Extracellular DEK Activates NRF2 Antioxidant Targets and Drives Expansion of Functional and ROS-Low Human Hematopoietic Stem Cells

Umbilical cord blood (CB) is a promising source of hematopoietic stem cells (HSCs) for hematopoietic cell transplantation (HCT), a life-saving treatment, due to lower incidence of graft versus host disease and its availability. CB HCT is limited by low numbers of HSCs, thus there is a strong need to increase numbers of engraftable HSCs, for instance by expanding functionally competent HSCs ex vivo. We have shown that extracellular DEK protein, an epigenetic regulator that is secreted in response to stress and inflammation, is a novel hematopoietic regulatory chemokine. Treatment with extracellular DEK in growth stimulating media significantly enhances the expansion of engraftable HSCs in mouse bone marrow (BM) and human CB (Capitano, et al. JCI 2019; 129(6):2555-2570).

To examine mechanisms driving enhanced expansion of functional HSCs after DEK treatment, we performed RNA-sequencing on CB HSCs, multipotent progenitor cells (MPPs), common myeloid progenitors (CMPs), and granulocyte macrophage progenitors (GMPs) after overnight DEK treatment in expansion media. All cell populations upregulated gene programs related to antioxidant responses upon DEK treatment. Indeed, 7/10 genes upregulated by DEK in all HSC/HPC populations (NMRAL2P, PIR, NQO1, SELENOW, FTL, RIT1, and QPCT) promote antioxidant activity. Overnight DEK treatment significantly reduced mean fluorescence of the activated reactive oxygen species (ROS) indicator CM-H2DCFDA in HSCs, MPPs, CMPs, and GMPs. Thus, DEK may induce antioxidant activity. We next found that all 10 genes commonly upregulated in HSCs/HPCs are either direct transcriptional targets of NRF2 (NMRAL2P, LUCAT1,PIR, NQO1, FTL, and QPCT) or are directly affected by changes in NRF2 activation (SELENOW, EBI3, TNFRSF9, and RIT1). Further, several genes highly upregulated by DEK treatment in HSCs (i.e., CCL2, KYNU, and CXCL8) are NRF2 targets. To examine if DEK activates NRF2 transcriptional activity, we utilized a transgene reporter assay where expression of red fluorescent protein (RFP) is driven by activation of promoter antioxidant response elements (ARE), to which active NRF2 binds. DEK treatment of hematopoietic derived cell lines (mouse MLL-AF9 AML and human HL-60 APL) expressing the transgene significantly increased RFP signal, indicating that DEK could activate NRF2. We further examined whether DEK treatment affects mitochondrial metabolism, a source of ROS and tightly regulated process in hematopoiesis. Seahorse XFe analysis, which measures metabolic flux, showed DEK treatment of MLL-AF9 cells significantly reduced mitochondrial oxygen consumption rate, suggesting a reduced reliance on oxidative phosphorylation as an energy source in DEK treated cells.

These data suggest that DEK treatment during expansion of primitive hematopoietic cells drives activation of NRF2 transcription, upregulation of antioxidant response genes, reduction of intracellular ROS levels, and may reduce the mitochondrial metabolic rate. These effects in a sense "mimic” the in vivo niche of HSCs, where they reside in low oxygen tensions, have low intracellular ROS, and exhibit low levels of oxidative phosphorylation. To test this hypothesis, we examined if DEK treatment has similar effects to maintaining HSCs in physiologic oxygen tensions (3%) compared to HSCs exposed to extraphysiologic ambient air (21%), which drives oxidative stress (Mantel, et al. Cell 2015; 161(7):1553-65). Isolating and maintaining mouse BM in physiologic oxygen tensions or treating mice in vivo with recombinant DEK preserved HSC numbers compared to exposing the BM cells to ambient air or treating with vehicle. However, combining in vivo DEK treatment with physiologic isolation of HSCs/HPCs had no additive or synergistic effect, suggesting that the overall effects of DEK treatment and maintaining low oxygen tensions may be similar.

Thus, extracellular DEK treatment leads to expansion of functionally competent HSCs and reduction of oxidative stress. These data have important implications for the regulation of HSCs/HPCs in stress conditions, where DEK is most likely to be secreted. Further, DEK may present a novel way to mimic physiologic oxygen conditions for enhancement of functional HSC expansion, and the elucidated mechanisms may be exploited for more efficient targeting of molecular programs critical for functional HSC competency, potentially providing new ways to enhance HCT.

No relevant conflicts of interest to declare.