Functional Validation of Cancer Stem Cell Theory
Ontario Cancer Institute ♦ University of Toronto ♦ Hospital for Sick Children ♦ McMaster University ♦ others..
Published in Nature Medicine, August 28th, 2011
John Dick’s lab has begun functionally validating the cancer stem cell theory in a mouse model and created the closest thing we have to a true leukemia stem cell (LSC) signature. The cancer stem cell theory operates under the premise that not all cells in a tumour can support malignant growth. Instead, only a small, and rare subset of cells at the top of a complex hierarchy is able to sustain the life of a tumour.
Researchers initially attempted to identify cancer stem cells through transplantation into xenograft models. Those fractions that could promote and sustain tumour growth were isolated and analyzed in order to create a cancer stem cell fingerprint. However, it has been discovered that the mouse xenograft models used to create these fingerprints were not sensitive enough. As a result, some cancer stem cells have gone unnoticed.
In order to ensure that the CSC characteristics identified through these experiments are clinically meaningful, they must be linked to patient outcome and survival. This is the first study to take such a robust approach to confirming CSC function, setting a precedent for future experiments in the CSC field. To illustrate how this can be achieved, researchers analyzed fractions of cells from acute myeloid leukemia (AML) tumours from 16 different patients.
The process began by creating four populations of AML cells using the major LSC surface markers (CD34, CD38). These populations were functionally validated through transplantation into a very sensitive xenograft mouse model. Functionally defined LSCs (tumour forming) were subsequently analyzed using global gene expression analysis. Bioinformatic analysis was then utilized to compare 25 LSC-enriched fractions to 29 fractions without LSCs.
Despite the fact that AML samples exhibited different karyotypes and patients were of varying backgrounds (French, British, American), this comparison generated a common LSC signature composed of 42 genes. This signature was then compared to a hematopoietic stem cell (HSC) signature derived from cord blood by similar functional analysis. As could be expected, the genetic profile of LSCs derived from the AML cell lines was quite similar to that of the HSC profile derived from cord blood stem cells; the differences in the LSC and HSC signatures representing genetic targets for treating AML.
The true testament to the value of this study is illustrated by the tests for clinical relevance that followed the formation of LSC and HSC signatures. Correlations between the LSC/HSC signatures and clinical outcome were evaluated using three comprehensive, clinically annotated, gene expression sets. Researchers found a significant negative correlation between complete remission and overall survival, and high expression of the LSR and HSC signatures.
This eloquent work by John Dick’s lab shows us that strong connections can be made between CSC signatures and clinical outcome. Researchers studying other cancer stem cell types should use this study as a model in which to investigate their identity and clinical relevance.