Posted by: Dan | August 2, 2006

Genetic Programming and the stem cell theory of cancer

Nature Reviews Cancer has an opinion piece of interest to this blog by Carla Boccaccio and Paolo Comoglio: Invasive growth: a MET-driven genetic programme for cancer and stem cells. This is significant both in light of the epithelial to mesenchymal transition (EMT) and the stem cell theory of neoplasms, both of which I think appear to be popular but incorrect models of cancer progression.

Boccaccio and Comoglio argue that:

Althought perceived as a subversion of tissue structure and function, the process of invasion and metastasis probably originates from a physiological property of undifferentiated stem and/or progenitor cells.

They note, too, that neoplastic cells undergo (superficially) similar phenotypic changes, and frequently utilize molecular determinants of morphogenesis. And that’s correct – to a point.

What they leave out is the recognition that the phenotypic change is not a true change in the differentiation state to a new cell lineage, but is instead an acquired deprogramming towards dysregulation of the genetic program or the epigenetic regulation of this program.

How significant is this distinction? Admittedly, not very. Boccaccio and Comoglio’s characterization of the biochemical elements of cancer progression, involving MET, HGF, and other factors, is spot on. Yet, I’m skeptical of their call to “identify, characterize rare cancer stem cells,” as if a good strategy is to discover a hidden cell type that is programmed to initiate neoplasms. Instead, a much better strategy for oncological therapy would be to expand and catalog the wide array of possible (epi)genetic predispositions for generating cellular dysfunction, and characterize the potential for these changes of occuring, as well as their relative contributions to neoplastic progression.


Indeed, their own Figure 2 conceptually portrays expansion of stem cell lineages (a) and tumor cell clonal expansion (b), which display two separate patterns – linear versus branching selection. Their caption, however, muddles and confuses the distinction between these two patterns.


  1. The idea, as I understand it, of the cancer stem cell is that neoplastic cells aren’t, in fact, dedifferentiated cells from the tissue of origin, but are rather the progenitor cells within the tissue, such as intestinal crypt cells and mesenchymal stem cells which fail to respond to the normal environmental cues that tell them they’re out of their niche. So they remain stem-like and self-renewing while making the transition from a niche-resident, self-renewing but slowly-cycling cell to a transitory amplifying cell which would normally go on to terminally differentiate as part of the regeneration process of the tissue. They sorta get stuck in the transitory amplifying cell stage.

  2. I may be overstating the dedifferentiation point, but part of the model is that differentiation and proliferation are opposing cell behaviors, and that the spindle-morphology of invasive tumor cells is a less-specialized cellular phenotype than its original tissue type, representing deregulation of morphology, cell-cell and cell-matrix interactions, programmed cell death, and proliferative functions.

    I argue that there are no pre-destined progenitor cells, as the use of the term cancer stem cells (CSCs) implies. Instead, what we find is an evolving ecosystem in a tumor, where we see variation in the behaviors of cells within a pre-cancerous lesion. Sometimes, in this lesion, we see failure to acquire angiogenic or immune-evasive characteristics, and the lesion does not persist; occaisionally, however, we see a cell successfully acquire a new characteristic, making it more successful than its neighboring cells, such that it soon takes over the cell population. I find this alternative explanation to be much better than labeling them CSCs.

    My main objection is that CSCs behave like part (a) in the figure I provided, but tumor cell populations behave more like part (b), and despite some similarities between the two, they’re not completely analogous behaviors.

  3. […] I’m a little confused, however: my contribution, Genetic Programming and the Stem Cell Theory of Cancer has not been included. Was I rejected by the carnival’s “editorial review board,” or did I simply not meet an aesthetic standard? …I’m just curious. […]

  4. […] Along the lines of my previous posts on Cancer Stem Cells (CSCs) and genetic programming comes a research report and commentary in the journal Nature echoing what I’ve been saying for a while: Krivtsov et al. (Transformation from committed progenitor to leukaemia stem cell initiated by MLL–AF9), and the commentary from Emmanuelle Passegué – A game of subversion. […]

  5. Read this month’s Sci Am. It makes a heck of a lot more sense and is much better written than this page.

  6. Fred,
    Could you please clarify, regarding what does not make sense, in your opinion?

    Also, why the fake (or incorrect) email address?

  7. Also, to the SciAm article that you mention, by Peter Duesberg – it has been discussed elsewhere and in more detail by Grrlscientist, and more critically by Orac.

    Chromosomal chaos as an origin of cancer isn’t a new theory, but Duesberg overreaches in arguing for a role for aneuploidy in causing cancer. It’s one cause, sure, but it’s a bit of an oversimplification (read Orac’s post).

    But regardless, Duesberg talks about just one cause of cancer, aneuploidy – I talked about another in the above post, dysregulation of stem cells. Both are very significant sources of neoplasia, to different degrees in various types of cancer.

  8. hi im a first year med student so dont have much knowledge to contribute but was wondering if someone could possible share some knowledge about why the dedifferentiation of cells to take on primordial stem cell characteristics aids the development/progression of cancer
    greatly appreciated

  9. Dedifferentiation, like “epithelial to mesenchymal transition” and “stem cell-like,” are terms used to generally describe cellular phenotypes, based on morphological and proliferative characteristics. For instance, normal and mature cells have rigid constraints on cell cycle activities, well-defined and consistent shapes, and so on. Cancer cells and stem cells, however, acquire the ability to divide more frequently, escape cell cycle checkpoints, avoid cell senescence, have irregular shapes, have disrupted cell-cell interactions, etc.

    So that may be a bit of a rambling reply, but generally the reasons why dedifferention aids in cancer progression fall into the following categories:

    1) Reduced constraints on cell division

    2) Reduced cell-cell contacts

    3) Disruption of normal tissue structure

    Merlo LM, Pepper JW, Reid BJ, Maley CC. Cancer as an evolutionary and ecological process. Nat Rev Cancer. 2006 Dec;6(12):924-35. Pubmed.
    Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000 Jan 7;100(1):57-70. Pubmed.


%d bloggers like this: