PARP-1-dependent RND1 transcription induced by topoisomerase I cleavage complexes confers cellular resistance to camptothecin

RHO GTPases regulate essential functions such as the organization of the actin cytoskeleton. The classic members cycle between an active GTP-bound and an inactive GDP-bound conformation whereas atypical members are predominantly GTP-bound. Besides their well-established role, the classic RHO GTPases RHOB and RAC1, are rapidly induced and/or activated by genotoxic stress and contribute to the DNA damage response. Here we used camptothecin, a selective topoisomerase I (TOP1) inhibitor that stabilizes TOP1 cleavage complexes (TOP1cc), to search for other potential early DNA damage-inducible RHO GTPase genes. We identified that an atypical RHO GTPase, RND1, is rapidly induced by camptothecin. RND1 induction is closely associated with the presence of TOP1cc induced by camptothecin or by DNA lesions that elevate TOP1cc levels such as UV and hydrogen peroxide. We further demonstrated that camptothecin increases RND1 gene transcription and mRNA stability. Camptothecin also increases poly(ADP-ribose) polymerase 1 (PARP-1) activity, whose inhibition reduces RND1 transcription. In addition, overexpression of RND1 increases PARP-1, suggesting a cross-talk between PARP-1 and RND1. Finally, RND1 protects cells against camptothecin-induced apoptosis, and hence favors cellular resistance to camptothecin. Together, these findings highlight RND1 as an atypical RHO GTPase early induced by TOP1cc, and show that the TOP1cc-PARP-1-RND1 pathway protects cells against apoptosis induced by camptothecin.


Introduction
The RHO GTPase family comprises 20 members in human, which can be divided into classic and atypical members 1 . Classic RHO GTPases, such as RHOB and RAC1, cycle between an active GTP-bound and an inactive GDP-bound conformation. Atypical RHO GTPases, such as RND1, are unable to hydrolyze GTP and are therefore in a constitutive active GTP-bound conformation 2,3 . Other atypical members, such as RHOU, and presumably also RHOV, have a high nucleotide exchange rate and hence are assumed to be mainly GTP-bound 4 . Consequently, the tight control of the expression of atypical RHO GTPases is important to precisely tune their activity. GTP-bound RHO GTPases bind to their effectors and regulate pivotal cellular functions, including the organization of the actin and microtubule cytoskeletons, cell adhesion and cell migration 5 .
Besides their canonical roles, the RHO GTPases RAC1 and RHOB have been implicated in the early response to DNA damage. Inhibition or deletion of RAC1 reduces the DNA damage signaling pathway upon UV light 6 or ionizing radiation 7 and, sensitizes cell to ionizing radiation 7 or to UV-light-induced apoptosis 6 . Unlike RAC1 that is primarily activated in response to DNA damage without change in expression 7,8 , RHOB is both induced and activated [9][10][11][12] . RHOB induction by genotoxic stress, such as UV light and the topoisomerase I (TOP1) inhibitor camptothecin (CPT), is rapid and relies on increased transcription and/or transcript stability 9,10 . Increased expression of RHOB promotes DNA repair and confers cell resistance to genotoxic stress 9 . At present, it is not known whether, besides RHOB, other RHO GTPases are early DNA damage-inducible genes, at the expression level.
TOP1 solves DNA topological problems that are generated during transcription and replication 13 . It relaxes DNA by forming transient TOP1 cleavage complexes (TOP1cc), which are TOP1-linked DNA single-strand breaks . After DNA relaxation, TOP1cc reverse rapidly, and TOP1 is released as the DNA religates. The transient TOP1cc can be trapped selectively by CPT and its derivatives irinotecan and topotecan, used to treat cancers, which bind at the TOP1-DNA interface 14 . Many DNA alterations including oxidative base damages 15,16 and UV lesions 17,18 also interfere with TOP1 nicking-closing reactions and give rise to elevated levels of TOP1cc (see Table 1 in ref. 13 ). Persistent TOP1cc can lead to the production of DNA double-strand breaks (DSBs) during replication [19][20][21] and transcription [22][23][24] , and ultimately to apoptotic cell death 25 .
An early response to long-lived TOP1cc is the interference with the progression of transcription 14,26 . Indeed, trapping TOP1cc by CPT inhibits transcription elongation with increasing efficiency as the genes become longer and contain more exons [27][28][29] . However, genes are differentially affected by CPT and a fraction of them, primarily the short and low-expressed genes, are upregulated 27,28 . The mechanisms by which CPT-induced TOP1cc trapping enhances transcription at some genes are largely unknown. Here, we identified RND1 as the first atypical RHO GTPase, which is rapidly induced at the gene level by CPT and DNA damaging agents that indirectly trap TOP1cc, such as hydrogen peroxide (H 2 O 2 ) and UV light. We found that persistent TOP1cc increase RND1 transcription by a mechanism that depends on poly(ADPribose) polymerase 1 (PARP-1) activity, providing one of the first examples of how stabilized TOP1cc can stimulate gene transcription. Lastly, we found that increased RND1 expression reduces CPT-induced apoptosis, highlighting a protective function for the TOP1cc-PARP-1-RND1 pathway.

Quantitative reverse transcription-PCR
Total RNAs were extracted using the RNeasy Plus mini kit (QIAGEN) according to the manufacturer's instructions and the concentration and purity of RNA were determined using the Nanodrop ND-1000. RNAs were reverse transcribed using the iScript cDNA synthesis kit (Bio-Rad). qPCR analyses were performed on a CFX96 real-time system device (Bio-Rad) by using IQ SYBR green Supermix (Bio-Rad) according to the manufacturer's instructions. All samples were analyzed in triplicate, and actin, GAPDH and 28 S mRNA were used as endogenous controls in the ΔΔCT analysis. The human (h) and mouse (m) primer pairs used were hRHOA-FW

WST-1 cell viability assays
GFP-positive and GFP-negative RND1 sorted cells (Astrios, Beckman) were immediately seeded in triplicate into 96-well microplates at a density of 1,000 cells per well. Twenty-four hours after plating, cells were treated with increasing concentrations of CPT (from 1.6 nM to 25 µM) and cultured for 72 h. The WST-1 reagent (Roche Diagnostics) was then applied for 1 h at 37°C. The formazan dye was quantified at 450 nm using a plate reader (FLUOstar Optima, BMG Labtech). Data were expressed as the percentage of cell survival (mean ± SD of treated cells normalized to the mean ± SD of untreated cells, which was set to 100%).

Clonogenic assays
Three hundred U2OS shCtrl or shRND1 cells were treated with increasing concentrations of CPT (from 1.25 nM to 20 nM). Ten days after CPT treatment, cells were fixed with 3.7% paraformaldehyde (Sigma-Aldrich) and stained with 1% crystal violet (Sigma). Colonies containing more than 50 cells were counted.

Nascent RNA transcripts analysis
Nascent RNAs were labeled and captured using the Click-iT Nascent RNA capture kit (Life Technologies) according to the manufacturer's instructions. A 2-h 5ethynyl uridine (EU) pulse at the concentration of 0.2 mM was performed to label nascent RNAs. Three to seven µg of total RNA was used for the Click reaction.

Meta-analysis of RND1 mRNA expression
This analysis was performed using the online Next-bioResearch tools (http://www.nextbio.com/). We collected RND1 mRNA expression fold-change after treatment with CPT or derivatives in different cancer cells (OCI-LY3 cells: diffuse large B cell lymphoma; MCF-7: breast cancer cells; PC3: prostate cancer cells; HCT116: colon cancer cells) or tissue (bone marrow from rats) from five gene expression datasets. GEO accession numbers of gene expression datasets in order of appearance: GSE63902 33 ; GSE51068 34 ; GSE18552 35 ; GSE5258 36 and GSE37352 27 .

Detection of TOP1 cleavage complexes
Cellular TOP1 cleavage complexes (TOP1cc) were detected as previously 22 , except that immunoblotting was revealed with a mouse anti-TOP1cc from Millipore (MABE1084) 37

Flow cytometry
For sub-G1 analysis, cells were fixed with 70% ethanol, incubated with RNase A (Sigma-Aldrich) and stained with propidium iodide (PI; Molecular probes). The stained cells were analyzed on a BD Accuri C6 flow cytometer (BD Biosciences). Analysis was performed with the BD Accuri C6 flow cytometer software.

RND1 transcripts are rapidly induced by CPT
To determine the RHO GTPases that are induced early in response to CPT, we treated human osteosarcoma U2OS cells for short times (1 h and 2 h), and analyzed RHO GTPase mRNA expression by reverse transcription followed by qPCR (RT-qPCR) (Fig. 1a). CPT efficiently induced TOP1cc in U2OS cells (Fig. 1b) as previously reported 38 . Among the RHO GTPase family, the two atypical members RND1 and RHOV, were increased by CPT with RND1 displaying an approximately 4 and 12 folds' increase after 1 h and 2 h, respectively, and RHOV an approximately 3 and 4 folds' increase (Fig. 1a). RHOB also increased under these conditions (Fig. 1a), as previously reported 9 . Among the two newly identified RHO GTPases, RND1 and RHOV, which are induced early by CPT (Fig. 1a), we further characterized RND1.
Kinetics of RND1 mRNA induction in U2OS cells showed that at 25 µM of CPT, RND1 increased within 1 h and reached a plateau after 2 h (Fig. 1c). To investigate whether the induction of RND1 mRNA was dose-dependent, cells were treated for 2 h with increasing CPT concentrations. RND1 induction was detected at 1 µM and increased with increasing concentrations of CPT (Fig. 1d). RND1 induction was also observed in other human cell lines (glioblastoma U87, colon carcinoma HCT116, primary lung WI38 hTERT), and mouse cell lines (melanoma B16F10, embryonic NIH3T3) treated with CPT (Fig. 1e). Under these conditions, CPT efficiently induced TOP1cc in all these cell lines (Fig. 1f). Meta-analysis of microarray databases further supports the increase of RND1 mRNA levels after short treatment with CPT or its water-soluble derivatives, topotecan and irinotecan, in human and rodent cell lines, and in tissues (Supplementary Figure 1A). In contrast, the two RND1 homologs, RND2 and RND3 were not induced after a short treatment with CPT in the cell lines analyzed (Fig. 1a, Supplementary Figure 1B, C).
As other groups 39 , we could not find or generate highaffinity antibodies that react specifically with endogenous RND1. Therefore, to determine whether the increase in RND1 transcript levels could be associated with an increase in RND1 protein levels in CPT-treated cells, we generated U2OS cells stably expressing low or high levels of V5-tagged RND1 transcripts (Fig. 1g). Cells with low and high levels of RND1 transcripts (Fig. 1g) expressed low and high levels of RND1-V5 protein (Fig. 1h), respectively, suggesting that increasing RND1 transcript expression also increases RND1 protein levels. Altogether, these results identify RND1 as a new early-inducible RHO GTPase gene in response to CPT.

RND1 transcripts are closely associated with the presence of TOP1cc
CPT has for sole cellular target the TOP1cc 14 . To assess whether TOP1cc stabilization by CPT primes the increase of RND1 mRNA levels, we examined whether other agents that induce TOP1cc would also induce RND1. Oxidative-and UV-mediated DNA lesions give rise to elevated levels of TOP1cc (see Table 1 in ref. 13 ). As a result, H 2 O 2 and UV light induce cellular TOP1cc (Supplementary Figure 2A) 18,40 . Figure 2a shows that both agents increased RND1 mRNA levels. Conversely, agents that do not induce TOP1cc, including the hypoxiamimicking agent cobalt (II) chloride (CoCl 2 ), the dihydrofolate reductase inhibitor methotrexate, and the tubulin inhibitor paclitaxel, did not increase RND1 (Fig. 2b), under conditions where they exert their expected biological effects (for CoCl 2  Because CPT-induced TOP1cc are reversible 41 , we further examined RND1 transcripts following CPT removal. After termination of the CPT treatment, RND1 mRNA returned to their baseline levels (Fig. 2c, top panel) as TOP1cc reversed (Fig. 2c, bottom panel). Because the stabilization of TOP1cc decreases TOP1 activity leading to topological stress 14 , we have examined RND1 induction in TOP1-depleted cells. Figure 2d shows that siRNAmediated depletion of TOP1 in U2OS and WI38 hTERT cells did not increase RND1 mRNA levels, suggesting that TOP1cc rather than inhibition of TOP1 activity promote RND1 induction. Collectively, these results indicate that the increase of RND1 transcripts is closely associated with the presence of TOP1cc.

CPT increases RND1 transcription and RND1 transcript stability
The early increase of RND1 mRNA in CPT-treated cells could depend on an increase in transcription and/or in transcript stability. Analysis of RND1 transcription by capture of nascent transcripts followed by RT-qPCR showed that CPT increased by approximately 20 folds the transcription of RND1 gene (Fig. 3a). This increase fully reversed after the removal of CPT (Fig. 3a), indicating that RND1 transcription is closely related to the presence of TOP1cc. To determine whether the increase of RND1 transcription in CPT-treated cells would depend on the activity of its promoter, we measured the activity of a luciferase reporter gene placed under the control of the RND1 minimal promoter either alone or together with a proximal or a distal enhancer region 42,43 . Fig. 3b shows that CPT did not increase luciferase activity in cells transfected with each of these constructs, suggesting that the increase in RND1 transcription by CPT might not primarily depend on an increased activity of its minimal promoter and the tested enhancers.
Next, we compared the stability of RND1 mRNA between untreated and CPT-treated cells. Experiments performed in the presence of the transcription inhibitor flavopiridol showed that the half-life of RND1 mRNA was greatly prolonged in CPT-treated cells (Fig. 3c). Similar results were obtained with the transcription inhibitor actinomycin D (Fig. 3d). Upon exposure to CPT, TOP1 has been reported to be degraded over time in a transcription-dependent manner 22,44,45 . Hence, the use of transcription inhibitors to analyze the lifespan of RND1 mRNA in CPT-treated cells is likely to sustain the levels of TOP1cc during the time course of these experiments, which might contribute to further increase the half-life of RND1 mRNA. Altogether, these data indicate that the increase of RND1 transcript levels in response to CPT is associated with an increase in both RND1 transcription and RND1 transcript stability.

PARP-1 increases RND1 transcription in response to TOP1cc
PARP-1 can promote gene transcription [46][47][48][49] and transcript stability 50 via the addition of poly(ADP-ribose) residues (PAR) on proteins, an activity named PARylation. Because a short time CPT treatment increases PARP-1 Fig. 3 CPT increases RND1 transcription and RND1 transcript stability. a U2OS cells were treated with 25 µM CPT for 1 h and washed and cultured in CPT-free medium (CPT + Washes) for 0.5 h to allow reversion of TOP1cc. At the end of each time point, 0.2 mM EU was added to the culture medium for 2 h, after which EU-labeled nascent RNAs were captured. Nascent RND1 RNAs were then analyzed by RT-qPCR (means ± SEM, n = 3; except for the "CPT + Washes" time point, for which n = 2). **P < 0.01 by unpaired t test. b Left panel: Diagram of pGL3-RND1 promoter constructs (minimal promoter alone, minimal promoter with a proximal or a distal enhancer in order of appearance). Right panel: Luciferase activity of U2OS cells treated with 25 µM CPT for 2 h (minimal and distal promoter) or 3 h (proximal promoter). For the positive control, luciferase activity of U2OS cells treated with 1 µM 5-aza-2'-deoxycytidine (5AZA) for 72 h and with 100 nM trichostatin A (TSA) for 24 h (means ± SD, n = 3), *P < 0.05, **P < 0.01 by one-way ANOVA. c U2OS cells were left untreated or were treated with 25 µM CPT for 2 h before the addition of the transcription inhibitor flavopiridol (FLV, 1 µM) for the indicated times. RND1 mRNA was then analyzed by RT-qPCR and normalized to the level at the time of FLV addition (means ± SD, n = 3). ***P < 0.001, ****P < 0.0001, by two-way ANOVA. The half-life (t 1/2 ) of RND1 mRNA is indicated. d Experiments were performed as in (c) with the transcription inhibitor actinomycin D (10 µg/ml). A representative experiment out of two is shown (means ± SD of triplicate samples). Ns, not significant. activity 31 , we examined whether PARP-1 could promote the increase in RND1 transcript levels in CPT-treated cells.
As reported 31 , CPT increased protein PARylation (Fig. 4a). Protein PARylation was reversible and returned to its baseline level after CPT removal (Fig. 4a), a similar effect to that of RND1 mRNA levels (Fig. 2c) and RND1 gene transcription (Fig. 3a). Then, we assessed whether inhibiting PARP-1 activity would prevent the induction of RND1 mRNA. The PARP-1 inhibitor veliparib partially prevented the induction of RND1 mRNA in response to CPT (Fig. 4b) under conditions where it prevented protein PARylation (Fig. 4a). Then, we asked whether PARP inhibition would decrease RND1 transcription and/or RND1 transcript stability. In CPT-treated cells, veliparib strongly inhibited RND1 transcription (Fig. 4c), while it did not decrease the half-life of RND1 transcripts (Fig. 4d). As previously reported 38 , veliparib did not affect TOP1cc levels in response to CPT (Fig. 4e, f), which further suggests that PARP-1 is downstream from TOP1cc to increase RND1 transcription. Altogether, these results suggest that, in CPT-treated cells, TOP1cc stabilization increases PARP-1 activity, which in turn increases the transcription of RND1 gene, leading to an increase of RND1 transcripts.

RND1 increases PARP-1 expression in a positive feedback loop
Next, we considered whether there is a cross-talk between PARP-1 and RND1 or whether the talk is limited to one direction in which PARP-1 induces RND1. To test this, we asked whether modulating RND1 expression would modulate PARP-1 expression. Downregulation of RND1 mRNA levels by shRNA in U2OS cells (Fig. 5a), decreased PARP-1 expression (Fig. 5b). Conversely, U2OS cells overexpressing RND1 (characterized in Fig. 1g, h), also overexpressed PARP-1 (Fig. 5c). These results suggest that PARP-1 activity increases RND1, which in turn increases PARP-1 expression in a positive feedback loop.
(see figure on previous page) Fig. 4 The PARP-1 inhibitor veliparib prevents CPT-induced RND1 transcription. a U2OS cells were treated with 25 µM CPT for 2 h and washed and cultured in CPT-free medium (CPT + Washes) for 2 h to allow reversion of TOP1cc. When indicated, cells were pretreated with 5 µM veliparib for 1 h. The expression of PAR and PARP-1 were analyzed by Western blotting. The top panel shows quantification of PAR normalized to PARP-1 (means ± SD, n = 2). b RT-qPCR analysis of RND1 mRNA in U2OS cells treated with 5 µM veliparib for 1 h before the addition of CPT for 2 h (means ± SD, n ≥ 3). **P < 0.01, ****P < 0.0001 by unpaired t test. c U2OS cells were treated with 5 µM of veliparib for 1 h before the addition of 25 µM CPT for 2 h. At the end of each time point, 0.2 mM EU was added to the culture medium for 2 h, after which EU-labeled nascent RNAs were captured. Nascent RND1 RNAs were then analyzed by RT-qPCR. Data were normalized to the level of CPT-treated cells, which was taken at 100%. A representative experiment out of two is shown (means ± SD of triplicate samples). d U2OS cells were treated with veliparib (5 µM, 1 h) followed by the addition of CPT (25 µM, 1 h). After which, the transcription inhibitor FLV (1 µM) was added for 4 h. RND1 mRNA expression was analyzed by RT-qPCR and normalized to the level at the time of FLV addition, which was set to 100 % (means ± SD, n = 3). Ns, not significant by two-way ANOVA. e, f U2OS cells were treated with 5 µM veliparib for 1 h before the addition of 25 µM CPT for 2 h, and TOP1cc were detected by probing two concentrations of genomic DNA (5 and 2.5 µg) with an anti-TOP1cc antibody. e Representative experiment. f Quantification of TOP1cc in veliparib + CPT-treated cells normalized to values from CPT-treated cells (means ± SEM, n = 3). Ns, not significant by unpaired t test.

DNA-PK-dependent DSB signaling prevents the induction of RND1 transcripts by CPT
Because CPT-induced TOP1cc can lead to the production of DSBs 19-21 22-24 , we examined the role of DSB signaling in the induction of RND1. ATM, ATR and DNA-PK are serine/threonine kinases that are readily activated by DSBs, and phosphorylate various DNA damage response proteins such as p53 51 . Consistent with that, CPT activated these three kinases in U2OS cells as demonstrated by autophosphorylation of ATM at S1981 (ATM-pS1981), phosphorylation of the ATR substrate Chk1 at S345 (Chk1-pS345), and autophosphorylation of DNA-PK at S2056 (Fig. 6a). To determine their potential role in RND1 induction, we assessed whether RND1 induction is modified by specific chemical inhibitors of these kinases in CPT-treated U2OS cells: the ATM inhibitor (ATMi) KU55933, the ATR inhibitor (ATRi) VE-821 and the DNA-PK inhibitor (DNA-PKi) NU7441 (Fig. 6a). Figure 6b shows that DNA-PKi, and in a lesser extend ATMi and ATRi, increased the induction of RND1 mRNA in response to CPT. Because p53 is phosphorylated and activated by these kinases 51 , we examined RND1 induction in p53+/+ and p53−/− HCT116 cells exposed to CPT. As expected, CPT induced p53 phosphorylation at S15 and increased p53 protein level in p53+/+ HCT116 cells (Fig. 6c, bottom panels). We found that p53 +/+ and p53−/− HCT116 cells both displayed similar induction of RND1 mRNA in response to CPT (Fig. 6c,  top panel). Collectively, our experiments suggest that the DNA-PK-dependent DSB signaling prevents the induction of RND1 transcripts by CPT. However, this DSB signaling pathway is not mediated by p53.

RND1 reduces the sensitivity of cells exposed to CPT
To assess the potential role of RND1 in the cellular response to TOP1cc stabilization, we first compared the sensitivity of shCtrl and shRND1 U2OS cells (characterized in Fig. 5a) to CPT treatment. Cells were treated with increasing concentrations of CPT, and CPT sensitivity was assessed by clonogenic assays. Figure 7a, b shows that shRND1 cells formed significantly less clones in response to CPT than shCtrl cells. The increased sensitivity of shRND1 cells to CPT was further associated with an increase of apoptotic marks such as sub-G1 population (Fig. 7c), and the cleavage of caspase-9, caspase-3 and PARP-1 (Fig. 7d). Conversely, overexpression of RND1 in Figure 3) decreased cell sensitivity to CPT as measured by WST-1 survival assays (Fig. 7e) and decreased apoptotic marks (Fig. 7f). Together, these results demonstrate that RND1 protects cells against CPT, likely by preventing apoptosis.

Discussion
Here we identified RND1 as an early inducible RHO GTPase gene in response to CPT. This is the first time that an atypical RHO is reported to respond early to DNA damaging agents. Our data support a model depicted in Fig. 8 in which CPT-induced TOP1cc stabilization increases PARP-1 activity that triggers RND1 transcription, which elevates the levels of RND1 transcripts (and likely also the protein). In turn, the increase of RND1 protein levels promotes an increase of PARP-1 protein levels, suggesting a positive feedback loop between PARP-1 and RND1 in response to CPT. The increase of RND1 induced by the TOP1cc-PARP-1 pathway protects cells against CPT, likely by inhibiting apoptosis. PARP-1independent pathways probably also contribute to the increase of RND1 transcript levels, as the inhibition of PARP-1 activity with veliparib does not completely suppress CPT-induced RND1 transcripts. Such pathways might involve an increased stability of RND1 transcripts as our analysis shows that CPT extends the half-life of RND1 mRNA in a PARP-1-independent manner. CPTinduced TOP1cc also induce DSBs, which activate DNA-PK that reduces the induction of RND1. DNA-PK could reduce the induction of RND1 by promoting TOP1 proteolysis as previously reported 22,52 . Albeit in a lesser extent, ATM, which can activate DNA-PK 22 , also reduces the induction of RND1. Consistent with that, ATM has also been reported to promote TOP1 proteolysis 53,54 .
Our study uncovers the close relationship between TOP1cc and the transcription of RND1. Indeed, CPT, which induces RND1 transcription, has for sole cellular target the TOP1cc 14 , and reversion of TOP1cc following termination of the CPT treatment readily restores the (see figure on previous page) Fig. 7 RND1 protects cells against CPT-induced apoptosis. a, b Colony formation assay in U2OS cells stably expressing shRNAs against RND1 (shRND1) or against a control sequence (shCtrl), and treated with increasing concentrations of CPT (from 1.25 to 20 nM). Percentages of colonies were assessed after 10 days by counting the number of colonies and normalized to that of untreated cells, which was set at 100% (means ± SD, n = 4), *P < 0.05, **P < 0.01, ***P < 0.001, by two-way ANOVA. c U2OS shRND1 or shCtrl cells were treated with 25 µM CPT for 24 h. Percentage of sub-G1 cell population was analyzed by flow cytometry. The top panel shows quantification of sub-G1 cell population (means ± SEM, n=3). *P < 0.05 by unpaired t test. Bottom: one representative experiment is shown. d Western blotting analysis of the indicated proteins in U2OS shRND1 or shCtrl cells treated with 25 µM CPT for 24 h. Caspase-9 CL : cleaved caspase-9, Caspase-3 CL : cleaved caspase-3, PARP-1 CL : cleaved PARP-1. Data shown are representatives from three experiments. (e) U2OS cells were transfected with pEGFP-RND1 plasmid. Forty-eight hours after transfection, GFP-and GFP + U2OS cells were separated by cell sorting. GFP− and RND1 GFP+ U2OS cells were treated with increasing concentrations of CPT (from 0.0016 µM to 25 µM). Seventy-two hours after treatment, cell survival was analyzed by a WST-1 assay. A representative experiment out of three is shown (means ± SD for triplicate samples). Ns not significant, **P < 0.01, ****P < 0.0001, by two-way ANOVA. f Similar experiments as in panel (d) in U2OS Ctrl and RND1-V5high cells. Ns not significant Fig. 8 Proposed model for the induction of RND1 in response to CPT. CPT stabilizes TOP1cc, which in turn induce DSBs and apoptosis 25 [1]. TOP1cc also activates a PARP-1-RND1 pathway that counteracts the induction of apoptosis [2]. DNA-PK-dependent DSB signaling prevents RND1 induction, possibly by promoting TOP1cc removal 22,52 [3].
baseline levels of RND1 transcription and RND1 transcripts. In addition, H 2 O 2 and UV light, which induce DNA lesions that interfere with TOP1 nicking-closing activity and give rise to elevated levels of TOP1cc [15][16][17][18]40 , also increase RND1 transcript levels. Besides TOP1 inhibitors and DNA alterations (see Table 1 in ref. 13 ), several other processes lead to persistent TOP1cc, including ribonucleotide incorporation into DNA [55][56][57] , genetic defects such as ATM defect 53,54 , and transcriptional activation 58 . Hence, the increased transcription of RND1 due to TOP1cc stabilization might be a frequent event occurring under both physiological and stress conditions.
An early response to CPT is the global inhibition of transcription 22,45 . However, genes are differentially affected by CPT and a fraction of them, primarily the short and low-expressed genes, are upregulated [27][28][29] . In accordance with this, RND1 is a short gene (8.7 Kbp), with a lowexpression in most healthy tissues apart from brain and liver 2 and in addition, RND1 expression is significantly downregulated in several aggressive tumors compared to normal tissues 39,59,60 . The mechanisms by which CPT enhances transcription of some genes are largely unknown. Here we reported that CPT induces PARP-1 activity, which in turn stimulates RND1 transcription. This effect is likely related to TOP1cc. Similar to CPT, H 2 O 2 and UV light induce persistent TOP1cc 18,40 , increase RND1 transcript levels (this study), and also increase PARP-1 activity 61,62 . Whether TOP1cc-induced PARP-1 activity is a common mechanism for CPT to promote gene transcription or whether it is restricted to RND1 gene remains to be investigated.
It is now well documented that PARP-1 regulates transcription 63 asides from its well-recognized role in DNA repair 64 . PARP-1 is enriched to the promoters of actively transcribed genes 65 and, stimulates transcription initiation by maintaining an 'open' chromatin environment through PARylation of core histones and exclusion of histone H1 from the DNA 48,65 , or inhibition of histone H3K4me demethylation by KDM5B 47 . PARP-1 could also promote transcription by stimulating transcription elongation. PARP-1 PARylates subunits of the negative elongation factors (NELF), NELF-A and NELF-E, which triggers the release of RNA polymerase II from its paused site for productive elongation 49 . Our results showing that increased RND1 transcription by CPT does not relies on increased activity of its promoter suggest that PARP-1 might primarily function in stimulating transcription elongation of RND1 gene. PARP-1-independent pathways probably also contribute to the increase of RND1 transcription, as PARP-1 inhibition does not completely suppress CPT-induced RND1 transcription. A previous study shows that the CPT derivatives topotecan can stimulate UBE3A transcription by downregulating the expression of its antisense transcript 66 . The non-coding RNA AGAP2-AS1 has been reported to inhibit RND1 transcription 67,68 , which raises the possibility that CPT could inhibit AGAP2-AS1 transcription, which in turn could increase RND1 transcription.
Lastly, our analysis shows that RND1 protects cells against CPT-induced apoptosis and hence favors cell resistance. These findings extend the role of RND1 beyond its original function in the disassembly of actin filament structures and loss of cell adhesion 2 as well as in embryonic development, where it promotes the formation and maturation of neuronal protrusions 69,70 and controls gastrulation movements 71 . In addition, RND1 behaves as a tumor suppressor gene. RND1 expression levels decrease in several aggressive tumors 39,59,60 , and RND1 loss in immortalized mammary cells can initiate breast tumorigenesis and promotes metastasis 39 . Even in tumor cell lines expressing low levels of RND1 such as MCF-7 cells 39 , U87 cells 59 and U2OS cells, RND1 could be transiently induced by TOP1cc to resist to CPT derivatives. This potential selective advantage of tumor cells suggests that inhibiting RND1-dependent signaling could sensitize them to CPT derivatives.