CITED2 Modulates Differentiation and Proliferation of Normal and Leukemic Hematopoietic Stem and Progenitor Cells Downstream of PU.1

Abstract 922

Recently, it was demonstrated that the transcriptional co-activator CITED2 has a conserved role in the maintenance of normal adult hematopoiesis. Little is known regarding the regulation of CITED2, its expression levels in leukemic stem cells (LSCs) and whether CITED2 expression contributes to leukemogenesis. Using microarray data, we found variable CITED2 expression in ∼90% of sorted CD34⁺ AML cells (n=46). Q-PCR on 12 of these samples indicated that 50% of the leukemic samples displayed higher expression of CITED2 as compared to mobilized peripheral blood (PB) or cord blood (CB)-derived CD34⁺ hematopoietic stem and progenitor cells (HSCPs). To verify whether this expression of CITED2 in leukemic cells is functional, we performed RNAinterference on primary CD34⁺ cells from acute myeloid leukemia (AML) patients (n=9). AML samples with high CITED2 expression were susceptible to lentiviral downregulation of CITED2 as evidenced by the loss of transduced cells from long-term leukemic cultures.

To investigate a potential role of CITED2 in leukemogenesis, lentiviral gain-of-function experiments were performed with CB-derived CD34⁺ HSCPs. These experiments indicate that CITED2 expression increases cell numbers up to 5-fold in long-term culture-initiating cell (LTC-iC) experiments. This cell expansion on MS5-stromal cultures was paralleled by a short-term maintenance of progenitors. Colony-forming-cell (CFC) assays demonstrated a 3.5 fold higher output of CFC colonies compared to control cultures up to 2 weeks of MS-5 co-culture. CITED2-expressing colonies were larger than control colonies, verifying the increased cell numbers in LTC-iC experiments.

To further dissect the effects of CITED2, transduced CD34⁺CD38⁻ HSCs and CD34⁺CD38⁺ progenitors were sorted for single-cell liquid cultures and cell-divisions were followed for 6 days. Analysis of 360 single CD34⁺CD38⁻ HSCs indicated that CITED2 overexpression leads to an enhanced quiescence (cells with no division) and decreased proliferation (cells that divide). However, tracking the divisions of 360 single CD34⁺CD38⁺ progenitor cells demonstrated the opposite: Cells overexpressing CITED2 divided more than control cells. These data are consistent with a role for CITED2 in leukemic stem cells (LSCs), where LSCs are thought to be more quiescent than leukemic progenitors.

Since leukemias are also characterized by a differentiation block, we subsequently analyzed the differentiation of CB-derived CD34⁺ HSCPs upon overexpression of CITED2. In LTC-iC experiments, a shift in myelo-monocytic differentiation was observed. Enhanced CITED2 expression led to an increased percentage of CD15-positive cells (64%) at the expense of CD14-positive cells (9%) compared to control cultures (35% and 32% respectively). This bias towards granulocytic differentiation was also observed on May-Grunwald-Giemsa (MGG) stains and was furthermore confirmed in CFC assays. Furthermore, when analyzing erythroid differentiation, a clear reduction in CD71^bright GPA⁺ cells could be observed in CITED2 expressing cells, compared to control cells (2% vs. 12% respectively), which was confirmed by CFC assays and MGG stains.

Towards clarifying the regulation of CITED2, we scanned the CITED2 promoter for transcription factor binding sites and identified several PU.1 binding sites. Gene expression comparison between PU.1 and CITED2 in a panel of primary AML samples indicated that, apart from FAB M2 AMLs, PU.1 expression is inversely correlated with CITED2 expression. To functionally investigate this inverse correlation, we performed chromatin immunoprecipitations (ChIP) and demonstrated that PU.1 is indeed able to bind to the PU.1 binding sites in the CITED2 promoter of CB CD34⁺ cells. Subsequent overexpression of PU.1 in CB CD34⁺ HSPCs led to a 2-fold reduction in CITED2 expression.

Taken together, we propose a model in which PU.1 tightly regulates CITED2 expression during normal myeloid differentiation. In certain AML subsets, this model would predict that low PU.1 expression results in failure to lower CITED2 expression below a certain threshold, which subsequently results in maintenance of LSC quiescence. Furthermore, the enhanced CITED2 levels result in an increased proliferation of downstream leukemic progenitors, where together with low PU.1 levels the normal myeloid differentiation program is perturbed contributing to leukemic development.