SF3B1 is the most frequently mutated splicing factor in cancer. Mutations in SF3B1 likely confer clonal advantages to cancer cells but they may also confer vulnerabilities that can be therapeutically targeted. SF3B1 cancer mutations can be maintained in homozygosis in C. elegans, allowing synthetic lethal screens with a homogeneous population of animals. These mutations cause alternative splicing (AS) defects in C. elegans, as it occurs in SF3B1-mutated human cells. In a screen, we identified RNAi of U2 snRNP components that cause synthetic lethality with sftb-1/SF3B1 mutations. We also detected synthetic interactions between sftb-1 mutants and cancer-related mutations in uaf-2/U2AF1 or rsp-4/SRSF2, demonstrating that this model can identify interactions between mutations that are mutually exclusive in human tumors. Finally, we have edited an SFTB-1 domain to sensitize C. elegans to the splicing modulators pladienolide B or herboxidiene. Thus, we have established a multicellular model for SF3B1 mutations amenable for high-throughput genetic and chemical screens. SF3B1 encodes a splicing factor frequently mutated in hematological malignancies, and with less frequency in other solid tumors. Each SF3B1 mutation is predominant in a cancer type, indicating cell type-specific pathogenic mechanisms. We have reproduced three cancer-related SF3B1 missense mutations in the C. elegans ortholog sftb-1 and they do not cause any overt phenotypes, but few alterations in splicing. However, the combination of these mutations has additive effects, hampering SFTB-1 functions and allowing the study of synthetic interactions that may result in a defective spliceosome structure. As a result of an RNAi screen, we identified other proteins of the U2 snRNP complex as targets for synthetic lethality. The sftb-1 mutant strains also allow the identification of mutations in other splicing factors that are mutually exclusive with SF3B1 mutations in tumors. SF3B1 cancer mutations are heterozygous, and interestingly, the most promising therapeutic venue is based on pladienolide B (PB) derivatives that hit the wild-type and the mutated protein. We have edited SFTB-1 to sensitize it to pladienolide B (PB) and herboxidiene (HB), therefore making C. elegans suitable for assaying the toxicity and efficacy of their chemical derivatives. Thus, our model represents a pre-clinical system to investigate both synthetic lethality and chemotherapies. Besides that, we have generated the first C. elegans strain in which splicing can be chemically modulated by PB and HB.
Cisplatin and derivatives are commonly used as chemotherapeutic agents. Although the cytotoxic action of cisplatin on cancer cells is very efficient, clinical oncologists need to deal with two major difficulties, namely the onset of resistance to the drug and the cytotoxic effect in patients. Here, we used Caenorhabditis elegans to investigate factors influencing the response to cisplatin in multicellular organisms. In this hermaphroditic model organism, we observed that sperm failure is a major cause of cisplatin-induced infertility. RNA sequencing data indicate that cisplatin triggers a systemic stress response, in which DAF-16/FOXO and SKN-1/NRF2, two conserved transcription factors, are key regulators. We determined that inhibition of the DNA damage-induced apoptotic pathway does not confer cisplatin protection to the animal. However, mutants for the pro-apoptotic BH3-only gene ced-13 are sensitive to cisplatin, suggesting a protective role of the intrinsic apoptotic pathway. Finally, we demonstrated that our system can also be used to identify mutations providing resistance to cisplatin and therefore potential biomarkers of innate cisplatin-refractory patients. We show that mutants for the redox regulator trxr-1, ortholog of the mammalian thioredoxin reductase 1 TRXR1, display cisplatin resistance. By CRISPR/Cas9, we determined that such resistance relies on the presence of the single selenocysteine residue in TRXR-1. This article has an associated First Person interview with the first author of the paper.
CRISPR/Cas and the high conservation of the spliceosome components facilitate the mimicking of human pathological mutations in splicing factors of model organisms. The degenerative retinal disease retinitis pigmentosa (RP) is caused by mutations in distinct types of genes, including missense mutations in splicing factors that provoke RP in an autosomal dominant form (s-adRP). Using CRISPR in Caenorhabditis elegans, we generated mutant strains to mimic s-adRP mutations reported in PRPF8 and SNRNP200. Whereas these inherited mutations are present in heterozygosis in patients, C. elegans allows the maintenance of these mutations as homozygotes, which is advantageous for genetic and drug screens. We found that snrp-200(cer23[V676L]) and prp-8(cer14[H2302del]) display pleiotropic phenotypes, including reduced fertility. However, snrp-200(cer24[S1080L]) and prp-8(cer22[R2303G]) are weak alleles suitable for RNAi screens for identifying genetic interactions, which could uncover potential disease modifiers. We screened a collection of RNAi clones for splicing-related genes and identified three splicing factors: isy-1/ISY1, cyn-15/PPWD1 and mog-2/SNRPA1, whose partial inactivation may modify the course of the disease. Interestingly, these three genes act as modifiers of prp-8(cer22) but not of snrp-200(cer24). Finally, a screen of the strong allele prp-8(cer14) with FDA-approved drugs did not identify molecules capable of alleviating the temperature-sensitive sterility. Instead, we detected drugs, such as dequalinium chloride, which exacerbated the phenotype, and therefore, are potentially harmful to s-adRP patients since they may accelerate the progression of the disease.
SF3B1 is the most frequently mutated splicing factor in cancer. Mutations in SF3B1 confer growth advantages to cancer cells but they may also confer vulnerabilities that can be therapeutically targeted. In contrast to other animal models, SF3B1 cancer mutations can be maintained in homozygosis in C. elegans , allowing synthetic lethal screens with a homogeneous population of animals. These mutations cause alternative splicing (AS) defects in C. elegans , as it occurs in SF3B1 -mutated human cells. In a screen, we identified RNAi of U2 snRNP components that cause synthetic lethality with sftb-1/SF3B1 mutations. We also detected synthetic interactions between sftb-1 mutants and cancer-related mutations in uaf-2/U2AF1 or rsp-4/SRSF2 , demonstrating that this model can identify interactions between mutations that are mutually exclusive in human tumors. Finally, we have edited an SFTB-1 domain to sensitize C. elegans to the splicing inhibitor pladienolide B. Thus, we have established a multicellular model for SF3B1 mutations amenable for high-throughput genetic and chemical screens.