Unraveling the Drivers of the Stress Granule Signature in Splicing Factor-Mutant Myeloid Malignancies

Splicing factor (SF) mutations are a hallmark of myeloid malignancies, but the limited efficacy of current splicing modulators highlights gaps in our mechanistic understanding of how SF mutations alter cellular functions. We recently applied sophisticated multi-omics and functional validation to show that U2AF1 mutations, through RNA binding and splicing alterations, exploit ribonucleoprotein granules, condensates of RNAs and RNA-binding proteins, to mount an enhanced stress response. Specifically, U2AF1 mutations in myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML) increase the availability of stress granule (SG) components leading to better cell survival under oxidative stress [Biancon G, et al. Mol Cell 2022].

To gain further insight into stress response and other mechanisms sustaining the clonal advantage of SF-mutant cells, we performed single-cell RNA-seq on MDS patient samples without SF mutations (n=5), with U2AF1 S34F mutation (n=3), or with SRSF2 P95H mutation (n=2) that carries an equally poor prognosis as U2AF1-mutant disease. Single-cell analysis through our newly developed tool, the Cell Marker Accordion [https://doi.org/10.1101/2024.03.08.584053], revealed differences in the MDS architecture. In comparison to SF-wildtype patients, both U2AF1- and SRSF2-mutant patients are characterized by an increase in hematopoietic multipotent progenitors, common lymphoid progenitors and monocytes, and by a decrease in megakaryocyte-erythroid progenitors. With the integration of single-cell mutation calling, we classified each cell from U2AF1 S34F patients as either WT or S34F and each cell from SRSF2 P95H patients as either WT or P95H. Of note, cell types with a reduced number of cells, such as megakaryocyte-erythroid progenitors in U2AF1-mutant patients and myeloid cells in SRSF2-mutant patients, have high mutation frequencies indicating the clonal advantage and differentiation blockade conferred by SF mutations. We then performed differential gene expression and gene set enrichment analyses comparing S34F or P95H cells vs cells from SF-wildtype patients. Considering the hematopoietic multipotent progenitors' population, we found 345 deregulated genes in S34F cells and 1071 deregulated genes in P95H cells. Functionally, these genes are involved in immune- and inflammatory-related pathways. S34F and P95H cells are also characterized by significant upregulation of SG-associated genes. The top 10 genes driving the high SG signature in hematopoietic multipotent progenitors are specifically involved in the regulation of translation, microtubules, transcription and cell growth.

We then aimed at identifying the main proteins and RNAs that drive the enhanced SG formation in SF-mutant conditions. We used U2AF1-mutant and wildtype AML cell lines, with or without arsenite stress, and we isolated SG cores by biochemical purification followed by mass spectrometry and RNA-seq. Mass spectrometry analysis identified 37 known SG proteins that are enriched into SG cores in S34F vs WT cells under arsenite stress. Proteins inside SGs are RNA-binding proteins that are fundamental for SG maturation but the sequestration into SGs affects their subcellular localization and therefore their availability and function. Main SG proteins in S34F cells include ribosomal proteins and translation initiation factors, involved in protein synthesis, and RNA helicases involved in RNA metabolism. RNA-seq analysis identified 1299 genes that are differentially expressed into SG cores in S34F vs WT cells under arsenite stress and, most importantly, 1501 genes that are differentially expressed without arsenite stress, suggesting that SF mutations alter the transcriptome to form pre-stress seeds and prepare the cell to cope with stress stimuli. Finally, intersecting RNAs that are enriched into SG cores (vs total cytoplasm) in S34F (vs WT) cells, with differential splicing data and published SG transcriptome data, we identified: under arsenite stress, 38 genes mainly involved in signal transduction; without arsenite stress, 60 genes mainly involved in post-translational modifications.Collectively, our multi-omics approach at single-cell and subcellular resolution has the potential to shed light on the mechanism behind enhanced stress response and improved cell fitness of SF-mutant cells in MDS and AML, providing novel therapeutic strategies.

No relevant conflicts of interest to declare.