A seldom-mentioned caveat biasing a great deal of cancer research? Tumor heterogeneity in mouse models of cancer?
5 February 2013
I’ve been writing often on the topic of tumor heterogeneity as of late, and today I became curious about the level of tumor heterogeneity in mouse models of human cancer vs. the amount seen in the clinic, and how this seldom-mentioned caveat might have biased a huge amount of cancer research and therapy development to date.
This morning a labmate presented a really neat, compact paper from Cancer Research in Journal Club: “Synthetic Lethal Interaction of Combined BCL-XL
and MEK Inhibition Promotes Tumor Regressions in KRAS Mutant Cancer Models” by Cororan et al. our of Massachusetts General Hospital Cancer Center. Journal Club is an activity we do once a week where scientists and clinicians from all over the Moores Cancer Center get together and go over a recently published original research paper. We discuss the techniques, the data, the implications, and how we can best apply this to our own research and treating cancer on a weekly basis over coffee and bagels. It’s one of my favorite times of the week.
Anyway, this morning the above paper was presented, which illustrates a very elegant model for combining two different drugs to target cancers driven by KRAS mutations, like many pancreatic and colorectal cancers. In a nutshell, one drug targets the fast growth of the cells, another takes the lock off the cancer cells’ self destruct switch.
Dr. Stupack pointed out a caveat of the study, in which cell lines isolated from tumors have likely undergone artificial selection to enhance these anti-cell-suicide traits; a type of selection that might not necessarily reflect that which occurs in patients, and may bias the cancer cells used in the model to be homogenous in a way that would promote response to the authors’ hypothesis. Not good.
However, the authors of the paper did complement their studies with a spontaneous model of pancreatic cancer driven by KRAS, which was argued to not be subject to the same caveats. Perhaps the cancer cells developed by the mice are not going to be selected for in the same capacity as human cells in artificial culturing conditions for years on end, but the course of the disease model itself is less than a year, which is not likely to have the time to develop the same heterogeneity as, say, a human cancer that takes decades to develop and undergoes many, many more rounds of cell division and selection for diverse pockets of meaner, hardier, of cancer cells.
In short, how could one engineer a mouse model of cancer to better reflect human cancers in terms of genetic heterogeneity? Mice only live about two years, and experimental cancer studies lasting decades are impractical on many fronts. Is there some way to engineer enhanced tumor heterogeneity into existing mouse models of human cancer?
Again, I will leave this article with an open question for researchers and collaborating curious minds (like yourself, dear reader!). I offer it as another short glimpse into my stream of thoughts on enhancing targeted cancer therapies.