Biology Reference
In-Depth Information
Differentiation of embryonic stem cells became
a subject of quantitative exploration as well [16] . Mutual
repression between Cdx2 and Oct4 has been suggested to
regulate differentiation to the trophectoderm lineage, while
mutual repression between Gata6 and Nanog regulates
differentiation to the endoderm lineage (see Figure 22.3 D).
One can see that the model integrates a pair of bistable
switches, each combining a single pluripotency factor and
a single lineage commitment regulator [19] along with
bimodal distributions for the core factor concentrations.
The model predicted a quite interesting concentration-
dependent ('bell-shaped') response of Gata6 to the
concentration of Oct4, resulting from direct activation and
indirect repression of Gata6 by Oct4. Analogous concen-
tration-dependent responses are known in other well-
studied systems, such as Drosophila embryogenesis [66] .
Mutual exclusion between the alternative cell fates may be
achieved, in part, via Oct4: activation of Cdx2 (trophecto-
derm primed cells) causes Oct4 repression and prevents
Oct4-mediated activation of Gata6, thus blocking endo-
dermal cell fate. Interestingly, the explored network
topology and the model suggested that Nanog over-
expression, not suppression of the lineage-specific factor
Gata6, is the optimal way of reprogramming endodermic
cells into iPS. Inferring differentiation and reprogramming
conditions from future quantitative models may potentially
supplement or even replace empirical studies in the
reprogramming field.
or myeloid lineages [69] . The switch is mediated by mutual
repression between the transcription factors Pu.1 and Gata1
(see Figure 22.3 E) [70] . In this case, however, the bistable
switch is placed directly between these factors, responsible
for the alternative cell fate commitment, and it is also
reinforced by yet another parallel switch downstream
(C/Epba
Fog1).
One can see that in both considered cases (see differ-
entiation of ESC above) differentiation events are typically
associated with mutual repression; however, the bistable
switches occupy quite different positions relative to each
other and to the core network. In the case of ESCs, their
communication is indirect and much 'softer'.
A simple three-gene model of hematopoiesis utilizing
Boolean networks has also been proposed [71] . In Boolean
terms, the stem cell genes should be 'ON' and the differ-
entiation genes 'OFF' in the dormant state; the genes
switch to opposite states when the cells are fully
committed. Under the Boolean network framework,
a hematopoietic stem cell represented by three genes can
fall into either a 'rest' or a 'cycling' attractor state
according to a set of rules governing the three genes.
Interestingly, this may correspond to switching between the
bistable and the oscillatory scenarios, described for ESCs
above (see Figure 22.3 A and B).
e
STRATEGIES FOR MODEL
CONSTRUCTION AND VALIDATION
Data Integration and Network Construction
The functional properties of various biological systems can
be uncovered by inferring the architecture of biological
networks and modeling GRN dynamics using either reac-
tion-based (BRN) or statistical-based (SIN) models
( Figure 22.4 ). With the advent of high-throughput
biotechnologies, efforts have been made to map and
reverse-engineer the GRNs of stem cells. The past few
years have witnessed the development of resources for
stem-cell-centered networks and databases for the broad
stem cell community. A number of regulatory interactions
are deposited in network-based repositories such as Pluri-
NetWork [72] , Plurinet [73] and iScMiD [36] . Additionally,
WikiPathways provides the CIRW portal that highlights
pathways contributed to and maintained by the stem cell
community [74] . In the context of transcriptome data,
databases such as FunGenES and StemBase contain gene
expression studies and others with interactive query tools
and a web interface [75,76] . In addition to gene expression
data, genome-wide protein
Models for Hematopoietic Stem Cells
Hematopoietic stem cells (HSCs) are undifferentiated cells
that give rise to all types of blood lineage progenitors and
terminally differentiated blood cells. HSCs are found in
myeloid tissues such as red bone marrow with frequency
~10 e 4 , relative to other cell types. Unlike with embryonic
stem cells, HSC culturing is not yet attainable, and purifi-
cation for medical applications is inefficient and very
expensive.
A large proportion of HSCs in mouse are present in
a non-proliferating or quiescent (dormant) state. Under
native conditions with a very low frequency HSCs may
enter the cell cycle, which typically leads to irreversible
differentiation of HSCs to blood lineage progenitors. In this
respect it is still not quite clear whether the dormant HSC
state is analogous to the pluripotent ESC state. Recently, an
architecture for the core HSC network, presumably
responsible for transition between the dormant and differ-
entiated states, has been suggested (see Figure 22.1 B) [37] .
Models describing this transition are not yet available, with
the exception of quantitative analysis of HSC pool
exhaustion and cellular aging due to proliferation [67,68] .
One well-studied model for hematopoiesis focuses on
a switch governing differentiation of HSCs into erythroid
gene-binding ChIP-seq/chip
data in stem cells are collected into the public domain in
NCBI's GEO [77] database, while databases such as ChEA
[78] and ESCDb [79] collect a number of processed ChIP-
seq/chip results and allow users to query for genes of
e
Search WWH ::




Custom Search