Processes 2019, 7,37
early entry of T-cells effectively eliminated the tumor and was dependent on CD137 (a co-stimulatory
protein that helps in tumor rejection [121]) expression in tumor vasculature.
Oncolytic virus therapy is a strategy that utilizes viral infection to kill cancer cells, but not normal
cells, with the potential of enhancing T-cell recruitment to the tumor and increasing their access to
cancer cells. Several computational models have examined the conditions of success for this type
of therapeutic in silico [122]. Walker et al. developed an agent-based model of pancreatic tumors
to study the synergy between chimeric antigen receptor (CAR) T-cell therapy and oncolytic virus
therapy [123]. CAR T-cell therapy is one type of adoptive cell transfer treatment involving genetically
engineered T-cells specifically targeting cancer cells, and has been the subject of several computational
models [124]. The agent-based model recapitulates treatment mechanisms including cancer specific
CAR T-cell recruitment to the tumor site via vasculature and the injection and spread of oncolytic virus.
Rohrs et al. demonstrated the ability of the model to track the dynamics of cancer cells and stromal
cells in space in the presence of the treatment combinations; optimization of the combination therapy
requires more accurate calibration [124].
Immune checkpoint inhibitors are used in cancer immunotherapy that enhances anti-tumor
immune response by targeting cancer immune evasion mechanisms. In many cancer types, tumor
neoantigens are sufficiently immunogenic to promote the expansion of antitumor immune cells [125];
however, these immune cells are not functional due to the inhibitory signals from molecules adaptively
induced during cancer development [24,126]. Among them, one of the most prominent mechanisms
is PD-1/PD-L1 interaction, where T-cells are suppressed through PD-1 signaling upon contact with
induced PD-L1 in the tumor microenvironment. Gong et al. developed an ABM of tumor-immune
interaction in 3D to study the spatio-temporal dynamics of cancer cells and cytotoxic T-cells [127]. In
this study, the inhibitor to the checkpoint molecules were modeled as a factor which modulates the
parameter governing the suppression of tumor specific T-cells by PD-L1+ cancer cells. They found
that patient responsiveness to such therapy could be associated with the level of mutational burden of
the cancer and antigen strength among patients. They also found that tumor growth is insensitive to
the vascular density of the tumor core. From these results, a scoring method was proposed to predict
anti-PDL1 treatment efficacy in patients.
3.5. Models Focusing on Tumor-Enhancing Immune Cells
While the immune system has evolved to kill off tumor cells, there are many ways in which
cancer cells can avoid immune detection. In addition, there is mounting evidence that immune cells
can stimulate tumor growth under certain conditions. Several agent-based models have focused on
understanding the tumor-enhancing contributions of the immune system. Enderling and colleagues
explored the interactions between tumor cell death and the immune system using a cellular automata
model focused on the interplay between cancer stem cells and the immune system [128]. They showed
that immune system-induced tumor cell death led to stem cell selection, and thus, more aggressive
tumors [128,129]. In this model, even though immune cells effectively killed off tumor cells, they also
affected progenitor cells. This resulted in the creation of a space for cancer stem cells to proliferate and
produce more cancer stem cells. This ultimately resulted in a larger stem cell population and a more
aggressive tumor.
Several studies have specifically focused on the tumor-enhancing contribution of immune cells,
such as tumor-associated macrophages (TAM). Macrophages are one of the most abundant immune
cells found in tumors, but their population is heterogeneous [130]. M1-type macrophages have
been shown to be tumor inhibiting, whereas M2-type macrophages have been shown to be tumor
enhancing [26]. One model looked at the transition from the M1 to M2 macrophage phenotype on
tumor growth and then predicted targeted therapies [131]. Knútsdóttir et al. used a hybrid model
to investigate epidermal growth factor (EGF) and macrophage colony-stimulating factor 1 (CSF-1)
signaling between macrophages and cancer cells during macrophage aggregation [132]. They found
that CSF-1/CSFR1 autocrine signaling affects the ratio of tumor cells to macrophages during tumor
49
early entry of T-cells effectively eliminated the tumor and was dependent on CD137 (a co-stimulatory
protein that helps in tumor rejection [121]) expression in tumor vasculature.
Oncolytic virus therapy is a strategy that utilizes viral infection to kill cancer cells, but not normal
cells, with the potential of enhancing T-cell recruitment to the tumor and increasing their access to
cancer cells. Several computational models have examined the conditions of success for this type
of therapeutic in silico [122]. Walker et al. developed an agent-based model of pancreatic tumors
to study the synergy between chimeric antigen receptor (CAR) T-cell therapy and oncolytic virus
therapy [123]. CAR T-cell therapy is one type of adoptive cell transfer treatment involving genetically
engineered T-cells specifically targeting cancer cells, and has been the subject of several computational
models [124]. The agent-based model recapitulates treatment mechanisms including cancer specific
CAR T-cell recruitment to the tumor site via vasculature and the injection and spread of oncolytic virus.
Rohrs et al. demonstrated the ability of the model to track the dynamics of cancer cells and stromal
cells in space in the presence of the treatment combinations; optimization of the combination therapy
requires more accurate calibration [124].
Immune checkpoint inhibitors are used in cancer immunotherapy that enhances anti-tumor
immune response by targeting cancer immune evasion mechanisms. In many cancer types, tumor
neoantigens are sufficiently immunogenic to promote the expansion of antitumor immune cells [125];
however, these immune cells are not functional due to the inhibitory signals from molecules adaptively
induced during cancer development [24,126]. Among them, one of the most prominent mechanisms
is PD-1/PD-L1 interaction, where T-cells are suppressed through PD-1 signaling upon contact with
induced PD-L1 in the tumor microenvironment. Gong et al. developed an ABM of tumor-immune
interaction in 3D to study the spatio-temporal dynamics of cancer cells and cytotoxic T-cells [127]. In
this study, the inhibitor to the checkpoint molecules were modeled as a factor which modulates the
parameter governing the suppression of tumor specific T-cells by PD-L1+ cancer cells. They found
that patient responsiveness to such therapy could be associated with the level of mutational burden of
the cancer and antigen strength among patients. They also found that tumor growth is insensitive to
the vascular density of the tumor core. From these results, a scoring method was proposed to predict
anti-PDL1 treatment efficacy in patients.
3.5. Models Focusing on Tumor-Enhancing Immune Cells
While the immune system has evolved to kill off tumor cells, there are many ways in which
cancer cells can avoid immune detection. In addition, there is mounting evidence that immune cells
can stimulate tumor growth under certain conditions. Several agent-based models have focused on
understanding the tumor-enhancing contributions of the immune system. Enderling and colleagues
explored the interactions between tumor cell death and the immune system using a cellular automata
model focused on the interplay between cancer stem cells and the immune system [128]. They showed
that immune system-induced tumor cell death led to stem cell selection, and thus, more aggressive
tumors [128,129]. In this model, even though immune cells effectively killed off tumor cells, they also
affected progenitor cells. This resulted in the creation of a space for cancer stem cells to proliferate and
produce more cancer stem cells. This ultimately resulted in a larger stem cell population and a more
aggressive tumor.
Several studies have specifically focused on the tumor-enhancing contribution of immune cells,
such as tumor-associated macrophages (TAM). Macrophages are one of the most abundant immune
cells found in tumors, but their population is heterogeneous [130]. M1-type macrophages have
been shown to be tumor inhibiting, whereas M2-type macrophages have been shown to be tumor
enhancing [26]. One model looked at the transition from the M1 to M2 macrophage phenotype on
tumor growth and then predicted targeted therapies [131]. Knútsdóttir et al. used a hybrid model
to investigate epidermal growth factor (EGF) and macrophage colony-stimulating factor 1 (CSF-1)
signaling between macrophages and cancer cells during macrophage aggregation [132]. They found
that CSF-1/CSFR1 autocrine signaling affects the ratio of tumor cells to macrophages during tumor
49