Valproic Acid Results In Maintenance but Not Expansion of Transplantable Hematopoietic Stem Cells From Human Umbilical Cord Blood

Abstract 827

Epigenetic modifications are considered to be important in determining the fate of hematopoietic stem cells (HSC). We previously demonstrated that the sequential addition of the chromatin-modifying agents (CMA) 5-aza-2′-deoxycytidine (5azaD) and trichostatin A (TSA) expands transplantable HSC (Araki et al. Blood 2007, Exp Hematol 2009). Others have shown that valproic acid (VPA), an HDAC inhibitor, also expands HSC (DeFelice et al. Cancer Res 2005). We thus compared the efficacy of 5azaD/TSA and VPA in promoting the ex vivo expansion of human cord blood (CB) HSC. Cells were incubated with cytokines alone (SCF, Flt3 ligand, TPO and IL-3) or with cytokines and either 5azaD/TSA or VPA, resulting in 2.2-fold, 10.7-fold or 65-fold expansion, respectively, of primitive CD34+CD90+ cells after 9 days (n=3, Cytokine alone vs. VPA p=0.004; Cytokine alone vs. 5azaD/TSA p=0.03; VPA vs. 5azaD/TSA p=0.003). Interestingly, the 10.7-fold expansion of CD34+CD90+ cells following 5azaD/TSA treatment correlated with a 10- and 10.5-fold expansion of short-term colony-forming cells (CFC) and long-term cobblestone area-forming cells (CAFC), respectively. However, the 65-fold expansion of CD34+CD90+ cells achieved with VPA treatment yielded only a 25.6- and 8.4-fold expansion of CFC and CAFC, respectively. These results suggest a marked discordance between the phenotype and function of CD34+CD90+ cells when they are expanded in VPA, but not in 5azaD/TSA. Thus, we examined the in vivo hematopoietic repopulation potential of CMA-expanded CB HSC by quantitating SCID mouse repopulating cells (SRC) using limiting dilution assays. The frequency of SRC was 1 in 22,000 in primary CB cells (n=29 mice), 1 in 123,315, in (cytokine) controls (n=16 mice), 1 in 21,720 with VPA-treatment (n=27 mice), and 1 in 3,147, in 5azaD/TSA-treated CD34+CD90+ cell cultures (n=22 mice). Unlike control, treatment with VPA prevents loss of SRC but only results in SRC maintenance, whereas 5azaD/TSA treatment leads to a 7-fold expansion of SRC. Furthermore, serial transplantation of bone marrow (BM) from primary recipients engrafted with unmanipulated CB cells resulted in engraftment in 2 of 5 secondary mice, while BM from mice engrafted with VPA-treated cells failed to display secondary engraftment (n=5 mice), whereas BM from mice engrafted with 5azaD/TSA-treated cells resulted in engraftment in 5 of 6 secondary mice. Hence, we conclude that treatment of CB CD34+ cells with 5azaD/TSA or VPA results in distinct SRC outcomes-expansion or maintenance, respectively. To dissect the molecular mechanisms that may mediate these distinct SRC fates, we examined genes implicated in HSC self-renewal including HoxB4, Bmi1, STAT3, Ezh2 and PU.1. These gene transcript levels were increased in CD34+ cells treated with either 5azaD/TSA or VPA when compared to control cultures as measured by real time quantitative PCR. In accordance with these studies, CHIP assays using antibody against acetylated histone H4 indicate increased acetylation of the promoters of HoxB4 and Bmi1 genes in both VPA- and 5azaD/TSA-treated cells. In addition, higher levels of HoxB4, Ezh2 and PU.1 proteins were observed in VPA- and 5azaD/TSA-expanded cells, compared to control cultures. Since VPA treatment does not result in SRC expansion, these observations raise questions as to the importance of the upregulation of these genes for HSC expansion. Since the pharmacologic activity of CMAs is short (hours) we hypothesize that temporal effects, including early epigenetic modifications, lead to changes in transcription factor expression, which directly or indirectly promote symmetric or asymmetric divisions ultimately resulting in expansion or maintenance of HSC. Importantly, our global microarray data (n=3) using a human genome affymetrix chip (U133 plus 2.0) revealed a set of differentially expressed genes present in 5azaD/TSA- but not in VPA-expanded CD34+ cells, thus uncovering a potential molecular signature for HSC expansion. Currently, we are examining the molecular interactions of these signature genes and the effects of silencing of these genes on HSC expansion or maintenance which should allow us to begin to unravel the molecular mechanisms involved. In summary our data indicate that treatment of HSCs with different CMAs results in distinct fates: expansion or maintenance of HSC, an observation of potential therapeutic importance.