Evaluation of phenotypic heterogeneity and the role of MRC2 in leukemic stem cells

Acute myeloid leukemia (AML) is a blood malignancy affecting people from all age categories
yet most frequently elderly people (older than 65-year-old). Therapy options have improved
over the last decades but AML still remain lethal in many cases despite many patients
achieving apparent remission. Indeed, patients often relapse and such reoccurrence
becomes harder to treat due to the development of therapy resistance and the selection of
specific more aggressive subclones. AML was the first malignancy to support the theory of
cancer stem cells and is widely described for its hierarchical structure. In fact, AML is known
to originate and relapse from a rare cell subpopulation of leukemic cells called leukemic stem
cells (LSC). AML and their LSC are very heterogeneous on the transcriptomic and
phenotypic level which makes it challenging to eradicate these cells. Currently, efforts are
made by many research groups to determine how to selectively and specifically target LSC.
Not only would this improve our knowledge on their biology but this could potentially greatly
improve clinical care.
In this thesis, we improved knowledge on LSC biology and highlighted a potential new
vulnerability. We discovered that MRC2 is overexpressed in AML and specifically enriched
cells with LSC properties. RNA-sequencing of MRC2+ vs. MRC2- cells revealed the former
population to have, on the transcriptomic level, a stemness signature that is highly
concordant with our in vitro and in vivo assays indicating positive cells to be more clonogenic
in colony forming unit (CFU) assay and patient-derived xenograft (PDX) models, respectively.
MRC2+ cells are functionally involved in collagen uptake, which engages cells metabolically.
In depth in silico analyses revealed proline dehydrogenase (PRODH) as overexpressed in
MRC2+ LSC. Importantly, PRODH inhibition impaired cell viability and stemness (with no
effect on healthy cells) and synergized with the BCL-2 inhibitor venetoclax. Furthermore,
TGF-β1 was discovered as a strong inducer of MRC2 expression/function in AML.
In a related second project, variables influencing leukemic engraftment in PDX models were
studied in an attempt to further improve these models that are essential for functional studies
of human AML. Noticeably, we discovered an improved leukemic engraftment in female vs.
male NSG animals when injected with the exact same cell dose of primary sample in both
primary and secondary transplant assays. This importantly made us reconsider how to
continue with PDX models and, by only relying on female NSG mice, we further discovered
that AML samples from the adverse risk group, with high FLT3-ITD ratio or with high LSC
content had shorter time to engraft in female NSG mice. Our extended incubation period
allowed us to appreciate engraftment in samples previously considered non-engraftable such
as from favorable molecular risk groups. These observations are helpful as they mimic well
clinical settings and furthermore allow comprehensive studies of AML biology in in vivo
settings.