EHA Library - The official digital education library of European Hematology Association (EHA)

Acute myeloid leukemia (AML) is a clonal hematologic malignancy arising from a rare population of leukemic stem cells (LSCs) that initiate and propagate the disease. LSCs are enriched within the CD34+CD38low/- compartment expressing the IL-3 receptor α-chain, CD123. In contrast, normal hematopoietic stem cells do not express CD123. Several studies have recently shown that a fraction of LSCs, residing in particular niches of the bone marrow, resists to chemotherapy and is associated with poor clinical outcome.

We were interested to explore cell-adhesion and migratory properties of LSC aiming at identification of new cell-surface marker able to define a Leukemic Initiating Cell population within the LSC compartment.

AML patient samples were analyzed by flow cytometry to characterize protein expression profile in the LSC compartment and then engrafted in NOD scid gamma chain deficient (NSG) mice. Clonogenic and cobblestone-area forming cell (CAFC) assays in vitro and gene-expression profiling were performed to characterize stemness and regenerative potential of defined LSCs subsets.

We confirmed in a large cohort of 34 AML de novo patients results showing correlation between mouse engraftment and poor AML disease prognosis. Engrafting patients had a significant poorer overall survival as compared to non-engrafting. Searching for molecular factors that could affect AML engraftment, we found that the junctional adhesion molecule-C (JAM-C) was expressed by a fraction of LSCs. JAM-C is an adhesion molecule that we have previously shown to be expressed by mouse and human hematopoietic stem cells. Functionally, JAM-C interaction with JAM-B expressed on stromal cells was involved in HSC retention in the bone marrow. Here, we showed that percentages of JAM-C expressing cells in CD34+CD38low/-CD123+ leukemic cells at diagnosis were significantly increased in engrafting AML patient samples as compared to non-engrafting samples. Since engraftment was correlated to survival, we then tested if the frequencies of circulating CD34+CD38low/-CD123+JAM-C+ cells could provide prognostic information. 60 AML de novo patient samples at diagnostic were analyzed. We found that high frequencies of JAM-C positive cells in the LSC compartment were correlated to poor overall survival and significant reduced leukemia-free survival. Along this line, the study of 10 paired AML samples at diagnosis and relapse revealed a significant increase of JAM-C positive cells in the LSC compartment in all patient samples at relapse as compared to diagnosis.  Functional properties of these cells were then tested. Clonogenic and CAFC assays showed that JAM-C+ LSCs grow into colonies whereas JAM-C- cells do not. Moreover, JAM-C+ LSCs established leukemia in vivo whereas JAM-C- cells didn’t engraft in mice. Finally we showed that JAM-C+ cells are predominantly in G0 phase and have a more immature gene expression profile compared to JAM-C- cells.

All together, these data indicate that JAM-C is a unique marker for primitive leukemic stem cells and strongly suggest that JAM-C expression defines a Leukemic Initiating Cell population within the LSC compartment. Given the increased expression of this adhesion molecule by LSCs during relapse, we propose that targeting JAM-C should be a promising adjuvant therapeutic strategy to inhibit retention of leukemic stem cells within bone marrow stromal niches.