Recent studies indicate that inflammatory molecules, including interferon, possess stimulatory effects which trigger excessive proliferation leading to hematopoietic stem cell (HSC) exhaustion and bone marrow (BM) failure. We hypothesized that transient activation of inflammatory molecules by chromatin modifying agents (CMA) stimulates proliferation of transplantable HSC in culture. We previously demonstrated that sequential in vitro addition of a hypomethylating drug, 5-aza-2’-deoxycytidine (5azaD) followed by addition of a histone deacetylase (HDAC) inhibitor, trichostatin A (TSA) during human umbilical cord blood CD34+ cell culture results in expansion of serially transplantable long term HSC (Araki et al. Blood 2007, Exp Hematol 2009). However treatment with valproic acid (VPA), a HDAC inhibitor, only expands short term HSC, which are not serially transplantable (Araki et al. Abst. 827, Blood 2010). This failure of VPA to expand long-term repopulating HSC is not however due to poor BM homing (homing efficiency VPA-expanded cells 1.68 ± 0.35%, n = 10 NOD/SCID mice, 5azaD/TSA-expanded cells 0.39 ± 0.056%, n = 7, p = 0.008). Molecular studies indicate that expression of genes implicated in HSC self-renewal (HoxB4, Bmi1, Ezh2, and GATA2) are equally increased in 5azaD/TSA or VPA expanded CD34+ cells (n = 3, control vs. 5azaD/TSA or control vs. VPA p < 0.05). Similarly, increased HoxB4, Ezh2 and PU.1 protein levels are observed in VPA or 5azaD/TSA expanded cells compared to controls. Our global microarray data (Affymetrix chip HG U133 plus 2.0, n = 3) reveals a set of 113 differentially expressed genes preferentially detectable in 5azaD/TSA expanded cells associated with transplantable HSC expansion in culture and 278 HSC maintenance genes differentially expressed in CD34+ cells expanded in VPA or 5azaD/TSA, but not in control. Differential expression of genes representative of HSC expansion or maintenance gene sets were validated by real time qPCR. Intriguingly, functional pathway analyses of the 113 HSC expansion genes indicate the importance of inflammation signaling pathways (genes such as S100A8, Alox5). Furthermore, global genomic DNA methylation analysis of uncultured vs expanded, enriched CD34+ cells by LINE-1 methylation assay indicates transient hypomethylation of 5azaD/TSA-expanded cells at day 3 (55.4 ± 2.2 vs pre–treatment levels 77.3 ± 1.9) which later returned towards baseline (69.7 ± 2.1), corresponding with recovery of DNMT1 expression. Notably, control cultures or CD34+ cells expanded in VPA do not display changes in LINE-1 methylation (78.5 ± 2.2 and 77.2 ± 2 respectively, n = 2, mean % ± SE of 4 CpG sites). Interestingly, ELISA assays of conditioned medium from 5azaD/TSA-expanded cells reveals that levels of the inflammatory mediator leukotriene B4 (LTB4; synthesized from arachidonic acid by the Alox5 gene products, 5-lipooxygenase and LTA4 hydrolase) are 426.87 ± 101.7 pg/ml, significantly greater than that of control cultures (109.5 ± 11.6 pg/ml; n = 3, p = 0.03). VPA-treated cultures had slightly higher levels than controls (346.8 ± 138.8 pg/ml; p = 0.16). Similarly 5azaD/TSA IL-1β levels were significantly greater than controls (10.1 ± 0.8 pg/ml vs 6.0 ± 0.6 pg/ml, n= 2, triplicate wells, p = 0.05). IL-1β levels from VPA cultures were 10.1 ± 4.5 pg/ml (p = n.s.). Notably, expansion of CD34+CD90+ cells in the presence of 5azaD/TSA is reduced by 47.6% by addition of dexamethasone, an anti-inflammatory agent, without affecting total cell numbers. Furthermore, demethylation of CpG sites of the promoters of inflammatory/stress response genes correspond with higher expression of these genes in CMA-expanded cultures (e.g., S100A8 and Cyp11A1; pre treatment: 76.42% ± 4.91% and 80.73 % ± 4.16%, post 5azaD/TSA treatment 54.5% ± 3.84% and 54.29% ± 3.72% respectively; n = 2, mean ± SE of 3 to 11 CpG sites methylation). S100A8 is a known agonist of toll-like receptor 4, which plays a role in the maintenance and proliferation of endothelial progenitor cells, in keeping with the influence of systemic infection on HSC cycling and of interferon on HSC proliferation. Our studies, for the first time, identify inflammatory molecules regulated by epigenetic mechanisms that are potentially important in governing HSC fate choices. Further studies should determine the mechanisms by which CMAs influence HSC expansion or maintenance through inflammatory and other stress response pathway genes.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.