Splicing Factor 3b Subunit 1 (Sf3b1) Mediates Mitochondrial Iron Overload In Myelodysplastic Syndromes With Ring Sideroblasts By Alternative Splicing Of Mitoferrin-1 (Slc25a37)

R. Mahfouz,John Barnard,Ali Tabarroki,Li Zhang,Edy Hasrouni,Heesun J. Rogers,Nanthawan Avishai,Jonathan E. Cowen,Reza Sharghi-Moshtaghin,Midori Hitomi,Daniel J. Lindner,Arthur H. Heuer,Alan E. Lichtin,Anjali S. Advani,Eitan Fibach,Eliezer A. Rachmilewitz,Matt Kalaycio,Mikkael A. Sekeres,Yogen Saunthararajah,Ramon V. Tiu
DOI: https://doi.org/10.1182/blood.V122.21.1555.1555
IF: 20.3
2013-01-01
Blood
Abstract:Splicing factor 3b subunit 1, SF3B1 contributes to the formation of ring sideroblasts (RS) in Myelodysplastic syndromes (MDS). Abnormalities in iron trafficking have been implicated in the pathogenesis of refractory anemia with RS (RARS) and RARS with thrombocytosis (RARS-T). In RARS/-T, light microscopy and traditional iron profiles do not accurately reflect intracellular iron status. SF3B1 mutant (MUT, n=25) and wild type (WT, n=8) RARS patients (pts) have no difference in iron profiles (ferritin (ng/mL): 1244 ng/mL ± 926 vs 1215 ± 1065; total iron binding capacity (ug/dL): 252 ± 80 vs 234 ± 50). We previously reported distinct differences in iron distribution between SF3B1 MUT and WT pts based on transmission electron microscopy (EM). Coarse iron deposits are present in the mitochondria of SF3B1 MUT while smaller iron deposits are found in WT. We hypothesized that SF3B1 mutations affect distinct downstream targets leading to the iron phenotype differences in MUT vs WT. To study these differences, we performed a series of EM, flow cytometric (FC) and RNA-sequencing experiments. Quantification of iron by scoring deposits of iron/grid showed higher amount of mitochondrial iron in MUT vs WT by EM. We used a FC based-approach to quantify the cytoplasmic and mitochondrial iron utilizing the properties of two cell-permeant compounds (calcein-AM and rhodamine-B) being retained in the cytoplasm and the mitochondria, respectively. In keeping with our EM studies, we found that SF3B1 MUT have a higher % of rhodamine-B+ cells compared to WT RARS pts and healthy subjects (77 vs 19 vs 0.87%; n=6) indicating that MUT accumulates more mitochondrial iron compared to WT. Cytoplasmic iron was not different between MUT and WT (57 vs 43%) but higher than healthy subjects (1.5%; p=0.02), suggesting that MUT store more mitochondrial iron compared to WT pts. Sf3b1+/- mice also showed higher iron stores in the mitochondria rather than in the cytosplasm (85 vs 16%; n=4). To understand the factors leading to increased mitochondrial iron, we analyzed the transcriptome of BM cells derived from a homogeneous group of SF3B1 MUT (n=3), WT RARS (n=3), and healthy subjects (n=3). Total RNA (1.5-3ug) was subjected to RNA-sequencing using Illumina HiSeq2000. Twenty-million sequencing reads were interrogated. A comprehensive bioinformatic analysis was conducted on three levels (exon usage, gene expression, pathway analysis). Global gene expression analysis detected significant gene expression changes in 59 genes (FDR < 0.2) between MUT and healthy subjects. Mitochondrial genes linked to iron pathophysiology were well represented in the analysis. A total of 354 genes with exon usage/ gene expression difference in at least 1 exon were investigated. One of the top candidate genes showing alternative splicing is SLC25A37 (chr:8p21.2), a mitochondrial iron importer that mediates Fe2+ incorporation into the mitochondria. SLC25A37 was associated with a 2-4 fold higher expression in MUT vs WT and healthy individuals (mean base=4145 vs 2451 vs 1058). We also investigated other genes important in iron metabolism including SLC25A38, PUS1, and GLRX5 but no differences in both exon usage and gene expression were noted supporting the different nature of acquired and congenital sideroblastic anemia (SA). Some mitochondrial genes important in iron trafficking showed no changes in exon usage but exhibited differences in gene expression suggesting that they are pathways independent of SF3B1 mutations but still contributing to iron metabolism. ALAS2 was down-regulated in MUT vs WT RARS (fold change (FC): 0.39) while both MUT and WT pts exhibited an increased level compared to healthy subjects (FC: 5.0 and 2.0). ABCB7 was down-regulated compared to healthy subjects in both groups (FC:0.45 and 0.42). Using bisulfite sequencing we found that MUT have significantly higher hypermethylation of ALAS2 compared to WT RARS pts (FDR<.01; p<.00036) suggesting that epigenetic modification may explain the reduced ALAS2 expression instead of a splicing defect. SLC25A37 has been linked to an erythropoietic protoporphyria variant but not to congenital or acquired SA. In summary, SF3B1 MUT have increased mitochondrial iron as shown by EM and FC vs WT RARS pts. SF3B1 mutations mediate increased mitochondrial iron and RS formation by the alternative splicing of an iron transporter gene, SLC25A37, a novel pathway of iron-overload in the pathogenesis of MDS with RS.
What problem does this paper attempt to address?