Acute myeloid leukemia (AML) and the myelodysplastic syndromes (MDS) are initiated and sustained by self-renewing malignant stem cells. Any curative therapy for AML or MDS must eliminate leukemic stem cells (LSCs) or MDS hematopoietic stem cells (HSCs), respectively, as these are likely the only self-renewing disease cells. Through transcriptional profiling and flow cytometric validation, we previously identified CD99 as a cell surface marker highly expressed on MDS HSCs (85%, n=63). Assessment of 78 paired diagnosis/relapse AML specimens revealed that CD99 is also frequently overexpressed in AML at diagnosis (86%) and relapse (92%). Consistent with CD99’s potential as an LSC marker, we showed that the LSC enriched CD34+CD38- fraction of AML consistently expressed higher levels of CD99 compared with more differentiated blasts (p<0.0001). We also found that CD99 expression was higher on leukemic blasts at relapse as compared with diagnosis (p=0.007).

To test whether CD99 expression enriches for functional LSCs, we transplanted the top and bottom 10% of CD99 expressing blasts within the LSC enriched CD34+CD38-CD90-CD45RA+ fraction of a CD34-positive primary AML specimen into sublethally irradiated NOD/SCID/IL-2Rgc-null (NSG) mice, observing that leukemia-initiating cell (L-IC) activity was restricted to the top 10% of CD99 expressors (L-IC frequency 1 in 24,401). Similar results were obtained utilizing a CD34-negative AML (L-IC frequency 1 in 105,688 in the top 10% of CD99 expressors), suggesting that CD99 can enrich for functional LSCs independent of CD34-expression status. CD99 is the first LSC marker reported to selectively enrich for LSCs, as we found other previously identified LSC markers (CD47, TIM3, CD44, and CD123) to be expressed at equal or decreased levels on LSC-enriched fractions compared to more differentiated blasts. Thus, high CD99 expression provides a novel method to purify LSCs, allowing for more refined studies of this critical cell type.

As therapeutically targeting CD99 with mAbs has been proposed in Ewing’s sarcoma and B cell acute lymphoblastic leukemia, we tested the ability of anti-CD99 mAbs to induce cell death in AML and MDS. Anti-CD99 mAbs (clones H036-1.1 and 12E7) were cytotoxic to primary AML blasts (n=5) and MDS CD34+ cells (n=2), as well as 17 myeloid leukemia cell lines in vitro. Normal HSCs and human umbilical vein endothelial cells express low and intermediate levels of CD99, respectively, but anti-CD99 mAbs were minimally toxic to these normal cell types, indicating a potentially large therapeutic window. Ex vivo incubation of primary AML blasts previously demonstrated to engraft NSG mice with H036-1.1 for 45 minutes prior to xenotransplantation completely abolished their engraftment. To rule out an effect on initial homing and lodging of LSCs, and to determine whether anti-CD99 mAbs can eradicate LSCs in vivo, NSG mice were transplanted with primary AML cells, allowed to engraft for two weeks, and then treated with 15 μg H036-1.1 or isotype control. One treatment with H036-1.1 was sufficient to completely abolish engraftment as measured in the BM five months post-treatment.

To determine the mechanism of anti-CD99 mAb induced cell death, we assessed the effect of anti-CD99 mAbs on Src-family kinase (SFK) activation. We demonstrated that CD99 negatively regulates SFK activation in AML, and that cytotoxic anti-CD99 mAbs induce rapid and robust SFK activation. To determine whether SFK activation is sufficient to induce cell death, we generated a constitutively active c-Src mutant (Y530F) and overexpressed it in anti-CD99 mAb “sensitive” and “resistant” cell lines (MOLM13 and K562, respectively). While MOLM13 cells showed a marked decrease in cell growth, increased apoptosis, and G0 cell cycle arrest, K562 cells demonstrated only a modest decrease in growth and S-phase entry, with no G0 arrest or increase in apoptosis. We thus propose that anti-CD99 mAbs induce apoptosis by inducing dysregulated SFK activation and oncogenic stress.

Taken together, our studies establish CD99 as a novel stem cell marker in AML and MDS, and as a promising therapeutic target in two diseases in need of novel therapies with validated activity against disease-initiating cells. Cytotoxic anti-CD99 mAbs appear to exert their effects by activating SFKs, revealing a novel therapeutic vulnerability in AML and MDS that has potential to be exploited using other modulators of this pathway.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.