Initial analysis of the phenotypic consequences of SF3B1 K700E mutation expression in vitro and in vivo

Background: Myelodysplastic syndromes (MDS) are clonal hematopoietic stem cell disorders characterized by dysplastic hematopoiesis in the bone mar- row and peripheral blood cytopenias. The splicing factor SF3B1 is the most commonly mutated gene in MDS and is mainly found in refractory anemia with ring sideroblasts (RARS). SF3B1 mutations are missense, heterozygous and clustered, suggesting they are gain-of-function with a causal link to the devel- opment of RARS. Nevertheless, in vitro studies are hampered by the inability to clone the full length SF3B1 cDNA and model the effect of its mutation in cell lines. Few reports on Sf3b1 knock-down mice have not convincingly recapitu- lated the features of its somatic mutations in RARS.
Aims: We therefore targeted the Sf3b1 locus of mouse embryonic stem cells (ESC) to create a conditional allele encoding the most common mutation Sf3b1K700E to assess the phenotypic consequences of its expression in vitro and in vivo. Methods: The C57BL/6N ESC line JM8 was targeted by homologous recom- bination. Albino pseudopregnant females were injected allowing for chimera screening based on coat color. One mouse carrying germline transmission was selected and its heterozygous offspring was crossed to Tg(Mx1:Cre) mice. Sf3b1K700E expression was then elicited in the bone marrow with pIpC injection to activate the Cre recombinase. In parallel, Cre was overexpressed in ESC to generate clones with constitutive Sf3b1K700E expression to perform in vitro work. WT and mutant ES cells were analyzed in basal conditions and after dif- ferentiation into haematopoietic cells. For the latter procedure, cells were allowed to grow into embryoid bodies that were subsequently dissociated with trypsin and cultured in a hematopoietic cytokines mix. Results: WT and mutant Sf3b1 ESC showed proliferation rates and colony morphologies similar to native ESC. Upon differentiation, >80% of cells expressed the pan-hematopoietic marker Cd45, with no difference between genotypes. Erythroid differentiation in particular was efficient as confirmed by an increase in transcription levels of the globin genes, that was reduced in Sf3b1 mutant cells. Nevertheless, gross morphology was similar in WT and mutant cells. Flow cytometry showed a marginal decrease in the percentage of Cd71low/Ter119hi cells for the mutant genotype, but no major differences in the frequency of expression markers of stem and progenitor, myeloid and megakaryocytic cells. By unsupervised clustering of gene expression profiles, we found that WT and mutant cells cluster closely together in the undifferenti- ated state, whereas differentiation causes mutant cells to acquire a dramatically different expression profile than WT cells. Mutant differentiated cells showed an excess of downregulated genes, among which known players in congenital sideroblastic anemias (Abcb7, Glrx5, and Scl25A38). Gene ontology analysis showed downregulation of several biological processes, including the respira- tory electron transport chain and regulators of mitochondrial ion transport as shown previously, but also processes involved in RNA metabolism like trans- lation initiation and pre-mRNA splicing. Initial analysis of conditional Sf3b1K700E expression in mice showed that heterozygous animals display no overt sign of disease at our median follow-up of 116 days. In turn, peripheral blood counts post-pIpC injection show that hemoglobin is significantly decreased at one (17.2 vs 14.2 gr/dl, p=0.001) and two months (16.7 vs 15, p=0.006) in mutant animals, with no difference in white cell and platelet counts. Analysis of bone marrow cells by flow cytometry showed a 2-fold decrease in Cd71low/Ter119hi (p=0.02) cells for the mutant genotype, suggestive of a late maturation defect. Lastly, iron stain of bone marrow cells showed an increment of iron-laden macrophages and occasional ringed sideroblasts in the mutant animals. Summary and Conclusions: Our data closely recapitulates initial observations in human samples and suggests mutated Sf3b1 expression leads to defective erythroid maturation. Our mouse model will allow better characterization of the molecular pathways altered by the mutation and the analysis of modifiers of this phenotype.