Molecular Insights at the Single Cell Level into the Reprogramming of Human Hematopoietic Stem Cells during Epigenetically Modified Ex Vivo Expansion

The rarity of human hematopoietic stem cells (HSCs) and the very limited ability to study these cells in vivo in the human stem cell niche are significant hurdles in understanding the regulatory networks and cues that control the balance between human HSC maintenance, survival, quiescence, self-renewal and differentiation. In murine studies, epigenetic mechanisms have been demonstrated to play important roles in regulating HSC fate (Cullen SM et al. Curr Top Dev Biol. 2014; Wang et. al. Cell Stem Cell. 2016). One key mechanism involves histone modification by acetylation, with the acetylation of histones and non-histone proteins being controlled by histone acetyltransferases and histone deacetylates (HDAC). Recent studies have demonstrated that combining histone deacetylase inhibitors (HDACi), with specific hematopoietic cytokine priming under serum-free conditions can significantly enhance the expansion ex vivo of phenotypically defined and/or NSG mouse engraftable human HSC (SCID repopulating ability cells or SRC) from cord blood (CB). Indeed, ex vivo expansion with VPA in combination with FL, TPO, SCF and IL-3 for 7 days was reported to enhance human cord blood SRCs by almost 300 fold more than cytokine priming alone and 35.9 fold more than unexpanded cord blood (Chaurasia et. al. J Clin Invest. 2014). Subsequent microarray analyses of bulk CD34+ hematopoietic progenitor cells (HSPCs) after VPA plus cytokine treatment were used to define epigenetic signatures (Gajzer et. al. Vox Sang. 2016).

Given that human phenotypically-defined HSCs (Lin-CD34+CD38-CD45RA-CD90+CD49f+) or long-term (LT) SRCs would be expected to increase 3-5 fold over 7 day cultures (estimated median doubling time of 36-48 hours), that LT-SRC are characterised by delayed G0 exit (with 1st division averaging 66-75h), that short-term SRC proliferate more rapidly, that HSCs are contained in both the CD90+ and CD90- subsets and that HSC develop in micro-environmental niches which provide additional regulatory cues ( Laurenti E et al. Cell Stem Cell. 2015; Knapp et. al. Stem Cell Reports. 2017; Zonari E et al. Stem Cell Reports. 2017), we hypothesised that HDACi in the presence of cytokines might not only expand the CD90+ subset without differentiation but also reprogram CD90- cells to CD90+ HSCs.

To test this hypothesis, we first cultured human CB CD133+ cells in SCF, FL and TPO with HDACi or vehicle addition on Nanex scaffolds (Gullo et. al. Bioinformatics. 2015). Total cell expansion over 5 days was consistently lower in HDACi cultures compared to the vehicle control, with HDACi significantly increasing the absolute numbers of phenotypically-defined HSCs. The HDACi-induced increase of HSCs was accompanied by a concomitant decrease of phenotypically defined Lin-CD38-CD34+CD45RA-CD90- cells. To confirm that cells within the CD90- population were able to give rise to CD90+ phenotypically-defined LT-HSC, we then sorted Lin-CD34+CD38-CD45RA-CD90- cells into serum-free medium containing cytokines and HDACi. After 2-day culture, the majority of CD90- cells became CD90+ culture (with no significant up-regulation of CD49f). Next, we sorted CD133+ cells, cultured for 5 days in cytokines with vehicle or HDACi, into CD90+ or CD90- subsets (Lin-CD34+CD38-CD45RA-CD49f+) and analysed their transcriptomes (100 cells per sample from 2 different donors) by RNA-sequencing using SMART-seq2 (Picelli et. al. Nature Methods. 2014). RNA-sequencing data of CD90+ cells expanded with vehicle or HDACi did not show statistically significant differences in their gene expression. However, we identified 55 genes that were differently expressed between CD90+ and CD90- (Lin-CD34+CD38-CD45RA-CD49f+) subsets but unchanged in the vehicle and HDACi conditions. Using single cell sorting and Fluidigm Biomark technology, we confirmed the differential expression of these genes between CD90+ and CD90- cells (Lin-CD34+CD38-CD45RA-CD49f+). In conclusion, we have defined gene signatures for human CD90+ and CD90- (Lin-CD34+CD38-CD45RA-CD49f+) subsets after treatment with the HDACi, and demonstrate that HDACi does not alter gene expression of the CD90+ subset but is able to both delay HSC differentiation and reprogram CD90- cells into CD90+ cells during in vitro culture.