Systems Immunology: From Cells and Molecules to a Dynamic Multi-Scale System - Systems Biology Concepts and Insights - page 470

(e.g., a stimulation which may be spatially or temporally
restricted) which would define each cell's initial condi-
tions and ultimately cellular behavior.
cImmSim is a stochastic cellular automata system based
on an the original ImmSim model proposed by Seiden and
Celada [59] . It incorporates a general textbook model of
innate and adaptive immunity, including several cell types,
states they can adopt with probabilistic transition rules,
antibodies and antigen. Cells may have unique properties,
such as their receptor specificity, specified by a bit string
(e.g., 001110), and an interaction between two such entities
occurs with a probability that is a function of bit string
similarity (modeling affinity). All the feasible interactions
among cells and molecules take place within a lattice with
cell-to-cell interactions occurring between neighboring
cells ordered on the lattice. Time is discretized, and at every
single time step entities (cells and molecules) may diffuse
to adjoin neighborhoods on the lattice. For example,
a B-cell entity in an 'exposing' state implies (but is not
directly modeled) that the cell has phagocytosed one
antigen and has already processed it. If the bind with the
MHCII molecule is successful then the cell is exposing the
MHCII molecule bond with one antigen peptide. If an
interaction with a T-helper cell in an 'Active' state occurs
(as determined by a stochastic event and position on the
lattice), the action is to update both the T-helper and B cells
instances to a 'Stimulated' state [60] . Millions of cells can
be simulated simultaneously, and their states computed in
parallel and asynchronously [60] . Simulating viral infec-
tions (e.g., HIV and Epstein

computers through software packages implementing the
formalism. It has been adapted for use in modeling
biological systems [64] . Behavior in Statecharts is
described using states and events that cause transitions
between states. States may contain substates that are
suitable for multilayered systems (e.g., molecules within
cells). In addition, Statecharts allow for 'orthogonal
states' where the same object (i.e., a cell) may exist in
different states. In modeling the immune system, this
functionality is co-opted to describe the differentiation of
a cell from one subtype to another (e.g., from a na¨ve cell
expressing CD4 to a memory cell when co-expressing the
cell surface protein CD45RO). Transitions in Statecharts
take the system from one statetoanother;thebiological
equivalent is the result of an interaction either between
two cells or between a cell and various molecules.
Although models here may be as simple as the one defined
using cellular automata, the high expressivity of the
programming language allows one to define much more
complicated models that include many states and many
substates, as well as spatial information without loss of
reliability ( Figure 25.3 A). A novel visualization layer sits
on top of these models and displays a visual image
of cells, molecules and their interactions as well as
enabling user manipulation of the simulation parameters
( Figure 25.3 B). The end result is a mathematically precise
presentation of the underlying simulation that is execut-
ableandisakintowatchingavideorecordingofacellor
atissuethroughamicroscope.
Using Statecharts, Efroni, Harel and Cohen modeled
first the biological process of T-cell maturation in the
thymus [65,66] . The migration of cells in the thymus
depends on many factors, including their receptor, the
chemokine gradients, epithelial cells, cell proliferation and
survival. The thymic environment, loaded with molecules
and cells, presents a challenge to many researchers from
different fields who have detailed knowledge of some of its
parts, but yet wish to comprehend either system-level
effects of a molecular level change or to identify the origin
of a system-level property. Efroni et al. integrated data
generated from reductionist biology from hundreds of
scientific publications into a specification model encoded in
Statecharts.
In silico testing the effects of loss-of-function mutations
in the CXCR4 and CCR9 chemokine receptors on cell
migration matched those observed in histological samples,
with the added advantage of being able to observe in real
time the dynamics of cell migration in thousands of cells. A
developing thymocyte must commit to becoming either
a CD4 þ or a CD8 þ cell in the thymus. The decision-making
process is obscure and the ratio of mature CD4 þ to CD8 þ is
unequal (roughly 2:1). Where information is not known, the
Statecharts framework enables the simultaneous specifica-
tion of alternative theories. With respect

Barr) using this model, Cas-
tiglione and colleagues showed that they could reproduce
disease natural history phenomena such as time from
infection to AIDS development in the model [61] , and then
use it to suggest the best time to administer antiretroviral
therapy (HAART) [62] .

e

Statecharts as a Rich Framework to Model
Immunity
Cellular automata-based models lack detailed biological
information on the modeled interactions. As such, they are
limited in the amount of insight they can yield, and an
observed disagreement between them and what is observed
in the actual system is common and often goes unex-
plained. Other models in use to model immune-related
phenomena are much richer in their specification of the
details of the biological process. These require gathering
large amounts of information from the literature and
encoding information gathered from text, tables and figures
into a machine-understandable format.
Themode inglanguageofStatecharts is a visual
formalism invented to aid the design of complex man-
made reactive systems [63] . It is mathematically well
defined, which makes it amenable to execution by to lineage