Crosstalk among cancer cells and between cancer cells and the tumour microenvironment plays a crucial role in tumour development and progression (1). This communication is a two-way process (inside-out and outside-in) and can be realized through both direct contact and indirectly. (2,3,4,5,6)
The importance of signaling pathways
Cells continuously release billions of signalling molecules. In healthy tissue they are (mostly) well-orchestrated, coordinating division and metabolic state of cells, working for the benefit of an organism.
However, disturbing signalling pathways can be deadly. For instance, one of the most important signalling pathways is so called hedgehog pathway. It is mostly active during embryonic development and childhood where it regulates cell differentiation and body development. However, its reactivation is involved in onset of many cancer types and accounts for approximately of 25% of cancer deaths (7).
Disrupting cancer-cell communication with drugs
Given the importance of cell communication for cancer survival, and the vast number of involved molecules and regulators, this domain is a rich source of potential anti-cancer targets. Novel drugs (e.g., Trastuzumab, Bevacizumab, Erlotinib) are designed to disrupt cell communication by either targeting signalling molecules, receptors or signalling pathways inside cancer cells. Numerous potential targets have been proposed (e.g., 8,9). However, while drug-induced communication disruption is focused on very specific targets, the open question is could we be more strategic in affecting cross-talk between cells? For example, what would happen if we precisely identify communication hubs and their interrelations and then plan how to attack them; or instead of shutting down elements of communication pathways could we jam the communication, thus overpowering the original signals?
Conceptualizing tumour communication channels using Information Theory
We do not yet have answers to those questions and to reach them we need a slightly different perspective. Application of the information theory would be the first step (for a great introductory overview for biologists, with application examples in oncology see 10).
Information theory, developed by Shannon and Weaver, precisely formalizes all components of a generalized communication system. The crux is a communication channel where an information source encodes the message which then goes through noisy channel until it reaches intended destination (receiver) where is decoded. Equipped with such conceptualization, we can then calculate channel capacity (limits of reliable transmission of a signal) and mutual information (the amount of information obtained about one variable by measuring the other variable), and eventually determine to what extent noise can interfere with communication depending on the properties of source, channel and receiver.
Could we analyze the interwoven tumour communication network as a whole?
However, that analysis would apply to a single channel type, and no matter how important this communication path is, it is only a part of interwoven, communication network composed of multiple and highly redundant channels and their “protocols”.
Such complexity leads to robustness, i.e. the ability of cells to maintain their functioning under a wide range of random perturbations (11). Much of the work has been done to estimate the robustness (and at the same time vulnerability) of biological systems using generalized approaches of systems theory and complex network analysis (12,13,14).
Their findings point out that the single point of attack on a robust and distributed system can hardly lead to the desired results (e.g. extermination of the entire tumour). Instead, they emphasize the need for a much deeper understanding of the interplay of cell-level and tumour-level dynamics in order to identify means of controlling such highly distributed nonlinear network systems.
Unfortunately, so far, those findings have been underutilized in the mainstream of clinical cancer research.
One possible reason is that mathematical results are often too general without outlining a clear path towards concrete application for a concrete clinical situation (tumour type, tumour stage, patient condition, etc). On the other hand – there are numerous ultra-specific mathematical models that overanalyse narrow specific cases: they fix up some convenient set of parameters, run simulations and offer analysis that again does not pay particular attention to varieties that stems from tumour types, tumour stages or patient conditions.
The way forward
In short, it seems that at least a part of the problem is a rift between mathematically and empirically minded research who do not fully understand each other’s needs. One possible way forward is to merge highly abstract level of analysis (information theory, complex networks, etc.) with specific cell-level data. For instance, recently Conforte et al. (15) analyzed protein-protein interaction data and RNA-sequencing data, using mathematical tools of information theory to identify the number of metabolic regulatory targets that should be inhibited so that therapy will be effective. However, even then, to better account for the realities of biological world, modelling findings should be regarded only as a part of probability distribution, within a phase space of tumour dynamics. So, we need more systemic investigation of ensembles of mathematical models, instead of one-fits-all solutions.
To conclude, tumour is a complex system and, in order to thrive, its constituting parts need to communicate with each other and with the environment.
Chances to knock it down for good by attacking a single communication channel, are miniscule. We need a more strategic perspective. However, to develop such perspective, an important update in approaching mathematical modelling of cancer is needed. Again, cooperation, tight collaboration and mutual understanding are the way forward.