC in mouse p53 null tumors,




Figure 1. Histogram of average distance of Wnt1
-
responsive cells from the nearest blood vessel
(In microns).


4.

Developed a transgenic mouse model that expresses cyan fluorescent protein from
virtually every cell in the body, mCherry (red fluorescent protein) in blood vessels, and
GFP in macrophages (Figure 2). Note that each CFP positive duct is almost completely
o
bscured by red blood vessels. The large blue mass in the center of the field is a lymph
node
. Note the high concentration of GFP+ macrophages surrounding the lymph node
.
We are in the process of crossing these three reporter genes into an
immunocompromize
d

(SCID/Bg) background for use in evaluating xenografts or p53 null
mouse tumors. By transplanting unlabeled tumors, tumor/stroma interactions can be
imaged and modeled.









0.00%
10.00%
20.00%
30.00%
40.00%
50.00%
0
5
10
15
20
25
30
35
40
45
Diameter (um)
Percentage of cells

Figure 3.

Three color fluorescent image of a normal mouse mammary gland.




Aim 1.2
Identify Pathways Regulating TICs

In the current funding period we have:

1.

Completed microarray analysis of all available mouse p53 null tumor models
, as well as

all available
xenograft models using bulk tumor. These arrays will be used as the
baseline by which the enriched TIC and “niche” cell populations will be measured for
enrichment of signaling pathways.

2.

Validated the ability of membrane based antibody proteomics arrays fo
r their ability to
yield data using small sample sizes typical of sorted TIC.

In the Proteomics Core of the Dan L. Duncan Cancer Center, a NCI
-
designated Cancer
Center, we have evaluated various antibody array platforms (ClonTech, Sigma, and
Raybiotech

Ful
lmoon Biosystem, and Raybiotech) and optimized conditions to achieve
low background and reliable assay variance of replicate samples.


Optimization requires
control of appropriate dye concentration after protein labeling, background reduction
methods, and
development of post
-
image acquisition data analysis software.

As a pilot
project we identified diffe
rentially expressed proteins in

serum samples from wild type
mice and a mouse knockout for JMJD1a, a member of the jumonji (jmj) domain
containing gene fam
ily of histone demethylases, using Raybiotech’s antibody array.
Using these serum samples, we successfully optimized the protocol and generated
reproducible results (R
2
=0.96) and identified differentially expressed proteins between
these two groups of samp
les. Differential expression was validated by Western Blot
analysis and/or by ELISA.


We have obtained similar results using other antibody arrays
tested.

3.

Initiated Reverse Phase Proteomic Array analysis on all available xenograft models
(under our U54 su
pplement). These proteomic arrays will identify signaling pathways
activated in the bulk tumor, and will be used as the baseline by which the enriched TIC
and “niche populations will be measured for enrichment of signaling pathways.


Aim 1.3
Iterative Func
tional Analysis of Candidate Genes

In the current funding period, we have:

1.

Completed a functional screen in a second cell line (BT459) in addition to the cell line
completed prior to start of the grant (SUM159). Comparison of these two screens yielded
5 genes in common that significantly downregulated mammosphere formation. We ar
e
initiating identical screens in two additional cell lines in order to increase confidence in
our screen, and to identify additional novel regulators of TIC.


Aim 1.4
Treatment Responses of TICs

In the current funding period, we have:

1.

Conducted

baseline
microarray analysis of all available p53 null mouse tumors and all
available human xenograft tumors by which to judge TIC response to treatment

(as
noted above in aim 1.2)
.

2.

Initiated microarray analysis of tumor samples treated with Notch signaling inhibit
or
(MK003), a Smoothened
-
targeted hedgehog signaling inhibitor (IPI926) and a GLI1/2
-
targeted hedgehog signaling inhibitor (GANT61).
These gene expression arrays will
identify signaling pathways altered by treatment in the bulk tumor and will be used as th
e
baseline by which the enriched TIC and “niche populations will be measured for
enrichment of signaling pathways as a function of treatments.

3.

Initiated RPPA analysis of tumor samples treated with Notch signaling inhibitor (MK003),
a Smoothened
-
targeted he
dgehog signaling inhibitor (IPI926) and a GLI1/2
-
targeted
hedgehog signaling inhibitor (GANT61).
These proteomic arrays will identify signaling
pathways altered by treatment in the bulk tumor, and will be used as the baseline by
which the enriched TIC and
“niche populations will be measured for enrichment of
signaling pathways as a function of treatments.


Component 2 (Computational)


Aim 2.1. 3D TIC niche microscopy image analysis


The 3D images of mouse breast cancer tissues are acquired from Comp 1. The tumor
initiating cells (TICs) are indicated by GFP
-
positive cells, blood vessels are shown in the red
channel (dyed by using Dextran), and the tumor cells are in the blue channel (d
yed by using
DAPI, see the Figure
4
). To quantify the spatial distributions of TICs, tumor cells and blood
vessels, following image analyses, 3D image reconstruction, blood vessel segmentation, tumor
cell and TIC segmentation, are being performed. To delin
eate the boundaries of blood vessels,
vascularity
-
oriented level set algorithm is developed to enhance the image contrast. The
iterative voting and gradient flow tracking 3D nuclei segmentation method are being developed
to refine the nuclei segmentation b
y adjusting the Bayesian inference approach [
1
].



Figure 4: Nuclei image (DAPI channel, upper left); Tumor initiating cell image (GFP channel,
upper right); Blood vessel image (Dextran channel, lower left); and the combined image (lower
right)




Aim 2.
2. TIC signal pathway analysis


Understanding the signaling mechanisms
regulating

breast TICs is important to design
of
efficient therapeutic and management strategies of breast cancer. A network
-
based signature
and a comprehensive signaling map for identification of candidates of drug repositioning and
combinations are proposed.

A network
-
based signature
analysis method we

call the


cancer
-
signaling bridges discovery method


was

developed based on an extended concept of network
motifs which can be used to expand the cancer drug
-
targets of known signaling pathways. The
profiles of TIC derived from CD44
+
/CD24
-
/low

breast canc
er cells and mammospheres cells
were

employed to establish network
-
based signatures. We also curate
d

a comprehensive signaling
map for breast TIC by using the available signaling transductions in BioCarta, KEGG, and IPA
which

include
s

seven signaling pathw
ays, PI3K/AKT, JAK/STAT, Notch, HH, Wnt, P53, and
ECM.



Aim 2.3. TIC mE modeling


(1)
TIC lineage model:

We have developed a lineage model for TICs, progenitor cells
and tumor cells to understand the dynamics of tumor formation. To understand
better
the
progression, heterogeneity and treatment response of breast cancer, especially in the context of
TICs, we developed a mathematical lineage model based on the cell compartment method. This
first step has been achieved within an ordinary differential equ
ations (ODE) modeling framework
[
3
]
, which accounts for TIC symmetric self
-
renew and asymmetric differentiation regulated by
the microenvironment. Some results indicate that treatments that inhibit tyrosine kinase activity
of EGFR may not only repress the
tumor volume, but also decrease the TICs percentages by
shifting TICs from symmetric divisions to asymmetric divisions. The model is the first step that
will be included in our existing multiscale model as source terms in the partial differential
equations

(PDE) modeling framework, and will aim at understanding the spatially
heterogeneous localization of the TIC and progenitor cell populations within the tumor.

(2)
Breast TIC multiscale model:

Further, we proposed a 3D multiscale tumor model
incorporating t
he TIC concept and multiple microenvironmental (mE) factors as well as the
above TIC lineage model to investigate how TIC content would influence the tumor evolution.
The simulation results showed that solid tumors initiated from TICs have enhanced prolife
ration
potential, albeit they take a longer time to reach their full
size
.

These findings are partially
demonstrated by our in vivo biological experiments. We also applied the tumor growth model to
study the response of tumor to drug treatment and found th
at drugs would seldom reach most of
the TICs
-
the latter proliferate relatively slowly than non
-
TICs and consequently tend to reside in
the interior of the tumor, which may explain why cancer stem cells are apt to escape from most
of therapies besides their

stronger resistance. These findings may bring insights for the design
of therapeutics.

(3)
Development of a novel multiscale algorithm:

One constraint of the PDE framework
to model TIC niche is its restriction to locally describe a large enough number of cells, which is
typically not the case when the TIC population represents a very small percentage of the total
population. We have adapt
ed and developed a
mathematical

multiscale framework that is able
to describe, in a consistent manner, phenomena that occur at the cellular level and generate
emerging behaviors at the tissue level. This framework, coming from Physics, is called
Dynamical
Density Functional Theory. In combination with the Equation free approach, it will
allow us to generically propose multiscale modeling for biological systems that we will
particularize to TIC niche modeling in order to predict localization of TICs with
in a

tumor. The
hybrid approach

we have developed is a necessary improvement of our numerical tools and
allows for a physically consistent transition across scales, where tissues are described as a
continuum, whereas a small number of cells is described by a d
iscrete approach where single
cell phenotypes play a role.

(4)
Breast
tumorigenesis

as perturbed mammary gland development:

Since breast
cancer can be seen as an abnormal regulation of physiological mammary gland development,
or a malfunction of the gland
, we are developing a lineage model of physiological mammary
gland development based on the data recently generated from Component 1. First, our model
aims at investigating the events involved in a normal mammary stem cell lineage and the
corresponding duc
t formation. In the current state, we have designed a model that reproduces
the key aspects of mammary gland development and will be calibrated by using the already
existing data to help identify the rules of normal development. Second, the perturbation of

these
rules will help us identify what disruption may lead to abnormal development, hyperplasia, and
eventually breast tumor.

(5)
Generic mechanisms of tumor development and their impact on tumor growth:
We
have developed several mathematical hypothesis
-
driven models for a better understanding of
the mathematical structures that are required to observe pattern formation, i.e. spatially
heterogeneous distribution of sub
-
populations within a tumor as experimentally observed for
TICs. These models are generi
c and aim at testing phenotypic properties of TICs. They focus,
respectively, on (a) Effect of angiogenesis on tumor growth
[
5
], (b) ME
-
regulated phenotypic
switch (proliferating and migratory) of cancer cells, which may give rise to spatially
heterogeneou
s phenotype localization [
6
], (c) TIC lineage spatio
-
temporal dynamics using a
cellular automaton framework
.


Component 3 (Educational)

The goals for component 3 are two
-
fold:

1.

To fund and train two multidisiplinary postdoctoral fellows

2.

To develop a summer
undergraduate program entitled the Multidisciplinary Summer
Undergraduate Training Program

In addition, we have added two new goals:

1.

T
o coordinate a bi
-
weekly R&D seminar series for U54 participants

across various
institutions in the Texas Medical Center


2.

To coordinate monthly “working group” meetings (one for “imaging” and one for
“signaling and modeling”)

In the current funding period, we have:

1.

Filled one of the postdoctoral fellow positions. This fellow is resident in the laboratory of
Dr. Vittorio
Christini

2.

Identified a strong candidate for the second position. This individual is scheduled to
defend his laboratory thesis work in March 2011. He has already begun performing
bioinformatic analyses under the direction of Dr
s
. Michael Lewis
,

Stephen Wong
,
and
Susan Hilsenbeck, and is already an active participant in the U54 seminar series.

3.

Laid the groundwork for the summer undergraduate program. The call for applicants to
this program will be released in February 2011.

4.

Arranged to host an ICBP summer int
ern for summer 2011.

5.

Initiated the R&D seminar series

6.

Initiated the working group meetings

Significance


We have successfully initiated the postdoctoral training aspect of the project, which should
facilitate data analysis and enhance modeling capabilities
. In addition, we have established an
atmosphere of cooperation and free and open communication between the two research
components of the project.


Plans for the coming year


1.

Hiring of the second postdoctoral fellow

2.

Complete the summer undergraduate progr
am

3.

Mentor the ICBP summer intern effectively

4.

Continue the R&D seminar series and working group meetings


Component 4 (Administration)

T
he specific aim of the administrati
ve core i
s to provide a flexible yet effective administrative
structure to support the

infrastructural and scientific aims, in view of the many faceted
interactions that must necessarily occur among the
CMCD

teams
. To accomplish this aim, we
have begun
develop
ing

and exec
uting

a management plan based on a balanced management
strategy that s
upports an environment of shared decision
-
making and mutual re
sponsibility
among the core PIs, w
hile providing the oversight and leadership necessary to produce quality
biomedical imaging work. We
have

manage
d

the overall
CMCD

project using sound basics,
including phased delivery, quick and concrete feedback, clear articulation of the project needs,
project tracking and oversight, effective governance, and inter
-
group coordination.
The PI,
Project Manager (PM), and Core PIs have held regular meetings to c
ommunicate project status
and future goals.

The Methodist Hospital
,

Baylor College of Medicine, and
UTHSC

at Houston are located
within walking distance of each other
in the
Texas Medical Center. This geographic proximity
provides great convenience for the synergy and interaction
among

the Methodist
-
Cornell
,
Baylor,

and
UTHSC teams. Many of the team members
have
been
collaborati
ng

in a
number

of
research projects on
breast and

other type of cancers, including those requiring new techniques
in computational biology, bioimaging, pathway inference, tumor invasion microenvironment
modeling, and computational modeling of drug treatment response
.

We have formed
an External Advisory
Panel (EAP)
of four
experts in the areas
of cancer cell

biology and computational biology (
Drs.
Franziska Michor, Joe Gray, Tim Huang, and Robert
Gatenby).
This panel
is also composed of members from other NCI U54 Centers, which should
help us collaborate

with these other teams and disseminate information more effectively. Our
first EAP meeting has been scheduled for Feb. 10
th

and 11
th

and we plan on having our
panelists be guest speakers. We have also invited a strong consultant team composed of well
-
kn
own experts in the cancer biology, stem cell biology, systems biology, bioinformatics, cancer
microenvironment modeling, and in
-
vivo stem cell labeling
,

includ
ing

Professor
Norbert
Perrimon

at Harvard Medical School, Professor
Dihua Yu
at UT MD Anderson Ca
ncer center,
Professors
Margaret Goodell

and
Daniel Medina

at Baylor College of Medicine
,
Professor
Michael Zhang at Cold Spring Harbor Laboratory, Professor Muhammad Zaman at UT Austin,
and Professor
Charles Lin

at
Massachu
setts General Hospital (MGH)
,
Harvard Medical School
.


T
o ensure effective management, a relatively s
mall leadership team (PI, PM, and Core
PIs)
has been
formed to

coordinat
e

primary project
s
, task
-
specific
-
projects and supporting core
activities. This team also bring
s

many facets of k
nowledge to bear upon the decision
-
making
process, enabling faster, more effective decisions to be made about shaping the direction of the
scientific research

of the CMCD.

This
has been

especially important in view of the demands of
working with research g
roups across multiple
disciplines and multiple
institutions. The small yet
representative nature of the team minimize
s

the c
ost of overhead and ensure
s

swifter
communications.


During the first funding cycle
the CMCD has been within
budget
,

generated
larg
e quantities
of data
,

and develop
ed

numerous collaborative efforts. Thus far we have:


-

Started a bi
-
weekly seminar series

-

Started monthly working group meetings for the Modeling Component

-

Started monthly working group meetings for the Imaging Component

-

S
cheduled our first annual 2
-
day symposium

-

Selected and scheduled our first external advisory panel meeting

-

Hosted a 2
-
day ICBP JI Meeting

-

Held monthly and ad hoc leadership meetings

-

Held small group meetings between the wet lab and modeling groups

-

Funded 5

pilot projects at $75,000 a piece with the help of match
ed

funding

from the
Methodist Hospital

-

Hired 4 post docs with the help of match
ed

funding

from the Methodist Hospital


-

Worked with 3 summer students

-

Setup a public website

-

Setup a private password
protected wiki site


C. Significance



Component 1


While a variety of signaling pathways have been implicated in control
l
ing TIC self
-
renewal, it is not known whether these pathways function in the TIC itself, or the surrounding
epithelial and/or
stromal cells. The availability of validated signaling reports will allow us to
define the identity of responsive cells directly, to localize these cells in space relative to other
cell types, and to literally watch them in real time as they grow and respo
nd to treatment. In
addition, should TIC be uniquely identified by presence or absence of expression of a given
reporter, we can purify these cells and evaluate them molecularly

and by high
-
content imaging.
Ideally, some commonalities will emerge. However,

we may also be able to define the range of
heterogeneity across multiple breast cancer subtypes. Finally, identification of novel regulators
of TIC function should lead to more effective targeting of the TIC and better our ability to monitor
treatment res
ponse in a meaningful way other than by tumor shrinkage.



Component 2


First
, our mathematical model studied the impact of TICs division pattern on the tumor
expansion, incorporated the effects of TIC niche, and integrated simplified effects of signaling
pathways as well. The simulated responses of tumor to drug treatments suggest that future
therapies should be either designed to effectively target the TIC niche, or block signaling
pathways.
Second
, our multiscale mathematical framework is necessary for t
he development of
a computational platform of tumor growth. It allows for the integration of data coming from
different scales into the models, e.g. molecular biomarkers of a given pathological process and
macroscopic tumor size. Such a multiscale approach

(i) allows for the accurate calibration of
parameters at various scales and (ii) provides an innovative tool for performing reliable
numerical simulations by handling the flow of information across scales in a robust and coherent
way.
Third
, the mathematical model can reproduce spatially heterogeneous sub
-
population
distribution within a tumor, as experimentally observed for the TICs; and the roadmap of
tumorigenic traits stemming from physiological mammary gland development will improve the

current design of breast cancer treatments.
Finally
,
tumorigenesis

modeling will provide a better
understanding of the interplay between tumor growth and angiogenesis development, and
suggesting hypothesis to develop novel anti
-
angiogenic therapy protocol
s.


Component 3


We have successfully initiated the postdoctoral training aspect of the project, which
should facilitate data analysis and enhance modeling capabilities. In addition, we have
established an atmosphere of cooperation and free and open commu
nication between the two
research components of the project.



Component 4


T
he administrati
ve core i
s
here
to provide a flexible yet effective administrative structure to
support the infrastructural and scientific aims, in view of the many faceted interactions that must
necessarily occur among the
CMCD

teams
. To accomplish this aim, we
have begun
develop
ing

and

exec
uting

a management plan based on a balanced management strategy that supports an
environment of shared decision
-
making and mutual re
sponsibility among the core PIs, w
hile
providing the oversight and leadership necessary to produce quality biomedical i
maging work.
We
have

manage
d

the overall
CMCD

project using sound basics, including phased delivery,
quick and concrete feedback, clear articulation of the project needs, project tracking and
oversight, effective governance, and inter
-
group coordination.
The PI, Project Manager (PM),
and Core PIs have held regular meetings to communicate project status and future goals.



D. Plans and tasks for year 2



Component 1

(1)
Using the validated reporters, t
racking and analyzing in
-
vivo 3D TIC mE

imaging
data; (2) TIC Signaling pathway analysis based on genomics and proteomics data; (3)
Bioinformatic analysis of TIC pathways based on shRNA data; (4) TIC
-
stromal cell
-
cell
interaction modeling and optimization strategy for targeting TIC niche therap
y (Figure
5
);
(5) Development of the multiscale mathematical framework by incorporating signal
pathway information and physical force information into current model; (6) Simulations of
the TIC development model and comparison with in vivo imaging data; (7)

Investigation
of the mammary gland development model and identification of the potential tumorigenic
events; (8) Prediction and analysis of the TIC development using the developed TIC
multiscale framework.


Component 2



Figure 5
: We propose to establis
h a TIC
-
centered, signaling pathway
-
based, multi
-
scale cell
-
cell
interaction regulated lineage model to investigate the long
-
term tumor recurrence risks of neo
-
adjuvant breast cancer treatments. The backbone model will simulate the self
-
renewal and
differe
ntiation of TICs and cancer progenitors(PCs), while system flexibility and robustness will
be addressed by multi
-
scale regulations via cell
-
cell interactions. Such intercellular
communications in cell microenvrionments are: 1) tumor stromal cells promote T
IC renewal via
Wnt, Hh, and Akt
-
triggering factors (not known yet); 2) tumor cells and other fully differentiated
cells suppress the growth of TICs and PCs via TGF
β

(not sure yet); and 3) cell fate control
under the microenvironments of Wnt, Hh, and Akt
-
triggering factor levels via Notch/Delta
jaxacrine (need biologists to check the idea). The key parameters are the renewal rate,
differentiation rate, and the renew/diff ratio of the TICs and PCs. The major experiment
observations as well as model outputs

are TICs, PCs, and TCs population, and the Wnt, Hh,
Notch, Akt, and TGF
β

signaling strength at multiple time points in each treatment group (IR,
chemotherapy, perifosine, etc). Key parameters will be determined based on experiment data,
then the system wi
ll be explored by parameter sensitive analysis to reveal vulnerable chains as
efficient drug targets,


and the model will predict best strategies for TIC
-
targeted neo
-
adjuvant
treatments and will be validated by follow
-
up experiments. This work will help u
s to better
understand the roles of the TIC microenvironments, especially the multiscale cell
-
cell
interactions, in terms of radio
-

and chemo
-

therapy resistance and long
-
term tumor recurrence,
to discover efficient drug targets, and to optimize the TIC
-
ta
rgeted treatment strategies.


Component 3


During the second year we plan on h
iring
a

second postdoctoral fellow

for the educational
component. We also plan on c
omplet
ing

the summer undergraduate program
, mentoring

the
ICBP summer intern
s, and continuing

the R&D seminar series and working group meetings
.



Component 4


Our plan for the second year is to continue tracking project milestones while building upon the
success of the first year. Our
five
pilot projects will be fully funded in the second year and we
therefore look forward to increasing the amount of data produced by this center. Our first
annual symposium will also take place after the completion of this progress report, but still in the
f
irst year’s funding cycle. We will also market our private password protected wiki site to
enhance participation and increase collaboration.



E. Publications and manuscripts


1: Li, F., Zhou, X., and Wong, S.T.C, (2010) Optimal Live Cell Tracking for C
ell Cycle Study
Using Time
-
lapse Fluorescent Microscopy Images
International Workshop on Machine Learning
in Medical Imaging (MLMI 2010)
. Springer Lecture Notes in Computer Science, Beijing, China,
124
-
131.


2
: Zhang M. Atkinson RL., Rosen JM, Selective Ta
rgeting of Radiation
-
Resistant Tumor
-
Initiating Cells.
Proc
Natl Acad Sci U S A
. 2010 Feb 23;107(8):3522
-
7. Epub 2010 Feb 3

PMID:
20133717


3
: Zhu, XW, Zhou, XB, Lewis, M, Xia, L, and
Wong, STC, Cancer stem cell, niche and EGFR
decide tumor development and treatment response: A bio
-
computational Simulation Study,
Journal of Theoretical Biology

269 (1
):38
-
149
.

P
MID: 20969880


4
: Visbal AP, Lewis MT.
Hedgehog signaling in the normal and neoplastic mammary gland.

Curr
Drug Targets
. 2010 Sep;11(9):1103
-
11. Review.
PMID: 20545610


5
: H. Hatzikirou, A. Chauviere, J. Lowengrub, J. De Groot and V. Cristini,
E
ff
ect
of
vascularization on glioma tumor growth. In Modeling Tumor Vasculature: Molecular, Cellular,
and Tissue Level Aspects and Implications, Trachette

L. Jackson Eds., Springer (In Press)


6
. K. Pham, A. Chauviere , H. Hatzikirou, X. Li, H. Byrne, V. Cristini and J. Lowengrub, Density
-
dependent quiescence in glioma invasion: instability in a simple reaction
-
d
iff
usion model for the
migration/proliferati
on dichotomy, Submitted (2010) to Journal of Biological dynamics






F. Project
-
generated Resources


In year 1, w
e have developed cell segmentation tool packages, TIC signal network inference
tools, TIC lineage models, TIC multiscale modeling tools, and
a number of tumor growth
models. These tools and models will be freely available after they are validated.

As these tools
are validated they will be
made
available on our public website, and will be shared with the
ICBP Datasets &

Software Working Group that meets monthly.