Leukemia Stem Cell Potential of Different Progenitor Subpopulations in Myeloid Blast Phase CML

Blast phase (BP)-CML remains the most critical area of unmet clinical need in the management of CML and novel, targeted therapeutic strategies are urgently needed. In the tyrosine kinase inhibitor (TKI) era, the rate of progression to BP is 1 to 1.5% per annum in the first few years after diagnosis, falling sharply when major molecular response is obtained. Around 10% of patients present with de novo BP-CML and despite the use of TKIs, median survival after the diagnosis of BP-CML is between 6.5 and 11 months.Therefore, improved understanding of the biology of BP-CML and novel therapies to prolong therapeutic responses are urgently sought.

Studies of myeloid malignancies show that acquisition of tumor-associated mutations occurs principally in a step-wise manner. Initiating mutations usually originate in an hematopoietic stem cell (HSC) to give rise to preleukemic stem cell populations that expand through clonal advantage. Further mutation acquisition and/or epigenetic changes then lead to blast transformation and disruption of the normal immunophenotypic and functional hematopoietic hierarchy. At this stage, multiple leukemic stem cell (LSC) populations (also termed leukemia initiating cell populations) can be identified. We previously showed, in AML, that the CD34⁺ LSC populations were most closely related to normal progenitor populations, rather than stem cell populations, but had co-opted elements of a normal stem cell expression signature to acquire abnormal self-renewal potential (Goardon et al, Cancer Cell, 2011). CD34⁺CD38^- LSCs were most commonly similar to an early multi-potent progenitor population with lympho-myeloid potential (the lymphoid-primed multi-potential progenitor [LMPP]). In contrast, the CD34⁺CD38⁺ LSCs were most closely related to the more restricted granulocyte-macrophage progenitor (GMP).

In chronic phase CML, the leukemia-propagating population is the HSC, and the progenitor subpopulations do not have stem cell characteristics. To date, studies to isolate LSC populations in BP-CML have been limited, identifying the GMP subpopulation only as a possible LSC source (Jamieson et al, NEJM, 2004). Furthermore, in vivo LSC activity has not been assessed. We therefore set out to assess the LSC characteristics of different primitive progenitor subpopulations in myeloid BP-CML both in vitro and in vivo.

We isolated different stem and progenitor cell subpopulations using FACS; HSC (Lin^-CD34⁺CD38^-CD90⁺ CD45RA^-), multipotent progenitor (MPP; Lin^-CD34⁺CD38^-CD90^-CD45RA^-), LMPP (Lin^-CD34⁺CD38^-CD90^-CD45RA⁺), common myeloid progenitor (CMP; Lin^-CD34⁺CD38⁺CD45RA^-CD123⁺), GMP (Lin^-CD34⁺CD38⁺CD45RA⁺CD123⁺) and megakaryocyte erythroid progenitor (MEP; Lin^-CD34⁺CD38⁺CD45RA^-CD123^-). The functional potential of these purified populations was examined in 13 patients by: (i) serial CFC replating assays to study progenitor self-renewal (n=10); (ii) In vivo xenograft studies using NSG mice with serial transplantation to identify populations with LSC potential (n=6). Our data conclusively demonstrate that functional LSCs are present in multiple immunophenotypic stem/progenitor subpopulations in myeloid BP-CML, including HSC, MPP, LMPP, CMP and GMP subpopulations. There was inter-patient variability in terms of both in vitro and in vivo functional properties. Fluorescence in situ hybridisation (FISH) was used to assess clonality in the different progenitor subpopulations and identify which populations contained cells with additional cytogenetic abnormalities (ACAs) with a view to improving our understanding of the clonal hierarchy. Interestingly, there were no significant differences in ACAs in the different progenitor subpopulations in the majority of samples studied, suggesting that clonal evolution tends to occur in the HSC compartment in myeloid BP-CML. Preliminary gene expression profiling studies of the different progenitor subpopulations, using Affymetrix Human Gene 1.0 ST Arrays, demonstrated highly variable gene expression, supporting the functional heterogeneity seen.

Taken together, our results demonstrate that myeloid BP-CML is a very heterogeneous disorder with variable LSC populations. Further interrogation of these populations will likely identify novel therapies which will specifically target the LSC.