A 3D Computer Model to Map Modifications in Cellular Metabolism into Different Tumor Phenotypes

In this work we propose an approach for studying tumor
development as an iterative process that searches for the
minimum set of modifications in cell metabolism that would lead
healthy tissue to develop an invasive tumor.

There are many models in the literature describing putative
steps necessary for a homogenous population of healthy
epithelial cells to generate an invasive tumor, however these
studies do not account for the cellular metabolic modifications
that drive and support cancer development and progression
(hyperplasia, high aerobic glycolysis and acid resistance).

We propose an integrated process for discovering the minimal
sets of modifications in cellular processes that lead tumors to
become invasive (growth of tumoral tissue, invasion of normal
epithelial cells and basement membrane).

The development of tumors such as Ductal Carcinoma In Situ
(DCIS) is a complex process where a genetically heterogeneous
population of tumoral cells is submitted to selective pressure
in a dynamic environment. Here, we represent DCIS as a 3D
computer model based on a tubular structure of 40 x 40 cells
dimension (diameter x length), composed of endothelial cells,
basement membrane and a layer of normal epithelial cells
(Figure 1).

One region of this layer of healthy cells is chosen to be
mutation prone, meaning that these cells have a higher than
normal mutation rate due to environmental changes such as
chronic inflammation or a mutation in DNA repair mechanisms.
This will lead to appearance of different phenotypes - a
process that will ultimately lead to evolution of a malignant
population.

Each individual cell was built with algorithms to perform
duplication or apoptosis, and three metabolic pathways:
Glycolysis, TCA cycle, and pentose phosphate cycle. These cells
were therefore programmed to duplicate, die, metabolize glucose
and oxygen and generate ATP and excrete lactate. In order to
account for dynamics in different time scales, the model is
simulated in three levels: the first one being the
reaction-diffusion of chemical species, the second is the
metabolism of each cell and the last level is the cell
duplication.

Within this model, one simulation is performed for each set of
phenotypic modifications in tumoral cells and the set of
modified cells that have successfully achieved invasiveness are
selected. For each of these sub-populations the search moves
one step further and details the reactions composing that
pathway in search for the minimum set of enzymes that might be
responsible the modifications in the phenotype previously
observed. This drill-down process proceeds until reaching the
genes responsible for the synthesis of the proteins that
catalyze or regulate these metabolic reactions. Finally, the
candidate genes found are compared to oncogenes known in
literature and assessed experimentally for validation of the
model.

This approach may improve understanding the intricate factors
behind the development of tumors as well as test for treatments
such as modification of acidity, oxygen and other substrates in
the tumor microenvironment.

On the left, a snapshot of the simulation of a tumor developing in an epithelial duct (3D view on top right) and a 2D view of a transversal slice of the model showing the PH, oxygen and glucose concentration gradients. In the 3D view, red dots are blood vessels, gray represent the basement membrane, pink are healthy epithelial cells and the others are tumoral cells with different phenotypes. On the right, histology of DCIS shows the expansion of tumor cells within the duct (from Gatenby and Gilles, 2004).