Understanding the mechanisms underlying the abnormal differentiation of human acute myeloid leukemia (AML) may reveal novel therapies to eradicate leukemic stem cells (LSC), which are often resistant to standard treatments and contribute to relapse. Cellular metabolism is recognized as a hallmark of cancer and is known to be distinct between hematopoietic stem cells (HSC) and downstream progenitors. In particular, sphingosine-1-phosphate (S1P) is a bioactive lipid produced from sphingolipid metabolism that regulates proliferation and survival and is implicated in HSC egress, lymphocyte trafficking and mouse lymphoid lineage determination primarily through binding to S1PR1, one of five S1P G-protein coupled receptors. However, sphingolipids are understudied in human LSC and HSC. We recently found that sphingolipid perturbation governs HSC self-renewal and influences lineage outcome. Here, we show that S1PR3 governs myeloid commitment of LSC in a subset of human AML and is thus an attractive therapeutic target.

Lipidomic analysis of LSC+ and LSC‒ fractions derived from AML patient samples and validated by xenotransplantation assays showed distinct sphingolipid alterations when compared to each other and to normal human cord blood (CB) CD34+CD38‒ stem cells and CD34+CD38+ progenitor populations. Interestingly, LSC+ fractions have increased S1P levels, prompting us to wonder if S1P signaling contributes to LSC maintenance. To gain a better understanding of the role of S1P in stemness, we examined S1P signaling in normal CB. Analysis of gene expression of S1PR1-5 and S1P transporters (SPNS2 and MFSD2B) within a comprehensive transcriptional roadmap of human hematopoiesis comprising thirteen normal populations from HSC to mature blood lineages demonstrated distinct expression patterns in specific blood lineages. S1PR1 and MFSD2B were most highly expressed in lymphoid and erythroid lineages, respectively, consistent with murine data. Notably, S1PR3 expression is specific to mature monocytes and granulocytes in human CB. S1PR3 protein was absent from the surface of long-term (LT) and short-term (ST)-HSC as determined by flow cytometry. Remarkably, lentiviral overexpression (OE) of S1PR3 was sufficient to promote myelopoiesis at the expense of erythropoiesis from LT- and ST-HSC in vitro in liquid culture, single cell assays and colony forming assays. To ascertain the mechanisms promoting myeloid commitment, we performed gene expression profiling by RNA-seq of LT- and ST-HSC following S1PR3 OE. This yielded a subset of shared genes similarly upregulated following S1PR3 OE relative to controls, including the known myeloid differentiation and AML-associated factors Early Growth Response 1 and 2 (EGR1/2) and Tribbles 1 (TRIB1). Thus, promiscuous expression of S1PR3 promotes a myeloid fate program in human HSC.

S1PR3 protein expression was higher in AML patient samples relative to human CB and bone marrow cells by flow cytometry. Bioinformatic analysis of three independent AML cohorts revealed that AML patient samples with high S1PR3 gene expression also had high EGR2 and TRIB1 expression. Limiting dilution xenotransplantation assays of LSC-containing fractions sorted based on surface expression of S1PR3 demonstrated lower LSC frequency in S1PR3high versus S1PR3low/- LSC fractions. Moreover, S1PR3 OE in LSC+ fractions virtually abolished leukemic engraftment in xenotranplantation assays. These results suggest that higher S1PR3 levels in LSC are associated with a more mature myeloid state and that further increasing S1PR3 levels disrupts LSC maintenance. Treatment of mice bearing primary AML xenografts with FTY720, a S1P mimetic that targets S1P receptors including S1PR3, decreased leukemia burden in a subset of patient samples tested, including those obtained from relapsed and treatment-refractory patients. 70% of AML samples tested showed decreased LSC frequency in serial repopulation assays following FTY720 treatment. Importantly, treatment with FTY720 did not alter normal hematopoietic xenografts, demonstrating the existence of a therapeutic window. Collectively, our results provide the first direct evidence that sphingolipids govern myeloid commitment in human HSC and LSC, and demonstrate the potential of S1PR3 as a novel therapeutic target in AML for eradication of LSC while sparing HSC.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution