And the prize for best paper title 2015 (so far) goes toooooooo……
Andrea Sottoriva, Christina Curtis and their coworkers for
A Big Bang model of human colorectal tumor growth.
Big Bang, Big Bang, … reminds me of (a) the prevailing cosmological model of how everything we know came about and (b) Sheldon Cooper. So maybe this is a genius paper that revolutionizes our basic understanding of cancer. It certainly is an eye-catching title.
What is the Big Bang model?
The Big Bang model is an alternative to the clonal expansion model, which is (has been?) the prevailing model of how cancer comes about.
|After the initial transformation, tumors grow as a single expansion of a mix of clones.||Potentially a sequence of clonal expansions throughout tumor development.|
|Absence of selective sweeps.||Development driven by selective sweeps.|
|Timing of an alteration is more important than any selective advantage it might bring.||Later clones can outcompete earlier clones.|
|Selection is unimportant (after initial transformation)||Selection favors some clones and kills others.|
|Clones do not compete.||Clones compete.|
|The relative clonal composition of the tumor is preserved.||The relative clonal composition of the tumor can vary dramatically over time.|
|Subclonal mutations appear early (right after transition to advanced tumor)||Subclonal mutations are the result of selection of de novo clones.|
|Likely to produce a spatial mix of clones (called variegation in the paper).||Likely to produce spatially distinct regions dominated by one clone.|
|Spatial relationships between cells do not necessarily recapitulate their clonal relationship.||Cells close to each other are likely to come from the same clone.|
|Uniformly high intra-tumor heterogeneity||Heterogeneity varies from region to region.|
|Diversity should be similar within different glands.||Very little within-gland diversity, high between-gland diversity.|
What evidence is there for the Big Bang model?
No one can watch a tumor grow and follow its genetic makeup in real time. The validation of the Big Bang model has thus to be indirect. Sottoriva et al focus on the key prediction of their model: the genomic ‘patterning’ of the tumor is different in a Big Bang model and in a clonal expansion model.
To describe the genomic patterning they analyze 349 individual glands from 15 colorectal tumors by SNP-array based copy number profiling, exome-sequencing, targeted deep sequencing, bulk-sequencing, the kitchen sink, ultra-deep single-molecule methylation tag sequencing and FISH. (It certainly reads to me like a paper that spent a long time in review and received many ‘helpful’ comments on what type of additional data would really be needed before it could be accepted.)
The first Results section showing evidence for variegation in single-gland copy number profiles already contains all the important elements of the story – the other data just corroborate this result. A typical piece of evidence for the Big Bang model are phylogenetic trees (many produced by MEDICC) that cluster together glands from different sides of the tumor.
How general is the Big Bang model?
What the authors validated were the predictions of the model, not the model itself. There could be other models than the Big Bang model which explain the data equally well or even better. (Sottoriva et al are discussing migration of late clones and tumor cell reseeding, but easily dismiss them.)
The Big Bang model down-plays the importance of selection in tumor development. With infinite resources and no space constraints I can see how selection might be less important, but as soon as resources and space become sparse each clone will fight for survival and the fitter clones will sweep through the population. So the Big Bang might be a very special feature of the colon, which is not a very dense organ and provides plenty of space.
In particular, I wonder how breast cancer fits into this picture. On the one hand it shows high intra-tumor heterogeneity with a mix of clones cell by cell –a similar variegation pattern as Sottoriva describe– on the other hand the life histories paper has made breast cancer the poster child of the clonal expansion model.
So maybe the Big Bang model is not yet a Theory of Everything.
Sottoriva A, Kang H, Ma Z, Graham TA, Salomon MP, Zhao J, Marjoram P, Siegmund K, Press MF, Shibata D, & Curtis C (2015). A Big Bang model of human colorectal tumor growth. Nature genetics, 47 (3), 209-16 PMID: 25665006