Abstract The splicing factor RNA-binding motif protein 10 (RBM10) is frequently mutated in lung adenocarcinoma (LUAD) (9-25%). Most RBM10 cancer mutations are loss-of-function, correlating with increased tumorigenesis and limiting the efficacy of current LUAD targeted therapies. Remarkably, therapeutic strategies leveraging RBM10 deficiency remain unexplored. Here, we conduct a CRISPR-Cas9 synthetic lethality (SL) screen and identify ~60 RBM10 SL genes, including WEE1 kinase. WEE1 inhibition sensitizes RBM10-deficient LUAD cells in-vitro and in-vivo. Mechanistically, we identify a splicing-independent role of RBM10 in regulating DNA replication fork progression and replication stress response, which underpins RBM10-WEE1 SL. Additionally, RBM10 interacts with active DNA replication forks, relying on DNA Primase Subunit 1 (PRIM1) that synthesizes Okazaki RNA primers. Functionally, we demonstrate that RBM10 serves as an anchor for recruiting Histone Deacetylase 1 (HDAC1) to facilitate H4K16 deacetylation and R-loop homeostasis to maintain replication fork stability. Collectively, our data reveal a role of RBM10 in fine-tuning DNA replication and provide therapeutic arsenal for targeting RBM10-deficient tumors. Subject terms: DNA replication, DNA damage and repair __________________________________________________________________ RBM10 is the most mutated splicing factor in lung cancer. The authors reveal a non-canonical role of RBM10 in regulating DNA replication stress response. They also identify an arsenal of RBM10 synthetic lethal genes, such as WEE1, that can be therapeutically harnessed with immediate applicability. Introduction Lung cancer has the highest mortality rate among all types of cancer worldwide^[34]1,[35]2. Non-small cell lung cancer (NSCLC) comprises 85% of all lung cancer cases and is subdivided into different subtypes including lung adenocarcinoma (LUAD) (40%)^[36]3,[37]4. Most LUAD patients are diagnosed at advanced or metastatic stage, when treatment options are limited to surgery, chemotherapy, and few targeted therapies^[38]4. EGFR and KRAS are the most mutated oncogenes in LUAD amongst Asian (60%) and Caucasian (33%) cohorts, respectively^[39]5,[40]6. Accordingly, several targeted therapies, including EGFR and KRAS inhibition, are currently used in the clinic for the treatment of advanced stages of LUAD harboring KRAS and EGFR mutations^[41]7,[42]8. However, effective treatment of LUAD remains elusive due to the genetic diversity of the disease and the development of therapeutic resistance. Thus, there is a critical need for identifying new therapeutic targets for personalized treatment of LUAD. Interestingly, RNA-binding proteins are broadly dysregulated in human cancers including LUAD and play a key role in carcinogenesis and metastatic progression^[43]9–[44]17. Therefore, alterations in RNA-binding proteins might provide excellent therapeutic targets for treating lung cancer. One very attractive and yet unexplored therapeutic target is the RNA-binding protein, RBM10. RBM10 is mapped to the X chromosome^[45]18, and its mutations cause TARP syndrome, an X‐linked disorder that leads to pre- and postnatal lethality in affected males^[46]19–[47]25. RBM10 is a key regulator of alternative splicing that mainly promotes exon skipping of its target genes^[48]26–[49]33. Also, RBM10 impacts chromatin structure and chromosome segregation independently of its splicing activity^[50]34–[51]37. Mounting evidence shows that RBM10 acts as a tumor suppressor, as its depletion increases cell proliferation and enhances mouse xenograft tumor formation^[52]6,[53]29,[54]32,[55]38–[56]43. Concordantly, RBM10 overexpression suppresses LUAD tumor growth in vivo^[57]6,[58]44. RBM10 exerts its tumor suppressive activity through various mechanisms^[59]29,[60]41,[61]45. For example, it was shown that RBM10 inhibits cell proliferation and tumor growth through alternative splicing regulation of NUMB and EIF4H genes^[62]6,[63]32,[64]40. RBM10 is the most mutated splicing factor in LUAD and is mutated in 9% of all LUAD cases^[65]5. Strikingly, this percentage rises to 21% in invasive subtypes of LUAD, and 25% in multiple primary LUAD^[66]46,[67]47. Moreover, RBM10 mutation rate is higher than KRAS and TP53 mutations in early stages of LUAD in patients of Asian origin^[68]48,[69]49. Most RBM10 mutations in LUAD correspond to loss-of-function and are associated with poor survival^[70]50,[71]51. Notably, RBM10 mutations mostly co-occur with EGFR and KRAS mutations and were shown to reduce the efficiency of tyrosine kinase inhibitors (TKIs) in LUAD patients harboring EGFR mutations^[72]6,[73]52. Altogether, these findings highlight the urgency of therapeutic targeting of RBM10 deficiency in LUAD. Herein, we perform a genome-wide CRISPR-Cas9 synthetic lethality (SL) screen in isogenic LUAD cell line harboring RBM10 cancer mutation and identify ~60 high-scoring RBM10 SL genes, including WEE1 and Aurora A kinases. We show that pharmacological inhibition of WEE1 selectively sensitizes RBM10-deficient LUAD cells, including patient-derived cells harboring RBM10 cancer mutations, in vitro and in mouse xenograft model, and this effect is further exacerbated when combined with Aurora A inhibition. Mechanistically, we identify a splicing-independent role of RBM10 in promoting DNA replication fork progression that underpins RBM10-WEE1 SL. We also show that RBM10 is associated with active replication forks, which is contingent upon PRIM1, an enzyme responsible for synthesizing RNA primers of Okazaki fragments. Functionally, we demonstrate that RBM10 loss disrupts the localization of HDAC1 at replication forks leading to elevated levels of H4K16ac and R-loops, which destabilizes the replication fork and induces replication stress. Collectively, our data reveal a hitherto unrecognized function of the RNA-binding protein RBM10 in fine-tuning DNA replication, and identify DNA replication stress as an SL pathway with RBM10 loss. Moreover, we provide a repertoire of RBM10 SL targets that can be harnessed therapeutically to eradicate RBM10-deficient tumors with immediate clinical applicability. Results Genome-wide CRISPR-Cas9 screen reveals synthetic lethal partners of RBM10 in LUAD cells Given that RBM10 is one of the most mutated genes in LUAD, we sought to identify genetic vulnerabilities to RBM10 loss. To achieve this, we conducted an unbiased genome-scale CRISPR-Cas9 SL screen in LUAD HCC827 cell line. HCC827 cells harbor a common oncogenic deletion in EGFR (E746-A750), and are a pertinent model to screen for genetic vulnerabilities to RBM10 loss since RBM10 and EGFR mutations often co-occur in LUAD, and RBM10 loss was shown to limit the response to common EGFR-targeted therapy^[74]6,[75]52. Therefore, we sought to generate isogenic HCC827 cells stably expressing comparable levels of flag-Cas9 endonuclease and differ only in the status of RBM10. Toward this end, we designed a sgRNA targeting RBM10 exon 2 to introduce a frameshift mutation V176Rfs*, that exhibits a complete absence of RBM10 protein, while constitutively expressing Cas9 protein (HCC827-Cas9^RBM10-KO) (Fig. [76]1a and Supplementary Fig. [77]1a, b). Interestingly, the introduced mutation closely resembles a previously characterized LUAD cancer mutation E177* that abrogates RBM10 splicing activity and leads to RBM10 protein loss^[78]51. Notably, HCC827-Cas9^RBM10-KO cells exhibit an increase in exon inclusion of RBM10 splicing targets NUMB and EIF4H, phenocopying the defective splicing phenotype of RBM10-deficient LUAD^[79]6,[80]32 (Fig. [81]1b and Supplementary Fig. [82]1c). Fig. 1. Genome-wide CRISPR-Cas9 screen reveals synthetic lethal partners of RBM10 in LUAD cells. [83]Fig. 1 [84]Open in a new tab a Immunoblot analysis for RBM10 and flag-Cas9 protein expression in isogenic RBM10-deficient HCC827 cells constitutively expressing flag-Cas9. β-actin is used as a loading control. The samples derive from the same experiment but different gels for β-actin and RBM10 and another for flag were processed in parallel. The positions of molecular weight markers are indicated to the right. Representative of at least 3 independent experiments. b RT-PCR analysis of NUMB exon 9 alternative splicing in parental (WT) and HCC827-Cas9^RBM10-KO cells. RNA was isolated from the indicated cell lines and analyzed by RT-PCR using primers flanking NUMB exons 8–10. Left: representative agarose gel image showing amplification of two NUMB variants that differ in exon 9 inclusion. Right: precent-spliced-in (PSI) quantification of NUMB exon 9 inclusion. Data are presented as mean ± s.d. (n = 3 independent experiments). P value was determined by unpaired two-tailed t-test. c Results of CRISPR-Cas9 SL screen in WT and HCC827-Cas9^RBM10-KO cells performed in triplicates. CRISPR Counts Analysis (CCA) score is plotted against the difference in gene essentiality score, expressed as Bayes Factor (BF), between RBM10-KO cells and WT cells. RBM10 SL genes are shown in blue. d Gene ontology analysis of RBM10 SL genes identified in the CRISPR-Cas9 SL screen. e Clonogenic survival of WT and HCC827-Cas9^RBM10-KO cells transduced with inducible vector expressing either scramble shRNA or shRNA targeting the indicated genes upon the addition of doxycycline (DOX). Top: Representative images of crystal violet staining of the indicated cell lines treated with DOX. Bottom: Quantification of clonogenic survival in DOX-treated cells normalized to untreated cells. Data are presented as mean ± s.d. (n = 3 independent experiments). P values were determined by unpaired two-tailed t-test. f STRING interaction network of RBM10 SL genes involved in replication stress response and cell cycle. The thickness of the connecting line indicates interaction confidence. Source data are provided as a Source Data file. Next, we conducted a genome-scale CRISPR-Cas9 SL screen in the isogenic HCC827-Cas9^WT and HCC827-Cas9^RBM10-KO using the TKOv1 sgRNA library, which contains 91,320 sgRNAs targeting 17,232 protein-coding genes^[85]53. Cells were collected at 0 and 15 days following TKOv1 library transduction and subjected to next-generation sequencing to measure sgRNA abundance (Supplementary Fig. [86]1d). To identify bona fide RBM10 SL genes that are selectively essential in RBM10-KO cells, we analyzed the screen results using three distinct scoring methods: Bayesian Analysis of Gene Essentiality (BAGEL2)^[87]54, Model-Based Analysis of Genome-wide CRISPR-Cas9 Knockout (MAGeCK)^[88]55, and CRISPR Count Analysis (CCA)^[89]56. These methods unveiled 210, 280, and 357 RBM10 SL genes, respectively (Fig. [90]1c, Supplementary Fig. [91]1e, and Supplementary Data [92]1). Of note, a comparative analysis of RBM10 SL genes among the three different assays, revealed 60 RBM10 SL genes that are common to all three analysis methods (Supplementary Fig. [93]1f). Gene-set enrichment analysis revealed significant enrichment of several pathways including mRNA splicing, cell cycle checkpoints, S-phase, and DNA replication (Fig. [94]1d). To functionally validate the screen results, we conditionally knocked down 5 candidate RBM10 SL genes annotated to various enriched pathways (SF3B1, WAPAL, WEE1, AURKA, PEX10), and 2 common essential (XRN2, LSM7), followed by clonogenic survival assay. While depletion of common essential genes had a similar effect in inhibiting the proliferation of both parental and RBM10-KO cells, depletion of candidate RBM10 SL genes selectively inhibited the proliferation of HCC827-Cas9^RBM10-KO cells, confirming successful identification of bona-fide RBM10 SL genes (Fig. [95]1e and Supplementary Fig. [96]1g). Interestingly, several RBM10 SL genes, such as WAPAL, WEE1, CHEK1, ATRIP, and MRE11, were identified using the three aforementioned scoring methods. These genes have known functions in DNA replication stress response and replication fork protection, suggesting DNA replication stress as a synthetic lethal pathway with RBM10 deficiency in LUAD (Fig. [97]1f). In support of this, RBM10 was previously identified as a candidate replication stress response gene by a genome-wide siRNA screen^[98]57, and a phosphorylation target of ATR^[99]58. Additionally, RBM10 was shown to interact with several proteins involved in DNA replication and repair in fission yeast^[100]37. Fittingly, analysis of publicly available cancer mutation data revealed that RBM10 deficiency is associated with increased tumor mutation burden and DNA damage response gene signature (Supplementary Fig. [101]1h, i). Collectively, these observations suggest a role of RBM10 in DNA replication and genomic stability. RBM10 promotes DNA replication fork progression and replication stress response Herein, we sought to determine the impact of RBM10 loss on replication fork progression. DNA fiber analysis showed that HCC827-Cas9^RBM10-KO cells display a significant replication fork slowdown, suggesting that RBM10 is required for proper fork progression (Fig. [102]2a). Similar results were also observed in a second RBM10-KO clone (KO2) isolated from HCC827-Cas9 cells, arguing against a clonal effect (Supplementary Fig. [103]2a). To confirm that the effect of RBM10 loss on replication fork progression is not cell-line specific, and since RBM10 mutations occur in p53-deficient and EGFR-proficient LUAD^[104]59, we tested the effect of RBM10 knockout on replication fork progression in NCI-H1299 LUAD cells (H1299^RBM10-KO) (Supplementary Fig. [105]2b, c). Results showed that RBM10 loss in H1299 cells also led to significant reduction in replication fork progression, suggesting that the effect of RBM10 on DNA replication is independent of common co-occurring mutations such as EGFR and TP53 (Fig. [106]2b). Additionally, immunostaining analysis revealed that replication fork slowdown upon RBM10 loss is accompanied by an increase in the levels of RPA32 on position 33 (pRPA32-S33) and phosphorylated histone H2A.X (γH2AX), primarily observed in S phase cells marked by EdU staining (Fig. [107]2c–e). Similarly, western blot analysis demonstrated a significant increase in pRPA32-S33, phosphorylated CHK1 on position 345 (pCHK1-S345), and γH2AX in RBM10 knockout (KO) cells synchronized at the S phase using a double-thymidine block (Supplementary Fig. [108]2d). Next, we investigated the effect of RBM10 loss on the recovery from hydroxyurea (HU) induced replication stress. Results showed that RBM10-deficient cells display a significant increase in replication stress and DNA damage markers after release from HU, with no significant effect on replication fork restart (Fig. [109]2f and Supplementary Fig. [110]2e). Corollary to this, RBM10-deficient cells are markedly more sensitive than RBM10-proficient cells to increasing HU concentrations (Fig. [111]2g). Collectively, our results demonstrate that RBM10 fosters efficient DNA replication fork progression and fork recovery during replication stress. Fig. 2. RBM10 promotes DNA replication fork progression and replication stress response. [112]Fig. 2 [113]Open in a new tab a, b Replication fork speed measurement using DNA combing assay in WT and RRBM10-KO HCC827 (a) or H1299 (b) cells. Horizontal bars represent mean value of replication fork speed ± SEM (HCC827-Cas9 cells: n[wt] = 223, n[RBM10-KO] = 209, n[RBM10-KO2] = 214; H1299 cells: n[wt] = 193, n[RBM10-KO] = 195 fibers). P value was determined by two-tailed Mann–Whitney test. c Representative immunofluorescence microscopy images of pRPA32-S33 and γH2AX foci in H1299^WT and H1299^RBM10-KO cells. EdU is used to mark S-phase cells. DAPI is used to stain nuclei. Scale bar, 20 µm. d, e Quantification of γH2AX (d) and pRPA32-S33 (e) foci in S-phase (EdU-positive) H1299^WT and H1299^RBM10-KO cells. Horizontal bars represent mean foci number per nucleus ± SEM (n = 95 cells for WT cells and n = 110 cells for RBM10-KO cells) and representative of three independent experiments. P value was determined by two-tailed Mann–Whitney test. f H1299^WT and H1299^RBM10-KO cells were treated with 2 mM hydroxyurea (HU) for 2 h and subjected to immunoblot analysis at the indicated times after release from HU. The samples derive from the same experiment but different gels for RBM10, pRPA32 S4/S8, γH2AX, H3 and another for RPA32 were processed in parallel. UT=untreated. The positions of molecular weight markers are indicated to the right. Representative of three independent experiments. g Short-term cell viability assay in parental (WT) and RBM10-KO HCC827-Cas9 cells treated with increasing concentrations of HU. Data are presented as mean ± s.d. (n = 3 independent experiments). Source data are provided as a Source Data file. RBM10 is associated with active DNA replication forks in a PRIM1-dependent manner To gain molecular insights into the involvement of RBM10 in replication fork progression, we first determined its sub-cellular localization. Biochemical fractionation revealed that RBM10 is enriched at the chromatin-bound fraction (Supplementary Fig. [114]3a), supporting the notion that RBM10 has a direct role in regulating replication fork progression. To further substantiate this, cells expressing RBM10 fused to EGFP were subjected to GFP-trap followed by mass spectrometry (MS) analysis to map RBM10 interactome. Results revealed 410 RBM10-interacting proteins annotated to several pathways. While the most notable enriched pathway is the spliceosome, we observed a significant enrichment of DNA replication components (Supplementary Fig. [115]3b and Supplementary Data [116]2). Intriguingly, our data suggest that RBM10 interacts with replication fork components and proteins involved in replication stress response (Supplementary Fig. [117]3c). Next, we validated the authenticity of RBM10 interactome and showed that RBM10 interacts with MCM5, PCNA, HDAC1, PARP1, and RAD51 (Fig. [118]3a). Fig. 3. PRIM1-dependent RBM10 association with DNA replication forks promotes HDAC1 recruitment to limit replication stress. [119]Fig. 3 [120]Open in a new tab a Top: HEK293T cells expressing EGFP-RBM10 or EGFP only were subjected to GFP-trap analysis. Bottom: GFP-trap was performed on HEK293T cells additionally expressing Flag-MCM5. The samples derive from the same experiment but different gels for each indicated antibody were processed in parallel. Positions of molecular weight markers are indicated. Immunoblots are representative of three independent experiments. b, c Representative images of three independent experiments showing RBM10:PCNA (b) and RBM10:EdU-biotin (c) proximity ligation assay (PLA) foci in H1299^WT and H1299^RBM10-KO cells. Scale bar, 20 µm. d Left: representative images of RBM10:EdU-biotin PLA in H1299^WT cells transfected with siRNA against PRIM1 (siPRIM1) or control siRNA (siCtrl). Scale bar, 50 µm. Right: quantification of RBM10-EdU-biotin PLA intensity per nucleus. Horizontal bars represent mean PLA intensity per nucleus ± SEM (n = 206 cells for siCtrl and n = 225 cells for siPRIM1) and representative of three independent experiments. P value was determined by two-tailed Mann–Whitney test. e As in d, except slides were treated with RNaseH prior to PLA (n = 222 cells for untreated slides and n = 207 cells for RNaseH-treated slides). Scale bar, 50 µm. f As in d except of using H1299^RBM10-KO cells expressing Flag-RBM10^WT, Flag-RBM10^ΔZNF1, or vector only (VO) (n[WT]=191, n[VO] = 183, n[RBM10-WT] = 354, n[RBM10- ΔZNF1] = 310, n[RBM10-ΔRRM2] = 130 cells). Scale bar, 20 µm. g Immunoblot analysis of H1299^RBM10-KO cells expressing Flag-RBM10^WT, Flag-RBM10^ΔRRM2, Flag-RBM10^ΔZNF1, or vector only (VO). The samples derive from the same experiment but different gels for each indicated antibody were processed in parallel. Band intensity of γH2AX (relative to H3) and pRPA32-S33 (relative to RPA32) was quantified. Data presented as mean ± s.d. (n = 3 independent experiments). P value was determined by unpaired two-tailed t-test. ns, not significant. h, i Left: Representative immunofluorescence images of HDAC1:EdU-biotin (h) and H4K16ac:EdU-biotin (i) PLA in H1299^WT and H1299^RBM10-KO cells. Cells were either left untreated (UT) or treated with HU. Scale bar, 20 µm. Right: Quantification of PLA intensity per nucleus. Horizontal bars represent mean PLA intensity per nucleus ± SEM (HDAC1:EdU PLA—n[WT-UT] = 106, n[WT-HU] = 144, n[KO-UT] = 119, n[KO-HU] = 129; H4K16ac:EdU PLA—n[WT-UT] = 110, n[WT-HU] = 134, n[KO-UT] = 125, n[KO-HU] = 109 cells) and representative of 3 independent experiments. P value was determined by two-tailed Mann-Whitney test. j Left: Representative images of R-loops detected by S9.6 antibody in H1299^WT and H1299^RBM10-KO cells. Scale bar, 10 µm. Right: Quantification of S9.6 signal in H1299^WT and H1299^RBM10-KO cells. Horizontal bars represent mean intensity per nucleus ± SEM (n[WT-] = 76, n[WT+RNaseH] = 88, n[KO-]=68, n[KO+RNaseH] = 86 cells) and representative of 3 independent experiments. P value was determined by two-tailed Mann–Whitney test. Source data are provided as a Source Data file. To further substantiate RBM10 interaction with replication forks, first we confirmed the interaction between endogenous RBM10 and PCNA in LUAD cells (Supplementary Fig. [121]3d). Second, we tested whether RBM10 is associated with active replication forks in cells using proximity ligation assay (PLA). Results showed that RBM10 is in close proximity to PCNA (Fig. [122]3b). Moreover, PLA combined with EdU click chemistry revealed that RBM10 is associated with EdU-labeled nascent DNA, marking sites of active DNA replication (Fig. [123]3c). Collectively, our data provide firm evidence that RBM10 interacts with active DNA replication forks under native conditions. Next, we aimed to gain molecular insights into the mechanism governing the association of RBM10 with DNA replication forks. Remarkably, a recent work has shown that PRIM1, an enzyme involved in synthesizing RNA primers for Okazaki fragments, facilitates the recruitment of 53BP1 to DNA replication forks^[124]60. Hence, we investigated whether a similar scenario exists with RBM10. To achieve this, control and PRIM1-depleted cells were subjected to RBM10-EdU PLA. Similar to 53BP1, our results revealed that PRIM1 depletion pronouncedly diminishes the association of RBM10 with active DNA replication forks, as demonstrated by the significant decrease in RBM10-EdU PLA signal (Fig. [125]3d and Supplementary Fig. [126]4a, b). Notably and in line with previous reports^[127]60,[128]61, the observed decrease in the PLA signal in PRIM1-deficient cells is not due to alteration in cell cycle phase or reduced incorporation of Edu (Supplementary Fig. [129]4c). Altogether, our data infer that Okazaki RNA primers play a pivotal role in mediating the recruitment of RBM10 to replication forks. In support of this, treatment with RNase H significantly disrupts the interaction between RBM10 and DNA replication forks, as evidenced by the marked decrease in the RBM10-EdU PLA signal following RNase H treatment (Fig. [130]3e). Consistent with this observation, an RBM10 deletion mutant lacking the first ZnF domain (RBM10^ΔZNF1), known to bind RNA molecules^[131]62,[132]63, lost its association with active DNA replication forks, as determined by PLA (Fig. [133]3f). Interestingly, RBM10 splicing mutant lacking RRM2 domain (RBM10^ΔRRM2)^[134]32,[135]40,[136]63 has no detectable effect on RBM10 association with DNA active replication fork. Collectively, these findings robustly support the notion that RBM10 ZnF1 domain and PRIM1 enzyme mediate RBM10 recruitment to DNA replication forks. The role of RBM10 in replication stress response is distinct from its splicing activity Since RBM10 modulates the splicing of hundreds of genes, we sought to test whether the splicing activity of RBM10 underlies its role in the DNA replication stress response. To address this, we reintroduced RBM10-deficient cells with a vector expressing either flag-RBM10-WT or RBM10^ΔRRM2 or RBM10^ΔZNF1. Subsequently, we assessed the ability of these RBM10 mutants to restore both splicing efficiency and the levels of replication stress markers to normal levels. Our results demonstrate that while RBM10^ΔZNF1 successfully restores the splicing efficiency of RBM10 target genes, NUMB and EIF4H, it still exhibits increased levels of replication stress markers compared to cells expressing RBM10-WT (Fig. [137]3g and Supplementary Fig. [138]4d, e). Conversely, RBM10-deficient cells expressing RBM10^ΔRRM2 mutant show defects in the splicing of RBM10 target genes but display normal levels of replication stress markers compared to control cells. These findings collectively support the notion that the role of RBM10 in DNA replication is independent of its splicing activity. To further validate this observation, we conducted transcriptome analysis of RBM10 in isogenic HCC827 and NCI-H1299 cells. Results revealed no discernible changes in the expression of core DNA replication and repair genes (Supplementary Fig. [139]4f and Supplementary Data [140]3). Our findings align with previously reported transcriptome analyses of RBM10 in various cell lines, demonstrating that RBM10 does not regulate the expression or splicing of core DNA replication and repair genes^[141]6,[142]32. RBM10 targets HDAC1 to active replication fork to fine-tune histone acetylation To elucidate the mechanistic role of RBM10 in DNA replication progression and replication stress response, we focused on RBM10-HDAC1 interaction that has been identified in the RBM10 interactome and validated through immunoprecipitation (Fig. [143]3a and Supplementary Fig. [144]3c). Also, fission yeast Rbm10 was previously shown to interact with Clr6, a homolog of human HDAC1/2^[145]37. In addition, previous reports have shown that HDAC1 is localized at active DNA replication forks, where it deacetylates histones to prevent replication fork collapse^[146]64–[147]67. Prompted by this, we sought to investigate whether RBM10 is essential for recruiting HDAC1 to DNA replication forks. Our results demonstrated a pronounced disruption in the association of HDAC1 with DNA replication in RBM10-deficient cells under normal and stress conditions (Fig. [148]3h). Consequently, the levels of H4K16 acetylation were increased in RBM10-deficient cells compared to control cells (Fig. [149]3i). In summary, these data suggest that RBM10 serves as an anchor for recruiting HDAC1 to active replication forks to facilitate the deacetylation of H4K16, a modification recently recognized as crucial for replication fork stability^[150]67,[151]68. Since a recent work demonstrated that histone deacetylation counteracts the levels of R-loops at active replication forks to mitigate transcription-induced replication stress^[152]68, we postulated that the increased levels of H4K16ac in RBM10-deficient cells is also accompanied by increased R-loops levels. Indeed, we observed that RBM10-deficient cells exhibit elevated levels of R-loops, indicating that replication stress observed upon RBM10 loss is mediated, at least in part, by R-loop accumulation (Fig. [153]3j). Additionally, since RBM10 interacts with RAD51 protein (Fig. [154]3a), which has a central role in homologous recombination (HR) repair of double-strand breaks (DSB), and its depletion impairs HR repair leading to replication stress^[155]69,[156]70, we sought to investigate the functional relevance of their interaction^[157]67,[158]68. We aimed to determine whether the replication stress in RBM10-deficient cells is due to defective HR repair. Toward this, we assessed the efficacy of HR repair of DSBs using the Cas9-mClover-LMNA assay^[159]71. Results indicate that RBM10 deficiency has no discernible effect on the percentage of mClover-positive cells (representing cells that repaired DSBs through HR) compared to control cells (Supplementary Fig. [160]4g). This observation suggests that the replication stress observed in RBM10-deficient cells is not attributed to defective HR repair. This data raises an interesting possibility that the RBM10-RAD51 interaction might be required for replication fork reversal rather than HR repair. WEE1 kinase inhibition selectively sensitizes RBM10-deficient LUAD cells Herein, we sought to exploit the role of RBM10 in replication stress to selectively target RBM10-deficient LUAD cells. Towards this, we focused on WEE1 kinase, since it is a high scoring RBM10 SL gene that was identified in the three scoring methods (BAGEL, MAGeCK and CCA) (Fig. [161]1b and Supplementary Fig. [162]1e, f), and is involved in several RBM10 SL pathways, such as S-phase and G2/M cell cycle checkpoints, and was shown to protect the stability of stalled replication forks^[163]72–[164]74. Moreover, WEE1 has a small molecule inhibitor, MK1775 (AZD1775), which is currently in several phase II clinical trials for treatment of advanced solid tumors (ClinicalTrials.gov), highlighting the clinical potential of targeting RBM10-deficient LUAD^[165]75–[166]77. To test whether the observed RBM10-WEE1 synthetic lethality (Fig. [167]1e) is dependent on WEE1 kinase activity, we pharmacologically inhibited WEE1 using MK1775 in parental HCC827-Cas9^WT and two HCC827-Cas9^RBM10-KO cell lines. Short- and long-term survival assays revealed that HCC827-Cas9^RBM10-KO cells exhibit pronounced sensitivity to MK1775 treatment compared to parental cells (Fig. [168]4a, b and Supplementary Fig. [169]5a). The hypersensitivity of RBM10-deficient cells to MK1775 was also observed in RBM10-KO (KO3) cells derived from naïve HCC827 cells that do not stably express Cas9, suggesting that RBM10-WEE1 synthetic lethality is not due to a clonal effect or cytotoxicity caused by stable expression of Cas9 endonuclease (Supplementary Fig. [170]5b, c). Moreover, when RBM10-deficient cells were complemented with a vector expressing either flag-RBM10-WT or the splicing mutant RBM10^ΔRRM2, the sensitivity to MK1775 was suppressed compared to cells expressing an empty vector (Fig. [171]4c). Conversely, complementation of RBM10-deficient cells with a vector expressing RBM10^ΔZNF1 mutant (which lacks association with DNA replication forks, yet preserves its splicing activity) failed to mitigate sensitivity to MK1775. These findings collectively suggest that the molecular mechanism underlying the RBM10-WEE1 SL is dependent on RBM10 association with DNA replication forks, but independent of its splicing activity. Fig. 4. WEE1 kinase inhibition selectively sensitizes RBM10-deficient LUAD cells. [172]Fig. 4 [173]Open in a new tab a Short-term cell viability assay and EC[50] determination in WT and two RBM10-KO clones (KO and KO2) of HCC827-Cas9 cells treated with increasing concentrations of MK1775. Data are presented as mean ± s.d. (n = 3 independent experiments). b Clonogenic survival of WT and two RBM10-KO clones treated with the indicated concentrations of MK1775. Left, representative images of plates stained with crystal violet. Right, quantification of clonogenic survival. Data are presented as mean ± s.d (n = 2 independent experiments). c As in a, except of using HCC827^RBM10-KO3 cells expressing flag-RBM10^WT, RBM10^ΔRRM2, RBM10^ΔZNF1, or vector only. d As in a, except of using H1299^WT and H1299^RBM10-KO cells. e Clonogenic survival of H1299^WT and H1299^RBM10-KO cells treated with the indicated concentrations of MK1775. Top: representative images of plates stained with crystal violet. Bottom: quantification of clonogenic survival. Data are presented as mean ± s.d. (n = 3 independent experiments). f, g As in a, except of using NCI-H1944 (f) or NCI-H1975 (g) cells expressing either flag-RBM10-WT or vector only. Source data are provided as a Source Data file. Similar to p53-proficient HCC827, our results showed that H1299^RBM10-KO cells are hypersensitive to MK1775 compared to control cells (Fig. [174]4d,e). To increase the therapeutic potential of our observations, we tested RBM10-WEE1 synthetic lethality in two RBM10-deficient patient-derived cell lines, NCI-H1944 (containing KRAS^G13D mutation) and NCI-H1975 (containing EGFR^L858R/T790M mutation) harboring p.A683Rfs*10 and p.G905Afs*7 truncating RBM10 mutations, respectively. Towards this, NCI-H1944 and NCI-H1975 cells were transduced with flag-RBM10-WT expression vector, which restored RBM10 splicing activity (Supplementary Fig. [175]5d, e), and were subsequently treated with MK1775. Results showed that RBM10 expression mitigates the sensitivity of RBM10-deficient cell lines to MK1775, suggesting that RBM10 loss occurring during LUAD tumorigenesis potentiates the sensitivity to WEE1 inhibition (Fig. [176]4f, g). Altogether, our results recapitulate RBM10-WEE1 synthetic lethality observed in our CRISPR-Cas9 SL screen and demonstrate that WEE1 inhibition selectively sensitizes RBM10-deficient LUAD tumor cells from different origins and independently of common co-occurring mutations such as EGFR, KRAS, and TP53, highlighting the clinical relevance of targeting RBM10 loss in LUAD using MK1775. DNA damage and premature mitotic entry underpin RBM10-WEE1 synthetic lethality We sought to decipher the molecular basis underlying the hypersensitivity of RBM10-deficient cells to WEE1 inhibition. We showed that H1299^RBM10-KO cells treated with increasing concentrations of MK1775 exhibit elevated levels of γH2AX, suggesting that DNA damage accumulation upon WEE1 inhibition is exacerbated in RBM10-deficient cells (Fig. [177]5a). Likewise, higher levels of stalled forks and replication stress markers including phosphorylated RPA32 on positions 4 and 8 (pRPA32-S4/S8), pRPA32-S33, and pCHK1-S345 were observed in MK1775-treated H1299^RBM10-KO cells compared to control cells (Fig. [178]5b and Supplementary Fig [179]6a, b). Since WEE1 inhibition in RBM10-deficient cells leads to excessive replication stress, we investigated the effect of MK1775 treatment on replication fork progression. Results showed that WEE1 inhibition exacerbates the decrease in replication fork rate in RBM10-deficient cells compared to parental cells (Supplementary Fig. [180]6c). To determine whether the increase in replication stress following WEE1 inhibition was due to an increase in DNA breakage, we used alkaline and neutral comet assays to measure the single-stranded DNA (ssDNA) breaks and double-stranded DNA (dsDNA) breaks, respectively. Results show that MK1775 treatment in RBM10-KO cells led to increased accumulation of both ssDNA and dsDNA breaks, with the increase in ssDNA breaks being more prominent (Fig. [181]5c, d and Supplementary Fig. [182]6d). Next, we utilized native BrdU staining to further confirm the increase in ssDNA in H1299^RBM10-KO cells upon MK1775 treatment (Fig. [183]5e, f). To address whether the accumulation of ssDNA upon WEE1 inhibition is due to ssDNA gaps arising during DNA replication, we used a modified DNA combing assay incorporating a digestion step with S1 nuclease^[184]78. Results revealed that labeled DNA tracks in MK1775-treated RBM10-KO cells showed higher sensitivity to S1 nuclease digestion compared to MK1775-treated parental cells, suggesting WEE1 inhibition in RBM10-KO cells leads to accumulated ssDNA gaps during DNA replication (Fig. [185]5g). Altogether, our results suggest that WEE1 inhibition exacerbates replication stress upon RBM10 loss and that replication stress underpins, at least partly, RBM10-WEE1 synthetic lethality. Fig. 5. DNA damage and premature mitotic entry underpin RBM10-WEE1 synthetic lethality. [186]Fig. 5 [187]Open in a new tab a Quantification of γH2AX staining in H1299^WT and H1299^RBM10-KO cells treated with MK1775. Data are presented as mean nuclear intensity ± SEM (n[WT-] = 708, n[KO-]=1038, n[WT+300nM] = 648, n[KO+300nM] = 589 cells) and representative of three independent experiments. P values were determined by two-tailed Mann–Whitney test. b Immunoblot analysis of DNA damage and replication stress markers in H1299^WT and H1299^RBM10-KO cells after MK1775 treatment. Histone H3 is used as a loading control. The samples derive from the same experiment but different gels for pRPA32 S4/S8, pCHK1-S345, γH2AX, another for RBM10, another for CHK1, another for H3, and another for RPA32 were processed in parallel. Positions of molecular weight markers are indicated. Data is representative of at least three independent experiments. c Representative images of alkaline comet assay in H1299^WT and H1299^RBM10-KO cells treated with MK1775. d Quantification of DNA damage represented by comet tail moment. Horizontal bars represent mean tail moment ± SEM (n[WT ]= 137, n[RBM10-KO] = 135, n[WT+300nM] = 151, n[RBM10-KO+300nM] = 154, n[WT+1000nM] = 135, n[RBM10-KO+1000nM] = 106 cells) and representative of three independent experiments. P values were determined by two-tailed Mann–Whitney test. e Representative image for native BrdU staining in H1299^WT and H1299^RBM10-KO cells treated with MK1775. Scale bar, 50 µm. f Quantification of native BrdU staining. Data are presented as mean nuclear intensity ± SEM (n[WT ]= 215, n[RBM10-KO] = 114, n[WT+MK1775] = 550, n[RBM10-KO+MK1775] = 381 cells) and representative of three independent experiments. P value was determined by two-tailed Mann–Whitney test. g DNA combing analysis with S1 nuclease treatment in H1299^WT and H1299^RBM10-KO cells treated with MK1775. Data are presented as mean CldU track length ± SEM (n[WT ]= 102, n[WT+MK1775] = 134, n[RBM10-KO] = 106, n[RBM10-KO+MK1775] = 94 fibers). P values were determined by two-tailed Mann–Whitney test. h Representative images showing premature mitotic entry in H1299^WT and H1299^RBM10-KO cells treated with MK1775. EdU incorporation and pH3(Ser10) were used to determine DNA synthesis and mitotic entry, respectively. Scale bar, 20 µm. i, j Quantification of premature mitotic entry events in parental (WT) and H1299^RBM10-KO (i) and HCC827^RBM10-KO (j) cells. Premature mitosis events were calculated as the percentage of EdU-positive cells from the pH3(Ser10)-positive (mitotic) cells and presented as mean ± s.d. (n = 6 independent experiments). P value was determined by unpaired two-tailed t-test. k Representative images of γH2AX and DAPI staining in H1299^WT and H1299^RBM10-KO cells treated with MK1775. Scale bar, 10 µm. l Quantification of cells with micronuclei in H1299^WT and H1299^RBM10-KO cells treated with MK1775. Data are presented as mean ± SEM (n = 3 independent experiments). P values were determined by unpaired two-tailed t-test. Source data are provided as a Source Data file. Next, we sought to determine the effect of WEE1 inhibition on cell cycle progression in RBM10-deficient cells. Towards this, parental and H1299^RBM10-KO cells were co-stained for pH3(Ser10), a marker of early mitosis, and EdU as a marker for S-phase following MK1775 treatment. Remarkably, unlike parental cells, a significant fraction (~45%) of MK1775-treated H1299^RBM10-KO cells were co-stained with EdU and pH3(Ser10), suggesting that WEE1 inhibition leads to premature mitotic entry in RBM10-deficient cells (Fig. [188]5h, i). Similar results were also obtained in parental and HCC827^RBM10-KO cells (Fig. [189]5j). Consequently, we observed that MK1775-treated H1299^RBM10-KO cells exhibit high levels of micronucleation, extensive nuclear γH2AX staining and apoptosis (Fig. [190]5k, l and Supplementary Fig. [191]6e). Altogether, we concluded that the sensitivity of RBM10-deficient cells to MK1775 is mediated by increased replication stress and premature mitotic entry leading to mitotic catastrophe and cell death. MK1775 inhibits the growth of RBM10-deficient LUAD tumors in vivo Given that the WEE1 inhibitor MK1775 is currently undergoing phase II clinical trials for treating several tumor types^[192]79, we sought to test the therapeutic potential of WEE1 inhibition against RBM10-deficient LUAD cells in vivo. Our results showed that MK1775 treatment prominently reduced the size of xenograft tumors derived from HCC827^RBM10-KO and H1299^RBM10-KO cells, but not their parental counterparts (Fig. [193]6a, b and Supplementary Fig. [194]7a, b). Consistent with the observed γH2AX phenotype in vitro, RBM10-deficient tumors exhibit increased levels of DNA damage in MK1775-treated mice (Fig. [195]6c, d and Supplementary Fig. [196]7c). Fig. 6. MK1775 inhibits the growth of RBM10-deficient LUAD cells in vivo. [197]Fig. 6 [198]Open in a new tab a, b Tumor weight of parental (WT) and HCC827-Cas9^RBM10-KO (a) and H1299^RBM10-KO (b) xenografts treated with either MK1775 or vehicle. MK1775 was administered once daily at 40 mg/kg for 15 days. Results are shown as mean tumor weight ± SEM (n = 6 mice). P values were determined by unpaired two-tailed t-test. c, d Quantification of γH2AX immunostaining in xenograft tumor sections from parental and HCC827-Cas9^RBM10-KO (c) and H1299^RBM10-KO (d) tumors treated with either MK1775 or vehicle. Data are presented as the average percentage γH2AX positive cells ± SEM of sections of three different tumors for each group. P values were determined by unpaired two-tailed t-test. e Tumor weight of NCI-H1944 and NCI-H1975 xenografts treated with either MK1775 or vehicle. MK1775 was administered once daily at 40 mg/kg for 15 days. Results are shown as mean tumor weight ± SEM (n = 5 mice for NCI-H1944, n = 6 mice for NCI-H1975). P values were determined by unpaired two-tailed t-test. f Clonogenic survival of H1299^WT and H1299^RBM10-KO cells treated with the indicated concentrations of alisertib. Left, representative images of plates stained with crystal violet. Right, quantification of clonogenic survival. Data are presented as mean ± s.d. (n = 3 independent experiments). P values were determined by unpaired two-tailed t-test. g Short-term cell viability assay and EC[50] determination in H1299^WT and H1299^RBM10-KO cells treated with increasing concentrations of alisertib with or without treatment with 300 nM MK1775. Data are presented as mean ± s.d. (n = 3 independent experiments). h Tumor growth inhibition in H1299^WT and H1299^RBM10-KO xenografts treated with MK1775 alone, MK1775 and alisertib combination, or vehicle. MK1775 alone or MK1775 and alisertib combination were administered once daily at 30 mg/kg for 15 days. Results are shown as percentage tumor growth inhibition relative to vehicle ± SEM (n = 6 mice for MK1775 alone and vehicle, n = 5 mice for MK1775 + alisertib combination). P values were determined by unpaired two-tailed t-test. Source data are provided as a Source Data file. Collectively, these findings strongly suggest that RBM10 loss leads to increased sensitivity to WEE1 inhibition in vivo, independently of EGFR or TP53 status. To further substantiate the clinical relevance of targeting RBM10-deficient LUAD, we tested the anti-tumor effect of MK1775 against tumor xenografts derived from patient-derived LUAD cells, NCI-H1944 and NCI-H1975, harboring naturally occurring RBM10 cancer mutations. Results showed that WEE1 inhibition leads to remarkable tumor growth inhibition in RBM10-deficient NCI-H1975 and NCI-H1944 compared to RBM10-proficient cells (Fig. [199]6e). Altogether, our results demonstrate the therapeutic efficacy of MK1775 as a single-agent in eradicating LUAD tumors harboring RBM10 deleterious cancer mutations. Aurora kinase A inhibition exacerbates the anti-tumor activity of WEE1 inhibition in RBM10-deficient LUAD Given that both WEE1 and Aurora A (AURKA) scored highly in RBM10 SL screen and both kinases are targeted by small molecule inhibitors undergoing clinical trials, we tested the efficacy of the combined Aurora A and WEE1 inhibition on the growth of RBM10-deficient LUAD. First, we observed that the Aurora A inhibitor, alisertib, had a profound effect on the clonogenic survival of RBM10 KO cells, concurrent with the CRISPR SL screen results (Fig. [200]6f and Supplementary Fig. [201]7d). Moreover, while alisertib alone led to a significant reduction in RBM10-deficient cell viability, combined MK1775 and alisertib treatment further exacerbated cell toxicity (Fig. [202]6g). Prompted by this, we tested the effect of the combined MK1775 and alisertib treatment on the growth of RBM10-deficient xenograft tumors. Results showed that the combined treatment leads to a profound and significant reduction in RBM10-deficient tumor size that is greater than the reduction upon MK1775 treatment alone (Fig. [203]6h). Therefore, we concluded that Aurora kinase A inhibition might be a viable therapeutic strategy to enhance the anti-tumor activity of WEE1 inhibition in RBM10-deficient LUAD tumors. Discussion We present a model that underscores a new function of RBM10 in both DNA replication progression and replication stress response (Fig. [204]7). RBM10 interactome analysis revealed that it interacts with various replication fork components that are required for proper replication fork progression, as well as recovery from replication stress. Also, we show that RBM10 depletion leads to replication fork slowdown, which is accompanied by an increase in DNA damage. Our data are in line with a genome-wide siRNA screen that identified RNA processing factors, including RBM10, as replication stress response genes^[205]57. Moreover, previous studies showed that RBM10 is associated with human pre-replication complex^[206]80 and interacts with several proteins involved in DNA replication and repair in fission yeast^[207]37. In this regard, multiple splicing factors such as SF3B1, U2AF1, and ZRSR2 are associated with replication stress in cancer^[208]81–[209]83, although their localization at active replication forks remains to be explored. Fig. 7. Model depicting the novel role of RBM10 in DNA replication and replication stress response. Fig. 7 [210]Open in a new tab (Top) RBM10 association with active DNA replication forks is dependent on PRIM1, which synthetizes the RNA primer of Okazaki fragments. RBM10 promotes the recruitment of HDAC1 to ongoing and stressed replication forks and ensures the deacetylation of H4K16, thereby limiting R-loop formation and maintaining fork stability. (Bottom) In RBM10-deficient cells, defective HDAC1 recruitment and H4K16 deacetylation at stressed replication forks contributes to R-loop accumulation, fork destabilization, and ssDNA gap formation leading to replication stress. High levels of replication stress render RBM10-deficient tumor cells sensitive to WEE1 kinase inhibition, leading to the accumulation of DNA damage and mitotic catastrophe resulting in tumor cell death. Figure created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license. In this study, we show that the RNA-binding protein RBM10 is recruited to active DNA replication forks in a PRIM1-dependent manner to fine-tune the levels of histone acetylation at active DNA replication fork. Given that RNase H treatment and PRIM1 knockdown alleviate RBM10 association with active replication forks, we propose a model suggesting that this association requires RNA molecules, presumably the RNA primers of Okazaki fragments. In this regard, we revealed that the RBM10 ZnF1 domain, known for its RNA-binding capability, mediates its association with replication fork. Interestingly, this ZnF1 domain is found in various proteins including MDM2, NEIL3, ZRANB3, and FUS that have been previously implicated in regulating DNA replication progression and stress response^[211]84–[212]87. Future investigations will be necessary to determine whether the association of these proteins with DNA replication forks is dependent on PRIM1 activity. Functionally, we demonstrate that RBM10 counteracts the formation of R-loops in S-phase cells, concurrently facilitating the association of HDAC1 with active replication forks to deacetylate H4K16. These findings align with recent studies that described the role of HDAC1/2 in chromatin remodeling at active replication forks to maintain normal replication fork progression and replication fork stability^[213]66–[214]68. Interestingly, it was shown that HDAC1/2 decrease the levels of the transcriptionally active H4K16 acetylation at replication forks, thereby reducing the levels of R-loops and mitigating transcription-induced replication stress^[215]68. Further work will be required to determine the impact of RBM10 depletion on the acetylation levels of other histone and non-histone residues that are implicated in DNA replication regulation. While our data suggest that RBM10 function in DNA replication is distinct from its splicing activity (Fig. [216]3g), we cannot exclude the possibility that RBM10 regulates the splicing or expression of genes that may influence replication fork progression. In this study, we performed a genome-wide CRISPR-Cas9 knockout screen employing isogenic LUAD cells carrying RBM10 cancer mutation, a type of nonsense mutation occurring with frequencies ranging from 9% to 25% in LUAD. We identified ~60 RBM10 SL genes, such as WEE1 and Aurora A kinases, which can be therapeutically exploited for eradicating RBM10-deficient LUAD tumors. Indeed, pharmacological inhibition of WEE1 using MK1775 markedly reduces RBM10-deficient cell proliferation and tumor growth. Mechanistically, RBM10-WEE1 SL is mediated, at least in part, by increased replication stress leading to premature mitotic entry and cell death. This finding supports the notion that the sensitivity of RBM10-deficient cells to WEE1 inhibition is dependent on the emerging role of RBM10 in DNA replication. Our data are therefore in agreement with previous studies showing synergistic interaction between WEE1 inhibition and replication stress-inducing agents^[217]88–[218]91. Given that MK1775 is currently in phase II clinical trials for the treatment of several types of tumors, we propose that it can be also used against RBM10-deficient LUAD with immediate clinical applicability. Interestingly, the sensitivity of RBM10-deficient cells to WEE1 inhibition is further exacerbated by the combined treatment with alisertib, an Aurora Kinase A inhibitor. This synergistic effect between these two inhibitors is in agreement with a previous study showing that simultaneous inhibition of Aurora A and WEE1 prompts cells to traverse the G2–M checkpoint in the presence of DNA damage and impaired spindle assembly, culminating in mitotic catastrophe. Consequently, it leads to a synergic anti-tumor effect in p53-deficient squamous cell carcinoma of the head and neck^[219]92. We hypothesize that a similar mechanism underlies the synergy observed in RBM10-deficient tumors, though future work is needed to confirm this hypothesis. Together, these observations provide a basis for testing this combination for clinical treatment of RBM10-deficient LUAD patients, thus increasing therapeutic benefits and minimizing the risk of developing resistance. Intriguingly, a previous report showed that RBM10 deficiency in LUAD is associated with enriched immune pathways, elevated tumor mutational burden, and increased human leukocyte antigens (HLA) expression, suggesting that RBM10 deficiency may enhance anti-tumor immunity in LUAD^[220]93. In agreement with this, our RBM10 transcriptome in HCC827 and H1299 cell lines show that RBM10 depletion leads to increased expression of proinflammatory cytokines such as IL1R2, TNFRSF1B, BMP2, IL7R and CCL2. Further work will be vital to determine whether WEE1 inhibition in RBM10-deficient tumors further amplifies the upregulation of genes associated with immune pathways, which might enhance the response to immune therapies. Unbiased genome-wide CRISPR-Cas9 SL screens revolutionized the discovery of novel SL interactions that can be exploited for developing new cancer therapies^[221]94,[222]95. Our CRISPR-Cas9 knockout screen revealed ~60 RBM10 SL genes involved in various cellular pathways, such as RNA splicing and cell cycle. These genetic interactions advance our understanding of the physiological functions of RBM10 and provide a repertoire of novel targets that can be harnessed to formulate new strategies for targeting RBM10-deficiency. Interestingly, RBM10 is frequently mutated in various human cancers such as endometrial (8%) and urothelial (5%) carcinomas^[223]59. Therefore, our findings might be of clinical interest to other types of human cancer harboring RBM10 mutations. Methods Ethics statement All animal studies and protocols in this study were approved by the Committee on the Ethics of Animal Experiments of the Technion, Israel Institute of Technology (IL-069-07-22). Animal monitoring was performed daily and according to predetermined scoring criteria. Animals were judged by appearance, behavior, or clinical signs (severe diarrhea, cachexia, illness which may lead to severe pain, distress, or death). Mice were euthanized before the tumors reached the maximum tumor volume allowed (~1500 mm^3). Mice in treatment groups were euthanized when the volume of tumors in the mock-treated group reached the maximum volume. Other endpoint criteria were defined as follows: Animals whose tumor load exceeds 10% of body weight, animals with 20% decrease in normal body weight, animals with tumors that have ulcerated through the skin, and animals judged by appearance, behavior, or clinical sign (severe diarrhea, cachexia, illness which may lead to severe pain, distress, or death) will be euthanized. Mice with tumors that didn’t reach the maximum allowed, or those that didn’t develop any tumors at all, will be euthanized following 12 weeks after cell inoculation. Cell lines and cell culture HCC827, NCI-H1299, NCI-H1944, and NCI-H1975 cell lines were obtained from ATCC and grown at 37 °C and 5% CO2 and cultured in RPMI-1640 medium (Gibco) supplemented with 10% heat-inactivated fetal bovine serum (Gibco), 2mM L-glutamine (Gibco), and 100unit/mL penicillin and 100 μg/mL streptomycin (Gibco). HEK293T cell line was obtained from ATCC and grown at 37 °C and 5% CO2 and cultured in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco) supplemented with 10% heat-inactivated fetal bovine serum (Gibco), 2mM L-glutamine (Gibco), and 100unit/mL penicillin and 100 μg/mL streptomycin (Gibco). Lentiviral transduction Lentiviral particles were generated by co-transfecting HEK293T cells plated in 10 cm plates with 1.64pmol target vector, together with 1.3pmol psPAX2 (Addgene #12260) and 0.72pmol pMD2.G (Addgene #12259) using 3:1 μg PEI to μg DNA ratio. Media containing the viral particles were collected 48 h post-transfection and filtered with 0.45 μm filters. The indicated cell lines were transduced with the lentiviral particles in the presence of 10 μg/ml polybrene (Sigma-Aldrich H9268). At 48 h post-infection, transduced cells were selected using the appropriate antibiotics. Generation of Flag-Cas9 expressing cells and RBM10 knockout cell lines To generate HCC827 cell line constitutively expressing Cas9, cells were transduced with Lenti‐Cas9‐2 A‐Blast (Addgene #73310) lentiviral particles followed by selection with 10 μg/ml blasticidin (Invivogen #ant-bl) for 72 h. A single clone stably expressing Flag-Cas9 fusion was selected and validated by western blot and immunofluorescence. To generate isogenic RBM10 knockout in HCC827 cells expressing flag-Cas9 (HCC827-Cas9), guide RNAs (gRNA) were designed and cloned into pSPgRNA plasmid (Addgene #47108) using BbsI restriction enzyme. HCC827-Cas9 cells were transfected with RBM10 gRNA1 (HCC827-Cas9^RBM10-KO and naïve HCC827^RBM10-KO3) or gRNA2 (HCC827-Cas9^RBM10-KO2) followed by clonal isolation. Single clones were screened for RBM10 knockout using western blot and validated by immunofluorescence and sequencing. The sequences of RBM10 gRNAs are included in Supplementary Table [224]1. To generate RBM10 knockout in naïve HCC827 and NCI-H1299 cells, RBM10 gRNA1 was cloned into pSpCas9(BB)−2A-GFP (PX458) (Addgene #48138) and transfected to cells. At 24 h after transfection, GFP-positive cells were sorted using BD LSRFortessa™ cell analyzer (BD Biosciences) and plated in 96-well plates at a dilution of one cell per well. Single clones were screened for RBM10 knockout using western blot and validated by immunofluorescence and sequencing. Genome-wide CRISPR-Cas9 synthetic lethal screen 240 million cells of HCC827-Cas9^WT and HCC827-Cas9^RBM10-KO were transduced in triplicates with TKOv1 sgRNA library, which contains 91,320 sgRNAs targeting 17,232 protein-coding genes^[225]53, at MOI = 0.3 and an average of 500-fold coverage of the library. 24 h post-transduction, cells were selected with 2 μg/mL puromycin for 2 days. Following selection, 80 × 10^7 cells were harvested as day 0 samples (T0) and the remaining cells were plated and passaged every 5 days. Cells were then collected at day 15 (T15) for isolation of genomic DNA (gDNA) together with the T0 samples. Genomic DNA extraction was performed with Blood & Cell Culture DNA Maxi Kit (Qiagen) according to the manufacturer’s instructions. gRNA inserts were amplified using two rounds of PCR reactions. First, a 600 bp region of the lentiviral library vectors containing the sgRNA region was amplified using ordinary PCR with Q5 high-fidelity DNA polymerase for 25 cycles. A total of 120 μg gDNA was used as template for each sample (60 reactions: 2 μg gDNA/50 μl reaction). Second, the amplified PCR products were subjected to another PCR reaction with Illumina next-generation sequencing adapters for 10 cycles and 200 bp product was amplified. The resulting amplicons were pooled and gel purified using the NucleoSpin Gel and PCR Clean‑up (Macherey-Nagel). Samples were sequenced on a MiSeq platform with ~10 million reads per sample (Illumina). Demultiplexing and mapping of sequencing reads to the gRNA library was performed using MAGeCK^[226]55. Fold-change for each gRNA was calculated in each replicate versus the T0 sample. 3 analysis methods were used to predict RBM10 synthetic lethal genes: (1) Bioinformatic and Bayesian Analysis of Gene Essentiality (BAGEL) algorithm^[227]54 was devised to calculate the essentiality score (Bayes factor—BF) of each gene in parental (WT) and RBM10-KO cells. RBM10 SL genes were defined as genes with BF > 5 (i.e. essential) in RBM10-KO cells and BF < 2 (i.e. non-essential) in WT cells. (2) CRISPR Counts Analysis (CCA)^[228]56 was used to calculate synthetic lethality score for each gene and genes with score > 0.8 were considered candidate RBM10 SL genes. (3) MAGeCK was used to calculate log fold-change (LFC) for each gene and genes with LFC < −1 in RBM10-KO cells with LFC difference > 1 between WT and RBM10-KO were considered RBM10 SL genes. Pathway enrichment analysis and gene ontology were conducted using ShinyGO^[229]96. STRING interaction network was performed using string-db.org. Conditional short hairpin RNA knockdown Short hairpin RNA (shRNA) oligonucleotides directed against the indicated genes were annealed and cloned into Tet-pLKO-puro lentiviral vector (Addgene #21915) digested with EcoRI and AgeI. Lentiviral particles were generated as described above and used to transduce parental and RBM10-KO HCC827-Cas9 cells, followed by selection with 1 μg/ml puromycin (Invivogen #ant-pr) for 3 days. Cells were maintained in complete medium in the presence of 0.6 μg/ml puromycin. To induce shRNA-mediated gene expression knockdown, cells were cultured in the presence of 0.2 μg/ml doxycycline (DOX) for 48 h prior to clonogenic survival assays. Gene expression knockdown was validated by RT-qPCR. The shRNA and qPCR primer sequences used in this study are included in Supplementary Table [230]1. Generation of expression vectors To generate EGFP-RBM10 expression vectors, RBM10 was amplified from pDest26-RBM10-wt plasmid^[231]32 (gift from Prof. Juan Valcárcel) and cloned into pEGFP-C1 expression plasmid. To generate RBM10 deletion mutants, all-round PCR was used to delete the relevant regions. To constitutively restore RBM10 expression in RBM10-deficient cells, full-length RBM10 and RBM10 deletion mutants were subcloned to Lenti‐Cas9‐2 A‐Blast plasmid followed by lentivirus generation and transduction. To generate Flag-MCM5 expression vector, MCM5 was cloned from cDNA into p3X-Flag-CMV10 expression vector. Primer sequences used for cloning and mutagenesis are included in Supplementary Table [232]1. Transfections Cell transfections with plasmid DNA or siRNA were performed using Polyethylenimine (PEI) and Lipofectamine RNAiMax, respectively, following the manufacturer’s instructions. siRNAs used in this study are: RBM10 siRNA #1 (ThermoFisher Scientific HSS112075, 5′-CAAACGCCGAGAGAAGUGCU-3′); RBM10 siRNA #2 (ThermoFisher Scientific HSS112075, 5′-TTTGCCAAGGGTTCTAAGAG-3′); PRIM1 siRNA (Euphera Biotech EHU105881); Stealth RNAi negative control (Invitrogen). RNA isolation, RT-PCR, and qPCR Total RNA was isolated from cells using the TRIzol reagent according to the manufacturer’s instructions (Ambion) and treated with RQ1 DNase (M6101, Promega). 1 μg RNA was used for cDNA synthesis using the qScript cDNA Synthesis Kit (Quanta) with random primers. RT-PCR for alternative splicing assay of NUMB and EIF4H genes was performed by standard PCR using Red Load Taq Master (Jena Bioscience) followed by gel electrophoresis in agarose gel stained with ethidium bromide. Band intensities were quantified using GelDoc software. Real-time qPCR for measuring mRNA levels was performed using Step‐One‐Plus real‐time PCR System (Applied Biosystems) and the Fast SYBR Green Master mix (Applied Biosystems) with three technical repeats for each PCR. Data analysis and quantification were performed using StepOne software V2.2 supplied by Applied Biosystems. Primer sequences used for RT-PCR and qPCR are included in Supplementary Table [233]1. Small molecule inhibitors and drugs MK1775 (A5755), Alisertib (MLN8237) (A4110), and hydroxyurea (B2102) were purchased from APExBio. Caffeine was purchased from Sigma-Aldrich (C8961). Immunoblotting Whole-cell protein extracts were prepared using hot lysis buffer (1%SDS, 5 mM EDTA, 50 mM Tris, pH 7.5) supplemented with protease inhibitor mixture (Calbiochem). Protein samples were diluted and boiled in 5X protein loading buffer (10% SDS, 500 mM DTT, 50%Glycerol, 250 mM Tris-HCL and 0.5%bromophenol blue dye) and separated on an SDS-polyacrylamide gel. Proteins were transferred to a PVDF membrane (Millipore) and immunoblotted with the indicated antibodies. The following primary antibodies were used for immunoblotting: RBM10 (Sigma-Aldrich cat.no. HPA034972, 1:10,000), FLAG-tag (Sigma-Aldrich cat.no. F7425, 1:4,000), β-actin (Sigma-Aldrich cat.no. A5441, 1:10,000), α-Tubulin (Santa Cruz Bio cat.no. sc-5286, 1:1,000), pRPA32-S4/S8 (Bethyl cat.no. A300-245A, 1:1,000), pRPA32-S33 (Bethyl cat.no. A300-246A, 1:1,000), RPA32 (Abcam cat.no. ab2175, 1:500), γH2AX (Cell Signaling Technology cat.no. 2577, 1:1000), pChk1-S345 (Cell Signaling Technology cat.no. 2348, 1:1,000), Histone H3 (Abcam cat.no. ab1791, 1:30,000), Rad51 (Santa Cruz Bio cat.no. sc-8349, 1:1,000), PARP1 (Enzo Life Sciences cat. no. ALX-210-895, 1:2,000), HDAC1 (GeneTex cat. no. GTX100513, 1:2,000), PCNA (Santa Cruz Bio cat. no. sc-56, 1:1,000), GAPDH (Abcam cat.no. ab8245, 1:1,000), GFP (Santa Cruz Bio cat.no. sc-9996, 1:1,000). The following secondary antibodies were used: anti-rabbit IgG-HRP (Jackson ImmunoResearch cat.no. 111–035-003, 1:20,000), anti-mouse IgG-HRP (Amersham cat.no. NXA931, 1:10,000). Immunoblot intensity quantification was performed using ImageJ. Immunoblots source data are available in the Source Data file. Biochemical fractionation Cells were lysed with Buffer A for 5 min at 4 °C. Cell lysates were centrifuged at 1500×g for 5 min at 4 °C and the supernatant (cytoplasmic fraction) was removed. Then, the pellet was incubated with Buffer B (3 mM EDTA, 0.2 mM EGTA, 1 mM DTT, PMSF, and protease inhibitor mixture) for 10 min on ice followed by centrifugation at 1700×g for 5 min at 4 °C to extract nuclear-soluble fraction. To prepare chromatin-bound fraction, pellet was resuspended with hot lysis buffer (1%SDS, 5 mM EDTA, 50 mM Tris, pH 7.5), boiled for 15 min, and sonicated with two 15 s pulses of 35% amplitude, centrifuged at maximum speed for 20 min at 12 °C, and the supernatant was recovered. Fractions were subjected to immunoblot analysis with the indicated antibodies. GFP-trap HEK293T cells expressing either EGFP only or EGFP-RBM10 plasmids were subjected to GFP-TRAP assay as previously described^[234]97. Where indicated, cells were also co-transfected with Flag-MCM5 plasmid. Briefly, EGFP-RBM10 or EGFP-only expressing cells were lysed in RIPA buffer (150 mM NaCl, 50 mM Tris pH 7.4, 1% NP40, 0.5% deoxycholate, 0.1% SDS) supplemented with protease inhibitor cocktail and then diluted 1:4 with binding buffer (150 mM NaCl, 50 mM Tris pH 7.4). Two milligrams of cell lysate was incubated for 4 h at 4 °C with 15 μl GFP-TRAP beads followed by western blot analysis. For GFP-trap followed by mass spectrometry (MS), HEK293T cells were transfected with EGFP-RBM10, lysed and subjected to GFP-trap as described above. Beads were washed 5 times in PBSx1 followed by elution and trypsin digestion. Digested peptides were then analyzed by mass spectrometry (MS) at the Smoler Proteomics Center in the department of Biology in the Technion. The samples were analyzed by LC-MS/MS using Q-Exactive plus mass spectrometer (Thermo Scientific), coupled to Easy nano LC-1000 capillary UHPLC (Thermo Scientific). Protein identification and intensity analysis was performed using maxquant^[235]98, searching against the human section of the Uniprot database. RBM10-interacting proteins enrichment analysis and gene ontology was conducted using ShinyGO^[236]96. Immunoprecipitation For endogenous RBM10 immunoprecipitation, H1299 cells were harvested and lysed in RIPA buffer (150 mM NaCl, 50 mM Tris pH 7.4, 1% NP40, 0.5% deoxycholate, and 0.1% SDS) supplemented with protease inhibitor cocktail. Immunoprecipitations were performed using 0.5 μg IgG or RBM10 antibody and protein A magnetic beads (GenScript #L00273) incubated with 2 mg lysate overnight at 4 °C. Immunoprecipitated samples were subjected to immunoblot analysis using the indicated antibodies. Immunofluorescence Cells were seeded onto glass coverslips and grown for 24 h followed by drug treatments as indicated in the figure legends. Cells were fixed with 4% (wt/vol) paraformaldehyde for 10 min, permeabilized with 0.5% Triton X-100 in PBS for 10 min, blocked with blocking buffer (4% (wt/vol) BSA, 0.15% Tween 20 and 0.15% Triton X-100 in PBS) for 1 h at room temperature and incubated with the indicated antibodies for 1 h at 37 °C. Excess antibody was washed three times with wash buffer (0.15% Tween 20 and 0.15% Triton X-100 in PBS × 1), and cells were stained with secondary antibodies for 1 h at room temperature in the dark, and then washed as above. Coverslips were mounted onto glass slides using VECTASHIELD® Antifade Mounting Medium with DAPI (Vectorlabs). For immunofluorescence staining of γH2AX, pRPA32-S33, and BrdU, cells were pre-extracted with 0.25% Triton X-100 in PBS for 15 min on ice prior to fixation. For EdU incorporation, cells were pulsed with 10 μM EdU (5-ethynyl-2-deoxyuridine, Invitrogen A10044) for 30 min at 37 °C, 5% CO2. EdU click reactions were performed after cell permeabilization by incubating cells with EdU staining buffer containing 2 mM CuSO4, 100 mM ascorbic acid and 1 μM AlexaFluor 647 azide (Invitrogen A10277) in PBS for 30 min at room temperature. Images were acquired using inverted Zeiss LSM-700 or LSM-710 confocal microscopes. Image analysis was performed using ImageJ software. The following primary antibodies were used for immunofluorescence: BrdU FITC (BD Biosciences, cat.no. 51-23614 1:1000), RBM10 (Sigma-Aldrich cat.no. HPA034972, 1:2000), pRPA32-S33 (Bethyl cat.no. A300-246A, 1:500), γH2AX (Millipore cat.no. 05-636, 1:2500), FLAG-tag (Sigma-Aldrich cat.no. F7425, 1:1000), pH3(Ser10) (Cell Signaling Technology cat.no. 9706, 1:1,000), Anti-DNA-RNA Hybrid Antibody, clone S9.6 (Sigma-Aldritch cat.no. MABE1095, 1:1000). The following secondary antibodies were used: anti-rabbit IgG Alexa Fluor®488 (Invitrogen cat.no. A-21206, 1:500), anti-mouse IgG Alexa Fluor®488 (Invitrogen cat.no. A-21202, 1:500), anti-rabbit Alexa Fluor®568 (Invitrogen cat.no. A10042, 1:500), anti-mouse Alexa Fluor®568 (Invitrogen cat.no. A10037, 1:500), anti-rabbit Alexa Fluor®647 (Invitrogen cat.no. A-31573, 1:500), anti-mouse IgG Alexa Fluor®647 (Invitrogen cat.no. A32787, 1:500). Proximity ligation assay Proximity ligation assay (PLA) was performed using the Duolink in situ red starter kit (Sigma-Aldrich) following the manufacturer’s instructions. Rabbit anti-RBM10 (Sigma-Aldrich HPA034972, 1:2000) and mouse anti-PCNA (Santa Cruz Bio cat. no. sc-56, 1:250) antibodies were incubated for 1 h at 37 °C prior to PLA analysis. PLA assay between RBM10 and EdU-labeled nascent DNA was performed as previously described^[237]99. Briefly, cells were pulse labeled with 100 μM EdU for 10 min at 37 °C, followed by pre-extraction in 0.25% Triton X-100 in PBS for 10 min on ice and fixation in 4% (wt/vol) paraformaldehyde. Cells were incubated in EdU click reaction buffer containing 2 mM CuSO4, 100 mM ascorbic acid and 10 μM biotin-azide (Invitrogen [238]B10184) in PBS for 1 h at room temperature. Cells were incubated in blocking buffer for 1 h at room temperature followed by incubation with rabbit anti-RBM10 (Sigma-Aldrich HPA034972, 1:2000), rabbit anti-HDAC1 (GeneTex GTX100513, 1:2000), Rabbit anti-H4K16ac (Abcam ab109463, 1:2000), rabbit anti-53BP1 (Novus biological NB100-305, 1:1000) and mouse anti-biotin (Jackson ImmunoResearch cat.no. 200-002-211, 1:2000) antibodies for 1 h at 37 °C. Where indicated, RNaseH treatment (New England Biolabs M0297) was performed following permeabilization by incubating cells in RNaseH buffer without (control) or with RNaseH (diluted 1:50) for 3 hours at 37 °C. Next, PLA assay was performed using the Duolink in situ red starter kit for experiments described in Fig. [239]3c, d (Sigma-Aldrich) or NaveniFlex Cell MR Atto647N kit (Navinci) following the manufacturer’s instructions. Images were acquired using inverted Zeiss LSM-700 confocal microscope. Image analysis was performed using ImageJ software. PLA intensity was calculated as the product of number of spots and the mean intensity of the spots per nucleus. Cell synchronization To synchronize cells at G1/S border using double-thymidine block, cells were treated with 2 mM thymidine overnight, released for 10 h and treated again with thymidine overnight. Cells were released into growth medium containing nocodazole (120 ng/ml) to enrich mitotic cells. Cell cycle profiles were analyzed by Cytek Aurora^© after staining with propidium iodide (Sigma-Aldritch). Alkaline and neutral comet assays Alkaline and neutral comet assay were performed using the CometAssay Kit (Trevigen, 4250-050-K) according to the manufacturer’s instructions. Briefly, 1 × 10^5 cells/mL were mixed with molten LMAgarose (at 37 °C) at a ratio of 1:10 (v/v), and 70 μL of this mixture was transferred onto a CometSlide. Slides were placed flat at 4 °C in the dark for 30 minutes. Cells were then lysed and subjected to alkaline or neutral electrophoresis with voltage of 1 V/cm. DNA was stained with SYBR green and slides were imaged using ZEISS LSM700 confocal microscope. Comet tail moment was scored using CometScore 2.0 software. Short-term growth delay assay For determining drug sensitivity, cells were seeded in 96-well plates in duplicates at a density of 2000–5000 cells per well. 24 h post-seeding, drugs were added at the indicated concentrations. Cell viability was measured 72 h after drug treatment using the CellTiter 96® AQueous One Solution Cell Proliferation Assay (Promega) following the manufacturer’s protocol, and absorbance was measured using Epoch Microplate Spectrophotometer (BioTek). Cell viability was normalized to the viability of untreated cells. Clonogenic survival assays For colony survival assay, cells seeded onto 6-well plates (500 cells per well) and left untreated or treated with the indicated doses of drugs. After 14 days, cells were fixed in methanol and stained with 0.25% crystal violet. Individual colonies were counted, and relative survival was calculated by normalizing to the number of colonies in control wells. Annexin-V apoptosis assay Apoptosis was assessed by annexin V-FITC kit (Invitrogen) according to the manufacturer’s instructions. Parental (WT) and RBM10-KO cells were treated with the indicated concentrations of MK1775 for 24 h prior to analysis. Samples (10,000 cells) were analyzed using Cytek Aurora^©. Results were calculated as the percentage of positive annexin V-FITC cells out of total cells counted. Endogenous homologous recombination assay Homologous recombination repair assay was performed using Cas9-mediated knock-in of green fluorescent mClover into the first exon of the LMNA gene, as previously described^[240]71. Briefly, cells were plated in 6-well plates and co-transfected with 1.6 μg pX330-LMNA-gRNA1 plasmid containing Cas9 and gRNA against LMNA exon 1 and 0.4 μg pCR2.1-CloverLamin plasmid containing HR donor sequence. In addition, 0.4 μg pDsRed-Monomer-C1 was included per transfection as a transfection control unless otherwise indicated in the figure legend. 12−16 h post transfection, culture medium was renewed and where indicated, Caffeine (4 mM) was added. Seventy-two hours post-transfection, cells were collected and analyzed by Cytek Aurora^©. HR efficiency is percentage mClover-expressing cells from DsRed-Monomer positive cells. DNA combing assays DNA combing analysis was performed as described^[241]100. Briefly, unsynchronized cells were pulse labeled with 25 μM Idu for 30 min, followed by 600 μM CldU for another 30 min, followed by 1.5 h growth without any label. For combing analysis with hydroxyurea (HU), cells were pulse labeled with 25 μM Idu for 30 min, followed by incubation with 4 mM HU for 2 h, followed by pulse labeling 600 μM CldU for another 30 min. Cells were harvested into 2% agarose plugs, incubated for 48 h in a digestion buffer (1% N Laurouylsarcosine, 0.2% Na Deoxycholate, 10 mM Tris-HCl pH 7.5, 100 mM EDTA, and 1 mg/ml proteinase K). Next, TE 50X was added overnight, and the plugs were dissolved into 50 mM MES pH 6 with 100 mM NaCl. The DNA was combed on Genomic vision combing slides, denatured, and stained with the following primary antibodies: Rat anti-BrdU (Abcam #ab6326, 1:200), mouse anti-BrdU (BD Biosciences #347580, 1:20), and anti-ssDNA antibody (DSHB #AB10805144, 1:20). The following secondary antibodies were used: Goat anti-mouse Alexa Fluor®647 (Jackson Immunoresearch #115-605-166, 1:400), goat anti-rat Alexa Fluor®594 (ThermoFisher Scientific #A11007, 1:400), and anti-mouse IgG2a Alexa Fluor®488 (ThermoFisher Scientific #A21131, 1:500). The slides were photographed with a fluorescent microscope and analyzed using Olympus CellSens program. For DNA combing analysis with S1 nuclease digestion, cells were treated as indicated and then pulse labeled with 25uM Idu for 30 min, followed by 600 μM CldU for 60 min. Subsequently, cells were permeabilized (10 mM PIPES, pH 6.8, 0.1 M NaCl, 0.3 M sucrose, 3 mM MgCl2, EDTA-free Protease Inhibitor Cocktail (Roche)) at room temperature for 10 min, followed by incubation with S1 nuclease (20 U/ml) (Promega M5761) in S1 buffer for 60 min at 37 °C. RNA sequencing Three biological replicates of RNA samples were purified from parental and RBM10-KO HCC827 and H1299 cells, or HCC827 cells transfected with control or RBM10 siRNA#1. RNA sequencing libraries were prepared using TruSeq mRNA library preparation kit. Sequencing was performed at The Crown Genomics institute of the Nancy and Stephen Grand Israel National Center for Personalized Medicine, Weizmann Institute of Science using a NovaSeq 6000 system with S1 flow cell to obtain 150 bp paired-end reads. The average read depth was 60 million reads per sample. The quality of the raw FASTQ files was assessed using FastQC software ([242]http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). For differential gene expression analysis, raw sequencing reads were aligned to GENCODE GRCh38 genome assembly using Salmon package^[243]101 and differential gene analysis was performed in R using the DESeq2 package^[244]102. Pathway enrichment analysis and gene ontology was conducted using ShinyGO^[245]96. Mouse xenograft model Animal studies and protocols were approved by the Committee on the Ethics of Animal Experiments of the Technion, Israel Institute of Technology (IL-069-07-22). Immunocompromised NOD.CB17-Prkdcscid/NcrCrl (NOD SCID) mice were purchased from Envigo. Mouse model of LUAD cell lines subcutaneous xenografts were established by transplanting cells into 6-week-old female NOD SCID mice. Mice were randomly assigned into control and treatment groups. Prior to initiating any experiments, mice were allowed 1 week acclimation to housing conditions at the Technion Animal Facility. All mice were housed in a strict pathogen-free environment. 2.5 × 10^6 of the indicated cells were resuspended in 100 μl PBS and mixed 1:1 with Matrigel (High Concentration—Corning) and injected subcutaneously into the right flank of 6 weeks-old female NOD SCID mice. Tumors were measured using digital calipers. Tumor volumes were calculated using the formula: 0.5 × length × width^2. When tumors reached ∼100mm^3, mice were randomly assigned to treatment groups (n = 6 mice for each group, unless otherwise indicated). Each group received either control (vehicle), MK1775 (40 mg/kg, p.o.), or MK1775 + alisertib (30 mg/kg each, p.o.) once daily for 15 days. Mice were euthanized when the tumors of control mice reached ∼1500 mm^3 and tumors were excised, weighed, and photographed. Immunohistochemistry Tumors excised from mice were fixed in 4% formalin at 4 °C overnight, washed three times with PBS, and immersed in 30% (wt/vol) sucrose solution at 4 °C overnight. Then, tumors were embedded in optimal cutting temperature compound (TissueTek) and frozen at −20 °C and. 10 μm cryosections were prepared and permeabilized using 0.5% Triton X-100 in PBS followed by immunofluorescence staining as described above, except primary antibody incubation with γH2AX antibody (Millipore cat.no. 05-636,1:500) were performed at 4 °C overnight. Statistics and reproducibility The number of replicates in each experiment is indicated in the figure legends and/or methods. In-vitro cell biology, CRISPR screen, and interactome experiments were performed with 3 replicates or more if indicated. Sample sizes for in-vitro experiments were not determined by statistical power analysis but were chosen according to standard practices in the field. Immunoblots were repeated at least 2 times, as indicated in the figure legends, with one representative experiment presented. For in-vivo xenograft experiments, the number of mice was calculated according to statistical power analysis and consideration of standard attrition rate, yielding n = 6 mice per treatment group. No data were excluded from in vitro experiments. One mouse inoculated with NCI-H1944 cells in Fig. [246]6e died 2 weeks prior to the experiment end-point and therefore was excluded from the analysis. All identical untreated samples were allocated randomly into treatment conditions. All mice were randomly allocated into treatment groups. For microscopy-based analysis, a fixed number of images were obtained randomly from each sample and analyzed together with control samples using the same analysis pipeline. Researchers were not blinded during data collection and analysis as the study design included appropriate controls and did not require blinding. Data collection and analysis of different conditions were performed at the same time and using identical methods. Microscopy and high-throughput acquisition and analysis was performed automatically without bias between different samples. Reporting summary Further information on research design is available in the [247]Nature Portfolio Reporting Summary linked to this article. Supplementary information [248]Supplementary Information^ (1.7MB, pdf) [249]Peer Review File^ (227.9KB, pdf) [250]Supplementary Data 1^ (8.7MB, xlsx) [251]Supplementary Data 2^ (145.9KB, xlsx) [252]Supplementary Data 3^ (10.6MB, xlsx) [253]Reporting Summary^ (90.4KB, pdf) Source data [254]Source Data^ (32.4MB, xlsx) Acknowledgements