RBPs in Cancer

Splicing factors, Aberrant interactions, & modeling

  • Splicing factor (SF) gene mutations are the most common class of mutations in myelodysplastic syndrome (MDS), but convergent downstream disease drivers are still elusive. In co-corresponding work with Eirini Papapetrou’s group at Mount Sinai, we made an unexpected finding that unifies the mechanism by which SFs SRSF2 and U2AF1 mutations drive oncogenesis which has led us to identify FDA-approved small molecules (MEK inhibitors) as a attractive therapeutic option for MDS and other SF-mutant neoplasms. Papapetrou’s group generated isogenic IPSC lines harboring mutations in these SFs and 3X FLAG tagged the mutant and wild-type allele in the isogenic series and differentiated these into hematopoietic stem/progenitor cells. Emily Wheeler, graduate student in my lab conducted allele-specific eCLIP, RNA-seq data generation, and performed the integrative analysis with patient- specific data to identify a long isoform of GNAS (GNAS-L) that is a new driver of MDS. GNAS-L encodes a hyperactive long form of the stimulatory G protein alpha subunit that activates ERK/MAPK signaling. SF-mutant MDS cells have activated ERK signaling, are sensitive to MEK inhibitors (Wheeler et al. Cancer Discovery, 2021). My lab is continuing to identify novel ways to regulate the splicing of GNAS-L and to understand how GNAS-L as an example of a powerful and therapeutically relevant intersection between splicing and signaling pathways. This is leading us to a new area of research.

  • Aberrant RNA binding protein-RNA interactions can promote cancer progression. My graduate student Jackie Einstein interrogated the function of RBPs in cancer using pooled CRISPR-Cas9 screening to identify 57 RBP candidates with distinct roles in supporting MYC-driven oncogenic pathways. We found that disrupting YTHDF2- dependent mRNA degradation triggers apoptosis in triple-negative breast cancer (TNBC) cells and tumors. eCLIP and m6a sequencing indicate that YTDHF2 interacts with mRNAs encoding proteins in the MAPK pathway that when stabilized induced global translation rates. We then utilized our single cell RiboSTAMP profiling of translating mRNAs reveals alterations in the translatome of single ells within YTHDF2-depleted solid tumors which contributes to stress-induced apoptosis in TNBC. Our work highlights the critical roles that RBPs have in cancer (Einstein et al. Molecular Cell 2021). This was amplified in Cancer Discovery, and a preview in Molecular Cell by Anne Willis at Cambridge, UK. In parallel, my lab has developed computational methods to directly identify m6a sites from long-read nanopore sequencing (Lorenz, RNA, 2020). We hope to combine long-read sequencing with RNA analysis and more synthetic lethality screens to identify known and new RBPs and their RNA substrates as new targets for a variety of different cancers. We have also contributed our RBP CRISPR library to work with Dong-er Zhang at UCSD on identifying regulators of RUNX1 isoform generation (Davis et al., Blood Advances 2021).

  • Cellular models aimed at understanding cancer biology do not recapitulate pathobiology including tumor heterogeneity, which is an inherent feature of cancer that underlies treatment resistance. In co-corresponding work with Frank Furnari at UCSD, we developed cancer “avatars” using genetically engineered human IPSCs to contain tumor-associated genetic driver mutations revealed by TCGA for glioblastoma that are differentiated into neural progenitors that are then orthotopic engraftment into mice. My postdoc Markmiller had introduced the genome engineering strategy and transferred that to Frank Furnari’s lab early in the project. My postdoc Alex Chaim analyzed these tumor avatars using single cell sequencing and analysis to show that these models harbor inter-tumor heterogeneity and extrachromosomal DNA amplification (Koga, Chaim et al., Nature Communications 202056). Furnari and I are continuing to collaborate on building a plethora of new avatar models and have submitted an R01 for this work.

Team Members

Learn more about our ALS and Stress Granule Projects

To understand and identify ways to alter RNA granule biology, we published our landmark paper using proximity labeling to identify ~150 new proteins within SGs, making the surprising finding that granules are preassembled prior to stress. We revealed components necessary in the formation, stability, and resolution of SGs in ALS-patient iPSC- derived neurons. Excitingly, their depletion suppressed toxicity in animal models (Markmiller et al., Cell 2018)

We have continued to develop means to understand and control the formation and abundance of stress granules. Stress granule formation is frequently accompanied by ubiquitin proteasome system (UPS) impairment and ubiquitylated protein accumulation. In co-corresponding work with Eric Bennett’s lab at UCSD, my postdoc Markmiller tackled the unaddressed relationship between the ubiquitin pathway and SGs. Using pharmacological inhibition of either the ubiquitin or NEDD-activating enzyme (UAE or NAE), we show that UAE inhibition results in rapid loss of global protein ubiquitylation using ubiquitin-specific proteomics. Very interestingly, inhibiting neither UAE nor NAE affected SG formation or disassembly, indicating that active protein ubiquitylation or neddylation is dispensable for SG dynamics. Finally, we show that SGs co-localize primarily with unconjugated ubiquitin rather than polyubiquitylated proteins and suggest that free ubiquitin may alter SG protein interactions (Markmiller et al., Cell Reports 2019).

In 2019, my graduate student Mark Fang published a phenotypic screen using a library of >6,000 compounds to identify pathways that impinge on RNA granule dynamics and discovered a class of small molecules that interrupt the recruitment of proteins into pre-formed RNA granules. Excitingly, several of these compounds extended survival of ALS primary neurons to that of healthy neurons. As part of this study, we also demonstrated that transient, non- lethal cellular stress followed by a recovery period can induce the formation of persistent cytoplasmic TDP-43 aggregates in ALS patient-derived PSC-MNs. These persistent aggregates do not contain canonical SG markers such as G3BP1 and do not frequently occur in unstressed PSC-MNs, demonstrating that transient stress can be used to robustly induce an aging-dependent phenotype in an ALS mutant-specific context (Fang et al., Neuron 2019).

In 2019, my postdoc Markmiller used the transient stress paradigm in combination with a global mRNA localization profiling approach to measure subcellular mRNA distribution landscapes across different cell types and in response to cellular stress in the absence or presence of ALS-associated mutations. We show that the global mRNA localization landscape is shaped by a set of cell type-independent rules based on basic transcript architecture and sequence features and are profoundly disrupted by cellular stress. Remarkably, while stress-induced mRNA localization changes are transitory in wildtype PSC-MNs, HNRNPA2B1-mutant PSC-MNs retain a drastically reshaped mRNA localization profile and display elevated cell death 72h post-stress. Our single, non-toxic stressor also directly induces disease- relevant molecular changes such as splicing-mediated loss of Stathmin-2 (Markmiller et al., Cell Reports 2021).

In an effort to identify novel genetic regulators of stress granules, we developed CRaft-ID, the first pooled CRISPR screening approach using standard imaging readouts that utilizes microRaft arrays, and identified unexpected SG- regulatory proteins. We found that RNA binding proteins that control RNA processing at multiple layers appear to control SG formation and/or abundance, including surprisingly, a nuclear protein snRNP200 involved in splicing (Wheeler et al., Nature Methods 2020). Our ongoing and future work will continue to study some of these proteins such as snRNP200 to understand mechanistically how depletion of these affect cytoplasmic initiation and recovery of SGs. We also continue to unravel dysfunctional RBP-RNA relationships in neurodegeneration.