Recombinant DEK Enhances Ex Vivo expansion of Human Cord Blood and Mouse Bone Marrow Hematopoietic Stem Cells in a CXCR2- and Hspg-Dependent Manner

DEK is a nuclear DNA-binding protein that has been implicated in the regulation of transcription, chromatin remodeling, and mRNA processing.  Endogenous DEK regulates hematopoiesis, as BM from DEK^-/- mice manifest increased hematopoietic progenitor cell (HPC) numbers and cycling status and decreased long-term and secondary hematopoietic stem cell (HSC) engrafting capability (Broxmeyer et al., 2012, Stem Cells Dev., 21: 1449; 2013, Stem Cells, 31: 1447).  Moreover, recombinant mouse (rm) DEK inhibits HPC colony formation in vitro.  We now show that rmDEK is myelosuppressive in vitro in an S-phase specific manner and reversibly decreases numbers (~2 fold) and cycling status of CFU-GM, BFU-E, and CFU-GEMM in vivo, with DEK^-/- mice being more sensitive than control mice to this suppression.  In contrast, in vivo administration of rmDEK to wild type and DEK^-/- mice enhanced numbers of phenotypic LT-HSC.  This suggests that DEK may enhance HSC numbers by blocking production of HPCs.  We thus assessed effects of DEK on ex vivo expansion of human CD34⁺ cord blood (CB) and mouse Lin- BM cells stimulated with SCF, Flt3 ligand, and TPO.  DEK significantly enhanced ex vivo expansion of rigorously-defined HSC by ~3 fold both on day 4 (~15 fold increase from day 0) and 7 (~29 fold increase from day 0) when compared to cells expanded without DEK.  Expanding HSC with DEK also resulted in a decrease in the percentage of apoptotic HSC.

Further studies were done to better define how DEK works on HSC and HPC.  As extracellular DEK can bind to heparan sulfate proteoglycans (HSPG), become internalized, and then remodel chromatin in non-hematopoietic cells in vitro (Kappes et al., 2011, Genes Dev., 673; Saha et al., 2013, PNAS, 110: 6847), we assessed effects of DEK on the heterochromatin marker H3K9He3 in the nucleus of purified mouse lineage negative, Sca-1 positive, c-Kit positive (LSK) BM cells by imaging flow cytometry.  DEK enhanced the presence of H3K9Me3 in the nucleus of DEK^-/- LSK cells, indicating that rmDEK can be internalized by LSK cells and mediate heterochromatin formation.  We also investigated whether inhibiting DEK’s ability to bind to HSPG would block the inhibitory function of DEK in HPC.  Blocking the synthesis of, the surface expression of, and the binding capability of HSPG blocked the inhibitory effect of DEK on colony formation.  Blocking the ability of DEK to bind to HSPG also blocks the expansion of HSC in ex vivo expansion assays, suggesting that DEK mediates its function in both HSC and HPC by binding to HSPG but with opposing effects.

To further evaluate the biological role of rmDEK, we utilized single-stranded anti-DEK aptamers that inactivate its function.  These aptamers, but not their control, neutralized the inhibitory effect of rmDEK on HPC colony formation.  Moreover, treating BM cells in vitro with truncated rmDEK created by incubating DEK with the enzyme DPP4 (DEK has targeted truncation sites for DPP4)  eliminated the inhibitory effects of DEK, suggesting that DEK must be in its full- length form in order to perform its function.  Upon finding that DEK has a Glu-Leu-Arg (ELR) motif, similar to that of CXC chemokines such as IL-8, and as DEK is a chemoattractant for mature white blood cells, we hypothesized that DEK may manifest at least some of its actions through CXCR2, the receptor known to bind and mediate the actions of IL-8 and MIP-2.  In order to examine if this is indeed the case, we first confirmed expression of CXCR2 on the surface of HSC and HPC and then determined if neutralizing CXCR2 could block DEK’s inhibitory function in HPC.  BM treated in vitro with rmDEK, rhIL-8, or rmMIP-2 inhibited colony formation; pretreating BM with neutralizing CXCR2 antibodies blocked the inhibitory effect of these proteins.  DEK inhibition of CFU-GM colony formation is dependent on Gai-protein-coupled receptor signaling as determined through the use of pertussis toxin, which is a mechanism unique to DEK, as we have previously reported that IL-8 and MIP-1a are insensitive to the inhibitory effects of pertussis toxin.  Blocking the ability of DEK to bind to CXCR2 also inhibited the expansion of HSC in an ex vivo expansion assay.  This suggests that DEK binds to CXCR2, HSPG or both to mediate its function on HPC and HSC, enhancing HSC but decreasing HPC numbers.  Therefore, DEK may be a crucial regulatory determinant of HSC/HPC function and fate decision that is utilized to enhance ex vivo expansion of HSC.