Evaluation of Cell Surface Markers and Targets in Leukemic Stem Cells (LSC) Reveals Distinct Expression Profiles, Unique Drug Effects, and Specific Checkpoint Regulation in AML LSC and CML LSC

In an attempt to identify novel cell surface markers and targets in leukemic stem cells (LSC) in acute myeloid leukemia (AML) and chronic myeloid leukemia (CML), we screened bone marrow (BM) samples of patients with AML (n=257), CML (n=134), and controls (n=256: other BM neoplasms, n=116; normal/reactive BM, n=29; idiopathic cytopenia, n=31; lymphoma patients without BM involvement, n=80) for expression of cell surface markers and targets on CD34⁺/CD38⁻ stem cells and CD34⁺/CD38⁺ progenitor cells (PC) by multi-color flow cytometry. In addition, we examined mRNA expression profiles in highly purified CD34⁺/CD38⁻ stem cells and CD34⁺/CD38⁺ PC by gene array- and qPCR analyses. Aberrant LSC expression profiles were identified in all patients examined. In patients with CML, CD34⁺/CD38⁻ LSC expressed an almost invariable aberration profile defined as CD25⁺/CD26⁺/CD56⁺/IL-1RAP⁺. By contrast, in patients with AML, CD34⁺/CD38⁻ cells variably displayed aberrant surface antigens, including CD25 (55%), CD96 (35%), CD371 (CLL-1) (75%), and IL-1RAP (60%). With the exception of a subset of FLT3 ITD+ patients (45% of all FLT3-mutated cases), AML LSC did not exhibit CD26. All other markers and targets identified on AML and/or CML LSC, including CD9, CD33, CD44, CD47, CD52, CD105, CD114, CD117, CD133, CD135, CD184, neural proliferation and differentiation control-1 antigen (NPDC1), and roundabout-4 (ROBO4), were also detectable on normal hematopoietic stem cells (HSC). However, several surface-targets, including CD33 and CD52, were expressed at higher levels on CD34⁺/CD38⁻ LSC compared to normal HSC, and antibody-mediated targeting resulted in their selective depletion and in a significantly reduced engraftment of LSC in NSG mice. Since certain surface antigens, like CD47 (IAP), CD243 (MDR1), and CD274 (PD-L1) may contribute to intrinsic drug resistance of LSC, we also examined the expression of these antigens in AML and CML LSC. Regardless of the type of leukemia, LSC were found to express CD47 in all patients examined. MDR-1 was found to be expressed on LSC in 11/39 patients with AML (28%) and 2/22 patients with CML (9%). PD-L1 was not detectable on LSC in untreated BM samples. Exposure of LSC to interferon-gamma (IFN-G, 100 U/ml, 48 hours) resulted in expression of PD-L1 on LSC in all patients with Ph⁺ CML (control: 100%; IFN-G: 183±40%, p<0.05). Unexpectedly, however, IFN-G was found to induce PD-L1 expression in only 1 out of 11 AML patients tested. IFN-G showed no effects on expression of CD47 or MDR-1 in LSC. We next examined the potential mechanisms of IFN-G-mediated expression of PD-L1 on LSC and screened for drugs capable of counteracting the expression of this resistance-inducing checkpoint molecule. Of all targeted drugs examined (n=10) only the BRD4/MYC inhibitor JQ1 was identified as a regulator of IFN-G-induced PD-L1 expression on CML LSC. In particular, JQ1 (2.5 µM, 48 hours) was found to suppress IFN-G-induced (100 U/ml, 48 hours) upregulation of PD-L1 in primary CML LSC in all donors tested (IFN-G: 183±40% versus IFN-G+JQ1: 136±27% of control) suggesting that checkpoint expression in LSC is regulated by epigenetic mechanisms and MYC activity. Together, we have established cell surface marker- and target expression profiles in CD34⁺/CD38⁻ AML LSC and CML LSC which should facilitate their enrichment and may support the development of LSC-eradicating treatment concepts. In addition, the unique expression profiles of LSC should provide a powerful basis for diagnostic LSC phenotyping in patients with CML and AML.