Tumor-Derived Vasculogenesis - Cellular Potts Models: Multiscale Extensions and Biological Applications - page 131

strength (by setting ch ;C = 0 in Equation (6.2)). As represented in Figure
7.13(A), the resulting simulation reproduces a clear disruption of network as-
sembly, with the formation of poorly structured vascular islands (lT T 0.14L T
and pct = 0.86), similar to those experimentally obtained by extinguish VEGF
gradients, for example, in the case of \normal" ECs, by adding saturating
amounts of exogenous morphogen [260, 261, 360]. This result shows that a
stimulation of cell adhesion and motility, without a chemotactic mechanism,
does not suce for the maturation of a capillary-like structure. In particular,
an accurate analysis of cell tracks shows that, although maintaining a certain
degree of directional persistence (due to the term H persistence in (6.1)), cell
movement is completely uncorrelated from the directions of chemical gradi-
ents (see Figure 7.13(B)). Migrating individuals feature also an unpolarized
morphology during the entire patterning, as reproduced in the same represen-
tative image: without experiencing the chemical stimulus, their membranes
are in fact not able to protrude in a preferred direction, but isotropically
fluctuate in the environment. This result shows that a stimulation of cell ad-
hesion and motility, without a chemotactic mechanism, does not suce for the
maturation of a capillary-like structure, as also demonstrated experimentally
by extinguish morphogen gradients in [360], and theoretically in other Potts
models [260, 261].
In order to review and compare the different results, in Figures 7.14 7.15,
the percentage of reduction of the total tubule length (i.e., pct) has been
compared in the different cases using a one-sided equal-variance t-test with a
Bonferroni correction as a scoring method [333]. The level of significance has
been taken P < 0.005. Indeed, it is satisfying to see that the model has realis-
tically reproduced the effects of currently available therapies (focusing on the
interferences with VEGF activity or with calcium machinery) and has been
also able to suggeste novel and interesting interventions. In particular, we have
made some predictions about the possible success of therapies blocking the
mechanisms of either the cytoskeletal remodeling or the increment in cellular
adhesion. Development of interventions inhibiting cell chemotaxis and persis-
tent movement might be also optimal strategies. All the proposed solutions
have the potential to dramatically reduce the angiogenic phenotype of tumor-
derived endothelial cells in unexpected multiple ways and have emerged by
opportune and biologically reasonable variations in model parameters: even
though it is possible to provide a post hoc explanation the effectiveness of each
of them, it is unlikely that they would have been manually discovered. Ob-
viously, more experiments, focused on the relevance of our findings, are then
needed to shed light and validate our conclusions. Eventually, it would be also
interesting the relevance of our results in vivo. However, in this case, it would
be very dicult to have a complete control of all the involved biological mech-
anisms, and the relative investigations would imply obvious complications due
to the recruitment of cell types other than endothelial cells and to potential
unexpected side effects [57, 65].