Complex networks

The architecture of tissues

Connections with neighbors help to define the function of cells in the organism

Leo RamosFollowing the boom in studies about the genome and proteins, researchers in the field of the biological sciences have now turned their attention back to the cellular differentiation processes involved in the development of each individual, from the embryo stage to the adult phase. One example is the work of the electronic engineer Luciano da Fontoura Costa, a professor at São Carlos Physics Institute, of the University of São Paulo (IFSC-USP).

He developed and implemented computational methods to analyze images of cells in the epithelium, the tissue that internally or externally covers the organs. The objective was to examine the connections between the cells and to outline the network of contacts of each one of them, turning this information into graphs – a type of diagram, represented as a set of points (nodes or vertices) connected by straight lines.

This work led to an article recently published in the scientific journal Nature Communications. In it, the researchers show how to identify, more precisely and with greater sensitivity, the start of the specialization of the cells. According to Costa, this study is the result of years of collaboration between him and the researchers Madan Babu and Luis Escudero, both from the Cambridge Molecular Biology Laboratory in England. “The main objective of this research was to investigate how the epithelium is organized in a broader, more systematic manner, not only using measures of the shape of each cell, but also the network of contacts between them,” explains Costa.

He carried out the geometric characterization of each cell registered in microscopic images of the epithelium of wings and eyes of chicken embryos and of the larvae of drosophila, generally referred to as fruit flies, obtained by his colleagues at Cambridge. Costa also put together the network of contacts between the cells and carried out multivariate analysis of the data, which is a statistical method that takes into account more than one random variable at the same time and which is used to model nature, making it possible, among other things, to categorize data, to test hypotheses and to look for patterns.

For each type of epithelium, images were collected from various individuals and, for each image, a characteristics vector was generated, consisting of measurements such as the means and standard deviations of the area of a cell viewed under the microscope, the number of arches connected to a single node in the network, the degree of interconnection between the neighbors of a single node, and the average number of neighbors that the neighbors of a single node have. “This made it possible to carry out a more comprehensive comparison of the epitheliums in various stages of development of different tissues, organs and species, in addition to the natural variation in the organization of this tissue from one individual to the next,” he explains.

The researchers did this using an approach that they call geometric and network representation of epithelial organization (the abbreviation in English is GNEO). By using this strategy, they were able to examine the epithelium’s organization taking into account the cells’ contact patterns. GNEO also enables one to quantify differences between epitheliums of organisms and different tissues, even when the size and shape of the cells that make them up are visually impossible to tell apart. “We have shown that epitheliums of organs and different species have distinct, quantifiable structures,” says Costa.

Form, connection and function
The paper published in Nature Communications represents a step forward in relation to earlier studies, which merely took into account the geometry and size of the cells. The model now also includes data on the connection between the cells, which is linked to the specific functions that they perform. Costa explains that at the start of the formation of the embryo, all the cells are equal – having a hexagonal format – and that the connections between them resemble a hive. When the cells begin to change shape, becoming more elongated, spherical or cube-like, it is an indication that the process of cellular differentiation or specialization is getting underway. This is a critical moment for the formation of tissues and organs.

The problem is that nobody knows what triggers this change. “It’s not the DNA, because it’s the same for all the cells,” states Costa. “It’s just that some become kidneys, others become hearts and others neurons. What determines this is something that still needs to be discovered in biology.” Therefore, with the mapping of the genomes, it has become essential to understand how each gene is activated or inhibited during development. According to Costa, control of the gene expression takes place under the influence of a number of factors, internal and external to the individual, such as gravity, differences in the concentration of molecules and temperature, among others. In addition to this, the structures within the organism itself during development affect gene expression in the cells around it in a non-uniform way, for example, by means of diffusion of  the signaling molecules.

This article sheds some light on this issue. For instance, it made it possible to identify what the major factor was that contributed to this differentiation between epitheliums. “We discovered that the cell area contributes little to the distinction between structures of different species,” says Costa. “Actually, it’s the characteristics of the contact network that supply the more discriminatory measurements in these cases. We discovered that during cellular differentiation, the relationship between neighbors is more important than the shape.” In other words, the contact network plays an extremely important role in supplying the development characteristics of the tissues, thus providing new information on how they differentiate.

According to Costa, they have a good idea of the contribution of genetic mechanisms (external signs and regular paths of the associated gene) and of cellular mechanics (own patterns resulting from the rate of cellular division) to the formation of multi-cellular structures and to the development of epithelial architecture in various model systems. However, the means to characterize and to quantify the similarities and differences in the organization of this tissue in a more precise, comprehensive way were lacking. “One of the main objectives of our work was to help fill this gap, by furnishing such means,” he states. “Additionally, our approach can also be applied to other biological samples, such as the connections between nerve, muscle and tumor cells, as well as providing input for regenerative medicine.”

The Project
Modeling using networks (graphs) and pattern recognition techniques: structure, dynamics and applications (nº 2005/00587-5); Modality Thematic Project; Coordinator Roberto Marcondes Cesar Junior – IME/USP; Investment R$ 384,090.51 (FAPESP)

Scientific article
ESCUDERO, L. M. et al. Epithelial organization revealed by a network of cellular contacts. Nature Communications. 8 Nov. 2011.