SF3B1 Mutations Promote Pervasive Metabolic Reprogramming and Enhanced Sensitivity to Glycolysis Inhibition

Background

The splicing factor gene SF3B1 is frequently mutated in clonal hematopoiesis (CH) and detected in approximately 30% of myelodysplastic syndromes (MDS). SF3B1 mutations promote widespread aberrant splicing, but the consequences of many splicing alterations are poorly understood. Moreover, the mechanisms by which SF3B1 mutations drive clonal expansion and disease progression remain elusive, and there is a lack of effective targeted therapies for the treatment of SF3B1-mutant MDS.

Aims

In this study, we characterized the molecular and functional consequences of aberrant splicing in SF3B1-mutant cells. We used these findings to investigate the mechanisms underpinning the clonal expansion of SF3B1-mutant cells and to identify therapeutic vulnerabilities for the treatment of SF3B1-mutant MDS.

Methods

CRISPR-Cas9 base editing was used to generate SF3B1-K700E and SF3B1-K700K K562 cells and previously reported mis-splicing events in the former were validated using RT-PCR. For competitive proliferation assays, BFP- and mCherry-labelled cells were mixed in equal proportions and cultured in normoxia (20% O₂) or hypoxia (1% O₂). Short-read and long-read RNA-sequencing was performed on polyA-enriched RNA from untreated cells in 20% O₂ or 1% O₂ and analyzed using EdgeR, rMATS, DESeq2, bambu, and stringtie2. Purified protein from untreated cells was prepared for tandem mass tagging (TMT) proteomics, whilst metabolites were analyzed using liquid chromatography-mass spectrometry (LC-MS). Metabolic measurements were performed using the Seahorse MitoStress assay.

Results

We generated isogenic K562 cells expressing a heterozygous SF3B1-K700E mutation, the most frequently observed SF3B1 mutation in CH/MDS, alongside SF3B1-K700K silent mutation controls. We show that SF3B1-K700E, but not K700K, K562 cells replicate the characteristic splicing alterations observed in patients with these mutations. Combined analysis of RNA-sequencing and proteomics revealed a striking dysregulation of metabolism in SF3B1-K700E cells. In particular, we observed mis-splicing and subsequent protein downregulation of DLST and UQCC1, key components of the tricarboxylic acid (TCA) cycle and electron transport chain, respectively. In line with this, SF3B1-K700E cells exhibit a marked reduction in peak aerobic respiration and accumulation of TCA cycle intermediates such as 2-oxogluturate (2OG). High levels of 2OG can also be converted to 2-hydroxygluturate (2HG), an oncometabolite critical to the pathogenesis of IDH1/2-mutant leukemia. Notably, we observed an 8-fold increase in 2HG levels in SF3B1-K700E cells, suggesting a potential mechanism by which SF3B1 mutations may drive leukaemogenesis.

The reduced aerobic respiration in SF3B1-K700E cells was associated with upregulation of the hypoxia response. Competitive culture of K700E vs K700K cells revealed a disadvantage of the former in normoxia, but an advantage under low oxygen tension, which better reflects conditions prevalent within the human bone marrow. Additionally, the increased reliance on glycolysis in SF3B1-mutant cells conferred enhanced sensitivity to glycolysis inhibitors such as 2-deoxyglucose (2DG).

Conclusion

We report that mis-splicing of key metabolic genes by SF3B1-K700E is associated with reduced aerobic respiration, 2HG accumulation and activation of the hypoxia response. This was associated with enhanced growth of K700E vs K700K cells in low oxygen tension levels equivalent to those prevailing in human bone marrow. We go on to show that glycolysis inhibition represents a potential therapeutic vulnerability of SF3B1-mutant cells which could be exploited in the treatment of SF3B1-mutant MDS.

Vassiliou:STRM.BIO: Consultancy; AstraZeneca: Research Funding.