Direct RNA Target Analysis Identifies Signatures That Define Msi2's Role in Sustaining HSC Cell Fate.

Abstract 2324

Hematopoietic stem cells (HSC) must maintain normal cell fate decisions between symmetric and asymmetric divisions as alterations can lead to hematopoietic malignancies. The MSI2 RNA binding protein is upregulated in patients with a poor clinical prognosis in acute myeloid leukemia and in the blast crisis phase of chronic myelogenous leukemia. The related RNA-binding protein Msi1 has been shown to block translation of specific target mRNAs by interacting with the 3'UTR. To understand the role of Msi2 in both normal and leukemic contexts, we are characterizing mice with a conditional deletion of Msi2 in the hematopoietic compartment. Msi2 conditional knockouts have reduced overall number of HSCs in the bone marrow and have decreased engraftment capacity in congenic transplants, however the molecular mechanisms through which Msi2 elicits these effects remains unclear.

To understand how Msi2 alters HSC self renewal, we utilized Mx1-Cre::Msi2^flox/flox conditional mice, enabling Msi2 inactivation via poly(I):poly(C) injection. We first tested if Msi2 deleted HSCs or MPPs had an intrinsic defect in cell fate decision and proliferation. We performed proliferation and colony assays on sorted HSCs and MPPs and found reduced cellular numbers in both the HSCs and MPPs with an increase in the frequency of more differentiated cells based on Mac/Gr1 staining after seven days. These reduced overall cell numbers in vitro may have resulted from a defect in maintaining the stem cell population or a decrease in proliferative capacity. To assess if there was a defect in the initial HSC cell division, we examined Numb protein levels and distribution as a surrogate readout for asymmetric cell division. Numb mRNA is a known target of Msi binding and translational repression. Surprisingly, in the Msi2 null HSC or MPPs we observed no difference in the overall staining of Numb indicating that Msi2 deficiency does not globally alter Numb levels. However, we did observe a decrease in the percentage of cells that underwent asymmetric Numb segregation in the MPPs and an increase in the percentage of cells that showed increased Numb staining in some of the daughter pairs in both HSCs and MPPs, indicating increased commitment away from the hematopoietic stem and progenitors.

Although controversial, Notch signaling has been implicated in self renewal of HSCs and as a critical downstream target of the Msi family through Numb inactivation. However, the Notch signaling pathway was not noted to be significantly altered in gene set analysis from microarrays performed on Msi2 deleted HSC enriched populations. Moreover using quantitative PCR for Notch target genes in sorted Lin^low, c-kit⁺ and Sca⁺(LSK) cells we found no statistical difference in the expression of Notch targets (Notch1, Notch2, Hes1 or Myc). These interesting negative results prompted us take a more global approach in characterizing the direct targets of Msi2 in hematopoietic cells.

Due to the requirement for a large number of cells, we utilized K562 cells overexpressing a FLAG-tagged version of MSI2 to identify its direct RNA-binding targets. Using recently developed technologies, we performed experiments with UV-cross-linked and immunoprecipitated MSI2 which was then followed by RNA-sequencing (HITS-CLIP) to identify the global direct binding. Our analysis characterized the distribution of binding across the genome. Additionally, gene set enrichment analysis (GSEA) indicated a positive correlation of genes that were upregulated in the Msi2 deleted LSKs and the top 3-fold bound RNA targets (2,713 genes). Using the entire RNA target list we identified gene set signatures including “Cell Cycle”, “Self-renewal” and “HSC to CMP” that matched our results in the microarray from Msi2 deleted LSKs.

In summary, direct RNA target analysis for MSI2 in human leukemia cells overlapped with self renewal and differentiation gene sets in mouse HSC enriched populations and correlated with the phenotypes we observed in isolated HSCs grown in vitro lacking Msi2. These results uncover the complexity of MSI2 RNA binding targets and have important implications for both normal and leukemic stem cell biology.