A year ago Sean Morrison’s group from the University of Michigan published results of experiments which raised questions about the validity of the cancer stem cell concept using a xenotransplant mouse model. I’d like to come back to this remarkable moment, because it clearly shows how the model can change the whole concept.
Initially all experiments describing cancer stem cells were done on xenotransplantation of human malignant cells into immunocompromised mice. For years NOD/SCID mice served this purpose and even though they aren’t completely devoid of an immune response, they generally worked quite well. Later, a new mouse strain, more severely immunocompromised, was developed – the NOD/SCID/IL-2R-gamma null (NOG) mouse. Ishikawa showed more robust engraftment of human leukemic cells in this model suggesting it was a more permissive environment and possibly a better model for studying cancer stem cells. When experiments to assess tumorigeneity and frequency of cancer stem cells were actually done using NOG mice, however, it changed everything.
Sean Morrison’s work is the clearest example of how just changing a model can bring a dramatic difference in result – the frequency of human melanoma-initiating cells went from about 1 in 1 million to 1 in 4 just by going from NOD/SCID to NOG. I pulled the data into the table below.
Figure 1: Example of how modification of xenotransplantation assay can increase the detectable frequency melanoma-initiating cells (credit: Elsa Quintana, Mark Shackleton and Sean Morrison, 2008):
| change in the model |
how was it done |
fold increased frequency of cancer stem cells |
| time assessment of tumorigeneity | longer time of assessment | 10 |
| xenotransplantation mouse model | NOG mice (more immuno-permissive than NOD/SCID) | 170 (!!!) |
| appropriate tumor microenvironment (niche) | cancer cells embedded in Matrigel | 20 |
Another example, from Domionique Bonnet’s group, shows that the initial markers, identifying leukemia-initiating cells does not apply in “new” – NOG model.
Quote from the original post on Hematopoiesis.info:
So, mouse xeno-model, antibody by which cells were treated, site of transplantation and route of administration, matrix and microenviroment, recepient conditioning… – everything is important!
Why do we need to be very careful about it? Because a xeno-model has a translational potential which can bring us to therapeutic protocols and clinical success.Also, as Morrison said, these studies do not invalidate the CSC hypothesis but they just show that not all of cancers fit in this hypothesis and the problem is more complex.
Lessons to learn:
1. The cancer stem cell concept is much more complex than we thought before and might be valid only for some cancers instead of being a universal theory of carcinogenesis.
2. Previous findings in the cancer stem cell field which were done in the NOD/SCID mouse model should be reassessed in NOG mice.
Watch the lecture given by Sean Morrison in MIT 2009 and see how model and assays could be important.
No related posts.
