A Multiagent-based Simulation of the Infection of the Macrophage by
Trypanosoma Cruzi in the Acute Phase of Chagas’ Disease: Influence of
the Initial Inoculum and Protozoan Escape Factor
Willian Cordeiro Farago
1
, Alcione de Paiva Oliveira
1
, Rodrigo Siqueira-Batista
2
,
Andr´eia Patr´ıcia Gomes
2
, Juliana Lopes Rangel Fietto
3
, F´abio Ribeiro Cerqueira
1
and Luiz Alberto Santana
2
1
Departamento de Inform´atica, Universidade Federal de Vic¸osa, 36570-900, Vic¸osa, Brazil
2
Departamento de Medicina e Enfermagem, Universidade Federal de Vic¸osa, 36570-900, Vic¸osa, Brazil
3
Departamento de Bioqu´ımica e Biologia Molecular, Universidade Federal de Vic¸osa, 36570-900, Vic¸osa, Brazil
Keywords:
Multiagent System, Immune System Simulation, Chagas’ Disease.
Abstract:
Chagas’ disease presents a wide distribution in Latin America. Some epidemiological studies show that there
are a prevalence of 16 to 18 million people affected by the disease and at least 100 million people at risk of be-
ing infected on these areas. Recently, due to the globalization, the disease, which was previously endemic only
in countries of Central America and South America, is now presenting cases in other regions such as North
America and Europe. Trypanosoma cruzi, the etiological agent, has the ability to promote changes in the tis-
sues of the vertebrate host with significant morbidity, according to the degree of infection. This article presents
a simulation of the human immune system using a multiagent system approach. More specifically, we propose
a simulation study aimed at the acute phase of Chagas’ disease, related to the macrophage-Trypanosoma in-
teraction. The simulation showed that the initial number of T. cruzi influences in the outcome of infection and
was found a relationship between the escape factor and the total elimination of T. cruzi.
1 INTRODUCTION
Chagas’ disease (CD) was first described by Carlos
Ribeiro Justiniano Chagas in 1909, in northern Minas
Gerais (Brazil). Recently, because of the globaliza-
tion, this disease that was endemic only in countries
of Central America and South America is now oc-
curring in other regions of the world such as North
America and Europe (Siqueira-Batista et al., 2007;
Rassi Jr et al., 2012). Trypanosoma cruzi (T. cruzi),
the etiological agent, is capable of causing a sig-
nificant damage in the vertebrate host tissues lead-
ing to serious consequences, according to the de-
gree of infection (Alcantara and Brener, 1980). Af-
ter infection, the individual develops an acute phase
of short duration. At this stage a highly aggressive
form of T. cruzi, the trypomastigotes, is present in
the bloodstream. This causes the immune system re-
sponse via humoral and cellular mechanisms of innate
immunity such as complement-mediated lysis, natu-
ral killer cell cytotoxicity and phagocytosis carried
out by the mononuclear phagocyte system (MPS),
mainly by macrophages (Coura and Borges-Pereira,
2010; Rodrigues et al., 2010). The parasites use
escape mechanisms to survive to the action of the
macrophages and are not completely eliminated lead-
ing to a chronic disease characterized by low par-
asitemias with the prevalence of intracellular para-
sitism of the amastigotes form (Coura and Borges-
Pereira, 2010; Rodrigues et al., 2010). The disease
presents two distinct phases, an acute phase which
is followed by a chronic phase that can be classified
as indeterminate, cardiac or digestive form (Siqueira-
Batista et al., 2007). Usually, one third of the pa-
tients presenting the indeterminate classification de-
velop symptomatic CD some decades after the ini-
tial infection (Siqueira-Batista et al., 2007; Coura and
Borges-Pereira, 2010).
It is believed that the T. cruzi load in the initial in-
fection is a major factor in the development of CD
(Borges-Pereira et al., 1988; Martins et al., 2012).
Cases reporting feces of triatomine with more than
1500 parasites have been described, but it is believed
that the quantities of T. cruzi needed to cause infec-
tion is far less than the inocula used in experimen-
tal infections (Coura and Borges-Pereira, 2010; Ro-
Farago, W., Oliveira, A., Siqueira-Batista, R., Gomes, A., Fietto, J., Cerqueira, F. and Santana, L.
A Multiagent-based Simulation of the Infection of the Macrophage by Trypanosoma Cruzi in the Acute Phase of Chagas’ Disease: Influence of the Initial Inoculum and Protozoan Escape
Factor.
In Proceedings of the 18th International Conference on Enterprise Information Systems (ICEIS 2016) - Volume 2, pages 29-38
ISBN: 978-989-758-187-8
Copyright
c
2016 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
29
drigues et al., 2010). Studies report that the vast ma-
jority of cases of acute CD in Brazil are asymptomatic
or mildly symptomatic, possibly due to low inoculum
and/or humoral response (Coura et al., 1983; Coura,
2013). A study encompassing almost three decades
of patients follow-up showed that the cases of more
severe chronic forms originated from cases with se-
vere acute phase. Thus it is possible to deduce that
the initial inoculum and/or the infecting strain of T.
cruzi influenced the evolution of the disease. Experi-
mental studies have shown that this load is generally
low (Dias, 1982).
Among the cells of the mononuclear phagocytic
system, the macrophage presents a key role in the
host response to T. cruzi. These cells have trypanoci-
dal and tripanostatic capacity and often determine the
degree of susceptibility of the infected host (Russo
and Starobinas, 1991). Activation of macrophages
in the acute phase of CD has long been described in
the literature. Activated macrophages are able to kill
the parasite. However, the resident macrophages do
not perform this action. The relevance of this try-
panocidal activity is not yet fully known (Celentano
and Gonz´alez Cappa, 1993). Macrophages phago-
cyte epimastigotes and trypomastigotes. The first
kind is digested, while the latter kind escapes from
the vacuole and is transformed into amastigotes that
multiply inside the macrophages. This phenomenon
occurs in the region where the penetration of try-
pomastigotes takes place, initiating the intracellular
stage of the parasite. Each amastigote by binary fis-
sion yields dozens of new amastigotes that after four
to six days are transformed into tripomastigotes that
disrupt the cell and reach other tissues. A simple bi-
nary division of amastigote is repeated every twelve
hours and produces about nine generations of para-
sites that disrupt the macrophage (Rey, 2008). Acti-
vation of macrophages leads to activation of T lym-
phocytes that secrete interferon γ, which in turn in-
duces macrophages to produce interleukin 1 (IL 1),
tumor necrosis factor (TNF), and nitric oxide (NO)
(Siqueira-Batista and Geller, 2008). The NO is ex-
tremely toxic for the T. cruzi and seems to be respon-
sible for tripanolitic action exerted by macrophages
(Rey, 2008). More detailed studies to improve the un-
derstanding of the immune system may contribute to
the treatment and prevention of the disease.
Multiagent system (MAS) has been successfully
applied in many complex systems. It allows the ex-
ploration of macroscopic behavior emerging from mi-
croscopic interactions. For this reason, MAS is con-
sidered by many authors the best approach to model
the immune system. Its main disadvantage, however,
is the high computational cost when a large number
of agents is used (hua Li et al., 2009). The simula-
tion using techniques of multiagent systems (MAS)
has been an important tool in the investigation of the
phenomena of interactions between the various cell
types of the immune system and pathogens such as
viruses, bacteria, and protozoa(Possi et al., 2011; Fol-
cik et al., 2007; Borges, 2012). These studies led to
designs with a bottom-up approach, i.e., the systems
are conceived defining the individual aspects of the
agents, such that the emergency of the collective be-
havior occurs naturally from the interactions of the
agents with themselves and with the environment. As
a result, it is possible to test hypotheses of interest in
a realistic manner, i.e., without programming the de-
sired scenarios directly, but, instead, obtaining them
as a consequence (Possi et al., 2011).
The group of Modeling and Simulation of the Im-
mune System (ModeSimmune) of the XXX Univer-
sity, in a joint effortwith the Departments of Medicine
and Computer Science, is developing a simulator
of the human immune system called AutoSimmune,
based on the simulator initially developed by (Fol-
cik et al., 2007). AutoSimmune has been developed
since 2010 and is now able to perform many simu-
lations of interest, including viral and bacterial in-
fections and events involved in autoimmune condi-
tions (Possi et al., 2011; Bastos et al., 2013). In this
work, the investigations are focused on CD. To sim-
ulate CD regarding the interaction between T. cruzi
and macrophage cells, it was necessary to model and
introduce the T. cruzi agent. Furthermore, we had to
modify the existing pattern of the macrophage agent
in the simulator so that it was able to respond to stim-
uli from T. cruzi as described by (Coura and Borges-
Pereira, 2010). We analyzed whether the initial num-
ber of T. cruzi parasites influences the outcome of in-
fection and also the relationship between the rate of
escape of the parasite from the macrophage phago-
cytic vacuoleand the final number of T. cruzi agent af-
ter an intracellular multiplication cycle of them (Possi
et al., 2011). Assuming that we have a reliable sim-
ulation of the real immune system, the analysis of
the results that we have obtained from our in silico
simulation provide us with strong evidences that new
hypotheses on the investigation of potential therapies
and/or drugs against CD might be of great impact in
this field of study.
The next section describes some work related to
the proposed system. Section 3 presents the material
and methods used in the simulation, including a de-
scription of the model. Section 4 shows the results of
running the simulation. Section 5 discusses the results
presented. Finally, section 6 presents the conclusions
of this work.
ICEIS 2016 - 18th International Conference on Enterprise Information Systems
30
2 RELATED WORK
With regard to novelty, there have been other agent-
based models used to study T. cruzi, including
((Galv˜ao et al., 2008),(Devillers et al., 2008),(Galv˜ao
et al., 2010) and (Rogers et al., 2010)), but none have
focused on the interactions of macrophages and T.
cruzi.
(Galv˜ao et al., 2008) developed of a two-
dimensional agent-based model for chronic chagasic
cardiomyopathy after stem cell transplantation. Their
model included six different types of agents: inflam-
matory cell, fibrosis area, cardiomyocyte, proinflam-
matory cytokine tumor necrosis factor-α, T. cruzi and
bone marrow stem cell. Although not explicit in the
paper, they have used a model based on cellular au-
tomata.
(Galv˜ao et al., 2010) developed a three-
dimensional multi-agent-based computational model
for the evolution of Chagas’ disease. Their model
included five different types of agents: inflammatory
cell, fibrosis, cardiomyocyte, fibroblast, and T. cruzi.
Note that the macrophages were not modelled. Fi-
brosis was a fixed agent and the other types of agents
could move through the empty space randomly by us-
ing the Moore neighborhood. The aim of the model
was to reproduce the acute and chronic phases of Cha-
gas’ disease and the volume occupied by all different
types of cells in the cardiac tissue.
(Rogers et al., 2010) created a model-based agents
focusing on the T. cruzi vector in the Texas region.
They incorporated data related to the dominant syl-
vatic vector and host species for the T. cruzi in Texas
into an agent-based model using NetLogo.
3 MATERIALS AND METHODS
AutoSimmune is an immune system simulator with
original focus in autoimmunity. In its basic version,
it simulates the bone marrow, thymus, lymph nodes,
blood circulation and parenchymal tissue region. The
regions are simulated as a discrete space in a form
of a two-dimensional grid in which each agent has
a position (i, j). More than one agent can occupy
the same position, which somehow simulates a 3D
space. The movement of the agent is done by chang-
ing its current position to a position in the Moore
neighborhood (Dewri and Chakraborti, 2005). Thus,
an agent cannot “jump” positions, it needs to move
one position at a time. In a cell structure in the form
of a two-dimensional grid, the Moore neighborhood
(of radius one) comprises the eight neighboring cells
to a central cell. If allowed in their specification,
an agent can move from one region to another by
means of special elements called portals, as proposed
in (Folcik et al., 2007). The simulation of substances
such as cytokines are performed by means of a layer
of data, named ValueLayer, provided by the Repast
framework
1
, over which the simulator was built. Sub-
stances, as they are released by cells, undergo a pro-
cess of diffusion, spreading in the surroundings of the
site in which they were released, decreasing its con-
centration, and also undergoing a process of decay,
reducing its amount with time (Possi et al., 2011).
The ValueLayer is an abstract layer of data that, at
the time of its creation, is associated with a region of
the grid. It is possible to combine multiple layers of
data at the same grid. Thus, an agent can know the
concentration of a given substance at that time instant
at position (i, j) (Possi et al., 2011).
The time is modeled using the concept of dis-
crete unit time provided by the framework, called
tick. Each component schedules their execution time,
informing when to start and the invocation interval
(Possi et al., 2011). As a result, all events are properly
synchronized. In the simulator, the affinity (which is
the recognition strength of an antigen by a receptor)
is simulated by the number of matching bits between
two bit sequences: One belonging to a cell receptor
and another belonging to the antigen. The greater
the length of the matching, the greater the affinity.
For the calculation of the affinity we used the method
suggested by (Floreano and Mattiussi, 2008), called
the “length of the longest common subsequence”,
whose goal is to compute, given two patterns of bit
sequences A and B, the size of the largest contigu-
ous subsequence of bits that are contained in A and B
simultaneously, in the same order.
3.1 The Macrophage Agent Model
The software agent that plays the role of macrophages
in the simulator AutoSimmune needed to be im-
proved, since some of its functions that are essential
in the interaction with the T. cruzi were not modeled
and implemented in previous versions. The rules of
this agent are illustrated by the state chart showed in
Fig. 1.
The macrophages in AutoSimmune accomplish
their activities in the tissue, represented in the sim-
ulator by the Tissue zone (based on “Zone 1” defined
by (Folcik et al., 2007)) which represents a slice of
a microscopic generic parenchymal tissue. This agent
migrates to the tissue when they detect the presence of
pro-inflammatory cytokines (MK1 group, after (Fol-
cik et al., 2007)) and enter the Tissue zone. The
1
repast.sourceforge.net
A Multiagent-based Simulation of the Infection of the Macrophage by Trypanosoma Cruzi in the Acute Phase of Chagas’ Disease: Influence
of the Initial Inoculum and Protozoan Escape Factor
31
Figure 1: Macrophage state chart.
agent follows the cellular stress signaling substance
until encountering the site of infection flagged by the
necrosis substance released after cellular lysis. In the
presence of necrosis, itcommutes to the macrophage
pro-inflammatory state, which produces and releases
cytokine MK1 and phagocytes dead cells, PAMP-
positive pathogens and antigen-antibody complexes,
i.e., antibodies bound to any antigen marking it for
phagocytosis.
Changes in the macrophage behavior when it
phagocytes some foreign particle were inserted in the
AutoSimmune model. In this case, a transition state
called Destroying was created. When a macrophage
phagocytes T. cruzi, it may occur that T. cruzi is
not destroyed and ends up infecting and multiply-
ing inside the macrophage. If the micro-organisms
escape from the parasitophorous vacuole and infect
the macrophage, the macrophage switches to infected
state, continuing with its actions until the number of
parasites is large enough to rupture the macrophage.
After rupture, the cell dies and goes to the Necro-
sis state. At this time, the parasites are released
into the medium and the dispersion of necrosis sub-
stance occurs until the macrophage agent is engulfed
by another cell. When the micro-organisms or par-
ticles are destroyed the agent macrophage processes
it, extracting their antigens and presenting them to
T lymphocytes featuring the Enabled state. Infected
macrophages that destroy any micro-organism will
change to the Infected/Enabled state with the same
functions of the Enabled state, but keeping the inter-
nal proliferation of pathogens.
3.2 The Trypanosoma Cruzi Model
Initially, T. cruzi infects and multiplies in cells of
the mononuclear phagocyte system (MPS), notably
macrophage. Later on, T. cruzi migrates to other
parts of the organism and invades other host cells, es-
pecially smooth muscle, skeletal, cardiac nerve, en-
abling their stay in the host (Siqueira-Batista et al.,
2007; Martins et al., 2012). As we want to sim-
ulate the primary immune response, we inserted T.
cruzi in the Tissue zone. The representative agent
of T. cruzi in AutoSimmune was modeled and imple-
mented according to the parasite’s main functions and
features described in the biological literature, keeping
the highest possible fidelity, so that we could emu-
late its behavior and allow the possibility for poste-
rior in vivo analysis of the immune response towards
the pathogenic agent and its peculiarities. The rules
of this agent are illustrated in Fig. 2.
Initially, the agent is in trypomastigote form and
randomly moves looking for other agents, which is
called Circling state. When it meets another agent, T.
cruzi checks if there is some affinity with it. If affinity
exists, this is a possible host cell and so the parasite
invades it. After invading the cell, T. cruzi turns into
amastigote and initiates a cycle of multiplications by
successive binary divisions with nine fissions mostly
producing 50-500 parasites depending on the strain of
T. cruzi and the host cell (Rey, 2008; Siqueira-Batista
et al., 2007). After this step, still inside the parasitized
cell, T. cruzi undergoes further transformation into the
trypomastigote form and is released from the cell in
order to penetrate into other cells and tissues.
ICEIS 2016 - 18th International Conference on Enterprise Information Systems
32
Figure 2: Trypanosoma cruzi state chart.
3.3 Definition of Variables
The environment for the interactions between agents
in the simulations was the Tissue zone with grid size
150 by 150 units of space. To further analyze the in-
teraction of the parasite with the immune system, a
period of 1825 ticks was fixed. The proposed value
for the tick in this study corresponds to six months of
infection. In this simulated environment, according to
the modeling decisions, the immune reaction should
not happen effectively, since the secondary immune
response has not been triggered yet and it is central to
restrain the levels of parasites (Siqueira-Batista et al.,
2007). During the simulations the following parame-
ters were evaluated:
Tick: Unit of time provided by the framework.
The relationship between computational tick and
real time required for analysis and interpretation
of the simulations was based on the multiplica-
tion of T. cruzi. Amastigotes multiply by binary
fission every 12 hours in vivo, according to Rey
(2008), and in AutoSimmune the same process
occurred within 5 ticks, so we defined that a tick
corresponds to 2 hours and 24 minutes.
Initial number of T. cruzi:Total number of T. cruzi
inserted in the Tissue zone two ticks after the be-
ginning of the simulation. This inoculated para-
site load was set to 300 in the first series of simu-
lations and 3000 in the second series.
Initial number of macrophages: Number of
macrophages present in the Tissue zone before the
parasite load being inserted.
Escape factor: Defines the chance of T. cruzi es-
caping from the phagocytic system and multiply-
ing within the macrophage. It can vary from zero
to one hundred percent of chance.
Active infection: It is the active entry with energy
expenditure of T. cruzi into macrophages. As the
main mechanism of invasion occurs by phagocy-
tosis, the chance of occurring active infection has
been fixed at 0.10 percent.
Breach number: The maximum number of para-
sites within the macrophage after multiplication
cycle. When the number of T. cruzi reaches this
limit, the cell breaks downand the parasites are re-
leased. In this work, we follow what is described
by (Rey, 2008), i.e., after invading macrophage,
T. cruzi multiplies through a series of binary divi-
sions, about nine, generating approximately 500
strains.
Tick when the numberof T. cruzi is zero: The time
when the number of parasites reaches zero.
Maximum number of T. cruzi: It is the maximum
number of parasites in the Tissue zone area at a
given time. Generally,in vivo, this moment occurs
in the acute phase of the disease during the first
weeks of infection.
Maximum number of macrophages: The highest
occurrence of macrophages in the Tissue zone.
Although initially there is a fixed number of
macrophages, when the parasites are introduced,
infecting and destroying cells, more macrophages
are recruited to the site of infection. The amount
depends on how the infection spreads through
the Tissue zone and not necessarily on existent
amount of T. cruzi.
A Multiagent-based Simulation of the Infection of the Macrophage by Trypanosoma Cruzi in the Acute Phase of Chagas’ Disease: Influence
of the Initial Inoculum and Protozoan Escape Factor
33
Number of T. cruzi at the end:Amount of T. cruzi
in Tissue zone at the end of 1825 ticks.
Number of Macrophage at the end: Amount of
macrophages in the Tissue zone at the end of 1825
ticks.
4 RESULTS
In order to validate the modeling and to study the in-
teractions between T. cruzi infection and the immune
system, two sequences of simulations varying the es-
cape factor were performed, one with initial number
of T. cruzi equal to 300 and another with this pa-
rameter equal to 3000. These values were based on
(Borges, 2012; Coura and Borges-Pereira, 2010) re-
view to investigate the variation of inoculums in vivo.
The values chosen for the escape factor were based on
its relationship with the number of T. cruzi at the end.
Each experiment has an increment of 0.1 or 0.01 in the
value of the escape factor of the previous experiment.
The range of 0.01 was used when the results showed
a random pattern. On the other hand, the value of 0.1
was used when the experiments behaved with a more
predictable pattern. Values above 1% were not simu-
lated because it has been observed that in such cases,
it is difficult for the macrophages to restrain the infec-
tion.
The purpose was to investigate three important is-
sues in the study of CD: Whether there is a situation of
aggression and inflammation in response to the para-
site, as described in the literature by (Siqueira-Batista
et al., 2007; Rochitte et al., 2007); Whether the ini-
tial number of T. cruzi influences the outcome of in-
fection, as demonstrated experimentally in (Borges,
2012); and what is the relationship between the es-
cape factor and the total elimination of T. cruzi. Ta-
bles 1 and 2 show the results obtained after 104 simu-
lations, 52 for initial number of T. cruzi equal to 300,
and the remaining for initial number of T. cruzi equal
to 3000, respectively.
With inoculum of 300 T. cruzi agents, the
macrophages eliminated all of them in 20 simulations
with maximum escape factor up to 0.52%. For the in-
oculum of 3000 in 52 simulations, this phenomenon
occurred 13 times with maximum escape factor of
0.37%.
In simulations with inoculum of 3000 T. cruzi
agents, the lowest escape factor where the number of
T. cruzi at the end did not reach zero was 0.24% and
the highest escape factor where the number of T. cruzi
at the end zeroed was 0.37%. This is a narrow range
(around 0.13%) where the outcome of the infection
varies.
Table 1: Simulations with inoculum equal to 300.
escape Tick max. T. cruzi Mφ max.
factor(%) T. cruzi=0 T. cruzi tick 1825 tick 1825 Mφ
0.00 248 1001 0 364 369
0.10 166 1020 0 438 440
0.20 208 1000 0 348 352
0.21 259 3779 0 396 403
0.22 208 830 0 340 342
0.23 123 813 0 353 354
0.24 345 1000 0 343 348
0.25 316 1115 0 374 376
0.26 195 804 0 336 337
0.27 - 6526 6 373 399
0.28 121 799 0 349 349
0.29 419 1269 0 330 339
0.30 - 23064 7 1395 1396
0.31 - 4330 14 854 854
0.32 279 1555 0 426 430
0.33 1205 1221 0 475 477
0.34 1766 2085 0 489 494
0.35 322 832 0 424 430
0.36 - 25272 3655 2230 2248
0.37 1757 1204 0 443 446
0.38 - 33981 8406 1526 1537
0.39 - 95795 30648 1086 1479
0.40 - 34447 3782 1674 1674
0.41 723 1556 0 419 426
0.42 - 124231 38481 3078 3097
0.43 - 88893 26064 1020 1471
0.44 144 798 0 308 310
0.45 - 95426 11837 1267 1267
0.46 425 868 0 287 300
0.47 - 20003 3433 1804 1804
0.48 - 85131 17386 1092 1515
0.49 - 11000 427 1221 1236
0.50 - 122758 60693 2659 3030
0.51 - 61358 15387 1928 2069
0.52 358 799 0 310 311
0.53 - 105997 22354 2842 3627
0.54 - 132487 30863 1870 2698
0.55 - 186631 101125 4574 5874
0.56 - 91251 39787 2590 2590
0.57 - 132072 47724 4837 5743
0.58 - 60626 18765 1128 1194
0.59 - 67964 24631 746 1531
0.60 - 6432 1425 823 831
0.61 - 2206 47 602 608
0.62 - 149570 90749 2478 3332
0.63 - 59905 27439 3927 3966
0.64 - 117307 15611 678 2239
0.65 - 9386 3829 1474 1474
0.70 - 174423 32523 3182 3589
0.80 - 107358 46055 2627 2900
0.90 - 147222 68838 3034 3806
1.00 - 148243 133889 4332 5149
There is a greater tendency showing that the
largest the escape factor the largest the probability of
T. cruzi to multiply leading to an increasing number
of parasites in the environment, as observed in chart 3.
This occurred in the simulations of this study though
ICEIS 2016 - 18th International Conference on Enterprise Information Systems
34
Table 2: Simulations with inoculum equal to 3000.
escape Tick max. T. cruzi Mφ max.
factor(%) T. cruzi=0 T. cruzi tick 1825 tick 1825 Mφ
0.00 179 5502 0 425 425
0.10 211 3869 0 343 343
0.20 471 4019 0 491 507
0.21 1168 4516 0 406 419
0.22 959 5830 0 450 463
0.23 1247 4030 0 319 341
0.24 - 4290 3 518 523
0.25 413 3640 0 386 391
0.26 955 3803 0 525 525
0.27 1174 4601 0 630 650
0.28 1477 4414 0 615 620
0.29 - 11574 10 726 746
0.30 - 18928 1 981 981
0.31 1815 5269 0 629 636
0.32 - 3574 1 498 499
0.33 - 5183 4 607 638
0.34 1624 4503 0 606 624
0.35 - 69472 5421 1345 1424
0.36 - 36352 283 2794 2845
0.37 648 3595 0 408 409
0.38 - 114117 6382 733 1736
0.39 - 123258 64354 2442 3649
0.40 - 9987 334 1917 1917
0.41 - 87750 7292 2533 2533
0.42 - 127525 10635 1755 1755
0.43 - 120302 10790 1148 1696
0.44 - 102800 43675 3173 3471
0.45 - 93822 20017 1319 1514
0.46 - 137813 14931 2579 2591
0.47 - 87098 15687 1420 1585
0.48 - 7586 840 1712 1735
0.49 - 50139 5537 3423 3423
0.50 - 86034 24686 1803 2327
0.51 - 145739 12674 1523 2187
0.52 - 184655 96294 5319 6856
0.53 - 25522 9219 1627 1651
0.54 - 181202 68261 6731 6731
0.55 - 58513 22986 2233 2511
0.56 - 62830 15419 890 1115
0.57 - 156409 60189 2914 4690
0.58 - 127694 25559 930 2776
0.59 - 188053 58005 1406 2919
0.60 - 117632 17953 1412 2113
0.61 - 87890 13474 787 1441
0.62 - 85169 85169 6838 6875
0.63 - 106039 45118 1980 2198
0.64 - 129197 79346 5794 5911
0.65 - 91751 28611 2663 2663
0.70 - 113388 51583 4487 4781
0.80 - 130328 72959 4799 5046
0.90 - 158986 88700 2710 4788
1.00 - 109838 63082 3380 3427
this event did not happen uniformly (charts 4 and 5).
In the simulations with inoculum of 3000, the low-
est maximum number of T. cruzi was 3574 and it oc-
curred for an escape factor of 0.32%, while the high-
est number of T. cruzi was 188053 with an escape fac-
tor of 0.59%.
5 DISCUSSION
We observed that, in the presented simulations, the
results did not differ significantly from in vitro stud-
ies. In this work, the initial number of T. cruzi influ-
enced the outcome of infection. Experimental stud-
ies with different species of insects showed that the
number of parasites deposited at the site of infection
and that actively penetrate the host at the time of suc-
tion and defecation by these vectors is variable. After
the contact between infected triatomines and humans,
the outcome will depend on variables that control
the chances of infection (Coura and Borges-Pereira,
2010; Rassi Jr et al., 2012; Coura, 2013) Some stud-
ies report that variables such as the infection rate of
triatomines, the time between the sting and defeca-
tion, the number of evacuations and the quantity dur-
ing this time interval, the number of parasites elimi-
nated, the percentage of infecting forms and their ca-
pacity of penetration, and the intensity of the itching
during the sting are of great relevance to the occur-
rence of the infectious process (Coura and Borges-
Pereira, 2010). (Borges-Pereira et al., 1988) observed
that among eight species of triatomines infected with
T. cruzi, the mean number of parasites per evacuation
was 140. In their study, the authors found an average
of 232 parasites per evacuation among the P. megistus
species. In our simulation study, besides the use of
3000 agents, we also used 300 T. cruzi agents, which
is similar to the study of (Borges-Pereira et al., 1988).
The results of this study support the assumption
that the escape factor and the total elimination of T.
cruzi are closely linked, since variations of tenths and
even hundredths in escape factor had a significant ef-
fect on the outcome of the simulations of the T. cruzi
macrophage interaction in the acute phase of CD.
As might be expected, the secondary immune re-
sponse is very important to prevent infection by the
parasite that causes CD since in most simulations only
the macrophage trypanocidal action was not enough
to halt the progression and action of the parasite. Still,
even not sufficient, the macrophages were able, in
cases where the escape factor was very low, to elim-
inate parasites, and in other cases detain the spread
of infection by stabilizing the amount of parasites. It
shows the macrophage importance in modulating the
immune response to this disease.
As noted by (Possi et al., 2011), AutoSimmune
is still in development. However, the simulator has
been used in an others studies with promising results
(Da Silva et al., 2012; Bastos et al., 2013), which in-
A Multiagent-based Simulation of the Infection of the Macrophage by Trypanosoma Cruzi in the Acute Phase of Chagas’ Disease: Influence
of the Initial Inoculum and Protozoan Escape Factor
35
Figure 3: Relationship between escape factor and maximum number of T.cruzi/Number of T. cruzi at the end to inoculum
equal to 300.
Figure 4: Relationship between escape factor and maximum
number of T. cruzi to inoculum equal to 300.
Figure 5: Relationship between escape factor and maximum
number of T. cruzi to inoculum equal to 3000.
creases the expectation that it will be possible to refine
our model, enabling the creation of a tool that sup-
ports hypothesis testing in biology and indicates new
insights for in vitro and in vivo research.
6 CONCLUSION
We conducted a study of an in silico simulation of in-
teractions between T. cruzi parasites and macrophage
cells, in the AutoSimmune. We did not find similar
studies in a wide search in PubMed about in silico
simulation of this interaction. The study presented in
this paper showed that the in silico model of the inter-
action between Trypanosoma cruzi and macrophage
is a plausible approximation to what is described in
the medical literature (Siqueira-Batista et al., 2007).
We could observe that the initial number of T. cruzi
parasites influences in the outcome of infection, as
demonstrated experimentally, and found a relation-
ship between the escape factor and the total elimina-
tion of T. cruzi.
Observing the results for the number of T. cruzi
equal to 300 and 3000, we noticed the importance of
the quantity of inoculated strains on the outcome of
Chagas’ disease. As in (Borges, 2012), it was ob-
served that the tissue parasite load is directly related
to the inoculum used for the infection.
In addition to the results that support the hypoth-
esis about the quantity of inoculated strains on the
outcome of Chagas’ disease and relationship between
the escape factor and the total elimination of T. cruzi,
other important contributions of this research are a
computational model for the acute stage of Chagas’
disease and models of behavior for the macrophage
and T. cruzi.
ICEIS 2016 - 18th International Conference on Enterprise Information Systems
36
As future work, we intend to compare results of
analysis in silico,in vitro and in vivo as well, to obtain
a consensus measure of the value of the escape fac-
tor, and its relation to the activation of the secondary
immune response, due to its importance (Siqueira-
Batista et al., 2007).
ACKNOWLEDGEMENTS
This research is supported in part by the funding agen-
cies FAPEMIG, CNPq, and CAPES. The author Al-
cione de Paiva Oliveira receives a grant from CAPES,
process n.0449/15-6.
REFERENCES
Alcantara, A. and Brener, Z. (1980). Trypanosoma cruzi:
Role of macrophage membrane components in the
phagocytosis of bloodstream forms. Experimental
parasitology, 50(1):1–6.
Bastos, C. A., Oliveira, A. d. P., Gomes, A. P.,
de Ara´ujo Possi, M., Santana, L. A., Cerqueira, F. R.,
and Siqueira-Batista, R. (2013). Simulac¸˜ao do sis-
tema imunol´ogico por meio de sistemas multiagentes:
um estudo da resposta imune na glomerulonefrite p´os-
infecciosa (gnpe) por streptococcus pyogenes. In XIII
Workshop de Inform´atica M´edica, pages 1093–110.
Borges, D. C. (2012). Diferentes in´oculos parasit´arios
determinam a modulac¸˜ao da resposta imune e o
desenvolvimento da infecc¸˜ao experimental por Try-
panosoma cruzi. Master’s thesis, Universidade Fed-
eral do Triˆangulo Mineiro, Uberaba - MG.
Borges-Pereira, J., Pessoa, I., and Coura, J. (1988).
Observac¸ ˜oes sobre as dejec¸˜oes e o n´umero de t. cruzi
eliminados por diferentes esp´ecies de triatom´ıneos
durante a alimentac¸˜ao. Mem Inst Oswaldo Cruz,
83(Suppl I):7.
Celentano, A. M. and Gonz´alez Cappa, S. M. (1993). In
vivo macrophage function in experimental infection
with trypanosoma cruzi subpopulations. Acta tropica,
55(3):171–180.
Coura, J. R. (2013). Chagas disease: control, elimination
and eradication. is it possible? Mem´orias do Instituto
Oswaldo Cruz, 108(8):962–967.
Coura, J. R., Anunziato, N., and Willcox, H. P. (1983). Mor-
bidade da doenc¸a de chagas. i–estudo de casos proce-
dentes de v´arios estados do brasil, observados no rio
de janeiro. Mem Inst Oswaldo Cruz, 78(3):363–372.
Coura, J. R. and Borges-Pereira, J. (2010). Chagas disease:
100 years after its discovery. a systemic review. Acta
tropica, 115(1):5–13.
Da Silva, C., De Paiva Oliveira, A., De Araujo Possi, M.,
Cerqueira, F., Gomes, A., Santana, L., and Siqueira-
Batista, R. (2012). Immune system simulation: Mod-
eling the mast cell. In Bioinformatics and Biomedicine
(BIBM), 2012 IEEE International Conference on,
pages 1–4.
Devillers, H., Lobry, J. R., and Menu, F. (2008). An
agent-based model for predicting the prevalence of
trypanosoma cruzi i and ii in their host and vector pop-
ulations. Journal of theoretical biology, 255(3):307–
315.
Dewri, R. and Chakraborti, N. (2005). Simulating recrys-
tallization through cellular automata and genetic algo-
rithms. Modelling and Simulation in Materials Sci-
ence and Engineering, 13(2):173.
Dias, J. C. P. (1982). Doenc¸a de Chagas em Bambu´ı, Mi-
nas Gerais, Brasil: estudo cl´ınico-epidemiol´ogico a
partir da fase aguda, entre 1940 e 1982. PhD thesis,
Universidade Federal de Minas Gerais. Faculdade de
Medicina.
Floreano, D. and Mattiussi, C. (2008). Bio-Inspired Arti-
ficial Intelligence: Theories, Methods, and Technolo-
gies. Intelligent Robotics and Autonomous Agents Se-
ries. Mit Press, Philadelphia.
Folcik, V., An, G., and Orosz, C. (2007). The Basic Immune
Simulator: An agent-based model to study the interac-
tions between innate and adaptive immunity. Theoret-
ical Biology and Medical Modelling, 4(1):39+.
Galv˜ao, V., Miranda, and Vivas, J. G. (2010). A three-
dimensional multi-agent-based model for the evolu-
tion of chagas disease. Biosystems, 100(3):225–230.
Galv˜ao, V., Miranda, J. G. V., and Ribeiro-dos San-
tos, R. (2008). Development of a two-dimensional
agent-based model for chronic chagasic cardiomyopa-
thy after stem cell transplantation. Bioinformatics,
24(18):2051–2056.
hua Li, X., xuan Wang, Z., yang Lu, T., and jiu Che, X.
(2009). Modelling immune system: Principles, mod-
els, analysis and perspectives. Journal of Bionic En-
gineering, 6(1):77 – 85.
Martins, A. V., Gomes, A. P., Gomes de Mendonc¸a,
E., Lopes Rangel Fietto, J., Santana, L. A.,
de Almeida Oliveira, M. G., Geller, M., de Fre-
itas Santos, R., Roger Vitorino, R., and Siqueira-
Batista, R. (2012). Biology of trypanosoma cruzi: An
update. Infectio, 16(1):45–58.
Possi, M. A., P., O. A., C., C., and R., C. F. (2011). An in-
silico immune system model for investigating human
autoimmune diseases. In XXXVII Conferencia Lati-
noamericana en Inform´atica CLEI.
Rassi Jr, A., Rassi, A., and Marcondes de Rezende,
J. (2012). American trypanosomiasis (chagas dis-
ease). Infectious disease clinics of North America,
26(2):275–291.
Rey, L. (2008). Parasitologia: parasitos e doenc¸as para-
sit´arias do homem nos tr´opicos ocidentais. Guanabara
Koogan.
Rochitte, C. E., Nacif, M. S., de Oliveira J´unior, A. C.,
Siqueira-Batista, R., Marchiori, E., Uellendahl, M.,
and de Lourdes Higuchi, M. (2007). Cardiac mag-
netic resonance in chagas’ disease. Artificial organs,
31(4):259–267.
Rodrigues, W. F., Miguel, C. B., Chica, J. E. L., and
Napimoga, M. H. (2010). 15d-pgj2 modulates acute
A Multiagent-based Simulation of the Infection of the Macrophage by Trypanosoma Cruzi in the Acute Phase of Chagas’ Disease: Influence
of the Initial Inoculum and Protozoan Escape Factor
37
immune responses to trypanosoma cruzi infection.
Mem´orias do Instituto Oswaldo Cruz, 105(2):137–
143.
Rogers, C., Nguy, K. N. T., and Amajor, O. A. (2010). Daily
behavior of trypanosoma cruzi hosts and vectors in
texas: An agent-based modeling approach in netlogo.
Russo, M. and Starobinas, N. (1991). Macrophage acti-
vation and resistance to trypanosoma cruzi infection.
Research in immunology, 142(2):144–146.
Siqueira-Batista, R. and Geller, M. (2008). Role of tcd4+
and tcd8+ lymphocytes in the immune response to try-
panosoma cruzi. RBM rev. bras. med, 65(7).
Siqueira-Batista, R., Geller, M., Martins, A., Bastos, O.,
Siqueira-Batista, R., Gomes, A., Corrˆea, A., and
Geller, M. (2007). Trypanosoma cruzi. Mol´estia de
Chagas.
ICEIS 2016 - 18th International Conference on Enterprise Information Systems
38