The controversy generated from recent murine studies as to whether hematopoietic stem cells (HSC) contribute to steady-state hematopoiesis emphasizes how limited our knowledge is of the mechanisms governing HSC self-renewal, activation and latency; a problem most acute in the study of human HSC and leukemia stem cells (LSC). Many hallmark stem cell properties are shared by HSC and LSC and therefore a better understanding of stemness regulation is crucial to improved HSC therapies and leukemia treatments targeting LSC.
Our previous work on LSC subsets from >80 AML patient samples revealed that HSC and LSC share a transcriptional network that represent the core elements of stemness (Eppert, Nature Med 2011; Ng, Nature 2016). Hence, to identify the key regulators of LSC/HSC self-renewal and persistence we selected 64 candidate genes based on expression in functionally validated LSC vs. non-LSC fractions and assessed their potential to enhance self-renewal in a competitive in vivo screen.
Here, we transduced cord blood CD34+CD38- cells with 64 barcoded lentiviral vectors to assemble 16 pools, each consisting of 8 individual gene-transduced populations, for transplantation into NSG mice. Strikingly, individual overexpression (OE) of 5 high scoring candidates revealed delayed repopulation kinetics of human HSC/progenitor cells (HSPC): gene-marking of human CD45+ and lin-CD34+ cells was reduced relative to input and control at 4w post transplantation, whereas by 20w engraftment of marked cells reached or exceeded input levels. For one of these candidates, C3ORF54/INKA1, we found that OE did not alter lineage composition neither in in vitro nor in vivo assays but increased the proportion of primitive CD34+ cells at 20w in vivo; moreover, secondary transplantation revealed a 4.5-fold increase in HSC frequency. Of note, serial transplantation from earlier time points (2w, 4w) revealed superior engraftment and hence greater self-renewal capacity upon INKA1-OE. Since we observed a 4-fold increase of phenotypic multipotent progenitors (MPP) relative to HSC within the CD34+ compartment (20w) we assessed whether INKA1-OE acts selectively on either cell population. The observation of latency in engraftment was recapitulated with sorted INKA1-OE HSC but not MPP.
Likewise, liquid culture of HSPC and CFU-C assays on sorted HSC showed an initial delay in activation and colony formation upon INKA1-OE that was completely restored by extended culture and secondary CFU-C, respectively. INKA1-OE MPP showed a slight increase in total colony count in primary CFU-C and increased CDK6 levels in contrast to reduced CDK6 levels in INKA1-OE HSC emphasizing opposing effects of INKA1 on cell cycle entry and progression in either population. Taken together, this suggests that INKA1-OE preserves self-renewal capacity by retaining HSC preferentially in a latent state, however, upon transition to MPP leads to enhanced activation.
Whilst INKA1 has been described as an inhibitor of p21(Cdc42/Rac)-activated kinase 4 (PAK4), no role for PAK4 is described in hematopoiesis. Nonetheless, its regulator Cdc42 is implicated in aging of murine HSPC by affecting H4K16 acetylation (H4K16ac) levels and polarity and has recently been described to regulate AML cell polarity and division symmetry. In our experiments immunostaining of HSPC subsets cultured in vitro and from xenografts indicates that INKA1-OE differentially affects epigenetics of these subsets linking H4K16ac to the regulation of stem cell latency.
In AML, transcriptional upregulation of INKA1 in LSC vs. non-LSC fractions and at relapse in paired diagnosis-relapse analysis (Shlush, Nature 2017) implicates INKA1 as a regulator of LSC self-renewal and persistence. Indeed, INKA1-OE in cells derived from a primary human AML sample (8227) with a phenotypic and functional hierarchy (Lechman, Cancer Cell 2016) revealed a strong latency phenotype: In vitro and in vivo we observed label retention along with a steady increase in percentage of CD34+ cells, transient differentiation block, reduced growth rate, G0 accumulation and global reduction of H4K16ac.
In summary, our data implicates INKA1 as a gate-keeper of stem cell latency in normal human hematopoiesis and leukemia. Studying the detailed pathways involved will shed light upon the mechanisms involved in HSC activation and latency induction and will help to harness these for novel therapeutic approaches.
Takayanagi:Kyowa Hakko Kirin Co., Ltd.: Employment.
Asterisk with author names denotes non-ASH members.