Advertisement

Tumor suppressor role of RBM22 in prostate cancer acting as a dual-factor regulating alternative splicing and transcription of key oncogenic genes

  • Juan M. Jiménez-Vacas
    Correspondence
    Reprint requests: Juan M. Jiménez-Vacas and Raúl M. Luque. Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), IMIBIC Building. Av. Menéndez Pidal s/n. 14004-Córdoba, Spain.
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Department of Cell Biology, Physiology, and Immunology, University of Córdoba, Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición, (CIBERobn), Cordoba, Spain
    Search for articles by this author
  • Antonio J. Montero-Hidalgo
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Department of Cell Biology, Physiology, and Immunology, University of Córdoba, Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición, (CIBERobn), Cordoba, Spain
    Search for articles by this author
  • Enrique Gómez-Gómez
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Urology Service, HURS/IMIBIC, Cordoba, Spain
    Search for articles by this author
  • Prudencio Sáez-Martínez
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Department of Cell Biology, Physiology, and Immunology, University of Córdoba, Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición, (CIBERobn), Cordoba, Spain
    Search for articles by this author
  • Antonio C. Fuentes-Fayos
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Department of Cell Biology, Physiology, and Immunology, University of Córdoba, Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición, (CIBERobn), Cordoba, Spain
    Search for articles by this author
  • Adrià Closa
    Affiliations
    The John Curtin School of Medical Research, Australian National University, Canberra, Australia

    EMBL Australia Partner Laboratory Network at the Australian National University, Canberra, Australia
    Search for articles by this author
  • Teresa González-Serrano
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Anatomical Pathology Service, HURS, Cordoba, Spain
    Search for articles by this author
  • Ana Martínez-López
    Affiliations
    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Anatomical Pathology Service, HURS, Cordoba, Spain
    Search for articles by this author
  • Rafael Sánchez-Sánchez
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Anatomical Pathology Service, HURS, Cordoba, Spain
    Search for articles by this author
  • Pedro P. López-Casas
    Affiliations
    Prostate Cancer Clinical Research Unit, Hospital Universitario 12 de Octubre, Instituto de Investigación Sanitaria Hospital 12 de Octubre (imas12), Madrid, Spain
    Search for articles by this author
  • André Sarmento-Cabral
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Department of Cell Biology, Physiology, and Immunology, University of Córdoba, Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición, (CIBERobn), Cordoba, Spain
    Search for articles by this author
  • David Olmos
    Affiliations
    Prostate Cancer Clinical Research Unit, Hospital Universitario 12 de Octubre, Instituto de Investigación Sanitaria Hospital 12 de Octubre (imas12), Madrid, Spain
    Search for articles by this author
  • Eduardo Eyras
    Affiliations
    The John Curtin School of Medical Research, Australian National University, Canberra, Australia

    EMBL Australia Partner Laboratory Network at the Australian National University, Canberra, Australia

    Catalan Institution for Research and Advanced Studies. Barcelona, Spain

    Hospital del Mar Medical Research Institute (IMIM), Barcelona, Spain
    Search for articles by this author
  • Justo P. Castaño
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Department of Cell Biology, Physiology, and Immunology, University of Córdoba, Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición, (CIBERobn), Cordoba, Spain
    Search for articles by this author
  • Manuel D. Gahete
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Department of Cell Biology, Physiology, and Immunology, University of Córdoba, Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición, (CIBERobn), Cordoba, Spain
    Search for articles by this author
  • Raul M. Luque
    Correspondence
    Reprint requests: Juan M. Jiménez-Vacas and Raúl M. Luque. Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), IMIBIC Building. Av. Menéndez Pidal s/n. 14004-Córdoba, Spain.
    Affiliations
    Maimonides Institute for Biomedical Research of Córdoba (IMIBIC), Cordoba, Spain

    Department of Cell Biology, Physiology, and Immunology, University of Córdoba, Cordoba, Spain

    Hospital Universitario Reina Sofía (HURS), Cordoba, Spain

    Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición, (CIBERobn), Cordoba, Spain
    Search for articles by this author
Open AccessPublished:September 08, 2022DOI:https://doi.org/10.1016/j.trsl.2022.08.016

      Abstract

      Prostate cancer (PCa) is one of the leading causes of cancer-related deaths among men. Consequently, the identification of novel molecular targets for treatment is urgently needed to improve patients’ outcomes. Our group recently reported that some elements of the cellular machinery controlling alternative-splicing might be useful as potential novel therapeutic tools against advanced PCa. However, the presence and functional role of RBM22, a key spliceosome component, in PCa remains unknown. Therefore, RBM22 levels were firstly interrogated in 3 human cohorts and 2 preclinical mouse models (TRAMP/Pbsn-Myc). Results were validated in in silico using 2 additional cohorts. Then, functional effects in response to RBM22 overexpression (proliferation, migration, tumorspheres/colonies formation) were tested in PCa models in vitro (LNCaP, 22Rv1, and PC-3 cell-lines) and in vivo (xenograft). High throughput methods (ie, RNA-seq, nCounter PanCancer Pathways Panel) were performed in RBM22 overexpressing cells and xenograft tumors. We found that RBM22 levels were down-regulated (mRNA and protein) in PCa samples, and were inversely associated with key clinical aggressiveness features. Consistently, a gradual reduction of RBM22 from non-tumor to poorly differentiated PCa samples was observed in transgenic models (TRAMP/Pbsn-Myc). Notably, RBM22 overexpression decreased aggressiveness features in vitro, and in vivo. These actions were associated with the splicing dysregulation of numerous genes and to the downregulation of critical upstream regulators of cell-cycle (i.e., CDK1/CCND1/EPAS1). Altogether, our data demonstrate that RBM22 plays a critical pathophysiological role in PCa and invites to suggest that targeting negative regulators of RBM22 expression/activity could represent a novel therapeutic strategy to tackle this disease.

      Graphical Abstract

      Abbreviation:

      AR (androgen receptor), BPH (benign prostatic hyperplasia), FDR (false discovery rate), FFPE (formalin-fixed, paraffin-embedded), GS (gleason score), IHC (immunohistochemistry), IPA (ingenuity pathway analysis), MD-PCa (moderately differentiated prostate cancer), N-TAR (non-tumor adjacent region), PCa (prostate cancer), PD-PCa (poorly differentiated prostate cancer), PI3K (phosphoInositide 3-kinase), PIN (prostatic intraepithelial neoplasia), RBM22 (RNA binding motif protein 22), RRM (rna recognition motif), TRAMP (transgenic adenocarcinoma of mouse prostate model)
      At a Glance Commentary

      Background

      The importance of alternative-splicing dysregulation in prostate cancer (PCa) is indisputable. However, the contribution of components of the molecular machinery controlling the splicing process remains poorly explored in PCa.

      Translational Significance

      Our results identified the RNA Binding-Motif 22 (RBM22) as a key splicing-factor regulating clinically relevant pathological functions with therapeutic potential in PCa. Indeed, RBM22 expression was downregulated in PCa, and its overexpression reduced tumor aggressiveness in vitro and in vivo through cell-cycle arrest in PCa-cells. Additionally, RBM22 high expressing PCa cells downregulated CDK1/ATR, and consequently are candidate for PARP-inhibition. Altogether, these findings indicate that RBM22 could represent a predictive-biomarker/therapeutic-target in PCa.

      Introduction

      Prostate cancer (PCa) is the commonest cancer in men and a leading cause of cancer-related death in developed countries.
      • Bray F
      • Ferlay J
      • Soerjomataram I
      • Siegel RL
      • Torre LA
      • Jemal A.
      Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
      PCa is highly heterogeneous, encompassing from slow-growing/indolent to aggressive/clinically-significant tumors, which hampers its clinical management.
      • Tolkach Y
      • Kristiansen G
      The heterogeneity of prostate cancer: a practical approach.
      In this scenario, although low aggressive localized lesions (ie Gleason score 6) are curable, the treatment of highly aggressive locally advanced PCa and metastatic PCa remains challenging.
      • Sandhu S
      • Moore CM
      • Chiong E
      • Beltran H
      • Bristow RG
      • Williams SG
      Prostate cancer.
      ,
      • Sartor O
      • de Bono JS
      Metastatic prostate cancer.
      Indeed, despite the initial robust responses to new generation AR-signaling targeting agents (ie abiraterone, enzalutamide, apalutamide and darolutamide) and taxanes (ie docetaxel and cabazitaxel), advanced PCa eventually become resistant, then representing a lethal disease.
      • Sartor O
      • de Bono JS
      Metastatic prostate cancer.
      ,
      • Westaby D
      • Fenor de La Maza MLD
      • Paschalis A
      • et al.
      A new old target: androgen receptor signaling and advanced prostate cancer.
      In this scenario, therapeutically exploitable molecular drivers of highly aggressive PCa comprise aberrations in androgen receptor (AR), phosphoinositide 3-kinase (PI3K), and DNA repair signaling pathways, among others.
      • Sandhu S
      • Moore CM
      • Chiong E
      • Beltran H
      • Bristow RG
      • Williams SG
      Prostate cancer.
      ,
      • Mateo J
      • Seed G
      • Bertan C
      • et al.
      Genomics of lethal prostate cancer at diagnosis and castration resistance.
      ,
      • Fraser M
      • Sabelnykova VY
      • Yamaguchi TN
      • et al.
      Genomic hallmarks of localized, non-indolent prostate cancer.
      Although specific subsets of patients benefit from these personalized strategies, the high biologic diversity, and the intra- and inter-patient heterogeneity of PCa hamper their long-term efficacy.
      • Cooper CS
      • Eeles R
      • Wedge DC
      • et al.
      Analysis of the genetic phylogeny of multifocal prostate cancer identifies multiple independent clonal expansions in neoplastic and morphologically normal prostate tissue.
      ,
      • Boutros PC
      • Fraser M
      • Harding NJ
      • et al.
      Spatial genomic heterogeneity within localized, multifocal prostate cancer.
      Therefore, a more profound knowledge of highly aggressive PCa biology is required to identify new therapeutic targets to tackle this lethal disease.
      In this sense, nowadays it is well-recognized that the splicing process and its regulation are highly relevant for understanding every hallmark of cancer, to the point that splicing alterations constitute another cancer hallmark.
      • Ladomery M
      Aberrant alternative splicing is another hallmark of cancer.
      • Oltean S
      • Bates DO
      Hallmarks of alternative splicing in cancer.
      • Rahman MA
      • Krainer AR
      • Abdel-Wahab O
      SnapShot: splicing alterations in cancer.
      Recent and increasing evidence from our group and others has documented that the machinery controlling the splicing process is often altered in different cancer types, which translates in an augmented pathobiological versatility (eg through generation of distinct/novel alternative splicing variants).
      • Jimenez-Vacas JM
      • Herrero-Aguayo V
      • Montero-Hidalgo AJ
      • et al.
      Dysregulation of the splicing machinery is directly associated to aggressiveness of prostate cancer.
      • Jimenez-Vacas JM
      • Herrero-Aguayo V
      • Gomez-Gomez E
      • et al.
      Spliceosome component SF3B1 as novel prognostic biomarker and therapeutic target for prostate cancer.
      • Hormaechea-Agulla D
      • Jimenez-Vacas JM
      • Gomez-Gomez E
      • et al.
      The oncogenic role of the spliced somatostatin receptor sst5TMD4 variant in prostate cancer.
      • Hormaechea-Agulla D
      • Gahete MD
      • Jimenez-Vacas JM
      • et al.
      The oncogenic role of the In1-ghrelin splicing variant in prostate cancer aggressiveness.
      • Fuentes-Fayos AC
      • Vazquez-Borrego MC
      • Jimenez-Vacas JM
      • et al.
      Splicing machinery dysregulation drives glioblastoma development/aggressiveness: oncogenic role of SRSF3.
      • Fuentes-Fayos AC
      • Perez-Gomez JM
      • GG ME
      • et al.
      SF3B1 inhibition disrupts malignancy and prolongs survival in glioblastoma patients through BCL2L1 splicing and mTOR/ss-catenin pathways imbalances.
      Likewise, proteins involved in the regulation of splicing (splicing factors [SFs]) play an oncogenic role in PCa by regulating mRNA splicing as well as other mRNA-metabolism processes and might represent exploitable therapeutic targets.
      • Paschalis A
      • Sharp A
      • Welti JC
      • et al.
      Alternative splicing in prostate cancer.
      ,
      • Takayama KI
      Splicing factors have an essential role in prostate cancer progression and androgen receptor signaling.
      More specifically, RNA binding motif (RBM) family members have particularly become a focus of interest in PCa field, given the essential role that some of these elements play in the progression of several tumor-types, including PCa.
      • Zeng Y
      • Wodzenski D
      • Gao D
      • et al.
      Stress-response protein RBM3 attenuates the stem-like properties of prostate cancer cells by interfering with CD44 variant splicing.
      • Zhao L
      • Li R
      • Shao C
      • Li P
      • Liu J
      • Wang K
      3p21.3 tumor suppressor gene RBM5 inhibits growth of human prostate cancer PC-3 cells through apoptosis.
      • Yamamoto R
      • Osawa T
      • Sasaki Y
      • et al.
      Overexpression of p54(nrb)/NONO induces differential EPHA6 splicing and contributes to castration-resistant prostate cancer growth.
      • Takayama KI
      • Suzuki T
      • Fujimura T
      • et al.
      Dysregulation of spliceosome gene expression in advanced prostate cancer by RNA-binding protein PSF.
      Among them, RBM22 stands out as a key member of the RBM family, since it is involved in the: (1) Activation of the spliceosome, by interacting with the U6 internal stem-loop and the pre-mRNA intron,
      • Rasche N
      • Dybkov O
      • Schmitzova J
      • Akyildiz B
      • Fabrizio P
      • Luhrmann R
      Cwc2 and its human homologue RBM22 promote an active conformation of the spliceosome catalytic centre.
      ,
      • Rasche N
      • Dybkov O
      • Schmitzová J
      • Akyildiz B
      • Fabrizio P
      • Lührmann R
      Cwc2 and its human homologue RBM22 promote an active conformation of the spliceosome catalytic centre.
      and (2) Regulation of gene expression, presumably acting as a transcription factor.
      • Van Nostrand EL
      • Freese P
      • Pratt GA
      • et al.
      A large-scale binding and functional map of human RNA-binding proteins.
      ,
      • Xiao R
      • Chen JY
      • Liang Z
      • et al.
      Pervasive chromatin-RNA binding protein interactions enable RNA-based regulation of transcription.
      Remarkably, although RBM22 has been associated with certain tumor-types, by regulating survival, mitosis and differentiation processes pathways,
      • Soubise B
      • Jiang Y
      • Douet-Guilbert N
      • Troadec MB
      RBM22, a key player of Pre-mRNA splicing and gene expression regulation, is altered in Cancer.
      its role in PCa remains completely unknown.
      Therefore, in the present study we used samples from 3 different well-characterized cohorts of PCa patients, various cellular (in vitro) and preclinical (in vivo) models, and a wide set of functional and molecular assays to evaluate RBM22 levels and its implication in the development, progression, and aggressiveness of PCa.

      Material and methods

      Patients and samples

      The experimental study of human PCa samples was approved by the Reina Sofia University Hospital Ethics Committee and conducted in accordance with the principles of the Declaration of Helsinki. The regional Biobank coordinated the collection, processing, management, and assignment of the biological samples used in the present study according to the standard procedures established for this purpose. Written informed consent was obtained from all patients. Three different cohorts of prostate samples were collected and analyzed:
      • -
        Cohort 1) formalin-fixed, paraffin-embedded (FFPE) PCa tissues (n = 84) and their non-tumor adjacent region (N-TAR; used as control tissues; n = 84), taken from radical prostatectomies from patients diagnosed with clinically localized PCa (Table I). The presence or absence of tumor was defined according to pathology review of haematoxylin/eosin-stained tissue. This cohort was used for RNA expression analysis.
        Table IDemographic, biochemical and clinical parameters of the patients from cohort 1
        Patients [n]84
        Age, years (median [IQR])61 (57-66)
        PSA levels, ng/mL (median [IQR])5.2 (4.2-8.0)
        Gleason score ≥ 7 (n (%))76 (90.5%)
        pT ≥ 3a (n (%))59 (70.2%)
        PI (n (%))72 (85.7%)
        VI (n (%))8 (9.52%)
        Recurrence (n (%))35 (41.7%)
        Metastasis [n (%)]0 (0%)
        PI: Perineural invasion; PSA: Prostate specific antigen; pT: Pathological primary tumor staging; VI: Vascular invasion.
      • -
        Cohort 2) fresh samples that were obtained by: (1) core needle biopsies from patients with significant PCa (n = 42; Table II); and (2) cystoprostatectomies from patients without PCa (n = 9; used as control tissue). The presence or absence of tumor was histologically confirmed by uro-pathologists. This cohort was also used for RNA analysis.
        Table IIDemographic, biochemical and clinical parameters of the patients from cohort 2
        Patients [n]42
        Age, years (median [IQR])75 (6981)
        PSA levels, ng/mL [median (IQR])62.0 (36.2254.5)
        Gleason score ≥ 7 (n (%))42 (100%)
        Metastasis (n [%])28 (66.7%)
        PSA: Prostate specific antigen; SigPCa: Significant PCa defined as Gleason score ≥ 7.
      • -
        Cohort 3) An additional set of samples was used for immunohistochemistry (IHC) analysis, which included tissue samples from benign prostatic hyperplasia (n = 11), prostatic intraepithelial neoplasia (n = 6), and PCa with GS 6 (n = 5).
      The clinical parameters collected from each patient were Gleason Score (GS; analyzed by uro-pathologists following the modified 2014 ISUP criteria), T-Stage, perineural invasion, lymphovascular invasion and presence of metastases at diagnosis (determined by computed tomography and bone scan).
      In addition, gene expression and clinical data from 3 additional and available in silico cohorts of patients (ie Lapointe, TCGA and SU2C/PCF cohorts) were downloaded from cBioPortal and GEO datasets.
      • Gao J
      • Aksoy BA
      • Dogrusoz U
      • et al.
      Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal.
      • Cerami E
      • Gao J
      • Dogrusoz U
      • et al.
      The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data.
      • Lapointe J
      • Li C
      • Higgins JP
      • et al.
      Gene expression profiling identifies clinically relevant subtypes of prostate cancer.
      Cancer Genome Atlas Research Network
      The molecular taxonomy of primary prostate Cancer.
      • Abida W
      • Cyrta J
      • Heller G
      • Prandi D
      • Armenia J
      • Coleman I
      • et al.
      Genomic correlates of clinical outcome in advanced prostate cancer.

      Cell lines and reagents

      Cell lines derived from normal prostate epithelium (RWPE-1) and PCa (LNCaP, 22Rv1 and PC-3) were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA) and cultured in a humidified incubator with 5% CO2 at 37°C according to ATCC recommendations, as previously described.
      • Jimenez-Vacas JM
      • Herrero-Aguayo V
      • Montero-Hidalgo AJ
      • et al.
      Dysregulation of the splicing machinery is directly associated to aggressiveness of prostate cancer.
      ,
      • Hormaechea-Agulla D
      • Jimenez-Vacas JM
      • Gomez-Gomez E
      • et al.
      The oncogenic role of the spliced somatostatin receptor sst5TMD4 variant in prostate cancer.
      ,
      • Hormaechea-Agulla D
      • Gahete MD
      • Jimenez-Vacas JM
      • et al.
      The oncogenic role of the In1-ghrelin splicing variant in prostate cancer aggressiveness.
      Cell line identity was validated by short tandem repeats sequences (STRs) analysis. All cell lines were tested for mycoplasma by PCR, as previously reported.
      • Jimenez-Vacas JM
      • Herrero-Aguayo V
      • Montero-Hidalgo AJ
      • et al.
      Dysregulation of the splicing machinery is directly associated to aggressiveness of prostate cancer.
      ,
      • Hormaechea-Agulla D
      • Jimenez-Vacas JM
      • Gomez-Gomez E
      • et al.
      The oncogenic role of the spliced somatostatin receptor sst5TMD4 variant in prostate cancer.
      ,
      • Hormaechea-Agulla D
      • Gahete MD
      • Jimenez-Vacas JM
      • et al.
      The oncogenic role of the In1-ghrelin splicing variant in prostate cancer aggressiveness.
      Dihydrotestosterone (DHT, S4757, Selleckchem, USA) was used at 10 nM.

      RNA-isolation, retrotranscription and real-time quantitative PCR (qPCR)

      RNA from FFPE samples, fresh tissues and cell-lines was isolated as previously reported.
      • Sáez-Martínez P
      • Jiménez-Vacas JM
      • León-González AJ
      • Herrero-Aguayo V
      • Montero Hidalgo AJ
      • Gómez-Gómez E
      • et al.
      Unleashing the diagnostic, prognostic and therapeutic potential of the neuronostatin/GPR107 system in prostate cancer.
      • Jimenez-Vacas JM
      • Gomez-Gomez E
      • Montero-Hidalgo AJ
      • et al.
      Clinical utility of ghrelin-O-Acyltransferase (GOAT) enzyme as a diagnostic tool and potential therapeutic target in prostate cancer.
      • Gómez-Gómez E
      • Jiménez-Vacas JM
      • Pedraza-Arévalo S
      • et al.
      Oncogenic role of secreted engrailed homeobox 2 (EN2) in prostate Cancer.
      Briefly, Maxwell 16 LEVRNA FFPE Kit (Promega, Madison, USA) was used in the Maxwell MDx 16 Instrument (Promega, Madrid, Spain) to isolate RNA from FFPE samples. Additionally, AllPrep DNA/RNA/Protein Mini Kit (Qiagen) and TRIzol Reagent (Thermo Fisher Scientific, Waltham, MA, USA) were used to isolate RNA from fresh tissues and PCa cell lines, respectively. RNA was DNase-treated using RNase-Free DNase Kit (Qiagen). Total RNA concentration and purity were assessed using Nanodrop One Spectrophotometer (Thermo Fisher Scientific, Madrid, Spain). Total RNA was retrotranscribed using random hexamer primers and the cDNA First Strand Synthesis kit (Thermo Scientific). Details regarding the development and validation of the primers for real time qPCR were previously reported by our laboratory.
      • Sáez-Martínez P
      • Jiménez-Vacas JM
      • León-González AJ
      • Herrero-Aguayo V
      • Montero Hidalgo AJ
      • Gómez-Gómez E
      • et al.
      Unleashing the diagnostic, prognostic and therapeutic potential of the neuronostatin/GPR107 system in prostate cancer.
      • Jimenez-Vacas JM
      • Gomez-Gomez E
      • Montero-Hidalgo AJ
      • et al.
      Clinical utility of ghrelin-O-Acyltransferase (GOAT) enzyme as a diagnostic tool and potential therapeutic target in prostate cancer.
      • Gómez-Gómez E
      • Jiménez-Vacas JM
      • Pedraza-Arévalo S
      • et al.
      Oncogenic role of secreted engrailed homeobox 2 (EN2) in prostate Cancer.
      Detailed information about the primers used herein (ie human RBM22, ACTB and GAPDH as well as mouse Rbm22 and Ppia) can be found in Supplemental Table 1. A normalization factor (calculated with ACTB and GAPDH expression levels, using GeNorm 3.3
      • Vandesompele J
      • De Preter K
      • Pattyn F
      • et al.
      Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes.
      ) and Ppia expression levels were used to adjust mRNA expression levels of the human and mouse genes, respectively.

      Immunohistochemistry

      IHC analysis was performed on samples from cohort 1 and cohort 3. The IHC protocol followed in this study was previously described.
      • Jimenez-Vacas JM
      • Herrero-Aguayo V
      • Montero-Hidalgo AJ
      • et al.
      Dysregulation of the splicing machinery is directly associated to aggressiveness of prostate cancer.
      Briefly, deparaffinized sections were incubated overnight (4°C) with the anti-RBM22 antibody (ab59157, Abcam Cambridge, UK) at 1:250 dilution, followed by incubation with an anti-rabbit horseradish peroxidase-conjugated secondary antibody (#7074; Cell Signalling). Finally, sections were developed with 3,39-diaminobenzidine (Envision system 2-Kit Solution DAB) and contrasted with haematoxylin. Independent pathologists performed histopathologic analyses indicating low, moderate, and high intensities of nuclear staining, following a blinded protocol.

      Stable transfection of RBM22

      LNCaP, 22Rv1 and PC-3 cells were stably transfected as previously reported.
      • Hormaechea-Agulla D
      • Jimenez-Vacas JM
      • Gomez-Gomez E
      • et al.
      The oncogenic role of the spliced somatostatin receptor sst5TMD4 variant in prostate cancer.
      ,
      • Hormaechea-Agulla D
      • Gahete MD
      • Jimenez-Vacas JM
      • et al.
      The oncogenic role of the In1-ghrelin splicing variant in prostate cancer aggressiveness.
      Briefly, cells were transfected with 1µg of RBM22 plasmid (OHu02939, GenScript) using Lipofectamine-2000 (Gibco, Barcelona, Spain) following manufacturer's instructions. Moreover, cells transfected with 1µg of pcDNA3.1 empty plasmid were used as control (called mock from now on). Transfected cells were selected by the continuous addition of geneticin at 1% (Gibco).

      Cell proliferation assay

      Cell proliferation was assessed by alamar-blue assay (Bio-Source International, Camarillo, CA, USA) as previously reported,
      • Jimenez-Vacas JM
      • Herrero-Aguayo V
      • Montero-Hidalgo AJ
      • et al.
      Dysregulation of the splicing machinery is directly associated to aggressiveness of prostate cancer.
      ,
      • Hormaechea-Agulla D
      • Jimenez-Vacas JM
      • Gomez-Gomez E
      • et al.
      The oncogenic role of the spliced somatostatin receptor sst5TMD4 variant in prostate cancer.
      ,
      • Hormaechea-Agulla D
      • Gahete MD
      • Jimenez-Vacas JM
      • et al.
      The oncogenic role of the In1-ghrelin splicing variant in prostate cancer aggressiveness.
      in response to RBM22 overexpression in LNCaP, 22Rv1 and PC-3. Specifically, 3,000 cells per well were seeded in 96-well plates, serum-starved overnight, and then fluorescence (540 nm excitation and 590 nm emission) was measured (after 3hour incubation with 10% resazurin) at 24, 48 and 72 hour using the FlexStation III system (Molecular Devices, Sunnyvale, CA, USA). All the data were normalized to values obtained in day 0 and represented as fold change compared to mock-transfected cells.

      Cell migration assay

      Cell migration was evaluated in PC-3, given its high invasiveness nature.
      • Su ZZ
      • Lin J
      • Shen R
      • Fisher PE
      • Goldstein NI
      • Fisher PB
      Surface-epitope masking and expression cloning identifies the human prostate carcinoma tumor antigen gene PCTA-1 a member of the galectin gene family.
      Specifically, 500,000 cells were seeded in 12-well plates. When confluence was reached, a scratch was made in each well, using a 100 µL tip and complete media was replaced by no serum media. Images were taken immediately after doing the scratch and 12 hour after. Wound healing was calculated as the area observed 12 hour after the wound was made vs the area observed just after wounding.

      Tumorspheres formation assay

      Tumorsphere formation assay was carried out, in LNCaP and PC-3 in response to RBM22 overexpression, as previously reported.
      • Jimenez-Vacas JM
      • Herrero-Aguayo V
      • Gomez-Gomez E
      • et al.
      Spliceosome component SF3B1 as novel prognostic biomarker and therapeutic target for prostate cancer.
      ,
      • Del Rio-Moreno M
      • Alors-Perez E
      • Borges de Souza P
      • et al.
      Peptides derived from the extracellular domain of the somatostatin receptor splicing variant SST5TMD4 increase malignancy in multiple cancer cell types.
      Briefly, 2,000 cells/well were seeded in Corning Costar 24-well ultra-low attachment plates (Merck, Madrid, Spain) with DMEM F-12 medium supplemented with 20 ng/mL EGF (Sigma-Aldrich, Madrid, Spain). The number of tumorspheres was determined after 14 days of incubation with ImageJ software.

      Colony formation assay

      To determine the clonogenic capacity of LNCaP and PC-3 cells in response to RBM22 overexpression, 2,000 cells were seeded into 6-well plates. Then, the medium was removed, the colonies washed with PBS and stained with crystal violet for 30 min and air-dried. The number of individual colonies and the percent of area covered with colonies by colony area were determined by ImageJ software (colony area plugin).

      RNA sequencing (RNAseq) in RBM22-overexpressing PCa cells

      RNA integrity of total RNA (500ng) from RBM22-overexpressing PC-3 cells (n = 3) and mock PC-3 cells (n = 3) was assessed using the Agilent 2100 Bioanalyzer. RNAseq was performed at the Genomics Core Unit of The National Centre for Cancer Research (CNIO), Madrid, Spain. Briefly, PolyA+ fraction was purified and randomly fragmented, converted to double-stranded cDNA and processed through subsequent enzymatic treatments of end-repair, dA-tailing, and ligation to adapters (NEBNext Ultra II Directional RNA Library Prep Kit for Illumina, NEB #E7760), as recommended by the kit manufacturer. Adapter-ligated library was completed by PCR with Illumina PE primers. The resulting purified cDNA libraries, with an average size of 400bp, were applied to an Illumina flow cell for cluster generation and sequenced on an Illumina instrument following manufacturer's protocols. Image analysis, per-cycle base-calling and quality score assignment were performed with Illumina HiSeq Control Software. Conversion of BCL files to FASTQ format was performed with the bcl2fastq Software (Illumina). Quantification at transcription level was developed using the Gencode transcriptome release 27 (GRCH38.p10)
      • Frankish A
      • Diekhans M
      • Ferreira AM
      • et al.
      GENCODE reference annotation for the human and mouse genomes.
      in transcripts per million (TPM) units with Salmon (v 0.7.2).
      • Patro R
      • Duggal G
      • Love MI
      • Irizarry RA
      • Kingsford C
      Salmon provides fast and bias-aware quantification of transcript expression.
      Gene level quantification was obtained by transforming transcript TPMs to counts per gene using the tximport library function from Bioconductor.
      • Soneson C
      • Love MI
      • Robinson MD
      Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences.
      Counts were transformed to log2 counts per million (logCPM) and filtered genes with mean logCPM < 0 were filtered out. Normalization was performed using TMM method from edgeR package and differentially expressed genes (DEGs) were analyzed using LIMMA package using limma-voom function adjusted using SVA package to remove batch effects.
      • Ritchie ME
      • Phipson B
      • Wu D
      • et al.
      limma powers differential expression analyses for RNA-sequencing and microarray studies.
      DEGs were functionally analyzed using Ingenuity Pathway Analysis (IPA, QIAGEN Redwood City, CA, USA, www.qiagen.com/ingenuity) to determine disease and function pathways, canonical pathways, and master regulators. The metrics entered were z-score and adjusted P-value, and default settings were used for the analysis. Pathways with z-score > 2 were considered “activated” and those with z-score < −2 were considered “inhibited”. SUPPA generateEvents was used to generate alternative splicing events defined from protein-coding transcripts and covering the annotated ORFs.
      • Trincado JL
      • Entizne JC
      • Hysenaj G
      • et al.
      SUPPA2: fast, accurate, and uncertainty-aware differential splicing analysis across multiple conditions.
      ,
      • Alamancos GP
      • Pages A
      • Trincado JL
      • Bellora N
      • Eyras E
      Leveraging transcript quantification for fast computation of alternative splicing profiles.
      The relative inclusion of an event was calculated as a Percent Spliced In (PSI) value with SUPPA psiPerEvent from the transcript TPM values obtained before. We applied a linear regression model to estimate the significance of the splicing changes and adjusted the P-value by calculating a false discovery rate (FDR). We considered significant the changes with an FDR corrected P-value < 0.05.

      Preclinical models of PCa

      Experiments with mice were carried out according to the European Regulations for Animal Care under the approval of the university/regional government research ethics committees.
      For tumor growth experiments, ten-week-old male athymic BALB/cAnNRj-Foxn1nu mice (Janvier Labs, Le Genest-Saint-Isle, France) were subcutaneously grafted in both flanks with 3 × 106 viable mock-transfected (n = 5 mice; n = 10 tumors) or RBM22-stably transfected PC-3 cells (n = 5 mice; n = 10 tumors) that were resuspended in 100 mL of basement membrane extract (Trevigen, Gaithersburg, MD, USA). Tumor growth was monitored once per week for 2 months using a digital caliper. After euthanization of mice, each tumor was kept at -80°C for later RNA extraction by using TRIzol reagent (Thermo Fisher Scientific) or protein extraction using SDS-DTT buffer as previously reported.
      • Hormaechea-Agulla D
      • Jimenez-Vacas JM
      • Gomez-Gomez E
      • et al.
      The oncogenic role of the spliced somatostatin receptor sst5TMD4 variant in prostate cancer.
      ,
      • Hormaechea-Agulla D
      • Gahete MD
      • Jimenez-Vacas JM
      • et al.
      The oncogenic role of the In1-ghrelin splicing variant in prostate cancer aggressiveness.
      Additionally, in order to determine the expression of RBM22 during PCa progression in vivo, we used the transgenic adenocarcinoma of mouse prostate (TRAMP) mice, heterozygous for the PB-Tag transgene, maintained in a C57BL/6 background and crossed with non-transgenic FVB mice to obtain transgenic (C57BL/6 × FVB) F1 males. TRAMP mice were sacrificed at 13, 21, and 30 weeks of age, when it has been demonstrated that these mice develop PIN, moderately differentiated PCa and poorly differentiated PCa, respectively.
      • Gelman IH
      How the TRAMP model revolutionized the study of prostate Cancer progression.
      Findings were then confirmed in a second PCa mouse model: the Hi-Myc (ARR2/Pbsn-Myc) mouse strain maintained in an FVB background which was obtained originally from NCI and backcrossed to C57BL/6 for more than 7 generations at Dr Olmos lab at CNIO to obtain pure genetic background. Hi-Myc mice were sacrificed at 4 months (No tumor), 6–8 months (predominantly PIN) and 12–15 months (predominately invasive carcinoma). PCR genotyping in both models was performed in genomic DNA extracted from the tail vein blood by PCR. TRAMP model was genotyped using the primers recommended by The Jackson Laboratory for TRAMP (Sense: TACAACTGCCAACTGGGATG; Antisense: CAGGCACTCCTTTCAAGACC) and those recommended by NCI′ to examine the transgene (sense: AAACATGATGACTACCAAGCTTGGC; antisense: ATGATAGCATCTTGTTCTT AGTCTTTTTCTTAATAGGG). All prostate tissues were processed and divide in specular fragments for formalin-fixation and paraffin inclusion or stored as fresh-frozen tissue at −80°C and used for gene expression and protein analysis.

      nCOUNTER analysis

      nCounter PanCancer Pathways Panel kit (GXA-PATH1-12; NanoString Technologies) was used and performed at the Laboratory of Genetics at UCAIB (IMIBIC) to simultaneously examine the expression of 730 genes associated with cancer (ie 606 genes representing all major cancer pathways and 124 key cancer driver genes), as previously described.
      • Sarmento-Cabral A
      • F LL
      • Gahete MD
      • Castano JP
      • Luque RM
      Metformin reduces prostate Tumor growth, in a diet-dependent manner, by modulating multiple signaling pathways.
      Briefly, the quality of all samples (RBM22-overexpressing xenograft tumors [derived from PC-3 cells; n = 3] and mock-overexpressing tumors [derived from PC-3 cells; n = 3]) was analyzed using the Agilent 2100 Bioanalyzer. Then, 100ng of RNA from all the samples were loaded in the nCounter PanCancer Pathways Panel kit and the experiment was run in the nCounter Analysis System (NanoString Technologies), following manufacturer's protocol. The data were analyzed using the nSolverAnalysisSoftware 3.0.22 from NanoStringTechnologies. Data were normalized using 40 genes as housekeeping genes as previously described.
      • Sarmento-Cabral A
      • F LL
      • Gahete MD
      • Castano JP
      • Luque RM
      Metformin reduces prostate Tumor growth, in a diet-dependent manner, by modulating multiple signaling pathways.
      All specific target sequences and panel details are available on the manufacture's webpage.

      Statistical analysis

      All the experiments were performed in at least 3 independent experiments (n≥3). Statistical differences between 2 variables were calculated by unpaired parametric t'test and nonparametric Mann Whitney U test, according to normality, assessed by Kolmogorov-Smirnov test. For differences among 3 variables, 1-Way ANOVA analysis was performed. Statistical significance was considered when P < 0.05. A trend for significance was indicated when P values ranged between >0.05 and <0.1. All the analyses were assessed using GraphPad Prism 8 (GraphPad Software, La Jolla, CA, USA).

      Results

      RBM22 levels are significantly decreased in PCa and inversely associate with key clinical and molecular features of aggressiveness

      RBM22 mRNA levels were slightly but consistently lower in PCa tissues as compared to non-tumor prostate tissues in 2 independent cohorts of samples available in our laboratory (cohort 1 and 2; Fig 1A). Consistently, downregulation of RBM22 in PCa was also observed in 2 independent cohorts of patients available in silico, Lapointe and TCGA datasets
      • Lapointe J
      • Li C
      • Higgins JP
      • et al.
      Gene expression profiling identifies clinically relevant subtypes of prostate cancer.
      ,
      Cancer Genome Atlas Research Network
      The molecular taxonomy of primary prostate Cancer.
      (Fig 1B). Genetic status of RBM22 was interrogated in the TCGA dataset (primary PCa samples, PanCancer Atlas)
      Cancer Genome Atlas Research Network
      The molecular taxonomy of primary prostate Cancer.
      and in SU2C/PCF cohort (mCRPC samples).
      • Abida W
      • Cyrta J
      • Heller G
      • Prandi D
      • Armenia J
      • Coleman I
      • et al.
      Genomic correlates of clinical outcome in advanced prostate cancer.
      Specifically, although genomic alterations of RBM22 were not common in these cohorts, RBM22 mRNA levels were associated with RBM22 copy number (Supplemental Figure 1A and 1B). On the other hand, RBM22 mRNA levels did not consistently associate with any molecular subtype or common genomic aberration of PCa (Supplemental Figure 1C and 1D). Further clinical and molecular interrogation of tumor biopsies revealed that lower RBM22 expression levels were associated with extraprostatic extension and perineural invasion capacity (Fig 1C), but not with disease free survival or overall survival in TCGA and SU2C/PCF cohorts, respectively (Supplemental Figure 2). Moreover, the expression levels of RBM22 tended to be inversely correlated with those of KLK3 and PCA3 (Fig 1D), while non-significant correlations/associations were found when analyzing Gleason Score, plasma PSA levels, age or presence of metastasis (Supplemental Figure 3A and 3B). Likewise, expression levels of RBM22 were not significantly correlated to those of AR, AR-V7 or MKI67 in our cohort of samples (Supplemental Figure 3C).
      Fig 1
      Fig 1RBM22 levels in human samples. A, Comparison of RBM22 mRNA levels in two cohorts of patients: cohort-1) formalin-fixed paraffin-embedded (FFPE) samples from PCa tissues vs nontumor adjacent regions (N-TAR) (left panel; n = 84); and cohort-2) freshly collected biopsy samples from patients with significant PCa (n = 42) vs non-tumor prostate samples (right panel; n = 9). RBM22 levels were determined by qPCR. Data are represented as mean ± SEM of mRNA levels adjusted by normalization factor (calculated from ACTB and GAPDH expression levels). B, Comparison of RBM22 mRNA levels from Lapointe (left panel; n = 112) and TCGA (right panel; n = 376) datasets.
      • Lapointe J
      • Li C
      • Higgins JP
      • et al.
      Gene expression profiling identifies clinically relevant subtypes of prostate cancer.
      ,
      Cancer Genome Atlas Research Network
      The molecular taxonomy of primary prostate Cancer.
      Data are represented as mean ± SEM of mRNA levels Log2 of the ratio of the median of Channel 2 (635 nm) to Channel 1 (532 nm) in Lapointe cohort, and RSEM (Batch normalized from Illumina HiSeq_RNASeqV2) in TCGA cohort. C, Association between RBM22 mRNA levels and clinical parameters (extraprostatic extension and perineural invasion) in PCa samples from cohort-2 (n = 42). Data are represented as mean ± SEM of mRNA levels adjusted by normalization factor (calculated from ACTB and GAPDH expression levels). D, Correlation between RBM22 mRNA levels and expression levels of KLK3 and PCA3. Correlations are represented by mean (connecting line) and error bands (pointed line) of expression levels. E, Comparison of RBM22 protein levels by immunohistochemistry (IHC) between a representative set of benign prostatic hyperplasia (BPH; n = 11), prostatic intraepithelial neoplasia (PIN; n = 6) and prostate cancer (PCa; n = 47) samples (left panel). Representative images (200x magnification) are shown in the right panel. F, Association of RBM22 protein levels with Gleason Score (GS 6 [n = 5] and GS≥7 [n = 42]). G, Association of RBM22 protein levels with presence of metastasis at diagnosis (no metastasis [n = 19] and metastasis [n = 28]). Representative images (200x magnification) are shown in the right panel. Data are represented as mean ± SEM of IHC staining scaled from low
      • Bray F
      • Ferlay J
      • Soerjomataram I
      • Siegel RL
      • Torre LA
      • Jemal A.
      Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
      to high
      • Sandhu S
      • Moore CM
      • Chiong E
      • Beltran H
      • Bristow RG
      • Williams SG
      Prostate cancer.
      intensity.
      Then, RBM22 protein levels were analyzed in a set of biopsies samples which included benign prostatic hyperplasia (BPH), prostatic intraepithelial neoplasia (PIN) and PCa tissues (Fig 1E). Consistent with the previous data observed at the mRNA level, the protein levels of RBM22 were significantly lower in PCa samples as compared to non-tumor and BPH tissues (Fig 1E). Moreover, consistent also with its role as splicing and transcription factor, RBM22 staining was exclusively detected in the nucleus of the cells (Fig 1E). Interestingly, RBM22 protein levels were especially low in PCa samples with Gleason score ≥ 7 (Fig 1F). Likewise, RBM22 staining tended to be (P = 0.082) lower in diagnostic biopsies from patients with metastasis compared with those without metastasis (Fig 1G).

      Low RBM22 levels are associated with poorly differentiated prostate cancer in the TRAMP model

      The TRAMP mouse model mimics the progression of human PCa by developing PIN, moderately-differentiated prostate cancer (MD-PCA) and poorly-differentiated prostate cancer (PD-PCA) over time.
      • Gelman IH
      How the TRAMP model revolutionized the study of prostate Cancer progression.
      Therefore, this model was used to further interrogate RBM22 levels during PCa progression. We first corroborated that the TRAMP model developed PIN, MD-PCa and PD-PCa at 13, 21 and 30 weeks of age, respectively (Fig 2A). Our results showed that RBM22 mRNA levels were significantly lower in MD-PCa and PD-PCa samples compared to PIN samples (Fig 2B). Consistently, a progressive decrease in RBM22 protein levels was also observed when comparing PIN, MD-PCa, and PD-PCa samples from TRAMP mice (Fig 2C). These findings were also reproduced in the Hi-Myc model (Fig 2,C–E).
      Fig 2
      Fig 2A, Schematic representation of the generation of the preclinical transgenic adenocarcinoma of mouse prostate (TRAMP) model. Representative images of prostatic intraepithelial neoplasia (PIN), moderately-differentiated PCa (MD-PCa) and poorly differentiated PCa (PD-PCa) obtained at 13, 21 and 30 weeks of age from TRAMP mice are shown in the bottom panel, respectively. Images of PIN and MD-PCa are taken with 100x magnification and PD-PCa with 200x magnification. B, Comparison of RBM22 mRNA levels from PIN (n = 5), MD-PCa (n = 4) and PD-PCa (n = 5) derived from TRAMP mice. Data are represented as mean ± SEM of mRNA levels adjusted by CICLO expression levels. Asterisks (* P < 0.05; *** P < 0.001) indicate statistically significant differences between groups. C, Comparison of RBM22 protein levels from PIN (n = 5), MD-PCa (n = 4) and PD-PCa (n = 5) derived from TRAMP mice. Protein levels were normalized by total protein loading (Ponceau staining). Asterisks (** P < 0.01) indicate statistically significant differences between groups. Western-Blot images of RBM22 protein levels are included in the right panel.

      Overexpression of RBM22 reduced relevant functional parameters of PCa aggressiveness in vitro

      Different human PCa cell-line models (LNCaP, 22Rv1 and PC-3) and a normal-like prostate cell-line (RWPE-1) were used to perform functional experiments. First, RBM22 expression levels were found to be significantly lower in all the PCa cell lines used herein compared to the normal- prostate cell line RWPE-1 (Fig 3A), indicating that LNCaP, 22Rv1 and PC-3 cell-models were appropriate PCa models to study the functional role of RBM22.
      Fig 3
      Fig 3Functional consequences of RBM22 overexpression in prostate cancer-derived cell lines. A, Comparison of RBM22 mRNA expression levels between non-tumor (RWPE-1) and tumor (22Rv1, LNCaP and PC-3) prostate-derived cell lines. Data are represented as mean ± SEM of mRNA levels adjusted by normalization factor (calculated from ACTB and GAPDH expression levels). B, Validation of RBM22 overexpression (by qPCR) in LNCaP, 22Rv1 and PC-3 cell-lines. Data (mean ± SEM) are represented as percentage of mock-transfected cells (indicated with the dotted line at 100%). RBM22 expression levels (mRNA) were adjusted by normalization factor (calculated from ACTB and GAPDH expression levels). C, Proliferation rate of LNCaP, 22Rv1 and PC-3 cells in response to RBM22 overexpression compared to mock-transfected cells (indicated with the dotted line at 100%). Effect of RBM22 overexpression in LNCaP D, and PC-3 E, colony number (left panel) and area (right panel) compared to mock-transfected cells. Representative images of LNCaP and PC-3 colonies are depicted on bottom panel. Effect of RBM22 overexpression in LNCaP F, and PC-3 G, tumorsphere number (left panel) and area (right panel) compared to mock-transfected cells. Representative images of LNCaP and PC-3 tumorspheres are depicted on bottom panel. H, Migration rate of PC-3 cells in response to RBM22 overexpression compared to mock-transfected cells. Representative images are depicted on right panel. Asterisks (* P < 0.05; ** P < 0.01; *** P < 0.001) indicate statistically significant differences between groups.
      Stable overexpression of RBM22 in LNCaP, 22Rv1 and PC-3 cell lines by transfecting the cells with an RBM22-overexpressing plasmid (Fig 3B) resulted in a significant decrease in the proliferation rate at 24-, 48-, and 72-hour in 22Rv1 cells, and at 48-, and 72-hour in LNCaP and PC-3 cells compared with mock-transfected control cells (Fig 3C). Similarly, RBM22-overexpressing LNCaP and PC-3 cells generated fewer and smaller colonies than mock-transfected control cells (Fig 3,D and E, respectively). Likewise, the number of tumorspheres was also significantly decreased in response to RBM22 overexpression in LNCaP and PC-3 cells, while their size was not affected (Fig 3, F and G, respectively). Finally, given its high invasiveness nature, the role of RBM22 on migration capacity was explored in PC-3 cells.
      • Su ZZ
      • Lin J
      • Shen R
      • Fisher PE
      • Goldstein NI
      • Fisher PB
      Surface-epitope masking and expression cloning identifies the human prostate carcinoma tumor antigen gene PCTA-1 a member of the galectin gene family.
      Our results showed that RBM22 overexpression markedly decreased the migration rate of PC-3 cells after 12 hours (Fig 3H). All these results revealed that RBM22 overexpression affected different critical functional endpoints associated with the development, progression, and aggressiveness of PCa cells.

      Overexpression of RBM22 impairs PCa progression in vivo by the modulation of critical signaling pathways in PCa

      To validate the RBM22 antitumor actions in vivo, we generated another preclinical model to monitor the growth of xenograft PC-3 tumors overexpressing RBM22 compared to mock-transfected PC-3 cells (Fig 4A). Specifically, analysis of tumor growth showed that PCa progression in vivo was completely blunted in RBM22-overexpressing tumors vs control (mock-transfected) tumors (Fig 4B). Consistently, overexpression of RBM22 in vivo resulted in smaller tumors as compared to mock-cells induced tumors (Fig. 4C). It should be mentioned that RBM22 overexpression was validated at the end of the experiment in the harvested tumors (Fig. 4D).
      Fig 4
      Fig 4Functional and molecular response to RBM22 overexpression in a PC-3 xenograft model. A, Generation of a preclinical-xenograft PCa-model by subcutaneously inoculating PC-3 cells in one flank cells overexpressing RBM22 and in the other flank mock-transfected cells (10 tumors/condition). Comparison between the growth over time B, and weight at the end of experiment C, of xenograft tumors derived from mock-transfected cells and RBM22-overexpressing cells. Representative images of tumors are depicted in top panel. D, Validation of RBM22 overexpression by qPCR. mRNA levels were adjusted by normalization factor (calculated from ACTB and GAPDH expression levels). Data (mean ± SEM) are represented as percentage of mock-transfected xenograft tumors. Asterisks (*** P < 0.001) indicate statistically significant differences between groups. E, Hierarchical heatmap generated using the expression levels of significantly altered cancer-related genes (154 genes) in RBM22-overexpressing (green) vs mock-transfected (red) xenograft tumors. F, Volcano plot showing differentially expressed genes (DEGs) in RBM22-overexpressing vs mock xenograft tumors. Top 10 significantly altered genes are depicted. Red dots represent statistically significant (P-value < 0.05) and black dots non-significant DEGs. G, Cancer-related pathways altered in RBM22-overexpressing tumors. Percentage of altered genes, number of dysregulated genes, and total number of genes per pathway are depicted. These expression levels were obtained and analyzed using nCounter PanCancer platform.
      Overexpression of RBM22 in vivo altered the expression of 154 out of 770 genes analyzed using the nCounter PanCancer Pathways Panel kit (Fig 4, E and F). Specifically, cell cycle-apoptosis pathway was the most altered signaling pathway, followed by PI3K pathway and driver gene pathway (Fig 4G). Additionally, other critical signaling pathways were also altered, including MAPK, Ras, transcription misregulation, Wnt and JAK-STAT pathways, among others (Fig 4G).

      Molecular landscape in response to RBM22 overexpression revealed that RBM22 regulates the splicing pattern of critical genes and the activity of oncogenic signaling pathways in PCa cells

      Given the pivotal role of AR in PCa progression
      • Westaby D
      • Fenor de La Maza MLD
      • Paschalis A
      • et al.
      A new old target: androgen receptor signaling and advanced prostate cancer.
      together with previous studies reporting a causal link between SFs and AR pathway,
      • Paschalis A
      • Sharp A
      • Welti JC
      • et al.
      Alternative splicing in prostate cancer.
      ,
      • Munkley J
      • Li L
      • Krishnan SRG
      • et al.
      Androgen-regulated transcription of ESRP2 drives alternative splicing patterns in prostate cancer.
      ,
      • Paschalis A
      • Welti J
      • Neeb AJ
      • et al.
      JMJD6 Is a Druggable oxygenase that regulates AR-V7 expression in prostate Cancer.
      the potential relationship between RBM22 and AR was investigated herein in LNCaP (hypersensitive to AR stimulation, AR +/AR-V7 ) and 22Rv1 (AR +/AR-V7 +) cells.
      • Horoszewicz JS
      • Leong SS
      • Kawinski E
      • et al.
      LNCaP model of human prostatic carcinoma.
      • Sramkoski RM
      • Pretlow 2nd, TG
      • Giaconia JM
      • et al.
      A new human prostate carcinoma cell line, 22Rv1.
      • Sowalsky AG
      • Figueiredo I
      • Lis RT
      • et al.
      Assessment of Androgen receptor splice variant-7 as a biomarker of clinical response in castration-sensitive prostate cancer.
      Specifically, DHT treatment did not dysregulate RBM22 levels in LNCaP cells (Supplemental Figure 4A). In addition, RBM22 overexpression did not alter AR nor AR-V7 levels in LNCaP and 22RV1 cells (Supplemental Figure 4B-D). Then, an unbiased approach was followed, a RNAseq analysis in RBM22-overexpressing PC-3 cells. This transcriptome analysis revealed that RBM22 overexpression dysregulated the splicing pattern of 397 genes by altering 669 alternative splicing events, being alternative first exon and exon skipping the most common events (41% and 30%, respectively; Fig 5A and Supplemental Figure 5). Moreover, according to its role as a potential transcription factor, the overexpression of RBM22 altered the expression levels of 4245 genes with an adjusted P-value < 0.01 (Fig 5B; Supplemental Table 2). Specifically, disease and function pathways analyses showed that cell cycle progression was the most significantly inactivated pathway in response to RBM22 overexpression (Fig 5C). In addition, RBM22 overexpression decreased the activity of other key pathways tightly related to cell cycle, including cell proliferation of tumor cell lines, metabolism of DNA, M phase, Cytokinesis, M phase of tumor cell lines, Mitosis, DNA replication and cancer cells, among others (Fig 5C, Supplemental Table 3). On the other hand, overexpression of RBM22 was associated to high activity of autophagy, development of cytoplasm, invasion of cells and cellular homeostasis among others (Fig 5C, Supplemental Table 3).
      Fig 5
      Fig 5Molecular and signaling landscape in response to RBM22 overexpression in PCa cells. A, Distribution of splicing events significantly altered in RBM22-overexpressing vs mock PC-3 cells (AF: Alternative First Exon; SE: Exon skipping; A5: Alternative 5′ Splice Site; A3: Alternative 3′ Splice Site; AL: Alternative Last Exon; RI: Intron retention; MX: Mutually Exclusive Exon). B, Volcano plot showing differentially expressed genes in RBM22-overexpressing vs mock-overexpressing PC-3 cells. Red dots represent statistically significant and black dots non-significant differentially expressed genes (DEGs), respectively. C, Activation status of disease and function pathways enriched in RBM22-overexpressing PC-3 cells. Top 10 most-activated and top 10 least-activated pathways are depicted. D, Activation status of canonical pathways enriched in RBM22-overexpressing PC-3 cells. Red blots represent canonical pathways significantly inhibited and green blots represent canonical pathways significantly activated in RBM22-overexpressing PC-3 cells. Black dots represent non-significantly enriched pathways. (E-G), Analysis of the expression levels of master regulators in response to RBM22 overexpression. E, Differentially expressed master regulators in RBM22-overexpressing PC-3 cells (data from RNAseq). Red blots represent downregulated and green dots upregulate genes in RBM22-overexpressing PC-3 cells. F, Master regulators differentially expressed in response to RBM22 overexpression in 3 independent biological replicates from PC-3 and 22Rv1 cells. G, Expression levels of differentially expressed master regulators in PC-3 xenografts overexpressing RBM22. Data (mean ± SEM) are represented as fold change of mock-transfected xenograft tumors. Asterisks (* P < 0.05; ** P < 0.01; *** P < 0.001) indicate statistically significant differences between groups. “(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)”
      Consistently, the transcriptional landscape predicted 216 canonical pathways that were altered in RBM22-overexpressing PC-3 cells. Specifically, RBM22 overexpression decreased the activation of key oncogenic pathways involved in cell cycle regulation including cell cycle control of chromosomal replication, kinetochore metaphase signaling pathway, mitotic role of polo-like kinase, and semaphorin pathway, among others (Fig 5D; Supplemental Table 4). Similarly, RBM22 overexpression also decreased the activation of antioncogenic pathways, including ephrin receptor, antiproliferative role of SSTR2, integrin signaling, thrombin signaling, and autophagy pathway, among others (Fig 5D; Supplemental Table 4).
      Transcriptional changes in RBM22-overexpressing cells were further analysed to interrogate the expression levels of master regulators and therefore identify the main molecular drivers underlying the antitumor role of RBM22 in PCa cells. Specifically, RNAseq data showed that EPAS1, WT1, AURKB, DNMT3B, MYBL2, CDK1, PRKCD, WWC1, HIC1, BHLHA15, ATR and CAV1 were downregulated, while ZMIZ2, TSPYL2, SGK1, LEF1, FOXN2, TCF4, and MITF were upregulated in response to RBM22 overexpression in PC-3 cells (Fig 5E). Changes in the expression levels of ATR, AURKB, CAV1, CDK1, EPAS1 and HIC1 were further validated by RT-qPCR in PC-3 and 22Rv1 using 3 independent biological replicates (Fig 5F; Supplemental Figure 6A-6C). Finally, the decrease in the expression levels of ATR, CDK1 and EPAS1 in response to RBM22 expression was also validated in the preclinical xenograft in vivo mouse-model (Fig 5G).

      Discussion

      RNA binding motif (RBM) proteins represent a complex family comprised by proteins that can either fuel cancer progression or suppress tumor development and/or progression.
      • Li Z
      • Guo Q
      • Zhang J
      • et al.
      The RNA-binding motif protein family in cancer: friend or foe?.
      ,
      • Dvinge H
      • Kim E
      • Abdel-Wahab O
      • Bradley RK
      RNA splicing factors as oncoproteins and tumour suppressors.
      Molecularly, although the main function of RBM proteins is the regulation of splicing process,
      • Li Z
      • Guo Q
      • Zhang J
      • et al.
      The RNA-binding motif protein family in cancer: friend or foe?.
      some of them can also regulate transcription, therefore controlling a plethora of signaling pathways.
      • Xiao R
      • Chen JY
      • Liang Z
      • et al.
      Pervasive chromatin-RNA binding protein interactions enable RNA-based regulation of transcription.
      ,
      • De Maio A
      • Yalamanchili HK
      • Adamski CJ
      • et al.
      RBM17 interacts with U2SURP and CHERP to regulate expression and splicing of RNA-processing proteins.
      This is the case of RBM22, which has an RNA Recognition Motif (RRM) but also a Zinc-Finger like domain.
      • Xiao R
      • Chen JY
      • Liang Z
      • et al.
      Pervasive chromatin-RNA binding protein interactions enable RNA-based regulation of transcription.
      As other members of the RBM family, the role of RBM22 in cancer seems to be context-dependent, inasmuch as RBM22 has been found to be overexpressed in glioblastoma multiforme,
      • Fuentes-Fayos AC
      • Vazquez-Borrego MC
      • Jimenez-Vacas JM
      • et al.
      Splicing machinery dysregulation drives glioblastoma development/aggressiveness: oncogenic role of SRSF3.
      but lost in myelodysplastic syndrome (ie precursor of acute myeloid leukemia).
      • Boultwood J
      • Pellagatti A
      • Cattan H
      • et al.
      Gene expression profiling of CD34+ cells in patients with the 5q- syndrome.
      Surprisingly, although many splicing factors have been found to be dysregulated and play a key role in the development, progression and treatment resistance of PCa,
      • Paschalis A
      • Sharp A
      • Welti JC
      • et al.
      Alternative splicing in prostate cancer.
      ,
      • Takayama KI
      Splicing factors have an essential role in prostate cancer progression and androgen receptor signaling.
      ,
      • Jimenez-Vacas JM
      • Herrero-Aguayo V
      • Montero-Hidalgo AJ
      • et al.
      Dysregulation of the splicing machinery is directly associated to aggressiveness of prostate cancer.
      the levels and potential pathophysiological actions that RBM22 might play in this tumor type remain completely unknown. Therefore, we herein evaluated the levels of RBM22 and interrogated its potential functional role in PCa, one of the top leading causes of cancer-related deaths among male population in developed countries,
      • Bray F
      • Ferlay J
      • Soerjomataram I
      • Siegel RL
      • Torre LA
      • Jemal A.
      Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
      which is still lacking global therapeutic targets that could lead to more effective therapeutic approaches. In this sense, our results showed that the expression of RBM22 was downregulated in PCa samples from 3 cohorts of samples available at our laboratory and 2 publicly available datasets. Interestingly, RBM22 protein levels were especially low in highly aggressive PCa samples (ie Gleason score ≥ 7, extra prostatic extension and perineural invasion), suggesting a potential tumor suppressor role for this protein in PCa. Consistently, RBM22 decreased (at mRNA and protein level) gradually from PIN to poorly differentiated PCa in samples from TRAMP mice (a transgenic model that closely mimics human PCa progression)
      • Gelman IH
      How the TRAMP model revolutionized the study of prostate Cancer progression.
      , thus reinforcing the link between low RBM22 levels and PCa progression and aggressiveness.
      Given that RBM22 was significantly downregulated in all the PCa models tested herein (ie cell lines as well as tumors from the TRAMP and Hi-MYC mice), our approach to study the potential pathophysiological role of RBM22 in PCa consisted in overexpressing RBM22 in LNCaP and 22Rv1 (AR driven cells) as well as in PC-3 (AR-independent cells). Specifically, RBM22 overexpression decreased aggressiveness features of these PCa cell-lines, including cell proliferation and migration. In addition, RBM22 overexpression also strikingly decreased PCa-stem/progenitor cells in terms of tumorspheres and colonies number and/or area, both relevant functional results that may help to explore the PCa-onset and how to overcome the well-known PCa-resistance to different/current drugs.
      • Verma S
      • Shankar E
      • Kalayci FNC
      • et al.
      Androgen Deprivation induces transcriptional reprogramming in prostate Cancer cells to develop stem cell-like characteristics.
      ,
      • Zhang L
      • Jiao M
      • Li L
      • et al.
      Tumorspheres derived from prostate cancer cells possess chemoresistant and cancer stem cell properties.
      Consistently, we also demonstrated the antitumor actions of RBM22 in vivo by generating a preclinical model, since tumor progression was completely blocked in xenografted RBM22-overexpressing tumors (ie tumor grew dramatically slower and tumor weight was significantly lower than mock-transfected control tumors). Thus, all these robust in vitro/in vivo results, together with the observations using 5 different human cohorts, unveiled an important pathophysiological role of RBM22 in PCa. Specifically, these results indicate that RBM22 plays an antitumor role in PCa cells, which is consistent with previous studies showing that RBM3, RBM5 and RBM25 act as tumor suppressors in PCa and are inversely related to PCa aggressiveness.
      • Zeng Y
      • Wodzenski D
      • Gao D
      • et al.
      Stress-response protein RBM3 attenuates the stem-like properties of prostate cancer cells by interfering with CD44 variant splicing.
      ,
      • Zhao L
      • Li R
      • Shao C
      • Li P
      • Liu J
      • Wang K
      3p21.3 tumor suppressor gene RBM5 inhibits growth of human prostate cancer PC-3 cells through apoptosis.
      ,
      • Ge Y
      • Schuster MB
      • Pundhir S
      • et al.
      The splicing factor RBM25 controls MYC activity in acute myeloid leukemia.
      In this sense, it has been reported that 5q33.1, the chromosome region harboring RBM22 gene, is commonly lost in Afro-American patients as compared to Caucasian patients.
      • Rose AE
      • Satagopan JM
      • Oddoux C
      • et al.
      Copy number and gene expression differences between African American and Caucasian American prostate cancer.
      In this context, it is widely known that prostate tumors derived from Afro-American patients are usually more aggressive than those derived from Caucasian patients;
      • Powell IJ
      Epidemiology and pathophysiology of prostate cancer in African-American men.
      ,
      • Reddy S
      • Shapiro M
      • Morton Jr, R
      • Brawley OW
      Prostate cancer in black and white Americans.
      whether RBM22 loss is a player involved in Afro-American PCa oncogenesis and aggressiveness remains unknown, but our data open a new research avenue on that direction and suggest that RBM22-overexpression could be a novel therapeutic avenue with relevant pathophysiological/clinical-potential, to combat this devastating disease.
      Therefore, our data lay the foundations to further study RBM22 biology in order to unveil the molecular and signaling mechanisms underlying the decrease in its expression levels in advanced PCa, and therefore identify novel therapeutic targets to tackle this lethal disease. In this regard, although RBM22 copy number alterations associated with RBM22 mRNA levels in the cohorts analyzed herein, many diploid samples exhibited low mRNA levels of RBM22, similar to those observed in samples harboring RBM22 shallow or deep deletions. Moreover, differences in RBM22 levels between PCa and non-tumor tissues were more pronounced at protein levels compared to mRNA levels. These data might suggest that the downregulation of RBM22 observed in PCa could be due in part to RBM22 shallow deletions, but in a more extent to an alteration in the transcriptomic rate/efficiency of RBM22 gene (eg, epigenetic changes) and/or to the dysregulation of RBM22 translation efficiency or protein stability. In this sense, many studies have reported a potential key role that proteins involved in ubiquitination and deubiquitination of SFs (E3-ligases and deubiquitinating enzymes) play to regulate splicing process.
      • Ka HI
      • Lee S
      • Han S
      • et al.
      Deubiquitinase USP47-stabilized splicing factor IK regulates the splicing of ATM pre-mRNA.
      • Das T
      • Park JK
      • Park J
      • et al.
      USP15 regulates dynamic protein-protein interactions of the spliceosome through deubiquitination of PRP31.
      • Song EJ
      • Werner SL
      • Neubauer J
      • et al.
      The Prp19 complex and the Usp4Sart3 deubiquitinating enzyme control reversible ubiquitination at the spliceosome.
      However, further studies are needed to fully address the molecular mechanisms driving RBM22 downregulation in PCa.
      Mechanistically, RBM22 has been found to shuttle between the nucleus and the cytoplasm in response to cellular stress conditions in other cell-types (regulating Calcium signaling through the interaction with ALG-2 and SLU7).
      • Janowicz A
      • Michalak M
      • Krebs J
      Stress induced subcellular distribution of ALG-2, RBM22 and hSlu7.
      ,
      • Montaville P
      • Dai Y
      • Cheung CY
      • et al.
      Nuclear translocation of the calcium-binding protein ALG-2 induced by the RNA-binding protein RBM22.
      Nevertheless, our data showed that RBM22 protein staining was restricted to the nuclei of the tumor cells, which suggest that the main processes regulated by RBM22 in PCa cells might be alternative splicing and gene transcription. According to this observation we aimed to identify the splicing events differentially processed and the genes differentially expressed in response to RBM22 overexpression. To that aim, RNA from RBM22-overexpressing and mock-transfected PC-3 cells were used. Our data demonstrated that RBM22 overexpression triggers a profound alteration in the splicing pattern of 856 genes, which is plausible considering the pivotal role that RBM22 plays in the activation of the spliceosome.
      • Rasche N
      • Dybkov O
      • Schmitzova J
      • Akyildiz B
      • Fabrizio P
      • Luhrmann R
      Cwc2 and its human homologue RBM22 promote an active conformation of the spliceosome catalytic centre.
      Interestingly, among all the types of splicing events, exon skipping, and alternative last exon were the most-commonly altered ones in response to RBM22 overexpression. On the other hand, Van Nostrand et al. showed that intron retention and exon skipping were the top-altered alternative splicing events in response to RBM22 knockdown in K-562 and HepG2 cells,
      • Van Nostrand EL
      • Freese P
      • Pratt GA
      • et al.
      A large-scale binding and functional map of human RNA-binding proteins.
      which might suggest that RBM22 modulation in terms of splicing landscape could be cell-type dependent, and might be one of the reasons underlying the distinct role of RBM22 in different cancer-types.
      Apart from its role as a regulator of the splicing process, it has been reported that RBM22 can also regulate gene transcription, presumably acting as a transcription factor through its Zinc-Finger like domain.
      • Xiao R
      • Chen JY
      • Liang Z
      • et al.
      Pervasive chromatin-RNA binding protein interactions enable RNA-based regulation of transcription.
      In line with this, we found 4245 genes differentially expressed in response to RBM22 overexpression in PCa cells, pointing out its role as transcription regulator. Then, to unveil the molecular mechanisms underlying the antitumor role of RBM22 in PCa cells, pathways enrichment analysis was done with the transcriptional landscape of RBM22-overexpressing PC-3 cells, using IPA software. Specifically, RBM22 overexpression was associated with the dysregulation of various critical pathways, including the inhibition of cell cycle progression. In fact, results from nCounter PanCancer Pathways Panel array performed in RBM22-overexpressing and mock-transfected xenograft tumors corroborated the relation of RBM22 with cell cycle progression. Therefore, our data suggest that high RBM22 levels lead to cell cycle arrest in PCa cells. Indeed, bioinformatics analyses using IPA software identified CDK1 as one of the master regulators driving the effects of RBM22 in PCa, which is widely known to be an essential player driving cell cycle.
      • Santamaria D
      • Barriere C
      • Cerqueira A
      • et al.
      Cdk1 is sufficient to drive the mammalian cell cycle.
      Specifically, CDK1 interacts with CCNB1, which we also found to be downregulated in response to RBM22 in vitro and in vivo, to form the maturation-promoting factor, a complex that is necessary for the proper control of the G2/M transition phase of the cell cycle.
      • Gavet O
      • Pines J
      Progressive activation of CyclinB1-Cdk1 coordinates entry to mitosis.
      In addition, EPAS1 (also known as HIF2a) was also found herein to be another master regulator of RBM22 effects in PCa. Remarkably, EPAS1 has been reported to be upregulated in PCa and positively correlated with androgen receptor expression.
      • Boddy JL
      • Fox SB
      • Han C
      • et al.
      The androgen receptor is significantly associated with vascular endothelial growth factor and hypoxia sensing via hypoxia-inducible factors HIF-1a, HIF-2a, and the prolyl hydroxylases in human prostate cancer.
      Importantly, EPAS1 transcriptionally upregulates tumorigenic hypoxia-responsive genes (ie VEGF), representing a key factor driving angiogenesis, but being also involved in cell proliferation.
      • Keith B
      • Johnson RS
      • Simon MC
      HIF1alpha and HIF2alpha: sibling rivalry in hypoxic tumour growth and progression.
      Finally, our data demonstrated that downregulation of ATR, an essential regulator of genome integrity,
      • Cimprich KA
      • Cortez D
      ATR: an essential regulator of genome integrity.
      is another key molecular event in response to RBM22 overexpression (see graphical abstract).
      Remarkably, these molecular changes driven by RBM22 dysregulation might be clinically exploitable. Specifically, it is widely known that CDK1 induces DNA homologous recombination repair (HR) by recruiting Rad51 to DNA damage sites.
      • Johnson N
      • Cai D
      • Kennedy RD
      • et al.
      Cdk1 participates in BRCA1-dependent S phase checkpoint control in response to DNA damage.
      Consistently, it has been reported that CDK1 inhibition reduces HR activity and causes synthetic lethal damage in BRCA proficient cells combined with PARP inhibitors.
      • Johnson N
      • Li YC
      • Walton ZE
      • et al.
      Compromised CDK1 activity sensitizes BRCA-proficient cancers to PARP inhibition.
      In addition, PARP and ATR dual inhibition has been shown to synergistically reduce tumor aggressiveness, and even to overcome PARP inhibition resistance, in vitro and in vivo in different tumor-types, including PCa.
      • Yazinski SA
      • Comaills V
      • Buisson R
      • et al.
      ATR inhibition disrupts rewired homologous recombination and fork protection pathways in PARP inhibitor-resistant BRCA-deficient cancer cells.
      • Kim H
      • Xu H
      • George E
      • et al.
      Combining PARP with ATR inhibition overcomes PARP inhibitor and platinum resistance in ovarian cancer models.
      • Neeb A
      • Herranz N
      • Arce-Gallego S
      • et al.
      Advanced prostate cancer with ATM loss: PARP and ATR inhibitors.
      Therefore, given that we demonstrated herein that stably overexpression of RBM22 led to a dramatic decrease in CDK1 and ATR expression levels in PCa cells, it is tempting to suggest that RBM22-overexpressing tumors might benefit from the therapeutic intervention with PARP inhibitors (eg, Olaparib, rucaparib). However, this hypothesis should be taken with caution since to date only HR gene mutations (especially in BRCA1/2) has been associated with response to PARPi in mCRPC.
      • Murai J
      • Huang SY
      • Das BB
      • et al.
      Trapping of PARP1 and PARP2 by clinical PARP inhibitors.
      ,
      • de Bono J
      • Mateo J
      • Fizazi K
      • et al.
      Olaparib for metastatic castration-resistant prostate cancer.
      Therefore, despite the promising preclinical data,
      • Johnson N
      • Li YC
      • Walton ZE
      • et al.
      Compromised CDK1 activity sensitizes BRCA-proficient cancers to PARP inhibition.
      • Yazinski SA
      • Comaills V
      • Buisson R
      • et al.
      ATR inhibition disrupts rewired homologous recombination and fork protection pathways in PARP inhibitor-resistant BRCA-deficient cancer cells.
      • Kim H
      • Xu H
      • George E
      • et al.
      Combining PARP with ATR inhibition overcomes PARP inhibitor and platinum resistance in ovarian cancer models.
      • Neeb A
      • Herranz N
      • Arce-Gallego S
      • et al.
      Advanced prostate cancer with ATM loss: PARP and ATR inhibitors.
      no clinical studies have proven the predictive value of any transcriptional HRR signature yet.
      Taken together, our data unveiled new conceptual and functional avenues in PCa, with potential clinical implications, by demonstrating that RBM22 has a critical functional role in the pathophysiology of PCa since its overexpression impaired key pathophysiological processes in PCa-biology (ie proliferation, migration, tumorspheres and colonies formation) likely by modulating critical oncogenic signaling pathways associated with PCa initiation/progression/aggressiveness. Moreover, our results pave the way to study negative regulators of RBM22 that could become potential therapeutic targets, as well as invites to suggest that RBM22 levels might represent a biomarker of response to PARP inhibitors, offering a clinically relevant therapeutic opportunity that should be further explored and ultimately tested in humans.

      Acknowledgments

      Conflicts of Interest: All authors have read the journal's policy on disclosure of potential conflicts of interest and have none to declare.
      This research was funded by the Spanish Ministry of Science, Innovation, and Universities (Research‐Grant: PID2019‐105564RB‐I00 ; Predoctoral contracts: FPU16/05059 , FPU17/00263 , FPU18/02485 ), Health Institute Carlos III ( DTS20-00050 ), Junta de Andalucia [ BIO-0139 ; Consejería de Transformación Económica, Industria, Conocimiento y Universidades ( P20_00442 ); Consejería de Salud y Familias, co-funded by European Union (Programa Operativo FEDER de Andalucía 2014-2020, “ Andalucía se mueve con Europa ”: PEER-0048-2020 )], and CIBERobn (CIBER is an initiative of Instituto de Salud Carlos III, Ministerio de Sanidad, Servicios Sociales e Igualdad, Spain). Funding for open access charge: University of Córdoba/CBUA. Some images were taken from Servier Medical Art (smart.servier.com) under the Creative Commons Attribution 3.0 Unported License. Special thanks to the staff of Biobank of the IMIBIC and of the experimental animal service (SAE) of the UCO/IMIBIC. We are gratefully indented to all the patients and their families for generously donating the samples and clinical data for research purposes. All authors have read the journal's authorship agreement.
      Author contributions are as follow: JMJV contributed to the conception of the work, design of the work, acquisition, analysis and interpretation of the data and drafted the work; AJMH contributed to the conception of the work, design of the work, acquisition, analysis and interpretation of the data and drafted the work; EGG contributed to the acquisition, analysis and interpretation of the data; PSM contributed to the acquisition and analysis of the data; ACFF contributed to the acquisition and analysis of the data; AC contributed to the analysis of the data; TGS contributed to the acquisition and analysis of the data; AML contributed to the acquisition and analysis of the data; RSS- contributed to the acquisition and analysis of the data; PPLC contributed to the acquisition of the data; ASC contributed to the acquisition of the data; DO contributed to the acquisition and interpretation of the data and substantively revised the work; EE contributed to the acquisition and interpretation of the data and substantively revised the work; JPC contributed to the acquisition and interpretation of the data and substantively revised the work; MDG contributed to the design of the work, acquisition and interpretation of the data and substantively revised the work; RMLH contributed to the conception of the work, design of the work, acquisition, analysis and interpretation of the data, drafted and substantively revised the work.
      Data statement: The datasets used and/or analyzed during the current study are available in this article and/or available from the corresponding author on reasonable request.

      Appendix. Supplementary materials

      • Supplemental Figure 1: RBM22 genomic alterations in PCa. A) Number of patients with different copy number alterations in RBM22 (ie deep deletion, shallow deletion, copy gain and amplification) using the TCGA and SU2C/PCF cohorts. B) Association between RBM22 mRNA levels (min to max boxplot, with median) and RBM22 copy number alterations. C and D) Association between RBM22 mRNA with molecular subtypes and common genetic alterations of PCa, in TCGA cohort (C) and SU2C/PCF cohort (D). Data were downloaded from cBioportal.

      • Supplemental Figure 2: Association between RBM22 mRNA levels with disease-free survival (DFS) or overall survival (OS) in PCa patients. Relationship between RBM22 mRNA expression levels and DFS in the TCGA cohort (A), and OS in the SU2C/PCF cohort (B). Statistical significance was calculated with Long-Rank test. Patients divided into low RBM22 levels (Q1, red line), medium RBM22 levels (Q2+Q3, green line) and high RBM22 levels (Q4, blue line). Molecular and clinical data was downloaded from cBioportal.

      • Supplemental Figure 3: Non-significant correlations between RBM22 expression levels parameters of PCa aggressiveness. (A) Correlations of RBM22 expression levels and Gleason Score (left panel), plasma PSA levels (central panel) and age (right panel) of PCa patients from cohort 2. (B) Association between RBM22 expression levels and the presence of metastasis at diagnosis of PCa patients from cohort 2. (C) Correlations of RBM22 expression levels and AR (left panel), AR-v7 (central panel) and MKI67 (right panel) expression levels of PCa patients from cohort 2. Correlations are represented by mean (connecting line) and error bands (pointed line) of expression levels. In the case of the association, data are represented as mean ± SEM.

      • Supplemental Figure 4: Relationship between RBM22 and AR pathway. A) Expression levels of RBM22 in response to DHT treatment in LNCaP cells. B) Expression levels of AR in response to RBM22 overexpression in LNCaP cells. C-D) Expression levels of AR and AR-V7 in response to RBM22 overexpression in 22Rv1 cells. Data (mean ± SEM) are represented as fold change (percentage) to vehicle-treated controls (A) or mock-overexpressing cells (B-D).

      • Supplemental Figure 5: Characterization of the alternative splicing events from RBM22-overexpressing vs mock PC-3 cells RNAseq. Total splicing events detected (black) and significantly different events (grey) are classified depending on their type, showing different frequencies (%) between both conditions. A3: Alternative 3′ Splice Site; A5: Alternative 5′ Splice Sites; AF: Alternative First Exon; AL: Alternative Last Exon; MX: Mutually Exclusive Exons; RI: Retained Intron; SE: Skipping Exon.

      • Supplemental Figure 6: Expression levels of master regulators in response to RBM22 overexpression. Expression levels of master regulators in PC-3 (A), 22Rv1 (B), and PC-3 xenografts preclinical model (C). Data (mean ± SEM) are represented as fold change to mock-overexpressing cells/tumors. Asterisks (** P < 0.01; *** P < 0.001) indicate statistically significant differences between groups.

      References

        • Bray F
        • Ferlay J
        • Soerjomataram I
        • Siegel RL
        • Torre LA
        • Jemal A.
        Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
        CA Cancer J Clin. 2018; 68: 394-424
        • Tolkach Y
        • Kristiansen G
        The heterogeneity of prostate cancer: a practical approach.
        Pathobiology. 2018; 85: 108-116
        • Sandhu S
        • Moore CM
        • Chiong E
        • Beltran H
        • Bristow RG
        • Williams SG
        Prostate cancer.
        Lancet. 2021; 398: 1075-1090
        • Sartor O
        • de Bono JS
        Metastatic prostate cancer.
        N Engl J Med. 2018; 378: 645-657
        • Westaby D
        • Fenor de La Maza MLD
        • Paschalis A
        • et al.
        A new old target: androgen receptor signaling and advanced prostate cancer.
        Annu Rev Pharmacol Toxicol. 2022; 62: 131-153
        • Mateo J
        • Seed G
        • Bertan C
        • et al.
        Genomics of lethal prostate cancer at diagnosis and castration resistance.
        J Clin Invest. 2020; 130: 1743-1751
        • Fraser M
        • Sabelnykova VY
        • Yamaguchi TN
        • et al.
        Genomic hallmarks of localized, non-indolent prostate cancer.
        Nature. 2017; 541: 359-364
        • Cooper CS
        • Eeles R
        • Wedge DC
        • et al.
        Analysis of the genetic phylogeny of multifocal prostate cancer identifies multiple independent clonal expansions in neoplastic and morphologically normal prostate tissue.
        Nat Genet. 2015; 47: 367-372
        • Boutros PC
        • Fraser M
        • Harding NJ
        • et al.
        Spatial genomic heterogeneity within localized, multifocal prostate cancer.
        Nat Genet. 2015; 47: 736-745
        • Ladomery M
        Aberrant alternative splicing is another hallmark of cancer.
        Int J Cell Biol. 2013; 2013463786
        • Oltean S
        • Bates DO
        Hallmarks of alternative splicing in cancer.
        Oncogene. 2014; 33: 5311-5318
        • Rahman MA
        • Krainer AR
        • Abdel-Wahab O
        SnapShot: splicing alterations in cancer.
        Cell. 2020; 180: 208-2e1
        • Jimenez-Vacas JM
        • Herrero-Aguayo V
        • Montero-Hidalgo AJ
        • et al.
        Dysregulation of the splicing machinery is directly associated to aggressiveness of prostate cancer.
        EBioMedicine. 2020; 102547
        • Jimenez-Vacas JM
        • Herrero-Aguayo V
        • Gomez-Gomez E
        • et al.
        Spliceosome component SF3B1 as novel prognostic biomarker and therapeutic target for prostate cancer.
        Transl Res. 2019; 212: 89-103
        • Hormaechea-Agulla D
        • Jimenez-Vacas JM
        • Gomez-Gomez E
        • et al.
        The oncogenic role of the spliced somatostatin receptor sst5TMD4 variant in prostate cancer.
        Faseb J. 2017; 31: 4682-4696
        • Hormaechea-Agulla D
        • Gahete MD
        • Jimenez-Vacas JM
        • et al.
        The oncogenic role of the In1-ghrelin splicing variant in prostate cancer aggressiveness.
        Mol Cancer. 2017; 16: 146
        • Fuentes-Fayos AC
        • Vazquez-Borrego MC
        • Jimenez-Vacas JM
        • et al.
        Splicing machinery dysregulation drives glioblastoma development/aggressiveness: oncogenic role of SRSF3.
        Brain. 2020; 143: 3273-3293
        • Fuentes-Fayos AC
        • Perez-Gomez JM
        • GG ME
        • et al.
        SF3B1 inhibition disrupts malignancy and prolongs survival in glioblastoma patients through BCL2L1 splicing and mTOR/ss-catenin pathways imbalances.
        J Exp Clin Cancer Res. 2022; 41: 39
        • Paschalis A
        • Sharp A
        • Welti JC
        • et al.
        Alternative splicing in prostate cancer.
        Nat Rev Clin Oncol. 2018; 15: 663-675
        • Takayama KI
        Splicing factors have an essential role in prostate cancer progression and androgen receptor signaling.
        Biomolecules. 2019; 9: 131
        • Zeng Y
        • Wodzenski D
        • Gao D
        • et al.
        Stress-response protein RBM3 attenuates the stem-like properties of prostate cancer cells by interfering with CD44 variant splicing.
        Cancer Res. 2013; 73: 4123-4133
        • Zhao L
        • Li R
        • Shao C
        • Li P
        • Liu J
        • Wang K
        3p21.3 tumor suppressor gene RBM5 inhibits growth of human prostate cancer PC-3 cells through apoptosis.
        World J Surg Oncol. 2012; 10: 247
        • Yamamoto R
        • Osawa T
        • Sasaki Y
        • et al.
        Overexpression of p54(nrb)/NONO induces differential EPHA6 splicing and contributes to castration-resistant prostate cancer growth.
        Oncotarget. 2018; 9: 10510-10524
        • Takayama KI
        • Suzuki T
        • Fujimura T
        • et al.
        Dysregulation of spliceosome gene expression in advanced prostate cancer by RNA-binding protein PSF.
        Proc Natl Acad Sci U S A. 2017; 114: 10461-10466
        • Rasche N
        • Dybkov O
        • Schmitzova J
        • Akyildiz B
        • Fabrizio P
        • Luhrmann R
        Cwc2 and its human homologue RBM22 promote an active conformation of the spliceosome catalytic centre.
        EMBO J. 2012; 31: 1591-1604
        • Rasche N
        • Dybkov O
        • Schmitzová J
        • Akyildiz B
        • Fabrizio P
        • Lührmann R
        Cwc2 and its human homologue RBM22 promote an active conformation of the spliceosome catalytic centre.
        EMBO J. 2012; 31: 1591-1604
        • Van Nostrand EL
        • Freese P
        • Pratt GA
        • et al.
        A large-scale binding and functional map of human RNA-binding proteins.
        Nature. 2020; 583: 711-719
        • Xiao R
        • Chen JY
        • Liang Z
        • et al.
        Pervasive chromatin-RNA binding protein interactions enable RNA-based regulation of transcription.
        Cell. 2019; 178 (107–21.e18)
        • Soubise B
        • Jiang Y
        • Douet-Guilbert N
        • Troadec MB
        RBM22, a key player of Pre-mRNA splicing and gene expression regulation, is altered in Cancer.
        Cancers (Basel). 2022; 14: 643
        • Gao J
        • Aksoy BA
        • Dogrusoz U
        • et al.
        Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal.
        Sci Signal. 2013; 6: l1
        • Cerami E
        • Gao J
        • Dogrusoz U
        • et al.
        The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data.
        Cancer Discov. 2012; 2: 401-404
        • Lapointe J
        • Li C
        • Higgins JP
        • et al.
        Gene expression profiling identifies clinically relevant subtypes of prostate cancer.
        Proc Natl Acad Sci U S A. 2004; 101: 811-816
        • Cancer Genome Atlas Research Network
        The molecular taxonomy of primary prostate Cancer.
        Cell. 2015; 163: 1011-1025
        • Abida W
        • Cyrta J
        • Heller G
        • Prandi D
        • Armenia J
        • Coleman I
        • et al.
        Genomic correlates of clinical outcome in advanced prostate cancer.
        Proc Natl Acad Sci U S A. 2019; 116: 11428-11436
        • Sáez-Martínez P
        • Jiménez-Vacas JM
        • León-González AJ
        • Herrero-Aguayo V
        • Montero Hidalgo AJ
        • Gómez-Gómez E
        • et al.
        Unleashing the diagnostic, prognostic and therapeutic potential of the neuronostatin/GPR107 system in prostate cancer.
        J Clin Med. 2020; 9: 1703
        • Jimenez-Vacas JM
        • Gomez-Gomez E
        • Montero-Hidalgo AJ
        • et al.
        Clinical utility of ghrelin-O-Acyltransferase (GOAT) enzyme as a diagnostic tool and potential therapeutic target in prostate cancer.
        J Clin Med. 2019; 8: 2056
        • Gómez-Gómez E
        • Jiménez-Vacas JM
        • Pedraza-Arévalo S
        • et al.
        Oncogenic role of secreted engrailed homeobox 2 (EN2) in prostate Cancer.
        J Clin Med. 2019; 8: 1400
        • Vandesompele J
        • De Preter K
        • Pattyn F
        • et al.
        Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes.
        Genome Biol. 2002; 3 (RESEARCH0034)
        • Su ZZ
        • Lin J
        • Shen R
        • Fisher PE
        • Goldstein NI
        • Fisher PB
        Surface-epitope masking and expression cloning identifies the human prostate carcinoma tumor antigen gene PCTA-1 a member of the galectin gene family.
        Proc Nat Acad Sci of the U S A. 1996; 93: 7252-7257
        • Del Rio-Moreno M
        • Alors-Perez E
        • Borges de Souza P
        • et al.
        Peptides derived from the extracellular domain of the somatostatin receptor splicing variant SST5TMD4 increase malignancy in multiple cancer cell types.
        Transl Res. 2019; 211: 147-160
        • Frankish A
        • Diekhans M
        • Ferreira AM
        • et al.
        GENCODE reference annotation for the human and mouse genomes.
        Nucleic Acids Res. 2019; 47: D766-DD73
        • Patro R
        • Duggal G
        • Love MI
        • Irizarry RA
        • Kingsford C
        Salmon provides fast and bias-aware quantification of transcript expression.
        Nature methods. 2017; 14: 417-419
        • Soneson C
        • Love MI
        • Robinson MD
        Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences.
        F1000Res. 2015; 4: 1521
        • Ritchie ME
        • Phipson B
        • Wu D
        • et al.
        limma powers differential expression analyses for RNA-sequencing and microarray studies.
        Nucleic Acids Res. 2015; 43: e47
        • Trincado JL
        • Entizne JC
        • Hysenaj G
        • et al.
        SUPPA2: fast, accurate, and uncertainty-aware differential splicing analysis across multiple conditions.
        Genome Biol. 2018; 19: 40
        • Alamancos GP
        • Pages A
        • Trincado JL
        • Bellora N
        • Eyras E
        Leveraging transcript quantification for fast computation of alternative splicing profiles.
        RNA. 2015; 21: 1521-1531
        • Gelman IH
        How the TRAMP model revolutionized the study of prostate Cancer progression.
        Cancer Res. 2016; 76: 6137-6139
        • Sarmento-Cabral A
        • F LL
        • Gahete MD
        • Castano JP
        • Luque RM
        Metformin reduces prostate Tumor growth, in a diet-dependent manner, by modulating multiple signaling pathways.
        Molecular cancer research: MCR. 2017; 15: 862-874
        • Munkley J
        • Li L
        • Krishnan SRG
        • et al.
        Androgen-regulated transcription of ESRP2 drives alternative splicing patterns in prostate cancer.
        Elife. 2019; 8 (e47678)
        • Paschalis A
        • Welti J
        • Neeb AJ
        • et al.
        JMJD6 Is a Druggable oxygenase that regulates AR-V7 expression in prostate Cancer.
        Cancer Res. 2021; 81: 1087-1100
        • Horoszewicz JS
        • Leong SS
        • Kawinski E
        • et al.
        LNCaP model of human prostatic carcinoma.
        Cancer Res. 1983; 43: 1809-1818
        • Sramkoski RM
        • Pretlow 2nd, TG
        • Giaconia JM
        • et al.
        A new human prostate carcinoma cell line, 22Rv1.
        In Vitro Cell Dev Biol Anim. 1999; 35: 403-409
        • Sowalsky AG
        • Figueiredo I
        • Lis RT
        • et al.
        Assessment of Androgen receptor splice variant-7 as a biomarker of clinical response in castration-sensitive prostate cancer.
        Clin Cancer Res. 2022; 28: 3509-3525
        • Li Z
        • Guo Q
        • Zhang J
        • et al.
        The RNA-binding motif protein family in cancer: friend or foe?.
        Front Oncol. 2021; 11757135
        • Dvinge H
        • Kim E
        • Abdel-Wahab O
        • Bradley RK
        RNA splicing factors as oncoproteins and tumour suppressors.
        Nat Rev Cancer. 2016; 16: 413-430
        • De Maio A
        • Yalamanchili HK
        • Adamski CJ
        • et al.
        RBM17 interacts with U2SURP and CHERP to regulate expression and splicing of RNA-processing proteins.
        Cell Rep. 2018; 25 (726–36 e7)
        • Boultwood J
        • Pellagatti A
        • Cattan H
        • et al.
        Gene expression profiling of CD34+ cells in patients with the 5q- syndrome.
        Br J Haematol. 2007; 139: 578-589
        • Jimenez-Vacas JM
        • Herrero-Aguayo V
        • Montero-Hidalgo AJ
        • et al.
        Dysregulation of the splicing machinery is directly associated to aggressiveness of prostate cancer.
        EBioMedicine. 2020; 51102547
        • Verma S
        • Shankar E
        • Kalayci FNC
        • et al.
        Androgen Deprivation induces transcriptional reprogramming in prostate Cancer cells to develop stem cell-like characteristics.
        Int J Mol Sci. 2020; 21: 9568
        • Zhang L
        • Jiao M
        • Li L
        • et al.
        Tumorspheres derived from prostate cancer cells possess chemoresistant and cancer stem cell properties.
        J Cancer Res Clin Oncol. 2012; 138: 675-686
        • Ge Y
        • Schuster MB
        • Pundhir S
        • et al.
        The splicing factor RBM25 controls MYC activity in acute myeloid leukemia.
        Nat Commun. 2019; 10: 172
        • Rose AE
        • Satagopan JM
        • Oddoux C
        • et al.
        Copy number and gene expression differences between African American and Caucasian American prostate cancer.
        J Transl Med. 2010; 8: 70
        • Powell IJ
        Epidemiology and pathophysiology of prostate cancer in African-American men.
        J Urol. 2007; 177: 444-449
        • Reddy S
        • Shapiro M
        • Morton Jr, R
        • Brawley OW
        Prostate cancer in black and white Americans.
        Cancer Metastasis Rev. 2003; 22: 83-86
        • Ka HI
        • Lee S
        • Han S
        • et al.
        Deubiquitinase USP47-stabilized splicing factor IK regulates the splicing of ATM pre-mRNA.
        Cell Death Discov. 2020; 6: 34
        • Das T
        • Park JK
        • Park J
        • et al.
        USP15 regulates dynamic protein-protein interactions of the spliceosome through deubiquitination of PRP31.
        Nucleic Acids Res. 2017; 45: 4866-4880
        • Song EJ
        • Werner SL
        • Neubauer J
        • et al.
        The Prp19 complex and the Usp4Sart3 deubiquitinating enzyme control reversible ubiquitination at the spliceosome.
        Genes Dev. 2010; 24: 1434-1447
        • Janowicz A
        • Michalak M
        • Krebs J
        Stress induced subcellular distribution of ALG-2, RBM22 and hSlu7.
        Biochimica et biophysica acta. 2011; 1813: 1045-1049
        • Montaville P
        • Dai Y
        • Cheung CY
        • et al.
        Nuclear translocation of the calcium-binding protein ALG-2 induced by the RNA-binding protein RBM22.
        Biochimica et biophysica acta. 2006; 1763: 1335-1343
        • Santamaria D
        • Barriere C
        • Cerqueira A
        • et al.
        Cdk1 is sufficient to drive the mammalian cell cycle.
        Nature. 2007; 448: 811-815
        • Gavet O
        • Pines J
        Progressive activation of CyclinB1-Cdk1 coordinates entry to mitosis.
        Dev Cell. 2010; 18: 533-543
        • Boddy JL
        • Fox SB
        • Han C
        • et al.
        The androgen receptor is significantly associated with vascular endothelial growth factor and hypoxia sensing via hypoxia-inducible factors HIF-1a, HIF-2a, and the prolyl hydroxylases in human prostate cancer.
        Clin Cancer Res. 2005; 11: 7658-7663
        • Keith B
        • Johnson RS
        • Simon MC
        HIF1alpha and HIF2alpha: sibling rivalry in hypoxic tumour growth and progression.
        Nat Rev Cancer. 2011; 12: 9-22
        • Cimprich KA
        • Cortez D
        ATR: an essential regulator of genome integrity.
        Nat Rev Mol Cell Biol. 2008; 9: 616-627
        • Johnson N
        • Cai D
        • Kennedy RD
        • et al.
        Cdk1 participates in BRCA1-dependent S phase checkpoint control in response to DNA damage.
        Mol Cell. 2009; 35: 327-339
        • Johnson N
        • Li YC
        • Walton ZE
        • et al.
        Compromised CDK1 activity sensitizes BRCA-proficient cancers to PARP inhibition.
        Nat Med. 2011; 17: 875-882
        • Yazinski SA
        • Comaills V
        • Buisson R
        • et al.
        ATR inhibition disrupts rewired homologous recombination and fork protection pathways in PARP inhibitor-resistant BRCA-deficient cancer cells.
        Genes Dev. 2017; 31: 318-332
        • Kim H
        • Xu H
        • George E
        • et al.
        Combining PARP with ATR inhibition overcomes PARP inhibitor and platinum resistance in ovarian cancer models.
        Nat Commun. 2020; 11: 3726
        • Neeb A
        • Herranz N
        • Arce-Gallego S
        • et al.
        Advanced prostate cancer with ATM loss: PARP and ATR inhibitors.
        Eur Urol. 2021; 79: 200-211
        • Murai J
        • Huang SY
        • Das BB
        • et al.
        Trapping of PARP1 and PARP2 by clinical PARP inhibitors.
        Cancer Res. 2012; 72: 5588-5599
        • de Bono J
        • Mateo J
        • Fizazi K
        • et al.
        Olaparib for metastatic castration-resistant prostate cancer.
        N Engl J Med. 2020; 382: 2091-2102