Biomedical Engineering Reference
In-Depth Information
change in response to mechanical distortion, this integrated structural and in-
formation-processing network would be perfectly designed to sense the diffuse
signals that emanate from a concerted rearrangement of the cytoskeleton in re-
sponse to mechanical stress or physical changes in cell shape. In fact, both
molecules that physically associate with the cytoskeleton in the focal adhesion
site and at other locations throughout the cell have been shown to change their
activity in response to applied mechanical stress or cell distortion (23).
Because the wiring of the signaling networks produce attractors that corre-
spond to only a limited number of distinct cell fates, the cell may naturally and
reliably sense a broad spectrum of signals and simultaneously orchestrate multi-
ple molecular responses to produce coherent behavioral programs. In other
words, the existence of attractors representing distinct cell behaviors facilitates
the evolution of a form of regulation that connects signals devoid of molecular
specificity like mechanical forces to the internal regulatory machinery that gov-
erns specific cell fates.
4.2.4. Experimental Evidence for Attractors in Gene Regulatory Networks
The characteristic dynamics of cell fate control, the mutual exclusivity of
different phenotypes, and their robustness in living cells all suggest that distinct
cell fates represent attractors that emerge in the dynamic network of gene regu-
latory interactions. But can we directly view the structure of the attractor land-
scape network at the molecular level without knowledge of the precise wiring
diagram of the underlying genome-wide regulatory? To map out this state space,
it would be necessary to simultaneously measure the activation state of the ge-
nome-wide set of molecular activities that are responsible for cell fate switching.
The arrival of technologies for the massively parallel monitoring of genes now
opens this possibility to follow trajectories of cell states in a high-dimensional
state space of the regulatory network. Gene expression profiling using DNA
microarrays allows the parallel measurement of the level of >10,000 mRNAs in
cells and tissues; this represents a surrogate measure for genome-wide gene ac-
tivation profiles, and hence for cell states.
One way to uncover the existence of a high-dimensional attractor in real
cells, where unlike in computer simulations we cannot systematically sample the
states in state space, is to approach it from different directions of the state space
and demonstrate the convergence of the high-dimensional trajectories (Figure 4).
For instance, expression profiling to probe such trajectories in state space could
be used to monitor cells that are induced by two different (biochemically dis-
tinct) stimuli to undergo the very same cell fate switch, e.g., from a proliferative
state to a differentiated state. As noted above, such scenarios are common, and
as such already suggest that the differentiated state is a stable attractor. If the
two trajectories in gene expression space first diverge but then converge with
