Abstract 1884

Acute myeloid leukemia (AML) is a common and aggressive hematologic malignancy affecting both children and adults which continues to have high mortality rates as well as high morbidity from toxic therapies. New treatments are needed to improve cure rates and decrease morbidity. A niche-based high throughput screen done in a murine system identified candidate small molecules potentially toxic to leukemic stem cells (LSCs) while sparing normal hematopoietic stem cells (HSCs) and bone marrow stroma (Hartwell KA, Miller, PG et al., in preparation). One such compound, SB-216641, demonstrated dose-dependent activity against leukemia in both a cell autonomous and non-autonomous manner, by modifying niche–based support. SB-216641 is a selective serotonin receptor antagonist specific for the 5-HT1B receptor, highlighting a pathway not previously investigated in the context of AML or leukemia stem cell biology.

We examined the effects of this candidate small molecule on 7 human primary AML samples. CD34+ cells were isolated from these samples with immunomagnetic beads. Using the colony forming assay to assess kill of progenitor cells, all samples had ≥99% cell kill at 25 μM (10 times the IC-50 found in the murine system). We then assessed the compound's effect on LSCs using the cobblestone area forming cell (CAFC) assay, a standard in vitro stem cell assay. The leukemic cells were pulse treated for 18 hours and washed to remove residual SB-216641 prior to placement on MS-5 murine stroma and therefore only the direct effect on the leukemic cells was measured in this assay. CAFCs were read out at week 5, or week 2 when the sample was FLT3-ITD+ (Chung KY et al, Blood 2005, Vol 105, 77–84). We first tested five samples at 25 μM. All samples formed cobblestone areas in the control setting (46–200 CAFCs/106 cells plated). Four samples had no CAFC formation with SB-216641 and the remaining sample had >95% decrease in CAFC formation. We then performed serial dilutions using the CAFC assay in the human primary samples as well as in HSCs derived from cord blood to obtain the IC-50 for human AML and to ensure that our differential cell kill of LSCs versus normal HSCs held true in the human samples. IC-50 for the human primary leukemias was found to be 630 nanomolar and at 10 μM all leukemic samples were fully killed with 100% survival of normal human HSCs [see figure 1]. As a confirmatory study, using HL60 and U937 human AML cell lines transduced with GFP-luciferase, 500 cells were preincubated with SB-216641 at 25 μM or DMSO control and then injected IV into Nod Scid IL2R-gamma null (NSG) mice and imaged at 5 weeks. In both cell lines, the control mice had engraftment and the mice that received treated cells had no engraftment. HL60 cells were then preincubated with SB-216641 at lower doses (10 and 5 μM) and injected into NSG mice and imaged at 3 weeks. Again, the control mice had engraftment and the mice that received treated cells had no engraftment.

5-HT1B receptor antagonists have not previously been known to be active against AML or leukemic stem cells. Some hematopoietic cells including platelets express serotonin receptors and T-cells specifically have been found to express the 5-HT1b receptor. Selective 5-HT1B receptor antagonists have found to have apoptotic effects in vitro against cell lines of other cancers and may be involved in MAP kinase and P13K/Akt signaling pathways. SB-216641 is a highly promising compound which warrants further investigation. Its high toxicity to LSCs and sparing of normal HSCs make it appealing for possible clinical use in the future.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.