scieee Open visual document viewer

TDP2 suppresses genomic instability induced by androgens in the epithelial cells of prostate glands

Al Mahmud, Md Rasel; Ishii, Kenichiro; Bernal Lozano, Cristina; Delgado Sainz, Irene; Toi, Masakazu; Akamatsu, Shusuke; Fukumoto, Manabu; Watanabe, Masatoshi; Takeda, Shunichi; Cortés Ledesma, Felipe; Sasanuma, Hiroyuki

Abstract

Androgens stimulate the proliferation of epithelial cells in the prostate by activating topoisomerase 2 (TOP2) and regulating the transcription of target genes. TOP2 resolves the entanglement of genomic DNA by transiently generating double-strand breaks (DSBs), where TOP2 homodimers covalently bind to 5′ DSB ends, called TOP2-DNA cleavage complexes (TOP2ccs). When TOP2 fails to rejoin TOP2ccs generating stalled TOP2ccs, tyrosyl DNA phosphodiesterase-2 (TDP2) removes 5′ TOP2 adducts from stalled TOP2ccs prior to the ligation of the DSBs by nonhomologous end joining (NHEJ), the dominant DSB repair pathway in G0/G1 phases. We previously showed that estrogens frequently generate stalled TOP2ccs in G0/G1 phases. Here, we show that physiological concentrations of androgens induce several DSBs in individual human prostate cancer cells during G1 phase, and loss of TDP2 causes a five times higher number of androgen-induced chromosome breaks in mitotic chromosome spreads. Intraperitoneally injected androgens induce several DSBs in individual epithelial cells of the prostate in TDP2-deficient mice, even at 20 hr postinjection. In conclusion, physiological concentrations of androgens have very strong genotoxicity, most likely by generating stalled TOP2ccs.

Full text

450 | Genes Cells. 2020;25:450–465. wileyonlinelib a y.com/jou nal/g c Genes o Cells Recei ed: 28 Feb ua y 2020 | Re ised: 30 Ma ch 2020 | Accep ed: 30 Ma ch 2020 DOI: 10.1111/g c.12770 ORIGINAL ARTICLE TDP2 supp esses genomic ins abili y induced by and ogens in he epi helial cells o p os a e glands Md. RaselAl Mahmud1 | Kenichi oIshii2 | C is inaBe nal-Lozano3 | I eneDelgado-Sainz3 | MasakazuToi4 | ShusukeAkama su5 | ManabuFukumo o6 | Masa oshiWa anabe2 | ShunichiTakeda1 | FelipeCo és-Ledesma3,7 | Hi oyukiSasanuma1 1Depa men o Radia ion Gene ics, G adua e School o Medicine, Kyo o Uni e si y, Kyo o, Japan 2Depa men o Oncologic Pa hology, Mie Uni e si y G adua e School o Medicine, Tsu, Japan 3Cen o Andaluz de Biología Molecula y Medicina Regene a i a (CABIMER), CSIC–Uni e sidad de Se illa Uni e sidad Pablo de Ola ide, Se illa, Spain 4Depa men o B eas Su ge y, G adua e School o Medicine, Kyo o Uni e si y, Kyo o, Japan 5Depa men o U ology, G adua e School o Medicine, Kyo o Uni e si y, Kyo o, Japan 6RIKEN Cen e o Ad anced In elligence P ojec , Tokyo, Japan 7Topology and DNA B eaks G oup, Spanish Na ional Cance Resea ch Cen e (CNIO), Mad id, Spain This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. © 2020 The Au ho s. Genes o Cells published by Molecula Biology Socie y o Japan and John Wiley & Sons Aus alia, L d. Co espondence Felipe Co és-Ledesma, Cen o Andaluz de Biología Molecula y Medicina Regene a i a (CABIMER), CSIC– Uni e sidad de Se illa Uni e sidad Pablo de Ola ide, Se illa 41092, Spain. Email: elipe.co es@cabime .es Hi oyuki Sasanuma, Depa men o Radia ion Gene ics, G adua e School o Medicine, Kyo o Uni e si y, Yoshida Konoe, Sakyo-ku, Kyo o 606-8501, Japan. Email: hi oysasa@ g.med.kyo o-u.ac.jp Funding in o ma ion Takeda Medical Resea ch Founda ion; Mi subishi Founda ion, G an /Awa d Numbe : ID30123 Communica ed by: Aki a Shinoha a Abs ac And ogens s imula e he p oli e a ion o epi helial cells in he p os a e by ac i a - ing opoisome ase 2 (TOP2) and egula ing he ansc ip ion o a ge genes. TOP2 esol es he en anglemen o genomic DNA by ansien ly gene a ing double-s and b eaks (DSBs), whe e TOP2 homodime s co alen ly bind o 5′ DSB ends, called TOP2-DNA clea age complexes (TOP2ccs). When TOP2 ails o ejoin TOP2ccs gene a ing s alled TOP2ccs, y osyl DNA phosphodies e ase-2 (TDP2) emo es 5′ TOP2 adduc s om s alled TOP2ccs p io o he liga ion o he DSBs by nonho- mologous end joining (NHEJ), he dominan DSB epai pa hway in G0/G1 phases. We p e iously showed ha es ogens equen ly gene a e s alled TOP2ccs in G0/ G1 phases. He e, we show ha physiological concen a ions o and ogens induce se e al DSBs in indi idual human p os a e cance cells du ing G1 phase, and loss o TDP2 causes a i e imes highe numbe o and ogen-induced ch omosome b eaks in mi o ic ch omosome sp eads. In ape i oneally injec ed and ogens induce se e al DSBs in indi idual epi helial cells o he p os a e in TDP2-de icien mice, e en a 20h pos injec ion. In conclusion, physiological concen a ions o and ogens ha e e y s ong geno oxici y, mos likely by gene a ing s alled TOP2ccs. KEYWORDS and ogen, a ypical epi helial hype plasia, DNA double-s and b eak, p os a ic in aepi helial neoplasia, TDP2, opoisome ase 2 | 451 Genes o Cells AL MAHMUD e AL. 1 | INTRODUCTION Sex ho mones, es ogens and and ogens, s ongly s imula e he p oli e a ion o epi helial cells in he mamma y glands and p os a e, espec i ely (La Vigne a, Condo elli, Russo, Mo gia, & Caloge o,2016; Liang & Shang,2013). Ac i a ed es ogen ecep o s α/β (ERs) and and ogen ecep o (AR) quickly induce ansc ip ion o ER- and AR- esponsi e genes as ansc ip ion ac o s (Kokon is, Takaku a, Hay, & Liao, 1994; Shang, Hu, DiRenzo, Laza , & B own, 2000; Wang e al.,2011; Yang e al.,2016). The an agonis s agains hese ecep o s a e widely used as i s -line he apies o b eas and p os a e cance pa ien s (Ho wich e  al., 2013; Musg o e & Su he land,2009). The e a e wo majo mech- anisms o chemical cance ogenesis, he s imula ion o cel- lula p oli e a ion and he induc ion o mu agenesis (Loeb & Ha is,2008). I has been widely belie ed ha sex ho mones enhance oncogenesis h ough he o me mechanism bu no he la e (Hende son & Feigelson,2000). Howe e , ecen s udies ha e sugges ed ha and ogens can d i e oncogene- sis by ac i a ing opoisome ase II (TOP2), which gene a es DSBs du ing i s physiological ca alysis and can gene a e ch o- mosome ansloca ion (Gómez-He e os e al.,2017; Ha ne e al.,2010; Nelson, Ha ne , & Yegnasub amanian,2018). I emains unclea how many b eaks a e ac ually gene a ed by he physiological concen a ion o and ogens. Ac i a ed ERs and AR igge signal-dependen ea ly ansc ip ional esponses by ec ui ing TOP2 o hei p o- mo e and enhance segmen s (Man ille e  al., 2015; Pommie , Sun, Huang, & Ni iss,2016). TOP2 has been e- po ed o be in ol ed in RNA polyme ase II p omo e -pause elease upon physiological signals by and ogens, insulin, glu- coco icoids, N-me hyl-d-aspa a e (NMDA), e inoic acid, hea shock and se um ( e iewed in e s. Aus in e al.,2018; Madabhushi,2018). TOP2 o ms a homodime and esol es DNA ca enanes by ca alyzing he ansien o ma ion o ga ed DSBs, which is ollowed by he enzyma ic ejoining o he b oken s ands h ough in insic in amolecula liga- ion ac i i y (Gale & Oshe o ,1992; Ni iss,2009). TOP2 becomes co alen ly bound o he 5′ DNA end o he an- sien ly o med ga ed DSBs, gene a ing TOP2-DNA clea age complexes (TOP2ccs). The ca alysis by TOP2 occasionally becomes “abo i e” and emains unsealed, leading o he gene a ion o s alled TOP2ccs (Gómez-He e os e al.,2014; Hoa e al.,2016). Thus, he sex ho mones enhance ca cino- genesis, possibly h ough an inc ease in he numbe o s alled TOP2ccs and gene a ion o mu a ions. Indeed, we p e iously showed ha physiological concen a ions o es ogens e- quen ly gene a e s alled TOP2ccs in G0/G1 phases (Sasanuma e al.,2018). S alled TOP2ccs a e epai ed by he wo-s ep p ocess, he emo al o 5′ TOP2 adduc s ollowed by he di ec liga ion o DSBs by nonhomologous end joining (NHEJ). Ty osyl DNA phosphodies e ase-2 (TDP2) is he only known en- zyme ha is capable o accu a ely emo ing 5′ TOP2 adduc s om s alled TOP2ccs (Ledesma, Khamisy, Zuma, Osbo n, & Caldeco ,2009; Schellenbe g e al., 2017). NHEJ equi es p eceding emo al o 5′ TOP2 adduc s o i s di ec liga ion o he DSBs in G0/G1 phase (Gómez-He e os e al.,2013). The loss o TDP2 causes a enua ed ansc ip ional esponses o and ogens in p os a e cells, sugges ing ha exposu e o an- d ogens may equen ly cause he abo i e ca alysis o TOP2. Howe e , i emains elusi e how many s alled TOP2ccs a e ac ually gene a ed by physiological concen a ions o and o- gens. Ano he un esol ed ques ion is he ole played by TDP2 in he p e en ion o p os a ic hype plasia and oncogenesis. The po en ial causal ela ionship be ween he loss o TDP2 and he oncogenesis is sugges ed by TCGA da abase, which shows ha homozygous deep dele ion o he TDP2 gene is seen in 0.4% and 0.8% o he cance s a ising in he b eas and p os a e issues, espec i ely, bu no in o he cance ypes (Sasanuma e  al., 2018). Conside ing he ac ha TDP2 is ubiqui ously exp essed in he human issues (Fage be g e al.,2014; Yue e al.,2014), an un esol ed ques ion is why a e de ec s in TDP2 seen speci ically in he issues whe e cel- lula p oli e a ion is s imula ed by he sex ho mones, and o- gens as well as es ogens. In his s udy, we examined he geno oxic e ec o and o- gens on TDP2-de icien human p os a e cance LNCaP cell line and mouse p os a e. Physiological concen a ions o an- d ogens induced 4.7 and 16 DSBs in indi idual wild- ype and TDP2−/− cells a G0/G1 phase, espec i ely. This geno oxici y depends on bo h ac i a ed AR and TOP2. These da a indica e ha and ogens ha e s ong geno oxici y, he e icien induc- ion o s alled TOP2ccs in G1 phase. We also demons a e he s ong geno oxici y o and ogen in he p os a e epi he- lial cells o TDP2-de icien mice. The loss o TDP2 caused he abno mal p oli e a ion o epi helial cells ollowing h ee imes daily injec ion o and ogens in o 2-mon h-old mice and also esul ed in p og essi e p os a e hype plasia in 2- and 6-mon h-old mice. We p opose ha TDP2 supp esses abno - mal p oli e a ion o he epi helial cells in he p os a e gland by p omo ing he epai o and ogen-induced DSBs and en- su ing p ope ansc ip ional esponses o and ogens. 2 | RESULTS 2.1 | And ogens induce DSBs in se um- s a ed LNCaP cells To in es iga e he geno oxic e ec o and ogens on LNCaP human p os a e cance cells, which exp ess unc ional AR (Ho oszewicz e  al., 1983), we en iched G1-phase cells mo e han 90% by se um s a a ion o 48h (Figu e S1a) and analyzed only cyclin-A-nega i e, G1-phase cells. We 452 | Genes o Cells AL MAHMUD e Al. pulse-exposed he cells o 1nM o R1881, a syn he ic and o ally ac i e and ogenic s e oid (T an e al., 2009), o wo hou s and subsequen ly emo ed R1881 om he medium (Figu e1a). The po ency o R1881 a 1nM is compa able o ha o 10nM and ogens, a physiological se um con- cen a ion a e pube y (Co sini e al.,2016; Kemppainen e al.,1999). We hen analyzed γH2AX oci 2h a e he emo al o R1881. Rema kably, R1881- ea ed LNCaP cells displayed 5.2±0.6 (SD; s anda d de ia ion) γH2AX oci pe cell and 17- old induc ion o γH2AX oci o e he baseline (0.3±0.06 (SD)) in G0/G1-phase LNCaP cells (Figu e1b,c, Figu e S1b), indica ing he s ong geno oxic- i y o and ogens. 2.2 | And ogen-induced DNA damage is dependen on a unc ional and ogen ecep o We nex in es iga ed whe he unc ional AR is equi ed o and ogen-induced γH2AX- ocus o ma ion in se um- s a ed LNCaP cells. We exposed he LNCaP cells o R1881 oge he wi h a clinically ele an concen a ion o an inhibi o agains he and ogen ecep o , lu amide (Schellhamme e al.,1997). The exposu e o he an ago- nis comple ely ep essed R1881-induced γH2AX- ocus o ma ion in se um-s a ed LNCaP cells (Figu e 1d,e, Figu e S1c). This obse a ion indica es ha he ac i a ion o he and ogen ecep o is equi ed o and ogen-depend- en DSB o ma ion. 2.3 | And ogen-induced DNA damage depends on bo h TOP2α and TOP2β We hypo hesized ha and ogen-induced γH2AX oci we e caused by TOP2. The e a e wo iso o ms o TOP2, TOP2α and TOP2β (Aus in e  al., 2018; Madabhushi, 2018), which ha e an o e lapping ole in ansc ip ion (Sasanuma e  al., 2018). To explo e his hypo hesis, we gene a ed TOP2β−/− LNCaP cells (Figu e S1d) and deple ed TOP2α mo e han 20- old using shRNA compa ed wi h shCon ol- ea ed cells (Figu e S1e, ). The loss o TOP2β educed he numbe o γH2AX oci a 2h om 5.2±0.6 (SD) o 2.7±0.6 (Figu e1g, Figu e S1g), sugges ing ha a leas 50% o he and ogen-induced γH2AX oci ep esen s alled TOP2βccs. We p e iously showed ha TOP2α compen- sa es o he lack o TOP2β in he induc ion o DSBs by es ogens (Sasanuma e al.,2018). We deple ed TOP2α by using shRNA and ound he esul ing TOP2β−/−/shTOP2α cells comple ely supp essed he induc ion o γH2AX oci by R1881 du ing G1 phase (Figu e1 ,g, Figu e S1g). These esul s indica e ha and ogen-induced DSBs depend on TOP2α and TOP2β. 2.4 | Loss o TDP2 causes p olonged γH2AX- ocus accumula ion a e a pulse exposu e o p os a e cance cells o and ogens To in es iga e he ole o TDP2 in he epai o and ogen- induced DSBs, we dis up ed he TDP2 gene in LNCaP cells (Figu e S2a). We pulse-exposed se um-s a ed cells o 1nM o R1881 o 2h and subsequen ly moni o ed he esolu ion kine ics o γH2AX oci a 4h , 12h and 24h a e he ad- di ion o R1881 (Figu e2a). The pulse exposu e o R1881 induced 5.2±0.6 (SD) γH2AX oci in wild- ype cells and 18±2 (SD) γH2AX oci in TDP2−/− cells a 4h (Figu e2b,c, Figu e S2b). The numbe o γH2AX oci educed o a back- g ound le el in wild- ype cells by 24h (Figu e2b,c, Figu e S2b). In ma ked con as , TDP2−/− cells showed only a ~20% dec ease in he numbe o γH2AX oci om 4 o 24h ( om 18±2 o 14±2 oci). This esul clea ly indica es a signi i- can con ibu ion o TDP2 o he epai o and ogen-induced DSBs. To assess a unc ional in e ac ion be ween TDP2 and NHEJ, we inac i a ed NHEJ by ea ing se um-s a ed TDP2−/− cells wi h an inhibi o , NU7441 agains DNA-PK- dependen p o ein kinase ca aly ic subuni (DNA-PKcsi), which enzyme is a cen al componen o NHEJ (Black o d & Jackson,2017). The inhibi ion o NHEJ in wild- ype cells inc eased he numbe o and ogen-induced γH2AX oci om 5.2±0.6 o 21±2 pe cell (Figu e2b,c, Figu e S2b). Wild- ype cells ea ed wi h DNA-PKcsi displayed i ually no de- c ease in he numbe o and ogen-induced γH2AX oci om 4 o 24h a e he addi ion o and ogens (Figu e2b,c, Figu e S2b). The da a indica e ha NHEJ plays an essen ial ole in he epai o and ogen-induced DSBs du ing he G1 phase. S ikingly, he numbe o γH2AX oci a 24 h was e y simila be ween TDP2−/− cells and he DNA-PKcsi- ea ed cells (14±2 oci in TDP2−/− e sus 14±1 oci in DNA- PKcsi- ea ed wild- ype in Figu e2c). Mo eo e , he loss o TDP2 did no u he inc ease he numbe o γH2AX oci in DNA-PKcsi- ea ed cells (14±2 oci in TDP2−/− e sus 16±6 oci in DNA-PKcsi- ea ed TDP2−/− in Figu e2c). Conside ing he ole o TDP2 in he emo al o 5′ TOP2 adduc s om s alled TOP2ccs, his epis a ic ela ionship be- ween TDP2 and NHEJ indica es ha TDP2 con ibu es o NHEJ-media ed epai o and ogen-induced DSBs. In sum- ma y, and ogens ha e a e y s ong geno oxic po en ial, and collabo a ion be ween TDP2 and NHEJ plays a key ole in p e en ing genome ins abili y caused by and ogens. To e i y ha and ogen-induced DSBs seen in TDP2−/− cells ep esen s alled TOP2ccs, we c ea ed TDP2−/−/ TOP2β−/−/shTOP2α cells (Figu e S2c). The inac i a ion o bo h TOP2α and TOP2β comple ely supp essed and o- gen-induced DSBs a 4h in TDP2−/− cells (Figu e2d,e). Collec i ely, TDP2 is equi ed o he epai o TOP2- dependen DSBs induced by and ogens ia NHEJ-media ed | 453 Genes o Cells AL MAHMUD e AL. (a) (b) (d) ( ) (g) (c) (e) h h h h h h 454 | Genes o Cells AL MAHMUD e Al. epai in G1 phase. A physiological concen a ion o and o- gens gene a es ~20 DSBs in indi idual cells a G1 phase, and hese DSBs a e mos likely e lec ed s alled TOP2ccs. 2.5 | And ogens induce ch omosome b eaks in mi o ic ch omosome sp eads To con i m he geno oxici y o and ogens in cycling cells, we quan i ied he numbe s o abe an ch omosomes in mi- o ic sp eads ollowing 69-h con inuous exposu e o cycling cells o 1nM R1881. R1881 ea men inc eased ch omo- some abe a ions by 3.0- old ( om 2.0±0.4 o 6.0±0.8 oci) in wild- ype and 11- old ( om 2.0±1 o 22±2 oci) in TDP2−/− cells, espec i ely (Figu e2 ). In conclusion, a physiological concen a ion o and ogens p oduces a ew ab- e an ch omosomes e en in wild- ype cells, and TDP2 plays an impo an ole in he ai h ul epai o and ogen-induced ch omosome abe a ions. 2.6 | And ogens induce p ominen γH2AX oci in p os a e epi helial cells o TDP2−/− mice To e i y he geno oxici y o and ogens in i o, we adminis- a ed R1881 by in ape i oneal injec ion (ip) in o wild- ype and TDP2−/− mice a 2mon hs o age and moni o ed he num- be o γH2AX oci in he p os a e, whe e he exp ession le el o AR is ela i ely highe han o he issues in bo h humans and mouse (Figu e S3a,b) (El-Al y e al.,1999; Mi ose ich e al.,1999). The selec i e exp ession o AR is de ec ed in he p os a e epi helial cells, which highly exp ess cy oke a in 8/18 (Ishii, Imanaka-Yoshida, Yoshida, & Sugimu a,2008; Wang, Haywa d, Cao, Thaye , & Cunha,2001). R1881 injec- ion caused 1.3±1 and 6.3±3 γH2AX oci pe cell a 6h a e ip in he p os a e epi helial cells (cy oke a in 8/18-posi i e) o wild- ype and TDP2−/− mice, espec i ely (Figu e3a,b, Figu e S4a) (Toi anen & Shen,2017). Rema kably, TDP2−/− mice s ill displayed 7.0±2 γH2AX oci a 12h a e ip, while wild- ype mice displayed a backg ound le el o oci a 12h (Figu e3a,b, Figu e S4a). We also analyzed he spleen, which poo ly exp esses AR in bo h humans and mice (Figu e S3a,b) (Fage be g e al.,2014; Yue e al.,2014). As expec ed, essen ially no γH2AX oci we e induced by and ogens in he spleen o wild- ype o TDP2−/− mice (Figu e S4b,c). Taken oge he , and ogens a e highly geno oxic in bo h human p os- a e cance cells and he epi helial cells o he mouse p os a e pa icula ly in he absence o TDP2. We p e iously epo ed ha in ape i oneal injec ion o es ogens (17β-es adiol) induces a se e al imes highe num- be s o γH2AX oci a 6h a e ip in he epi helial cells o he mamma y glands in BRCA1-de icien mice in compa ison wi h wild- ype con ols (Sasanuma e al.,2018). In his s udy, we injec ed es ogens in o wild- ype and TDP2−/− mice. The numbe o γH2AX oci in mamma y epi helial cells was only sligh ly highe in TDP2−/− mice han wild- ype a 6h a e ip (Figu e3c,d). This esul indica es ha TDP2 does no play a e y impo an ole in he epai o es ogen-induced DSBs in he mouse mamma y gland. 2.7 | And ogens s imula e he p oli e a ion o p os a e epi helial cells in TDP2−/− mice o a g ea e ex en han in wild- ype mice Ac i a ed AR s imula es cellula p oli e a ion by con ol- ling ansc ip ion o AR- a ge genes (Wang e al.,2009). A ecen s udy indica es he ole o TDP2 in ensu ing p ope ansc ip ional esponse o and ogens. We he e examined he e ec o and ogens on he p oli e a ion o he epi helial cells in he mouse en al p os a es, which a e e y sensi- i e o and ogens (Ke & Sea le,1973; Sand o d, Sea le, & Ke , 1984; Sugimu a, Cunha, & Donjacou , 1986), o FIGURE 1 A syn he ic and ogenic s e oid, R1881-induced DSBs in wild- ype LNCaP cells is dependen on bo h unc ional AR and TOP2. (a) Schema ic diag am o expe imen al design o examine R1881-induced γH2AX- ocus o ma ion in human p os a e cance LNCaP cells. A e 48-h se um s a a ion, we incuba ed he wild- ype LNCaP cells wi h 1nM R1881 o 2h , washed-ou R1881, and u he incuba ed he cells wi hou R1881 o 2h . γH2AX oci we e analyzed a 4h a e he addi ion o R1881 ( ime 52h ). (b, c) Rep esen a i e images (b) and median numbe s o R1881-induced γH2AX oci (c). The nuclei a e ou lined. E o ba s show s anda d de ia ion (SD) o h ee independen expe imen s. The numbe o he coun ed cells is desc ibed in Table S4. A Single as e isk indica es p=2.0×10–4, which was calcula ed by an unpai ed wo- ailed es . The box plo s o γH2AX oci in “c” a e shown in Figu e S1b. Scale ba ep esen s 25μm. (d, e) Rep esen a i e image (d) and median numbe s o R1881-induced γH2AX oci (e). The wild- ype LNCaP cells we e exposed o R1881 oge he wi h lu amide o 2h ( ime 48–50h in “a”), ollowed by incuba ion in d ug- ee media o an addi ional 2h . The nuclei a e ou lined. E o ba s show s anda d de ia ion (SD) o h ee independen expe imen s. The numbe o he coun ed cells is desc ibed in Table S4. A single as e isk indica es p=6.5×10–6, which was calcula ed by an unpai ed wo- ailed es . The box plo s o γH2AX oci in “e” a e shown in Figu e S1c. Scale ba ep esen s 25μm. ( , g) Rep esen a i e images ( ) and median numbe s o R1881-induced γH2AX oci (g) in he indica ed geno ypes. We analyzed γH2AX oci wi h he same expe imen al p ocedu e in “a”. The nuclei a e ou lined. The gene a ion o he TOP2β−/− mu an cells is depic ed in Figu e S1d. E o ba s show s anda d de ia ion (SD) o h ee independen expe imen s. The numbe o he coun ed cells is desc ibed in Table S4. Single, double and iple as e isks indica e p=2.6×10–3, p=5.5×10–2 (no signi ican di e ence) and p=5.2×10–1 (no signi ican di e ence), espec i ely, which we e calcula ed by an unpai ed wo- ailed es . The box plo s o γH2AX oci in “g” a e shown in Figu e S1g. Scale ba ep esen s 25μm | 455 Genes o Cells AL MAHMUD e AL. (a) (b) (c) (d) ( ) (e) 456 | Genes o Cells AL MAHMUD e Al. 2-mon h-old mice. To his end, we daily injec ed R1881 o h ee days and examined he p os a e a day 4 (Figu e4a). To isualize cycling cells, we injec ed R1881 oge he wi h 5-e hynyl-2'-deoxyu idine (EdU), which is inco po a ed in o newly syn hesized DNA du ing S phase. As expec ed, subs an ial numbe s o EdU-posi i e (EdU+) cells we e de- ec able in he epi helial cells o he in es ine and spleno- cy es (Figu e S5a) due o he apid u no e o hese cells (CREAMER, Sho e , & Bam o h, 1961; Gelbe g, 2007; Kama h e al.,2000). The 3days’ injec ion o R1881 showed only a sub le bu signi ican inc ease in he p oli e a ion o he epi helial cells in wild- ype mice ( om 1.4±0.1 a day 0 o 1.8±0.1 a day 3) (Figu e4b,c). In con as , his injec ion in- c eased he pe cen age o EdU+ epi helial cells by h ee imes in TDP2−/− mice ( om 2.1±0.3 a day 0 o 5.2±0.2 on day 3). Likewise, he 3days’ injec ion o R1881 d ama ically inc eased he pe cen age o he cy oke a in 8/18-posi i e epi- helial cells (Figu e S5b) ha exp essed he p oli e a ing cell nuclea an igen (PCNA), a con en ional bioma ke o cy- cling cells, om 3.0±0.3% o 8.8±0.3% in TDP2−/− mice, compa ed wi h hose o wild- ype mice (Figu e4d). These esul s consis en ly indica e ha TDP2 p e en s he abno - mal p oli e a ion o p os a e epi helial cells in esponse o injec ed R1881. Conside ing he ole o TDP2 in he epai o s alled TOP2ccs, a de ec in hei ejoining migh change ansc ip ional esponses o and ogens and abno mally s im- ula e he p oli e a ion o he epi helial cells. 2.8 | De elopmen o hype plasia o p os a e epi helial cells in TDP2−/− mice The abo e da a indica ed ha and ogens induced se e al DSBs in indi idual p os a e epi helial cells and s ongly s im- ula ed hei p oli e a ion in he absence o TDP2. To e alua e he p oli e a ion a e o he epi helial cells, we s ained PCNA in he en al p os a e, compa ing be ween 2-mon h and 6-mon h-old mice (Figu e 5a). The pe cen age o PCNA+ epi helial cells was 2.5±0.1% and 3.0±0.3% in wild- ype and TDP2−/− mice a he age o 2mon hs, espec i ely. The pe cen age did no inc ease in wild- ype mice (2.8±0.3%), while he pe cen age inc eased o 11±5% in TDP2−/− mice a he age o 6mon hs. One possible scena io is ha he ac- cumula ion o mu a ions wi h aging in he absence o TDP2 migh inc ease he sensi i i y o he p os a e epi helial cells o and ogens leading o hei enhanced p oli e a ion (Figu e5b). We nex in es iga ed he consequence o he enhanced p oli e a ion, he hype plasia o epi helial cells in he p os- a e. We analyzed he en al p os a e o wild- ype and TDP2−/− mice a 2mon hs and 6mon hs o age. We coun ed he numbe o epi helial cells along he ixed dis ance o he lumen su ace in 10 duc s o each mouse. Wild- ype mice did no exhibi mo phological abno mali y o en al p os a e (Figu e5c). In con as , TDP2−/− mice a bo h 2 and 6mon hs o age exhibi ed hype plasia in epi helial cells o en al p os a e. The numbe o epi helial cells was wo old highe in 2-mon h-old TDP2−/− mice compa ed wi h 2-mon h-old wild- ype mice (Figu e5c,d). The hype plasia o he en al p os a e was wo old mo e equen ly obse ed in 6-mon h- old TDP2−/− mice han 2-mon h-old ones. Epi helial laye s we e olded as a consequence o he inc ease in he numbe o epi helial cells. In e es ingly, 6-mon h-old TDP2−/− mice exhibi ed he accumula ion o epi helial cells ha ing a dis- o de ed, mul ilaye ed o ganiza ion in mo e han 75% o he examined duc s. None heless, we did no de ec malignan u- mo s. Collec i ely, TDP2 p e en s he abno mal p oli e a ion o epi helial cells in he en al p os a e. 3 | DISCUSSION We he e e eal p e iously unapp ecia ed e y s ong geno- oxici y o a physiological concen a ion o and ogens equi alen o a se um concen a ion in males a e pube y. A 2-h pulse exposu e o and ogens induced app oxima ely i e TOP2-dependen DSBs pe cell in he G1 phase in an FIGURE 2 TDP2 unc ions in he epai o R1881-induced DSBs in LNCaP cells. (a) Schema ic diag am o expe imen al design o examine he epai kine ics o R1881-induced γH2AX oci in LNCaP cells. R1881 and DNA-PKcsi (NU7441) we e simul aneously added in o he medium a ime 48h . R1881 was only emo ed om he medium a ime 50h , bu DNA-PKcsi exposu e was con inued un il cells we e ha es ed o inac i a e NHEJ. γH2AX oci we e analyzed a he indica ed ime (0, +4, +12 and +24h a e addi ion o R1881). (b, c) Rep esen a i e images (b) and median numbe s o R1881-induced γH2AX oci (c) in he indica ed geno ypes. The gene a ion o TDP2−/− mu an cells is depic ed in Figu e S2a. The nuclei a e ou lined. E o ba s show s anda d de ia ion (SD) o h ee independen expe imen s. The numbe o he coun ed cells is desc ibed in Table S4. Single, double, iple and quad uple as e isks indica e p=1.2×10–4, p=3.3×10–5, p=5.7×10–1 (no signi ican di e ence) and p=6.4×10–1 (no signi ican di e ence), espec i ely, which we e calcula ed by an unpai ed wo- ailed es . The box plo s o γH2AX oci a e shown in Figu e S2b. Scale ba ep esen s 25μm. (d, e) Rep esen a i e images (d) and median numbe s o R1881-induced γH2AX oci (e) in he indica ed geno ypes. E o ba s show s anda d de ia ion (SD) o h ee independen expe imen s. The nuclei a e ou lined. The numbe o he coun ed cells is desc ibed in Table S4. Single and double as e isks indica e p=3.4×10–4 and p=7.0×10–3, espec i ely, which we e calcula ed by an unpai ed wo- ailed es . Scale ba ep esen s 25μm. ( ) R1881-induced ch omosome abe a ions in mi o ic ch omosome sp ead. Following 48-h incuba ion wi h media con aining cha coal- il a ed se um, cells we e u he incuba ed in media con aining cha coal- il a ed se um in he absence (“–”) o p esence (“+”) o R1881 (1nM) o 72h . E o ba s a e s anda d de ia ion (SD) o h ee independen analyses. The numbe o he coun ed cells is desc ibed in Table S4. A Single as e isk indica es p=3.1×10–3, which we e calcula ed by an unpai ed wo- ailed es | 457 Genes o Cells AL MAHMUD e AL. FIGURE 3 Geno oxici y o R1881 and 17β-es adiol in he p os a e and mamma y issues, espec i ely, o TDP2−/− mice. (a, b) Rep esen a i e images (a) and median numbe s o γH2AX- ocus-posi i e (b) cells in luminal epi helial cells o he p os a e a he indica ed ime a e ip wi h R1881. We killed h ee 2-mon h-old mice o each geno ype and coun ed only luminal cells s ained wi h cy oke a in-8/18 (CK-8/18), a ma ke o epi helial cells. The box plo o R1881-induced γH2AX oci is shown in Figu e S4a. The numbe o he coun ed cells is desc ibed in Table S4. Single and double as e isks indica e p=4.9×10–2 and p=1.3×10–2, which we e calcula ed by an unpai ed wo- ailed es . Scale ba ep esen s 25μm. (c, d) Rep esen a i e images (c) and he a e age numbe o γH2AX oci (d) cells in he epi helial cells o he mamma y gland a he indica ed ime a e ip wi h 17β-es adiol (E2). We sub ac ed he a e age numbe o oci in E2- ea ed epi helial cells a 6h om he a e age numbe o oci in E2-un ea ed epi helial cells. Ac ual numbe s o oci pe cell a e shown in Figu e S4d. The da a o wild- ype #1, 53BP1−/− and 53BP1−/−/BRCA1−/− mice we e e-calcula ed as he numbe o γH2AX oci pe cell om he da a ha has been p e iously published (Sasanuma e al.,2018). We coun ed only cells s ained wi h cy oke a in-8/18 (CK-8/18). We killed h ee mice each ca ying ei he he wild- ype #2 o TDP2−/− geno ype a wo mon hs o age and analyzed mo e han 100 epi helial cells o each mouse. The numbe o he coun ed cells is desc ibed in Table S4. Single, double and iple as e isks indica e p=1.0×10–3, p=1.1×10–2 and p=2.0×10–1 (no signi ican di e ence), espec i ely, which we e calcula ed by an unpai ed wo- ailed es . Scale ba ep esen s 25μm (a) (b) (c) (d) (h ) (h ) (h ) h 458 | Genes o Cells AL MAHMUD e Al. AR- ecep o -dependen manne . The loss o ei he NHEJ o TDP2 caused an inc ease in and ogen-induced DSBs o ~20 pe cell. S ikingly, his loss esul ed in 14 o 16 DSBs le un epai ed e en a 24h a e a 2-h pulse exposu e o and ogens, which is in ma ked con as wi h he comple e ejoining o and ogen-induced DSBs in wild- ype cells a 24h (Figu e2c). A ew and ogen-induced DSBs we e le un epai ed in indi idual p os a e epi helial cells o TDP2- de icien mice e en a 20h a e he injec ion o and ogen when and ogen-induced DSBs we e comple ely epai ed in wild- ype mice (Figu e 3b). These da a highligh he key ole in he TDP2-NHEJ axis (Gómez-He e os e al.,2013) in p o ec ing p os a e epi helial cells om and ogen-induced genome ins abili y. The geno oxici y o and ogens as well as es ogens can be media ed by he ollowing wo mechanisms. A physiological concen a ion (1 nM) o 17β-es adiol gene a es DSBs in wild- ype MCF-7 b eas cance cells by gene a ing s able RNA:DNA hyb id s uc u es known as R-loops and hei collision wi h DNA eplica ion o ks causes DSBs (S o k e al., 2016). Hence, and ogens can gene a e DSBs h ough he collision be ween R-loops and DNA eplica ion o ks. Since R-loops a e no a se ious h ea o genome ins abil- i y in G0/G1 phases ( e iewed in e . Aguile a & Ga cía- Muse,2012), and ogen-induced DSBs in G0/G1 phases a e gene a ed by a mechanism o he han R-loops. Recen s ud- ies ha e shown he second mechanism o and ogen-induced DSBs, TOP2-dependen DSBs (Ha ne e  al., 2010) ( e- iewed in e s. Madabhushi,2018; Nelson e al.,2018). We he e showed TDP2 plays a i al ole in he e icien epai o and ogen-induced DSBs (Figu es2e and 3b), sugges ing ha hey include 5′ TOP2 adduc s a hei ends. Indeed, FIGURE 4 R1881 acili a es he cellula p oli e a ion in he p os a e gland o TDP2−/− mice. (a) Schema ic diag am o expe imen al design o a daily injec ion o R1881 in 2-mon h-old mice. (b, c) R1881-induced cellula p oli e a ion by EdU inco po a ion in he mouse p os a e. Rep esen a i e images (b) o he EdU-posi i e (EdU+) cells and he a e age numbe s o R1881-induced EdU+ (c) luminal epi helial (CK-8/18- posi i e) cells in he p os a e gland o mice. The numbe o he coun ed cells is desc ibed in Table S4. EdU (30mg/kg body weigh ) and R1881 (15mg/kg body weigh ) oge he we e in ape i oneally (ip) injec ed in o he mice (days 0, 1 and 2 in “a”). The indica ed issues we e isola ed 24h a e he las ip (day 3 in “a”). Small in es ine and spleen we e analyzed as posi i e con ols in Figu e S5a. Single and double as e isks indica e p=2.9×10–3 and p=5.3×10–5, espec i ely, which we e calcula ed by an unpai ed wo- ailed es . Scale ba ep esen s 75μm. (d) R1881- induced cellula p oli e a ion by PCNA immunos aining in he mouse p os a e. The g aph shows he a e age numbe s o R1881-induced PCNA- posi i e (PCNA+) cells. We analyzed PCNA+ cells wi h he same expe imen al p ocedu e in “a”. The numbe o he coun ed cells is desc ibed in Table S4 Single and double as e isks indica e p=3.1×10–2 and p=1.1×10–5, espec i ely, which we e calcula ed by an unpai ed wo- ailed es . Rep esen a i e images o PCNA+ cells in he p os a e a e shown in Figu es S5b (a) (c) (d) (b) | 465 Genes o Cells AL MAHMUD e AL. Yang, S., Jiang, M., G abowska, M. M., Li, J., Connelly, Z. M., Zhang, J., … Yu, X. (2016). And ogen ecep o di e en ially egula es he p oli e a ion o p os a ic epi helial cells in i o and in i o. Onco a ge , 7, 70404–70419. Yue, F., Cheng, Y., B eschi, A., Vie s a, J., Wu, W., Ryba, T., … Ren, B. (2014). A compa a i e encyclopedia o DNA elemen s in he mouse genome. Na u e, 515, 355–364. h ps://doi.o g/10.1038/na u e13992 SUPPORTING INFORMATION Addi ional suppo ing in o ma ion may be ound online in he Suppo ing In o ma ion sec ion. How o ci e his a icle: Al Mahmud MR, Ishii K, Be nal-Lozano C, e al. TDP2 supp esses genomic ins abili y induced by and ogens in he epi helial cells o p os a e glands. Genes Cells. 2020;25:450–465. h ps:// doi.o g/10.1111/g c.12770