Last week’s Nature included a brief but interesting review on some recent research on the role of the microenvironment in cancer progression.
The article starts with the premise (which I certainly share) that a tumour is a collection of different types of cells, some of them tumour cells, that act together towards the progression of the disease. Among these cells there are stromal cells (fibroblasts, endothelial cells, immune cells) whose interactions with abnormal epithelial (tumour) cells can potentially determine tumour progression.
One of the ways in which this interaction is being studied is through gene expression profiles. Although the precise mechanisms might be unknown, certain patters of gene expression can be associated with specific tumour outcomes. Also there are a number of pathways in the epithelial cells that are related to the response to signalling from stromal cells, whose disruption can lead to tumour invasion of the basement membrane.
Studies like the ones covered in this review provide an alternative view of that in which tumour progression is exclusively characterised by a number of genetic mutations. Not only the path of the tumour progression depends on interactions with the microenvironment but the diversity of this microenvironment in general and some stromal cell types in particular could potentially have a big impact on that path. A key element could be the coevolution between the tumour cells and those in the stroma.
Saying that the stroma influences tumour evolution by contributing to select some tumour phenotypes over others is unlikely to be a controversial statement. But this influence could also work the other way around and it is an intriguing hypothesis: if there is evolution in some key stromal cell types like fibroblasts, what kind of evolution would that be? and if it is meaningful, how tumour evolution influences stromal evolution and viceversa?.
The notion of stromal evolution demands some explanation. At the end of the day evolution requires a number of ingredients that most people would not expect to find in healthy mammalian cells (asides from those of the immune system). Specifically, asides from the already mentioned diversity, evolution needs that some of the phenotypes have higher fitness than others. A high degree of fitness in a phenotype should lead to more individuals of the same phenotype. This is normally acomplished by means of differential reproduction by which fitter individuals, in the long run, tend to produce more offspring than the less fit ones.
Not wanting to turn this into a philosophical post, I will just mention that I was recently discussing with friends about stromal evolution when I was suggested that, maybe, this differential reproduction with inheritance is not that indispensable as I once thought. At least in the context of somatic evolution in mammalian cells where there is a limit to how much time evolution has to get things changed and where there cells are endowed with certain phenotypic plasticity (mainly constrained by environmental signals). The idea then would be that somatic evolution of fibroblasts proceeds as some of them change their phenotypic profile (probably in response to signals sent by tumour cells) and influence other fibroblasts to adopt theirs. This, in return, would alter the microenvironment in such a way as to affect the tumour and also the fibroblasts themselves, maybe triggering the emergence of a different phenotypic profile.
David, I believe you are describing the Baldwin Effect
I feel that this aspect of evolutionary dynamics is underappreciated in general, so it’s not surprising that nobody considered it in somatic evolution before you did.
Rafe, thanks for your comment.
It took me a little while to realise what you were talking about with the Baldwin effect. It is only recently that I’ve learnt about it from a friend but I have to admit I was not thinking about it when I wrote this post. I agree that in some (probably not so rate) circumstances evolution will select for individuals that are capable of learning and that this effect is potentially the explanation of the plasticity of many mammalian cells like the fibroblasts in the stroma.
It would an interesting hypothesis(although I am not sure that this would be a likely event) if someone could prove that tumour cells evolve a higher capacity of learning as the tumour progresses.
I think “learning” is a loaded phrase and perhaps it’s better to think of the Baldwin Effect as applying to evolutionary systems where the agents have a enough phenotypic plasticity such that in their lifetimes an adaptive behavior is likely to be stumbled upon. In such cases, the genetic predisposition to acquire that behavior is selected for in the standard fashion, up to the point where we would be willing to say that that the behavior itself has been selected for. Of course, we always ignore the fact that nothing is truly selected for directly, but rather is always mediated via the process of ontogeny.
To illustrated why learning is a loaded concept, I recently read that even the simplest single cells do in fact adapt via their genetic (and epi-genetic, proteomic, etc) networks in their own lifetimes. There’s nothing particularly special about neural networks in regards to learning (aka adaptation).
Like punctuated equilibrium, I suspect that the Baldwin Effect is very general and occurs in all biological evolving populations (since ontogeny = plasticity), though we are quick to dismiss its effect when there are simpler “primary selection” dynamics to point to.