In silico RNA isoform screening to identify potential cancer driver exons with therapeutic applications.

Alternative splicing is crucial for cancer progression and can be targeted pharmacologically, yet identifying driver exons genome-wide remains challenging. We propose identifying such exons by associating statistically gene-level cancer dependencies from knockdown viability screens with splicing profiles and gene expression. Our models predict the effects of splicing perturbations on cell proliferation from transcriptomic data, enabling in silico RNA screening and prioritizing targets for splicing-based therapies. We identified 1,073 exons impacting cell proliferation, many from genes not previously linked to cancer. Experimental validation confirms their influence on proliferation, especially in highly proliferative cancer cell lines. Integrating pharmacological screens with splicing dependencies highlights the potential driver exons affecting drug sensitivity. Our models also allow predicting treatment outcomes from tumor transcriptomes, suggesting applications in precision oncology. This study presents an approach to identifying cancer driver exon and their therapeutic potential, emphasizing alternative splicing as a cancer target.

Genomic deletions explain the generation of alternative BRAF isoforms conferring resistance to MAPK inhibitors in melanoma.

Resistance to MAPK inhibitors (MAPKi), the main cause of relapse in BRAF-mutant melanoma, is associated with the production of alternative BRAF mRNA isoforms (altBRAFs) in up to 30% of patients receiving BRAF inhibitor monotherapy. These altBRAFs have been described as being generated by alternative pre-mRNA splicing, and splicing modulation has been proposed as a therapeutic strategy to overcome resistance. In contrast, we report that altBRAFs are generated through genomic deletions. Using different in vitro models of altBRAF-mediated melanoma resistance, we demonstrate the production of altBRAFs exclusively from the BRAF V600E allele, correlating with corresponding genomic deletions. Genomic deletions are also detected in tumor samples from melanoma and breast cancer patients expressing altBRAFs. Along with the identification of altBRAFs in BRAF wild-type and in MAPKi-naive melanoma samples, our results represent a major shift in our understanding of mechanisms leading to the generation of BRAF transcripts variants associated with resistance in melanoma.

Alternative splicing decouples local from global PRC2 activity.

The Polycomb repressive complex 2 (PRC2) mediates epigenetic maintenance of gene silencing in eukaryotes via methylation of histone H3 at lysine 27 (H3K27). Accessory factors define two distinct subtypes, PRC2.1 and PRC2.2, with different actions and chromatin-targeting mechanisms. The mechanisms orchestrating PRC2 assembly are not fully understood. Here, we report that alternative splicing (AS) of PRC2 core component SUZ12 generates an uncharacterized isoform SUZ12-S, which co-exists with the canonical SUZ12-L isoform in virtually all tissues and developmental stages. SUZ12-S drives PRC2.1 formation and favors PRC2 dimerization. While SUZ12-S is necessary and sufficient for the repression of target genes via promoter-proximal H3K27me3 deposition, SUZ12-L maintains global H3K27 methylation levels. Mouse embryonic stem cells (ESCs) lacking either isoform exit pluripotency more slowly and fail to acquire neuronal cell identity. Our findings reveal a physiological mechanism regulating PRC2 assembly and higher-order interactions in eutherians, with impacts on H3K27 methylation and gene repression.

SF3B1 mutation-mediated sensitization to H3B-8800 splicing inhibitor in chronic lymphocytic leukemia.

Splicing factor 3B subunit 1 (SF3B1) is involved in pre-mRNA branch site recognition and is the target of antitumor-splicing inhibitors. Mutations in SF3B1 are observed in 15% of patients with chronic lymphocytic leukemia (CLL) and are associated with poor prognosis, but their pathogenic mechanisms remain poorly understood. Using deep RNA-sequencing data from 298 CLL tumor samples and isogenic SF3B1 WT and K700E-mutated CLL cell lines, we characterize targets and pre-mRNA sequence features associated with the selection of cryptic 3' splice sites upon SF3B1 mutation, including an event in the MAP3K7 gene relevant for activation of NF-κB signaling. Using the H3B-8800 splicing modulator, we show, for the first time in CLL, cytotoxic effects in vitro in primary CLL samples and in SF3B1-mutated isogenic CLL cell lines, accompanied by major splicing changes and delayed leukemic infiltration in a CLL xenotransplant mouse model. H3B-8800 displayed preferential lethality towards SF3B1-mutated cells and synergism with the BCL2 inhibitor venetoclax, supporting the potential use of SF3B1 inhibitors as a novel therapeutic strategy in CLL.

Structural basis for specific RNA recognition by the alternative splicing factor RBM5.

The RNA-binding motif protein RBM5 belongs to a family of multi-domain RNA binding proteins that regulate alternative splicing of genes important for apoptosis and cell proliferation and have been implicated in cancer. RBM5 harbors structural modules for RNA recognition, such as RRM domains and a Zn finger, and protein-protein interactions such as an OCRE domain. Here, we characterize binding of the RBM5 RRM1-ZnF1-RRM2 domains to cis-regulatory RNA elements. A structure of the RRM1-ZnF1 region in complex with RNA shows how the tandem domains cooperate to sandwich target RNA and specifically recognize a GG dinucleotide in a non-canonical fashion. While the RRM1-ZnF1 domains act as a single structural module, RRM2 is connected by a flexible linker and tumbles independently. However, all three domains participate in RNA binding and adopt a closed architecture upon RNA binding. Our data highlight how cooperativity and conformational modularity of multiple RNA binding domains enable the recognition of distinct RNA motifs, thereby contributing to the regulation of alternative splicing. Remarkably, we observe surprising differences in coupling of the RNA binding domains between the closely related homologs RBM5 and RBM10.

Insplico: effective computational tool for studying splicing order of adjacent introns genome-wide with short and long RNA-seq reads.

Although splicing occurs largely co-transcriptionally, the order by which introns are removed does not necessarily follow the order in which they are transcribed. Whereas several genomic features are known to influence whether or not an intron is spliced before its downstream neighbor, multiple questions related to adjacent introns' splicing order (AISO) remain unanswered. Here, we present Insplico, the first standalone software for quantifying AISO that works with both short and long read sequencing technologies. We first demonstrate its applicability and effectiveness using simulated reads and by recapitulating previously reported AISO patterns, which unveiled overlooked biases associated with long read sequencing. We next show that AISO around individual exons is remarkably constant across cell and tissue types and even upon major spliceosomal disruption, and it is evolutionarily conserved between human and mouse brains. We also establish a set of universal features associated with AISO patterns across various animal and plant species. Finally, we used Insplico to investigate AISO in the context of tissue-specific exons, particularly focusing on SRRM4-dependent microexons. We found that the majority of such microexons have non-canonical AISO, in which the downstream intron is spliced first, and we suggest two potential modes of SRRM4 regulation of microexons related to their AISO and various splicing-related features. Insplico is available on gitlab.com/aghr/insplico.

Pancreatic microexons regulate islet function and glucose homeostasis.

Pancreatic islets control glucose homeostasis by the balanced secretion of insulin and other hormones, and their abnormal function causes diabetes or hypoglycaemia. Here we uncover a conserved programme of alternative microexons included in mRNAs of islet cells, particularly in genes involved in vesicle transport and exocytosis. Islet microexons (IsletMICs) are regulated by the RNA binding protein SRRM3 and represent a subset of the larger neural programme that are particularly sensitive to SRRM3 levels. Both SRRM3 and IsletMICs are induced by elevated glucose levels, and depletion of SRRM3 in human and rat beta cell lines and mouse islets, or repression of particular IsletMICs using antisense oligonucleotides, leads to inappropriate insulin secretion. Consistently, mice harbouring mutations in Srrm3 display defects in islet cell identity and function, leading to hyperinsulinaemic hypoglycaemia. Importantly, human genetic variants that influence SRRM3 expression and IsletMIC inclusion in islets are associated with fasting glucose variation and type 2 diabetes risk. Taken together, our data identify a conserved microexon programme that regulates glucose homeostasis.

Specialization of the photoreceptor transcriptome by Srrm3-dependent microexons is required for outer segment maintenance and vision.

Retinal photoreceptors have a distinct transcriptomic profile compared to other neuronal subtypes, likely reflecting their unique cellular morphology and function in the detection of light stimuli by way of the ciliary outer segment. We discovered a layer of this molecular specialization by revealing that the vertebrate retina expresses the largest number of tissue-enriched microexons of all tissue types. A subset of these microexons is included exclusively in photoreceptor transcripts, particularly in genes involved in cilia biogenesis and vesicle-mediated transport. This microexon program is regulated by Srrm3, a paralog of the neural microexon regulator Srrm4. Despite the fact that both proteins positively regulate retina microexons in vitro, only Srrm3 is highly expressed in mature photoreceptors. Its deletion in zebrafish results in widespread down-regulation of microexon inclusion from early developmental stages, followed by other transcriptomic alterations, severe photoreceptor defects, and blindness. These results shed light on the transcriptomic specialization and functionality of photoreceptors, uncovering unique cell type-specific roles for Srrm3 and microexons with implications for retinal diseases.

A developmentally programmed splicing failure contributes to DNA damage response attenuation during mammalian zygotic genome activation.

Transition from maternal to embryonic transcriptional control is crucial for embryogenesis. However, alternative splicing regulation during this process remains understudied. Using transcriptomic data from human, mouse, and cow preimplantation development, we show that the stage of zygotic genome activation (ZGA) exhibits the highest levels of exon skipping diversity reported for any cell or tissue type. Much of this exon skipping is temporary, leads to disruptive noncanonical isoforms, and occurs in genes enriched for DNA damage response in the three species. Two core spliceosomal components, Snrpb and Snrpd2, regulate these patterns. These genes have low maternal expression at ZGA and increase sharply thereafter. Microinjection of Snrpb/d2 messenger RNA into mouse zygotes reduces the levels of exon skipping at ZGA and leads to increased p53-mediated DNA damage response. We propose that mammalian embryos undergo an evolutionarily conserved, developmentally programmed splicing failure at ZGA that contributes to the attenuation of cellular responses to DNA damage.

Parallel evolution of a splicing program controlling neuronal excitability in flies and mammals.

Alternative splicing increases neuronal transcriptomic complexity throughout animal phylogeny. To delve into the mechanisms controlling the assembly and evolution of this regulatory layer, we characterized the neuronal microexon program in Drosophila and compared it with that of mammals. In nonvertebrate bilaterians, this splicing program is restricted to neurons by the posttranscriptional processing of the enhancer of microexons (eMIC) domain in Srrm234. In Drosophila, this processing is dependent on regulation by Elav/Fne. eMIC deficiency or misexpression leads to widespread neurological alterations largely emerging from impaired neuronal activity, as revealed by a combination of neuronal imaging experiments and cell type-specific rescues. These defects are associated with the genome-wide skipping of short neural exons, which are strongly enriched in ion channels. We found no overlap of eMIC-regulated exons between flies and mice, illustrating how ancient posttranscriptional programs can evolve independently in different phyla to affect distinct cellular modules while maintaining cell-type specificity.

Human EWS-FLI protein recapitulates in Drosophila the neomorphic functions that induce Ewing sarcoma tumorigenesis.

Ewing sarcoma (EwS) is a human malignant tumor typically driven by the Ewing sarcoma-Friend leukemia integration (EWS-FLI) fusion protein. A paucity of genetically modified animal models, partially owed to the high toxicity of EWS-FLI, hinders research on EwS. Here, we report a spontaneous mutant variant, EWS-FLI1FS, that circumvents the toxicity issue in Drosophila. Through proteomic and genomic analyses, we show that human EWS-FLI1FS interacts with the Drosophila homologues of EWS-FLI human protein partners, including core subunits of chromatin remodeling complexes, the transcription machinery, and the spliceosome; brings about a massive dysregulation of transcription that affects a significant fraction of known targets of EWS-FLI in human cells; and modulates splicing. We also show that EWS-FLI1FS performs in Drosophila the two major neomorphic activities that it is known to have in human cells: activation of transcription from GGAA microsatellites and out competition of ETS transcription factors. We conclude that EWS-FLI1FS reproduces in Drosophila the known oncogenic activities of EWS-FLI that drive EwS tumorigenesis in humans. These results open up an unprecedented opportunity to investigate EWS-FLI's oncogenic pathways in vivo in a genetically tractable organism.

Roles and mechanisms of alternative splicing in cancer - implications for care.

Removal of introns from messenger RNA precursors (pre-mRNA splicing) is an essential step for the expression of most eukaryotic genes. Alternative splicing enables the regulated generation of multiple mRNA and protein products from a single gene. Cancer cells have general as well as cancer type-specific and subtype-specific alterations in the splicing process that can have prognostic value and contribute to every hallmark of cancer progression, including cancer immune responses. These splicing alterations are often linked to the occurrence of cancer driver mutations in genes encoding either core components or regulators of the splicing machinery. Of therapeutic relevance, the transcriptomic landscape of cancer cells makes them particularly vulnerable to pharmacological inhibition of splicing. Small-molecule splicing modulators are currently in clinical trials and, in addition to splice site-switching antisense oligonucleotides, offer the promise of novel and personalized approaches to cancer treatment.

Site-Specific mRNA Cleavage for Selective and Quantitative Profiling of Alternative Splicing with Label-Free Optical Biosensors.

Alternative splicing of mRNA precursors is a key process in gene regulation, contributing to the diversity of proteomes by the alternative selection of exonic sequences. Alterations in this mechanism are associated with most cancers, enhancing their proliferation and survival, and can be employed as cancer biomarkers. Label-free optical biosensors are ideal tools for the highly sensitive and label-free analysis of nucleic acids. However, their application for alternative splicing analysis has been hampered due to the formation of complex and intricate long-range base-pairing interactions which make the direct detection in mRNA isoforms difficult. To solve this bottleneck, we introduce a methodology for the generation of length-controlled RNA fragments from purified total RNA, which can be easily detected by the biosensor. The methodology seizes RNase H enzyme activity to degrade the upstream and downstream RNA segments flanking the target sequence upon hybridization to specific DNA oligos. It allows the fast and direct monitoring of Fas gene alternative splicing in real time, employing a surface plasmon resonance biosensor. We demonstrate the selective and specific detection of mRNA fragments in the pM-nM concentration range, reducing quantification errors and showing 81% accuracy when compared to RT-qPCR. The site-specific cleavage outperformed random RNA hydrolysis by increasing the detection accuracy by 20%, making this methodology particularly appropriate for label-free quantification of alternative splicing events in complex samples.

A novel protein domain in an ancestral splicing factor drove the evolution of neural microexons.

The mechanisms by which entire programmes of gene regulation emerged during evolution are poorly understood. Neuronal microexons represent the most conserved class of alternative splicing in vertebrates, and are critical for proper brain development and function. Here, we discover neural microexon programmes in non-vertebrate species and trace their origin to bilaterian ancestors through the emergence of a previously uncharacterized 'enhancer of microexons' (eMIC) protein domain. The eMIC domain originated as an alternative, neural-enriched splice isoform of the pan-eukaryotic Srrm2/SRm300 splicing factor gene, and subsequently became fixed in the vertebrate and neuronal-specific splicing regulator Srrm4/nSR100 and its paralogue Srrm3. Remarkably, the eMIC domain is necessary and sufficient for microexon splicing, and functions by interacting with the earliest components required for exon recognition. The emergence of a novel domain with restricted expression in the nervous system thus resulted in the evolution of splicing programmes that qualitatively expanded the neuronal molecular complexity in bilaterians.

Somatic alterations compromised molecular diagnosis of DOCK8 hyper-IgE syndrome caused by a novel intronic splice site mutation.

In hyper-IgE syndromes (HIES), a group of primary immunodeficiencies clinically overlapping with atopic dermatitis, early diagnosis is crucial to initiate appropriate therapy and prevent irreversible complications. Identification of underlying gene defects such as in DOCK8 and STAT3 and corresponding molecular testing has improved diagnosis. Yet, in a child and her newborn sibling with HIES phenotype molecular diagnosis was misleading. Extensive analyses driven by the clinical phenotype identified an intronic homozygous DOCK8 variant c.4626 + 76 A > G creating a novel splice site as disease-causing. While the affected newborn carrying the homozygous variant had no expression of DOCK8 protein, in the index patient molecular diagnosis was compromised due to expression of altered and wildtype DOCK8 transcripts and DOCK8 protein as well as defective STAT3 signaling. Sanger sequencing of lymphocyte subsets revealed that somatic alterations and reversions revoked the predominance of the novel over the canonical splice site in the index patient explaining DOCK8 protein expression, whereas defective STAT3 responses in the index patient were explained by a T cell phenotype skewed towards central and effector memory T cells. Hence, somatic alterations and skewed immune cell phenotypes due to selective pressure may compromise molecular diagnosis and need to be considered with unexpected clinical and molecular findings.

Structural basis for the recognition of spliceosomal SmN/B/B' proteins by the RBM5 OCRE domain in splicing regulation.

The multi-domain splicing factor RBM5 regulates the balance between antagonistic isoforms of the apoptosis-control genes FAS/CD95, Caspase-2 and AID. An OCRE (OCtamer REpeat of aromatic residues) domain found in RBM5 is important for alternative splicing regulation and mediates interactions with components of the U4/U6.U5 tri-snRNP. We show that the RBM5 OCRE domain adopts a unique ß-sheet fold. NMR and biochemical experiments demonstrate that the OCRE domain directly binds to the proline-rich C-terminal tail of the essential snRNP core proteins SmN/B/B'. The NMR structure of an OCRE-SmN peptide complex reveals a specific recognition of poly-proline helical motifs in SmN/B/B'. Mutation of conserved aromatic residues impairs binding to the Sm proteins in vitro and compromises RBM5-mediated alternative splicing regulation of FAS/CD95. Thus, RBM5 OCRE represents a poly-proline recognition domain that mediates critical interactions with the C-terminal tail of the spliceosomal SmN/B/B' proteins in FAS/CD95 alternative splicing regulation.

The spliceosome as a target of novel antitumour drugs.

Several bacterial fermentation products and their synthetic derivatives display antitumour activities and bind tightly to components of the spliceosome, which is the complex molecular machinery involved in the removal of introns from mRNA precursors in eukaryotic cells. The drugs alter gene expression, including alternative splicing, of genes that are important for cancer progression. A flurry of recent reports has revealed that genes encoding splicing factors, including the drug target splicing factor 3B subunit 1 (SF3B1), are among the most highly mutated in various haematological malignancies such as chronic lymphocytic leukaemia and myelodysplastic syndromes. These observations highlight the role of splicing factors in cancer and suggest that an understanding of the molecular effects of drugs targeting these proteins could open new perspectives for studies of the spliceosome and its role in cancer progression, and for the development of novel antitumour therapies.

Multi-domain conformational selection underlies pre-mRNA splicing regulation by U2AF.

Many cellular functions involve multi-domain proteins, which are composed of structurally independent modules connected by flexible linkers. Although it is often well understood how a given domain recognizes a cognate oligonucleotide or peptide motif, the dynamic interaction of multiple domains in the recognition of these ligands remains to be characterized. Here we have studied the molecular mechanisms of the recognition of the 3'-splice-site-associated polypyrimidine tract RNA by the large subunit of the human U2 snRNP auxiliary factor (U2AF65) as a key early step in pre-mRNA splicing. We show that the tandem RNA recognition motif domains of U2AF65 adopt two remarkably distinct domain arrangements in the absence or presence of a strong (that is, high affinity) polypyrimidine tract. Recognition of sequence variations in the polypyrimidine tract RNA involves a population shift between these closed and open conformations. The equilibrium between the two conformations functions as a molecular rheostat that quantitatively correlates the natural variations in polypyrimidine tract nucleotide composition, length and functional strength to the efficiency to recruit U2 snRNP to the intron during spliceosome assembly. Mutations that shift the conformational equilibrium without directly affecting RNA binding modulate splicing activity accordingly. Similar mechanisms of cooperative multi-domain conformational selection may operate more generally in the recognition of degenerate nucleotide or amino acid motifs by multi-domain proteins.

Differential 3' splice site recognition of SMN1 and SMN2 transcripts by U2AF and U2 snRNP.

Spinal Muscular atrophy is a prevalent genetic disease caused by mutation of the SMN1 gene, which encodes the SMN protein involved in assembly of small nuclear ribonucleoprotein (snRNP) complexes. A paralog of the gene, SMN2, cannot provide adequate levels of functional SMN because exon 7 is skipped in a significant fraction of the mature transcripts. A C to T transition located at position 6 of exon 7 is critical for the difference in exon skipping between SMN1 and SMN2. Here we report that this nucleotide difference results in increased ultraviolet light-mediated crosslinking of the splicing factor U2AF(65) with the 3' splice site of SMN1 intron 6 in HeLa nuclear extract. U2 snRNP association, analyzed by native gel electrophoresis, is also more efficient on SMN1 than on SMN2, particularly under conditions of competition, suggesting more effective use of limiting factors. Two trans-acting factors implicated in SMN regulation, SF2/ASF and hnRNP A1, promote and repress, respectively, U2 snRNP recruitment to both RNAs. Interestingly, depending on the transcript and the regulatory factor, the effects on U2 binding not always correlate with changes in U2AF(65) crosslinking. Furthermore, blocking recognition of a Tra2-beta1-dependent splicing enhancer located in exon 7 inhibits U2 snRNP recruitment without affecting U2AF(65) crosslinking. Collectively, the results suggest that both U2AF binding and other steps of U2 snRNP recruitment can be control points in SMN splicing regulation.