Splicing Factor 3b Subunit 1 (SF3B1) Heterozygous Mice Manifest a Hematologic Phenotype Similar To Low Risk Myelodysplastic Syndromes With Ring Sideroblasts

The link between SF3B1 mutation and the ring sideroblast (RS) phenotype in myelodysplastic syndromes (MDS) was solidified by the identification of RS in Sf3b1 heterozygous (Sf3b1^+/-) mice. The identification of SF3B1 mutations in refractory anemia with RS (RARS) and RARS with thrombocytosis (RARS-T) showed the importance of RNA splicing in MDS biology. Furthermore, it opened the possibility of targeted therapy using spliceosome inhibitors in RARS/-T. However, many questions remain unanswered in linking SF3B1 dysfunction to MDS biology like the downstream targets of this gene. The identification of a robust murine model is essential to study a specific molecularly defined disease-phenotype and develop targeted therapies. We identified occasional RS in the bone marrow (BM) of Sf3b1^+/- which are rarely found in current mouse models of MDS (Beachy SH, Hematol Oncol Clin North Am, 2010). However, aside from RS in the BM no other MDS features were found. Sf3b1^+/- mice were originally engineered as a means to study the interaction between polycomb (PcG) genes and other protein-complexes (Isono K, Genes Dev, 2005). Homozygous Sf3b1^-/- mice died at the stage of pre-implantation of the embryos while Sf3b1^+/- appear healthy. Several tools have been tested to model MDS in genetically engineered mice targeting key genes in MDS. However, the creation of an ideal mouse model resembling distinct morphologic MDS subtypes is still lacking. To define a mouse model useful for preclinical therapeutic studies, we evaluated the hematologic features of Sf3b1^+/- and Sf3b1^+/+ mice during a long term follow-up. Five Sf3b1^+/- and 5 Sf3b1^+/+ mice were followed over time until 12 months of age. Blood was drawn from the retro-orbital vein every month starting from 6 months of age. Using two-sample Wilcoxon test we compared standard hematologic parameters finding a significant difference over the time between Sf3b1^+/- and Sf3b1^+/+: hemoglobin (g/dL) 6.9 ±0.73 vs 10.0 ±1.6 (P=0.008), red blood cells (M/uL) 8.3±0.5 vs 5.9±1.0 (P=0.008), platelets (K/uL) 731±105 vs 579±93 (P=0.008), and mean corpuscular volume (fL) 47.8±1.5 vs 45.1±1.1 (P=0.032). We did not detect any significant difference in other parameters although lymphocytes were more represented vs neutrophils, eosinophils and monocytes in Sf3b1^+/- vs Sf3b1^+/+ (6.3K/uL ±3.1 vs 5.8 ±1.8; P=1). Analysis of the BM, showed no difference in cell number between Sf3b1^+/- (n=7) and Sf3b1^+/+ (n=7) (44.1±9.1 vs 43.2 ±11; P=0.62). However, distinct dyserythropoiesis such as nuclear budding or irregular nuclei in Wright-Giemsa and occasional RS in Prussian blue stains were noted in Sf3b1^+/- which were not present in Sf3b1^+/+. In support of the iron overload seen in SF3B1 mutant patients (pts), a similar observation was made in Sf3b1^+/- by light microscopy and rhodamine based- flow cytometry to quantify mitochondrial iron (Visconte, Abstract #64897). We also characterized the transcriptome of Sf3b1^+/- and Sf3b1^+/+. Total RNA was isolated from BM of age/gender matched mice, polyA cDNA was prepared from 3ug of RNA and Mouse RNA-sequencing was run on Illumina HiSeq2000. 200 exons were found differentially used in Sf3b1^+/- vs Sf3b1^+/+. Chromosome 1 contains the highest number of genes with at least 1 exon alternatively used similar to what we observed in SF3B1 mutant pts. In total 22 genes showed stronger differential expression in Sf3b1^+/- vs Sf3b1^+/+. Sf3b1 was down-regulated as expected (MFC: 0.74) in Sf3b1^+/-. Studies in Sf3b1^+/- mice show that Sf3b1 protein physically interacts with Class II PcG proteins (PRC1) which are relevant in MDS. When we interrogated PcG genes and others, we found lower mRNA levels of ezh2 (MFC: 0.06) and npm1 and tpr53 (MFC: 0.01 and 0.28) and no difference in asxl1 and runx1 (MFC:1.22 and 1.1) in Sf3b1^+/- vs Sf3b1^+/+. Jak2, dock8, and uhrf2 showed significant (P=.0003) higher expression in Sf3b1^+/-. MDS is a heterogeneous disease characterized by genetic and non-genetic causes. Introduction of secondary events implicated in MDS pathogenesis can modify the phenotype of Sf3b1^+/- mice. In sum, Sf3b1^+/- mice after 6 months of follow-up developed macrocytic anemia, thrombocytosis, RS and dyserythropoiesis akin to human RARS/-T. Furthermore, transcriptome analysis shows exon usage/ gene expression changes similar to human SF3B1 mutants lending support that Sf3b1^+/- can serve as a mouse model for studying the biology of human low risk MDS specifically that of RARS/-T.