WRN helicase safeguards deprotected replication forks in BRCA2-mutated cancer cells

The tumor suppressor BRCA2 protects stalled forks from degradation to maintain genome stability. However, the molecular mechanism(s) whereby unprotected forks are stabilized remains to be fully characterized. Here, we demonstrate that WRN helicase ensures efficient restart and limits excessive degradation of stalled forks in BRCA2-deficient cancer cells. In vitro, WRN ATPase/helicase catalyzes fork restoration and curtails MRE11 nuclease activity on regressed forks. We show that WRN helicase inhibitor traps WRN on chromatin leading to rapid fork stalling and nucleolytic degradation of unprotected forks by MRE11, resulting in MUS81-dependent double-strand breaks, elevated non-homologous end-joining and chromosomal instability. WRN helicase inhibition reduces viability of BRCA2-deficient cells and potentiates cytotoxicity of a poly (ADP)ribose polymerase (PARP) inhibitor. Furthermore, BRCA2-deficient xenograft tumors in mice exhibited increased DNA damage and growth inhibition when treated with WRN helicase inhibitor. This work provides mechanistic insight into stalled fork stabilization by WRN helicase when BRCA2 is deficient.

G ermline mutations in BRCA1 and BRCA2 account for the majority of hereditary breast and ovarian cancers 1 . In addition to defective double-strand break (DSB) repair 2 , compromised fork stability significantly contributes to heightened genomic instability and chemosensitivity in BRCA-deficient tumors 3,4 . BRCA2 safeguards genome integrity by protecting stalled replication forks from unscheduled nucleolytic degradation in human cells 4,5 . However, the duty of fork protection extends beyond BRCA1/2 and solicits the involvement of other stress response proteins including ubiquitinated-PCNA 6 , RAD51 paralogs 7 , FANCD2 8 , FANCA 8 , FANCJ 9 , BODL1 10 , CTIP 11 , Abro1 12 , WRNIP1 13 , EXD2 14 , RIF1 15 , and RAD52 16 . These fork stabilizing factors are unique and operate independently or synergistically with BRCA1 and BRCA2 to maintain fork integrity 11,14,17 . Understanding how fork processing proteins functionally collaborate during replication stress to stabilize forks has important implications for genomic stability and cancer therapy.
In recent years, replication fork stability has emerged as an underlying mechanism of acquired drug resistance in BRCAmutated cancers 3,18,19 . BRCA2-defective tumors become resistant to poly (ADP-ribose) polymerase (PARP) inhibitors or cisplatin by restoring fork stability without recovering homologous recombination (HR) mediated by BRCA2 3,20 . Thus, BRCAmutated cancer cells rely on alternative fork stabilization mechanisms for survival upon drug-induced replication stress 18 . Therefore, increasing fork instability burden through targeted inhibition of compensatory fork stabilizing factors is a promising therapeutic approach to target human cancers with BRCA2 lossof-function mutations. Supporting this notion, pharmacological PARP inhibition exacerbates pathological fork degradation and cell death in BRCA2-deficient cells exposed to the replication inhibitor hydroxyurea (HU) 21 .
WRN, a member of the RECQ family of DNA helicases 22 , plays essential roles at stalled forks to counteract replication stress, thereby ensuring genomic stability [23][24][25][26][27] . However, owing to its dual helicase and exonuclease activities and multiple protein interactions, the functions of WRN at perturbed replication forks are complex and diverse. While WRN helicase activity promotes DNA2-mediated replication restart 26 , the exonuclease function 28 or structural (noncatalytic) role of WRN 29 has been implicated in fork protection. How WRN operates at stalled forks upon deficiency of major fork stabilization mechanisms is largely unexplored.
In this study, we report that WRN helicase rescues stalled forks and limits hyper-degradation of nascent DNA in BRCA2deficient cells. At the biochemical level, we show that WRN ATPase/helicase function promotes fork restoration and inhibits MRE11-mediated nucleolytic degradation of reversed forks. Accordingly, WRN helicase complementation rescues fork restart-delay in BRCA2-depleted cells. Moreover, WRN helicase or exonuclease partially rescues hyper-degradation of stalled forks caused by BRCA2 loss. Attenuation of WRN helicase results in enhanced fork degradation and defective fork restart in BRCA2mutated cancer cells. Targeted WRN helicase inhibition causes rapid sequestration of WRN protein on chromatin and destabilizes stalled forks in BRCA2-deficient cells in a manner dependent on fork remodeling enzymes and the structure-specific nuclease MRE11. Consistent with these findings, BRCA2-mutated cancer cells exhibit increased DNA damage accumulation and elevated chromosomal instability upon WRN helicase inhibition. Pharmacological inhibition of WRN helicase preferentially kills BRCA2-deficient cancer cells and sensitizes them to the PARP inhibitor olaparib. In vivo, we show that WRN helicase inhibition retards BRCA2 −/− tumor growth and induces DNA damage in a xenograft mouse model. Altogether, the results demonstrate that WRN helicase plays a specialized role in stabilizing forks experiencing stress that is deprotected due to BRCA2 deficiency.

Results
WRN stabilizes replication forks in BRCA2-deficient cells and ensures fork recovery. Experimental evidence suggests that when a major fork protector such as BRCA2 is absent, other fork metabolizing proteins assume roles to stabilize stressed replication forks 8,17,30,31 . Given the apparent involvement but poorly understood function(s) of WRN during replication stress, we investigated its role in stalled fork stabilization when BRCA2 function is lost. Using the single-molecule DNA fiber assay, we first examined fork recovery following HU-induced replication stalling in WRN CRISPR knock-out (KO) (WRN −/− ) U2OS cells upon siRNA-mediated knockdown of BRCA2 (Fig. 1a, b, and Supplementary Fig. 1a). As depicted in Fig. 1b, BRCA2-depleted U2OS/WRN −/− cells transfected with either control empty vector (EV) or a WRN expression construct (Fig. 1a) were subjected to a fork restart assay 15,26 . To evaluate fork restart efficiency upon release from HU-induced stalling, we determined the relative IdU/CldU ratio in the defined genetic conditions. Compared to control siRNA-treated cells (Median IdU/CldU ratio = 0.8376), BRCA2-depleted cells transfected with EV showed a significant decrease in IdU/CldU ratio (Median IdU/CldU ratio = 0.6279) (Fig. 1b), suggesting compromised replication recovery upon BRCA2 loss in WRN-deficient cells. To assess whether WRN can rescue the observed fork restart defect induced by loss of BRCA2, we examined fork recovery in BRCA2-depleted cells complemented with either wild-type (WT) or catalytic domain site-directed mutants of WRN (Fig. 1a, b, and Supplementary Fig. 1a). Exogeneous expression of WRN in BRCA2-depleted U2OS/WRN −/− cells fully restored the IdU/CldU ratio (Median IdU/CldU ratio = 0.8406) (Fig. 1b), suggesting efficient recovery of stalled forks ensured by WRN under the cellular condition of BRCA2 deficiency. However, the expression of helicase-dead WRN-K577M was unable to rescue the stalled fork phenotype evidenced by no significant change in median IdU/CldU ratio (0.6233) in BRCA2-depleted cells (Fig. 1b). In contrast, expression of WRN exonuclease-deficient mutant (WRN-E84A) rescued the fork restart defect (median IdU/ CldU ratio = 0.7684). However, as compared to EV transfected cells, no significant change in median IdU/CldU ratio (0.6196) was observed in cells expressing a WRN mutant deficient in both helicase and exonuclease activities (WRN-K577M/E84A) (Fig. 1b). In agreement with these results, purified recombinant WRN-WT or WRN-E84A catalyzed fork restoration from a model reversed fork ("chicken-foot") DNA substrate in vitro whereas WRN-K577M failed to do so ( Fig. 1c and d), suggesting fork restoration is supported by WRN ATPase/helicase activity. These results suggest that restoration of intact replications forks by WRN helicase is critical for efficient replication resumption following fork stalling when BRCA2 function is lost.
Defective replication restart is linked to uncontrolled nucleolytic processing of stalled forks if they are not sufficiently protected 13,15,32 . WRN has been previously shown to protect camptothecin (CPT)-induced stalled forks from MRE11/EXO1mediated degradation in cells expressing WT BRCA2 28,29 . Unprotected reversed forks are extensively degraded by MRE11 exonuclease 33 . We reasoned that fork restoration by WRN helicase (Fig. 1d) may result in decreased reversed fork degradation by MRE11 nuclease because of the reduced availability of the reversed fork substrates. We, therefore tested recombinant MRE11 3´-5´exonuclease activity on a reversed fork substrate in the presence of recombinant WRN-E84A to monitor MRE11 digestion without interference caused by WRN exonuclease on the substrate (Supplementary Fig. 1b). Digestion of the reversed fork DNA substrate by MRE11 was found to be substantially reduced in the presence of WRN-E84A, suggesting WRN helicase activity curtails MRE11-mediated fork degradation in vitro.
Next, we sought to determine if WRN could restrain uncontrolled nucleolytic degradation of unprotected forks in BRCA2-depleted cells. We, therefore, assessed the ability of WT or catalytic mutants of WRN to complement elevated fork instability induced by BRCA2 loss upon replication stalling by   Fig. 1e and Supplementary Fig. 1c). For this, we performed fork stability assays in which BRCA2-depleted WRN KO cells exogenously expressing WT or WRN catalytic domain mutants were sequentially labeled with CldU and IdU nucleoside analogs followed by incubation with HU (4 mM) for 5 h ( Fig. 1e and Supplementary Fig. 1c). With this experimental set-up, relative shortening of the IdU label compared to the corresponding CldU label of an ongoing (CldU-IdU) replication tract (IdU/CldU tract length ratio) was used to determine the level of nascent DNA degradation at stalled forks. Complementation with WRN-WT partially rescued the hyper-degradation phenotype caused by BRCA2 depletion in U2OS/WRN −/− cells (Fig. 1e). Expression of either WRN-K577M or WRN-E84A catalytic mutant could partially rescue fork degradation phenotype in BRCA2-depleted cells, whereas the WRN-K577M/E84A double mutant did not (Fig. 1e). We obtained similar results in DLD1/WRN −/− cells in which WRN-WT, WRN-K577M, or WRN-E84A, but not WRN-K577M/E84A, partially reduced the extensive fork degradation caused by BRCA2 silencing (Fig. 1f and Supplementary Fig. 1d 35 , whereas DNA2 nuclease and WRN helicase are involved in limited reversed fork resection leading to fork restart in BRCA2-proficient cells 26 . However, restoration of WRN function in BRCA2-depleted cells did not further upregulate fork resection. Rather WRN partially rescued nascent DNA tract length ( Fig. 1e and f). Therefore, the rescue of fork restart defects in BRCA2-depleted cells upon WRN complementation (Fig. 1b) is consistent with the role of WRN helicase in the restoration of the active replication fork from the reversed fork state (Fig. 1d). As defects in replication fork restart may lead to reduced cell survival 15 , we examined the clonogenic potential of isogenic WT or WRN −/− U2OS cells depleted of BRCA2. Clonogenic survival assays demonstrated a significantly compromised survival of BRCA2-depleted WRN −/− cells (Supplementary Fig. 2b), suggesting that combined loss of WRN and BRCA2 is toxic to human cells. Of note, as judged by Western blot analysis of total cell lysate, BRCA2 level was reduced in WRN −/− U2OS cells ( Supplementary Fig. 2b) and DLD1 cells ( Supplementary Fig. 2c), suggesting that phenotypes associated with WRN deficiency may reflect in part a loss of BRCA2; however, this requires further study.
WRN helicase inhibition exacerbates replication fork instability in BRCA2-mutated cancer cells. Having demonstrated that WRN complementation partially rescued the hyperdegradation phenotype of BRCA2-depleted cells ( Fig. 1e and f), we tested the effect of WRN knockdown on nascent DNA degradation in the BRCA2-mutated ovarian cancer cell line PEO1 37 . Consistent with the well-documented role of BRCA2 in fork protection 3,4 , treatment of PEO1 cells with 4 mM HU for 3 h resulted in reduced fork stability (Median IdU/CldU = 0.6773) compared to WT BRCA2 revertant PEO4 cells 37 (Median IdU/ CldU = 0.8369) ( Fig. 2a and Supplementary Fig. 2d). However, nascent strand degradation was even further elevated (Median IdU/CldU = 0.6126) by WRN depletion in the BRCA2-mutated cells (Fig. 2a). In contrast, WRN knockdown in WT cells significantly rescued nascent DNA tract length (Median IdU/ CldU = 0.9176), consistent with the involvement of WRN in resection of stalled forks in BRCA2-proficient cells as described previously 26 .
We further examined fork stability in BRCA2-mutated cells exposed to a small molecule inhibitor of WRN helicase (WRNi) NSC617145 38 ( Fig. 2b and Supplementary Fig. 2e). Treatment with HU (4 mM for 3 h) alone resulted in a significant (p < 0.0001) level of fork degradation in PEO1 cells (Median IdU/CldU = 0.6833) as compared to isogenic PEO4 cells with restored BRCA2 function 37 (Median IdU/CldU = 0.9166) (Fig. 2b). However, the stability of the unprotected forks in BRCA2-mutated cells was severely compromised upon treatment with WRNi (4 µM), as we observed a significant decrease in median IdU/CldU ratio (0.3831) (Fig. 2b). RECQL1 overexpression did not affect fork instability induced by the WRNi in PEO1 cells ( Supplementary Fig. 3a), suggesting the in vivo abundance of RECQL1 does not alter the fork-destabilizing effects of NSC617145. Elevated nascent DNA degradation was also observed in BRCA2 −/− DLD1 cells pre-treated with WRNi ( Supplementary Fig. 3b). Notably, the exacerbated fork instability in BRCA2-mutated cells upon acute inhibition of WRN helicase by NSC617145 (Median IdU/CldU = 0.3831) was found to be markedly greater (Fig. 2b) than genetic depletion of WRN altogether (Median IdU/CldU = 0.6126) (Fig. 2a), suggesting that WRN helicase inhibition triggers unrestrained degradation of stalled forks in BRCA2-deficient human cancer cells. Because DNA2 nuclease operates in conjunction with WRN helicase to allow limited fork processing in BRCA2 WT cells 26 , we tested the stability of HU-stalled forks upon DNA2 inhibition in BRCA2mutated cells using a previously characterized small molecule inhibitor of DNA2 nuclease 39 (Supplementary Fig. 3c). Unlike WRN helicase inhibition, DNA2 inhibition did not exacerbate fork degradation in BRCA2-mutated PEO1 cells; moreover, the DNA2i did not synergize with the WRNi in its effect on fork stability ( Supplementary Fig. 3c). WRN helicase inhibition was previously reported to cause mitomycin C hypersensitivity for cells lacking FANCD2 38 , a protein important for fork stability and implicated in the Fanconi anemia pathway of interstrand crosslink repair 8 . Therefore, we tested whether WRN helicase inhibition triggers elevated fork instability in FANCD2-deficient cells. Unlike BRCA2-deficient cells, WRN helicase inhibition did not significantly exacerbate fork degradation in FANCD2 −/− PD20 cells ( Supplementary Fig. 3d), suggesting that the role of WRN helicase activity in fork maintenance is distinct when BRCA2 or FANCD2 is absent.
Loss of WRN or inhibition of its helicase activity compromises replication restart in BRCA2-mutated cancer cells. Next, we examined the efficiency of stalled fork recovery in WT or BRCA2mutated cells either depleted of WRN ( Fig. 2c and Supplementary  Fig. 4a) or exposed to the WRNi ( Fig. 2d and Supplementary  Fig. 4b). Consistent with previously published results 3,35 , BRCA2mutated PEO1 cells exhibited shorter restarted tract length (Median IdU/CldU = 1.2121) compared to the WT PEO4 cells (Median IdU/CldU = 1.4607) (Fig. 2c). In WT cells, knockdown of WRN reduced IdU/CldU tract length ratio (Median IdU/ CldU = 1.1083), suggesting a role of WRN in fork recovery as described previously 25,26 . Moreover, WRN depletion in BRCA2mutated cells showed an even more pronounced reduction in IdU/CldU ratio (Median IdU/CldU = 0.4984) compared to the isogenic WRN-depleted WT cells, indicating severely compromised fork recovery under the condition of WRN and BRCA2 codeficiency.
Compared to WT PEO4 cells (Median IdU/CldU = 1.4169), BRCA2-mutated PEO1 cells exposed to WRNi exhibited a nearly three-fold reduction in IdU/CldU ratio (Median IdU/CldU = 0.5539) (Fig. 2d), suggesting dramatically inefficient or delayed fork restart under conditions of BRCA2 deficiency and pharmacological inhibition of WRN helicase. To test the specificity of the WRNi, we performed fork restart assays after NSC617145 exposure using WRN −/− cells ( Supplementary  Fig. 4c) or RECQL1 −/− cells ( Supplementary Fig. 4d). While WRN deficiency markedly attenuated the defect in WRNiinduced fork restart ( Supplementary Fig. 4c), we observed a significant reduction in fork restart efficiency in RECQL1 −/− cells upon WRNi treatment ( Supplementary Fig. 4d), indicating specificity of the WRNi to modulate fork restart. Consistent with a published role of RECQL1 in fork restart 40 , we observed significantly reduced fork restart efficiency in RECQL1 −/− cells compared to isogenic RECQL1 −/− cells corrected by exogenous expression of RECQL1-WT ( Supplementary Fig. 4d). Exposure of RECQL1 −/− cells to NSC617145 exasperated the fork restart defect ( Supplementary Fig. 4d), suggesting a specific effect of the WRNi that is independent of RECQL1 status.
In the absence of functional BRCA2, stalled forks are progressively degraded by MRE11 nuclease 4 . Therefore, we sought to determine whether the observed nascent DNA degradation upon NSC617145 treatment is mediated by  MRE11. Fork stability in BRCA2-mutated PEO1 cells exposed to NSC617145 was substantially rescued upon treatment with the MRE11 nuclease inhibitor Mirin 41 , with a statistically significant (p < 0.0001) increase in median IdU/CldU ratio (Fig. 3b). IdU/ CldU ratio was not significantly altered in WT PEO4 cells upon MRE11 nuclease inhibition (Fig. 3b). Unlike the case for MRE11, pre-exposure to a small molecule inhibitor of DNA2 nuclease 39 did not rescue WRNi-induced fork degradation in PEO1 cells (Fig. 3b), suggesting that MRE11 plays a more dominant role in nucleolytic degradation of stalled forks in WRNi-treated BRCA2mutated cells. Consistent with this notion, MRE11 was preferentially enriched in the chromatin fraction obtained from PEO1 cells exposed to WRNi compared to PEO4 cells (Fig. 3c). In BRCA2-deficient cells, degradation of HU-stalled fork initiated by MRE11 is further extended by EXO1 nuclease 35 . However, we did not observe chromatin enrichment of EXO1 following WRNi treatment (Fig. 3c). MRE11 and EXO1 may independently act on stalled forks in BRCA-deficient cells, as combined loss of the two nucleases has been reported to further rescue fork stability in these cells 35 . Our results suggest that WRNi causes rapid accumulation of replication fork intermediates suitable for MRE11 engagement and subsequent degradation of nascent DNA if not protected by the fork-stabilizing protein BRCA2.
In response to replication stress, stalled forks frequently undergo reversal to form four-way chicken foot-like structures 42 . SNF2-like DNA translocases (SMARCAL1, ZRANB3, and HLTF) have been implicated in driving fork reversal under conditions of replication stress in human cells 20,42,43 . As reversed forks are vulnerable to nuclease-mediated degradation upon BRCA loss, attenuation of fork reversal has been shown to restore fork stability in BRCA-deficient cells 20,33 . We sought to determine whether observed fork degradation in BRCA2-mutated cells occurs downstream of SNF2 translocase-dependent fork reversal upon WRNi treatment. Therefore, we determined relative fork stability upon siRNA-mediated knockdown of SMARCAL1, ZRANB3, and HLTF in BRCA2-mutated PEO1 cells exposed to 4 µM NSC617145. PEO1 cells depleted of SMARCAL1, ZRANB3 or HLTF exhibited a significant (p < 0.0001) rescue in median IdU/CldU tract length ratio, indicating restoration of fork stability (Fig. 3d). These results suggest that WRNi-induced nascent DNA degradation by MRE11 in BRCA2-mutated cancer cells is preceded by DNA translocase-mediated fork remodeling.
WRN helicase inhibition causes fork stalling and asymmetry in BRCA2-deficient cells. We next assessed the effect of WRN helicase inhibition on replication fork progression. Following 15 min of CldU pulse labeling to mark actively replicating genomic regions, U2OS cells were subsequently labeled with IdU for another 60 min ( Supplementary Fig. 5a). WRNi (2 µM) was added concomitantly with IdU labeling. Quantification of the data obtained from fiber spread analysis revealed a significant (p < 0.0001) reduction in IdU/CldU ratio following NSC617145 exposure, suggesting a delay in fork progression. Thus, NSC617145 alone impedes replication fork progression in the absence of genotoxic stress. However, the compound did not affect fork progression in U2OS WRN −/− cells ( Supplementary  Fig. 5a), suggesting that the replication defects triggered by NSC617145 are WRN-dependent.
To assess the effect of WRN helicase inhibition on fork progression as a function of BRCA2 status, we examined fork progression rate in WT and BRCA2 −/− DLD1 cells exposed to NSC617145. IdU tract lengths were significantly reduced (p < 0.0001) in both genotypic backgrounds by NSC617145 treatment (Fig. 4a and Supplementary Fig. 5b). Replication fork asymmetry was quantified by measuring relative lengths of sister forks (SFs) emanating from individual bi-directional origins (Fig. 4b). Compared to untreated cells, WRNi-treated cells exhibited increased SF asymmetry demonstrated by a significant reduction in SF ratios for WT cells (p = 0.0046) and BRCA2 −/− cells (p < 0.0001), respectively (Fig. 4b), suggesting frequent replication stalling or collapse. Importantly, fork asymmetry was further increased (p = 0.0035) in BRCA2 −/− cells as compared to WT cells exposed to WRNi (Fig. 4b). Compared to WT cells, we observed a relatively greater percentage of asymmetric forks (SF ratio < 0.6) in BRCA2 −/− cells (Fig. 4c). To substantiate these results, we monitored replication fork dynamics in isogenic WT PEO4 and BRCA2-mutated PEO1 cells upon WRNi treatment. In addition to a significant reduction in fork progression rate ( Supplementary Fig. 5c), we observed increased fork asymmetry in WRNi-treated BRCA2mutated PEO1 cells ( Supplementary Fig. 5d, e), corroborating the results obtained with DLD1 cells.
WRNi NSC617145 directly binds WRN and traps it on chromatin. To further characterize the biological effect of NSC617145, we tested for a direct interaction of the 14 C-radiolabeled compound ( Supplementary Fig. 6a) to WRN in vitro and in human cells. 14 C-NSC617145 was incubated with purified recombinant WRN ( Supplementary Fig. 6b), and the mixture was applied to a nitrocellulose filter. Detection of radiolabel bound to the nitrocellulose demonstrated that there was approximately a 9.5-fold greater 14 C signal in binding mixtures containing WRN compared to no WRN (Fig. 5a). Binding mixtures containing the sequence-related RECQL1 helicase (Fig. 5a) or the Superfamily 2 helicase FANCJ ( Supplementary Fig. 6c) displayed a low 14 C signal comparable to that of the binding mixture lacking WRN, suggesting specific binding of 14 C-NSC617145 to WRN in vitro. To assess the binding of NSC617145 to WRN in vivo, U2OS WRN −/− cells transiently expressing FLAG-tagged WRN were incubated with 14 C-NSC617145 followed by immunoprecipitation using anti-FLAG antibody (Fig. 5b). A 3-fold greater 14 C signal in immunoprecipitates prepared from WRN-expressing cells was detected compared to those from cells expressing only the FLAG epitope ( Fig. 5b), indicating 14 C-NSC617145 binding to WRN expressed in human cells.
Because targeted WRN helicase inhibition triggered greater fork degradation than that induced by WRN depletion ( Fig. 2a and b), we hypothesized that cellular exposure to NSC617145 generates static WRN-DNA complexes that interfere with DNA replication. To address this, we assessed the amount of chromatin-bound WRN in cells exposed to NSC617145 and observed its dose-dependent enrichment (Fig. 5c). Compared to WT cells, NSC617145-induced chromatin enrichment of WRN was elevated in BRCA2-mutated cells, suggesting that WRN accumulates at forks when BRCA2 is absent (Fig. 5d). Furthermore, even in the absence of NSC617145, WRN showed enrichment in the chromatin fraction of BRCA2-mutated cells (Fig. 5d). To test if this was related to fork stalling or general DNA damage processing, we examined WRN immunofluorescence intensity in PEO1 and PEO4 replicating cells counterstained with the S-phase-specific marker PCNA ( Supplementary  Fig. 6d). As reported previously 44 , we observed that WRN primarily localized to nucleoli in the unperturbed condition. However, compared to WT cells, PCNA-positive BRCA2-mutated cells showed increased WRN intensity in nucleoplasmic foci, suggesting WRN accumulation at forks in the absence of BRCA2. Of note, consistent with published data that actively proliferating cancer cell lines display an elevated level of WRN 45 , we detected substantially enhanced WRN immunofluorescence in PCNA-positive cells as compared to PCNA-negative cells ( Supplementary Fig. 6d). NSC617145-induced chromatin binding of WRN was further enhanced in the presence of HU (Fig. 5e), suggesting greater accumulation of WRN-DNA complexes when replication forks stall due to nucleotide depletion in vivo. To assess whether WRN trapping by NSC617145 is reversible, we examined the relative amount of chromatin-bound WRN upon drug removal (Fig. 5f). Cellular exposure to NSC617145 led to a substantial increase in chromatin-bound WRN which remained even 2.5 h after drug removal, suggesting persistent trapping of WRN ( Fig. 5f). Importantly, neither DNA topoisomerase I (  Fig. 7c). To substantiate these results, we examined the focal accumulation of 53BP1, a marker of DSBs, in cells exposed to WRNi (Supplementary Fig. 7d). Percent 53BP1 positive nuclei were found to be approximately 3-fold greater (p = 0.0045) in BRCA2 −/− cells treated with WRNi (48.4%) compared to WT cells (16.6%) exposed to the WRNi ( Supplementary Fig. 7d), indicating enriched DNA damage accumulation in BRCA2-deficient cells upon WRN helicase inhibition. To determine if DNA damage induced by exposure to the WRNi occurs preferentially in replicating cells, we co-stained for γH2AX and PCNA (Fig. 6a). γH2AX was predominantly observed in PCNA-positive nuclei, suggesting replication-associated DNA damage (Fig. 6a). PEO1 cells exposed to WRNi exhibited γH2AX foci in >80% of the PCNA-positive nuclei (Fig. 6a), suggesting increased DNA damage in BRCA2mutated cells following WRN helicase inhibition.
Previously, it was shown that WRN loss of function results in replication fork collapse and subsequent formation of MUS81mediated DSBs at collapsed forks 23 . Moreover, MUS81dependent DSBs are generated following hyper-degradation of stalled forks in BRCA2-deficient cells 35 . Therefore, we assessed DSB formation upon WRN helicase inhibition by neutral Comet assay (Fig. 6b). Compared to WT PEO4 cells, BRCA2-mutated PEO1 cells displayed elevated DSBs upon WRNi treatment (Fig. 6b). However, increased DSB formation in BRCA2-mutated cells was found to be significantly abrogated upon MUS81 depletion ( Fig. 6b and Supplementary Fig. 7e). These observations suggest that WRN helicase inhibition results in MUS81dependent DSBs as a consequence of stalled fork hyperdegradation in BRCA2-mutated cells.
Two major DNA repair pathways, HR and NHEJ, compete for the processing of replication-associated DNA breaks. Cells deficient of major HR factors such as BRCA1/2 shift DNA repair towards the error-prone NHEJ pathway [46][47][48] . Accordingly, we anticipated that DNA damage accumulated in BRCA2-mutated cells upon WRN helicase inhibition might trigger aberrant activation of NHEJ. To address this hypothesis, we examined for NHEJ activation by DNA-PKcsThr2609 foci formation 49,50 in PEO1 and PEO4 cells treated with NSC617145. WT cells showed a small (~1.7-fold) but statistically significant (p = 0.0112) increase in the percentage of cells having DNA-PKcsThr2609 positive nuclei upon treatment with WRNi ( Supplementary  Fig. 8a). In contrast, we observed a >4-fold increase (p = 0.0040) in the percentage of DNA-PKcsThr2609 positive nuclei in BRCA2-mutated PEO1 cells upon WRN helicase inhibition, suggesting increased activation of the NHEJ pathway ( Supplementary Fig. 8a). To determine if the aberrant NHEJ activation occurs in replicating cells, we monitored DNA-PKcs phosphorylation at serine 2056 (DNA-PKcs pS2056) 51 in cells costained for PCNA (Fig. 7a). We observed DNA-PKcs pS2056 formation primarily in PCNA-positive nuclei, suggesting WRNi-induced NHEJ activation in cells actively performing DNA synthesis. Interestingly, DNA-PKcs pS2056 foci were reduced in BRCA2-mutated cells exposed to Mirin (Fig. 7a), suggesting induction of NHEJ is related to MRE11-mediated fork degradation. In BRCA2-deficient cells, hyper-degraded forks are rescued by break-induced replication (BIR) repair mechanism following MUS81-mediated cleavage 35 . Rad52, a key mediator of HR has been implicated in the repair of collapsed replication forks 52 and recovery of stalled forks 16 Fig. 8b). Either WRN helicase inhibition or RAD52 depletion resulted in significantly elevated fork stalling. RAD52 knockdown in WRNi-treated PEO1 cells further increased the percentage of stalled forks, suggesting a role of RAD52-dependent BIR to aid in the recovery of hyper-degraded forks. However, given the observed increase in NHEJ (Supplementary Fig. 8a and Fig. 7a), our results collectively suggest that in BRCA2-mutated cells, a subset of collapsed forks downstream of WRNi-induced hyper-degradation also engage NHEJ for repair.
Elevated NHEJ leads to chromosomal aberrations under conditions of HR deficiency 46,48,53 . Moreover, experimental evidence suggests that uncontrolled degradation and subsequently delayed restart of stressed replication forks results in chromosomal instability in HR-deficient cells 4,8,15 . We, therefore tested whether inhibition of WRN helicase triggers genomic instability in BRCA2-mutated cells (Fig. 7b). Analysis of metaphase chromosome spreads prepared from WRNi-treated WT PEO4 cells did not reveal any significant change (p = 0.3853) in average chromosomal aberrations (Fig. 7b). However, structural aberrations including chromatid breaks, gaps or endto-end fusions between heterologous chromosomes were elevated by~1.7-fold (p < 0.0001) in BRCA2-mutated PEO1 cells exposed to the WRNi (Fig. 7b). Inhibition of MRE11 nuclease alleviated gross chromosomal aberrations in WRNi-treated BRCA2mutated PEO1 cells (Fig. 7c), suggesting that WRNi-induced genomic instability was primarily attributed to MRE11-mediated fork degradation. We also observed a significant reduction in chromosomal abnormalities upon DNA-PK inhibition (Fig. 7c), suggesting NHEJ induction contributes to the heightened genomic instability in WRNi-treated BRCA2-mutated cells. Collectively, these observations suggest that blocking WRN helicase function pharmacologically elicits DNA damage and error-prone NHEJ, leading to genomic instability in cells lacking functional BRCA2.
WRN helicase inhibition kills BRCA2-defective cancer cells. Uncontrolled fork degradation and defective restart of stressed replication forks result in elevated genome instability leading to cell death 15,18,26,33,35 . Therefore, we tested the effects of WRN helicase inhibition on the survival of cancer cells lacking functional BRCA2. BRCA2 −/− DLD1 cells displayed increased sensitivity to the WRNi as determined by colony survival and cell viability assays (Fig. 8a, Supplementary Fig. 9a-b). Selective killing of BRCA2 −/− cells by PARP1 inhibitor (PARPi) olaparib treatment confirmed the BRCAness phenotype of these cell lines. We also observed a marked decrease in colony survival capacity of BRCA2-depleted U2OS cells exposed to increasing concentrations of NSC617145 ( Supplementary Fig. 9c). As a positive control, BRCA2-depleted U2OS cells were found to be selectively killed by olaparib, consistent with the well-established synthetic lethal interaction of PARP1 and BRCA2 54 . BRCA2-selective hypersensitivity of cancer cells towards the WRNi was further recapitulated in the breast cancer cell line MDA-MB-231 ( Supplementary  Fig. 10a) and ovarian cancer cell line SKOV-3 ( Supplementary  Fig. 10b), suggesting generality of the phenotype. At a 0.5 µM dose of NSC617145, BRCA2-depleted MDA-MB-231 and SKOV-3 cells exhibited >4-fold and >3-fold decrease in colony survival, respectively ( Supplementary Fig. 10a, b). Although acute inhibition of WRN helicase by NSC61745 was found to be synergistic with BRCA2 loss, we did not observe any significant effect on colony-forming ability of DLD1 BRCA2 −/− cells upon siRNAmediated knockdown of WRN ( Supplementary Fig. 10c). However, no WRN CRISPR KO clones of BRCA2 −/− DLD1 or BRCA2-mutated PEO1 cells were recovered, suggesting a synthetic lethal interaction of WRN and BRCA2. Presumably, DLD1 BRCA2 −/− cells depleted of WRN by RNA interference remain viable due to residual WRN.
We next assessed if clinically relevant BRCA2 pathogenic mutations rendered cancerous cells susceptible to WRN helicase inhibition. For this, we compared the relative viability of BRCA2mutated ovarian cancer cell line PEO1 with that of BRCA2 WT revertant PEO4 cells upon WRN helicase inhibition. Compared to WT PEO4 cells, we observed drastically reduced survival of BRCA2-mutated PEO1 cells exposed to WRNi (Fig. 8b and Supplementary Fig. 11a, b). At 0.75 µM WRNi, hardly any PEO1 colonies were detected whereas~70% of PEO4 colonies remained viable (Fig. 8b). Notably, BRCA2-mutated PEO1 cells were not as sensitive to olaparib as BRCA2 −/− DLD1 (Fig. 8a and Supplementary Fig. 9a, b), as evident from colony survival (Fig. 8b) and cell viability assay data (Supplementary Fig. 11b). These observations are consistent with previous findings that some BRCA2-mutated ovarian cancer cell lines such as PEO1 are partially resistant to PARPi 55 .
RAD51 functions downstream of BRCA2 in the homologydirected repair of DNA DSBs as well as in the fork protection pathway 56 . We, therefore anticipated that like BRCA2, RAD51 depletion could have similar effects on cellular fitness upon WRN helicase inhibition. Indeed, knockdown of RAD51 sensitized DLD1 cells to WRNi treatment as revealed by results obtained from colony survival (Supplementary Fig. 12a) and cell viability ( Supplementary Fig. 12b) assays. We were curious if NSC617145induced cell killing was dependent on WRN. To address this, we examined the survival of WRN-depleted cells after NSC61745 exposure. Colony survival in response to NSC617145 was significantly rescued in WRN-depleted HeLa and DLD1 cells ( Supplementary Fig. 13a, b), suggesting WRN-dependent cytotoxicity of the inhibitor. Substantially improved survival of WRN −/− U2OS cells exposed to NSC617145 ( Supplementary  Fig. 13c, d) further attested to the specificity of the compound. However, isogenic RECQL1-deficient and RECQL1-corrected cells exhibited similar sensitivity to the WRNi ( Supplementary   Fig. 6 WRNi triggers MUS81-dependent double-strand break formation in BRCA2-mutated cancer cells. a Immunofluorescence staining of γH2AX following treatment of PEO4 and PEO1 cells with WRNi (5 µM) for 6 h. Cells were co-stained with PCNA to identify replicating cells. Nuclei were stained with DAPI. DAPI, PCNA, and γH2AX are shown in blue, red, and green, respectively. Scale bars are shown. Bar graph represents quantification of % PCNApositive cells with ≥10 γH2AX foci under untreated (UT) and WRNi-treated conditions. Data represent mean ± SD of three independent experiments; twotailed unpaired t-test; p values are indicated. b Analysis of DSB formation using neutral Comet assay. Control or MUS81-depleted PEO4 and PEO1 cells were subjected to neutral comet assay following treatment with WRNi (5 µM) for 6 h. Representative images of comets obtained from individual experimental conditions are shown. Scale bars are shown. Scatter dot plot showing tail moment in individual samples. Data represent mean ± SD of n ≥ 120 comets analyzed under each condition over n = 2 independent experiments. Mean tail moments (horizontal red bars) ± SD are shown. ≥120 comets were analyzed for each condition using the OpenComet plugin in ImageJ. Two-tailed Mann-Whitney test; p values (p < 0.0001, p < 0.0001, p = 0.0030, p < 0.0001, p < 0.0001) are indicated. Source data are provided as a Source Data file. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26811-w ARTICLE Fig. 13e), suggesting a RECQL1-independent effect. Hypersensitivity to WRNi upon BRCA2 depletion was also found to be attenuated in WRN −/− cells (Supplementary Fig. 13f). These observations are consistent with our previously published results demonstrating WRN dependency of NSC617145 for its effect on cell survival 38 .
Effect of NSC617145 on BRCA2-deficient tumor growth and DNA damage in vivo. Given the observed hypersensitivity of a broad range of BRCA2-defective cancer cells to NSC617145, we tested the in vivo response of BRCA2-deficient tumors to the WRNi using xenograft tumor models in athymic nude mice (Fig. 8c). Mice carrying DLD1 WT or BRCA2 −/− tumors were treated with either NSC617145 or drug vehicle alone and relative tumor growth was monitored over time. Compared to control vehicle treatment, treatment with NSC617145 (25 mg/kg body weight) resulted in a marked decrease (p = 0.05) in BRCA2 −/− tumor growth (Fig. 8c). Importantly, NSC617145 treatment exhibited >3-fold higher growth inhibitory effect on BRCA2 −/− tumor (TGI = 34.70%) as compared to WT tumors (TGI = 10.28%). Thus, the in vivo data are consistent with results from cell-based assays demonstrating preferential killing of BRCA2deficient cancer cells by WRN helicase inhibition. Furthermore, NSC617145 treatment led to elevated DNA damage accumulation in BRCA2 −/− tumor cells as demonstrated by increased immunohistochemical staining for γH2AX (Fig. 8d) and 53BP1 (Fig. 8e) in tissue sections of BRCA2 −/− tumors compared to WT tumors.
Pharmacological WRN helicase inhibition sensitizes BRCA2mutated cancer cells to olaparib. BRCA1/2-mutated tumors often exhibit de novo or acquired PARPi resistance 57 . Particularly, ovarian cancers with recurrent BRCA-mutations show incomplete responses to clinical PARP inhibitors 58 . Consistent with the previously published results 55 , we also observed that BRCA2-mutated ovarian cancer cells PEO1 showed reduced sensitivity to the PARPi olaparib ( Fig. 8b and Supplementary  Fig. 11b). We anticipated that treatment with sublethal doses of WRNi could potentiate olaparib cytotoxicity in BRCA2-mutated cancer cells. Olaparib treatment was not found to be very effective in killing PEO1 cells as~60% colonies were still viable even at the highest concentration of drug tested (0.5 µM) ( Fig. 9a and Supplementary Fig. 14a). In contrast, co-treatment with 0.25 µM NSC617145 led to~86% reduction in PEO1 colony survival at a 10-fold lower olaparib concentration (0.05 µM) (Fig. 9a). Although the apparently synergistic cytotoxic effect of the WRNi and olaparib was also observed in WT PEO4 cells, it was much less pronounced as compared to PEO1 cells and~40% of the WT PEO4 colonies remained viable at even the highest dose of olaparib (0.5 µM) (Fig. 9a). We further assessed drug-drug interaction effects by analyzing the coefficient of drug interaction (CDI) and observed synergistic cell killing (CDI < 1) by the WRNi-olaparib combination in both PEO1 and PEO4 cells (Fig. 9b). However, at the lowest concentration of olaparib (0.05 µM), there was a much stronger synergism (CDI < 0.7) in PEO1 cells compared to PEO4 cells (Fig. 9b). Bliss synergy analysis further revealed synergistic cell killing by WRNi and olaparib as indicated by positive Excess over Bliss (EOB > 0) synergy scores (Fig. 9c). BRCA2 −/− colorectal cancer cells DLD1 cotreated with olaparib and NSC617145 also displayed a greater reduction in proliferation than that due to an additive effect of the two drugs acting independently (Supplementary Fig. 14b). However, NSC617145 treatment did not sensitize WT DLD1 cells to olaparib, suggesting BRCA2-selective synergistic cytotoxicity of WRNi and olaparib. These results demonstrate that pharmacological inhibition of WRN helicase potentiates PARPi cytotoxicity in BRCA2-defective cancer cells.

Discussion
Replication stalling, estimated to occur at 3,000 of 12,000 active forks per nucleus in human cells 59,60 , is a serious threat to genomic stability if not dealt with in a timely fashion 61 . In response to mild genotoxic stress, it is estimated that 15-30% of stalled forks are remodeled to reversed fork structures to allow template repair and prevent DSBs 60,62 . Stabilization of stalled replication forks is a complex process involving multiple fork remodelers and protectors. Although WRN is demonstrably active in stalled fork processing, its mechanism of action at stalled and deprotected forks has remained enigmatic. Here, we identified a crucial role of WRN helicase to rescue stalled replication forks in cancer cells under genetic conditions of BRCA2 deficiency. WRN is required to enact a timely restart following replication stalling and limit excessive degradation of stalled forks in BRCA2-deficient cancer cells. WRN ATPase/helicase efficiently restored a model replication fork from reversed fork state in vitro and rescued fork restart defects caused by BRCA2 loss. DNA fiber experiments further demonstrated that loss or inhibition of WRN helicase activity resulted in severely compromised fork restart in BRCA2-mutated cells, suggesting WRN helicase, but not exonuclease, is required for timely fork restart in BRCA2-deficient cells. In contrast, both helicase and exonuclease functions of WRN partially rescued nascent strand degradation under the BRCA2depleted condition. Unlike BRCA2-proficient cells, WRN depletion in BRCA2-deficient cells did not rescue nascent DNA tract length, but instead further enhanced nascent strand degradation upon replication stalling, suggesting WRN limits hyperdegradation of unprotected forks. In agreement with these results, we observed reduced nucleolytic cleavage of a synthetic reversed fork substrate by MRE11 nuclease in the presence of WRN helicase in vitro. Thus, our results support a model in which stalled fork restoration by WRN helicase limits hyperdegradation of reversed forks by reducing substrate availability for MRE11 degradation in BRCA2-deficient cells (Fig. 9d). Rad52-dependent BIR promotes stalled fork recovery under conditions of replication stress 16 . Moreover, in BRCA2-deficient cells, resected reversed forks are rescued by a POLD3-dependent BIR mechanism following MUS81-mediated fork cleavage 35 .
Thus, although not formally tested, it is plausible that WRN and RAD52-MUS81-POLD3 are two axes used by cells to restart forks in a BRCA2-deficient background.
We determined that WRN exonuclease partially rescued the hyper-degradation of unprotected forks in BRCA2-depleted cells. Previously, it was reported that in BRCA2-proficient cells WRN Fig. 7 Elevated nonhomologous end-joining and genomic instability in BRCA2-mutated cancer cells exposed to WRNi. a Representative immunofluorescence staining of DNA-PKcs p-S2056 foci in PEO4 and PEO1 cells following treatment with WRNi (5 µM) for 16 h. Cells were optionally treated with Mirin (25 µM). Cells were co-stained with PCNA to mark replicating cells. Bar graph represents quantification of % PCNA positive cells with ≥5 DNA-PKcs p-S2056 foci in individual experimental conditions. Data represent mean ± SD of three independent experiments; two-tailed unpaired t-test; p values are indicated. b Representative images of metaphase chromosome spread prepared from PEO1 and PEO4 cells treated or non-treated with WRNi (2 µM) for 48 h. Scatter plot showing relative chromosomal aberrations scored from the cells as described above. Data represent mean ± SEM of n ≥ 40 metaphase spreads analyzed under each condition over n = 3 independent experiments; Two-tailed Mann-Whitney test; p values (p = 0.3853, p < 0.0001, p < 0.0001) are indicated. c Analysis of chromosomal aberrations in BRCA2-mutated PEO1 cells following treatment with WRNi (5 µM) for 16 h. Cells were optionally treated with Mirin (25 µM) or DNA-PK inhibitor (DNA-PKi) NU7441 (10 µM). Quantification of chromosomal aberration analysis is shown. Data represent mean ± SEM of n ≥ 50 metaphase spreads analyzed under each condition over n = 3 independent experiments; Two-tailed Mann-Whitney test; p values (p < 0.0001, p < 0.000, p = 0.0023) are indicated. Structural aberrations including chromatid breaks/gaps, chromosome end-to-end fusions, dicentric and radial chromosomes are indicated by red arrows in b and c. Scale bars are shown in all representative images. In a, DAPI, PCNA, and DNA-PKcs p-S2056 foci are shown in blue, red, and green, respectively. Source data are provided as a Source Data file.
The current study provides molecular insight into the role of WRN at deprotected forks. BRCA2 ensures fork protection by loading Rad51 onto exposed single-stranded DNA at the site of the stalled fork. However, our findings suggest that WRN also plays a role in controlling nucleolytic processing when stalled fork structures (e.g., reversed fork) are deprotected. Furthermore, experimental results in our study demonstrate that inhibition of the structure-specific nuclease MRE11 helps to curtail excessive nucleolytic degradation of nascently synthesized DNA caused by WRN helicase inhibition in BRCA2-mutated cells. Controlling nucleolytic degradation at the stalled fork is paramount to fork restart so that DNA replication can be completed to its entirety in a timely manner. Collectively, our results suggest an important function of WRN catalytic functions (helicase and exonuclease) to prevail when there is a genetic deficiency of BRCA2.
Targeted inhibition of WRN helicase triggers heightened fork instability and reduces survival of BRCA2-mutated human cancer cells. Furthermore, we found that pharmacological inhibition of WRN helicase destabilizes replication forks to a greater extent than WRN depletion in vivo, suggesting an additional burden imposed by NSC617145-induced WRN helicase inhibition compared to loss of WRN altogether. Using biochemical and cell-based assays, we found that NSC617145 binds WRN and exerts its biological effects in a WRN-dependent manner, confirming its target specificity. Intriguingly, the selective and persistent sequestration of WRN to chromatin in cells exposed to WRNi suggests that WRN becomes trapped on genomic DNA. Therefore, the increased fork instability induced by NSC617145 is likely the combined effects of WRN helicase inhibition and static WRN-DNA complexes that interfere with fork progression. There are important classes of anticancer drugs including Topoisomerase inhibitors and PARP inhibitors that confer cytotoxicity by both catalytic inhibition and trapping of their target enzymes 54 . However, this is an instance of a small molecule inhibitor targeted against a human DNA helicase that operates by a trapping mechanism to interfere with replication fork dynamics. It is plausible that other protein partners are also sequestered with the trapped WRN-DNA complex; however, there was not a significant enrichment of DNA2 or Topoisomerase I, consistent with the specificity of the WRNi. Replication fork slowing induced by the WRNi was alleviated upon WRN depletion, suggesting static WRN-DNA complexes impeded fork progression as well.
We further demonstrated that WRNi-induced stalled forks undergo MRE11-mediated nucleolytic degradation in BRCA2mutated cancer cells. Nascent strand degradation occurs downstream of reversed fork formation catalyzed by the SNF2 fork remodeling enzymes SMARCAL1, ZRANB3, and HLTF 20,33 . Therefore, restoration of fork stability upon knockdown of SNF2 translocases in WRNi-treated BRCA2-mutated cells suggests that pharmacological inhibition of WRN helicase leads to nascent DNA degradation preceded by translocase-dependent fork remodeling (Fig. 9e). Consistent with this idea, WRN helicase inhibition resulted in a strong chromatin enrichment of MRE11 that contributes to fork degradation. WRN loss results in fork collapse and MUS81-dependent DSBs following replication stalling 23 . Also, in BRCA2-deficient cells, partially resected forks are cleaved by MUS81 endonuclease leading to transient accumulation of DSBs 35 . In agreement with this, we found that WRNi treatment contributes to DSBs in BRCA2-mutated cells in a MUS81-dependent manner, suggesting WRNi-induced stalled forks undergo MUS81 cleavage following nascent DNA degradation. Our results demonstrate that fork stalling represents a rapid and global response to WRN helicase inhibition in human cells and that the consequential stalled forks require BRCA2 for stabilization. Uncontrolled fork degradation destabilizes stalled forks leading to defective fork restart which contributes to irreversible fork collapse, genomic instability, and cell death 4,18,20,63 . Thus, our results suggest that in BRCA2-deficient cells, forks stalled by static WRN-DNA complexes collapse leading to elevated chromosomal instability and cell death (Fig. 9e).
Fork stability has been linked to acquired chemoresistance in BRCA2-mutated human tumors 3,19 . We demonstrate that WRN helicase inhibition results in preferential cell killing of multiple BRCA2-deficient cancer cell lines with distinct tissue origins, suggesting an avenue of exploration for therapy. Furthermore, mice administered the WRNi displayed elevated DNA damage and marked inhibition of BRCA2-deficient tumor growth in the xenograft model, in agreement with the hypersensitivity of BRCA2-deficient cancer cells to the WRNi. Drug-driven synthetic lethality, fueled by the seminal discovery of PARPi-induced selective cytotoxicity in BRCA1-or BRCA2-deficient cancer cells 64,65 , represents a paradigm for developing therapeutics by exploiting tumor-specific genetic deficiencies 18,54,[64][65][66] . Thus, our results provide a rationale for targeting BRCA2 deficiencies in human cancers by pharmacological inactivation of DNA repair/ replication stress response helicases implicated in fork stability maintenance. Moreover, WRN helicase inhibition markedly potentiated olaparib cytotoxicity in BRCA2-mutated ovarian cancer cells. This finding has a potentially important therapeutic implication as acquired resistance to PARP inhibitors is one of the major challenges in treating BRCA2-mutated ovarian cancers.
Based on our experimental findings, pharmacological inhibition of WRN helicase is a potential therapeutic strategy for tumors with BRCA2 loss-of-function mutations. Recent works suggest that WRN is a druggable target in microsatellite-unstable human cancers [67][68][69][70][71][72] . Whether WRN's role at microsatellites directly relates to its function to stabilize replication forks under genetic conditions of fork deprotection remains to be seen. WRN and perhaps other replication stress response helicases represent potential targets in therapeutic strategies that rely on druginduced or genetic synthetic lethality to combat cancer 72 . Fig. 8 WRNi kills BRCA2-deficient cancer cells and retards BRCA2 −/− xenograft tumor growth in mice. a Colony survival of WT and BRCA2 −/− DLD1 cells exposed to WRNi (left panel) or olaparib (right panel). WT and BRCA2 −/− DLD1 cells were plated on a six-well plate at 1,500/well and exposed to the indicated doses of WRNi or olaparib 4 h after cell seeding. b Line graphs showing relative colony survival of PEO1 and PEO4 cells exposed to increasing doses of WRNi (left panel) or olaparib (right panel). PEO1 and PEO4 cells were plated on a six-well plate at 1,000 cells/well and treated with the indicated doses of WRNi or olaparib 4 h after cell seeding. In a and b, cells were grown in presence of WRNi or olaparib for 12 days before colony staining. Data represent mean ± SD of three independent experiments. c Line graph showing a relative growth rate of BRCA2 −/− (upper panel) and WT (lower panel) xenograft tumors in athymic nude mice treated with either drug vehicle or NSC61745. After the tumors reached the palpable size, mice were injected intraperitoneally with either drug vehicle or NSC617145 (25 mg/kg body weight) every alternate day for a total of 14 days. Mean tumor volume ± SD (n = 6) at each time point is shown. Tumor growth inhibition (TGI) is indicated. p-values were determined using two-tailed unpaired Welch's t-test. Recombinant FLAG-WRN proteins were expressed in insect cells. FLAG-WRN was subcloned from the pFLAGWRN plasmid (WRN CDK1 24 ) into the pFastBacHTB construct (Thermo Fisher Scientific) via 5′ NotI and 3′ KpnI sites, making the pFastBacHTB-FLAG-WRN plasmid. Site-directed mutagenesis was carried out by Genewiz, Inc. to generate pFastBacHTB-FLAG-WRNE84A, pFastBacHTB-FLAG-WRNK577M and pFastBacHTB-FLAG-WRNE84A/K577M plasmids. All plasmids were sequenced to confirm their identity. Miniprep plasmids were transformed into DH10Bac to generate baculovirus DNA which was transfected into 1 × 10 6 Sf9 insect cells using Cellfectin II transfection reagent (Thermo Fisher Scientific) according to the manufacturer's protocol. Five days post-transfection, the media containing the baculovirus was isolated and used to reinfect Sf9 insect cells for 5 additional days generating a higher titer virus. Media containing the baculovirus was stored at 4°C. High titer baculovirus for pFastBacHTB-FLAG-WRN, pFastBacHTB-FLAG-WRNE84A, pFastBacHTB-FLAG-WRNK577M, and pFastBacHTB-FLAG-WRNE84A/K577M were used to infect Hi5 insect cells (Thermo Fisher Scientific) at an approximate MOI of 10. Hi5 cells were harvested 48 h post-infection by washing in 1× PBS with Complete Ultra Protease Inhibitor Cocktail (Roche) and centrifugation at 1500 RPM for 5 min and pellets placed in −80°C until lysed. Cell pellets containing approximately 1 × 10 8 cells were resuspended in 10 ml of Lysis Buffer (50 mM Tris pH 7.4, 150 mM NaCl, 0.4% NP40, 10% glycerol, 5 mM BME, and Complete Ultra Protease Inhibitors (Roche)) and rotated at 4°C for 45 min. Lysates were centrifuged at 20,000 RPM for 10 min. Supernatants were filtered through a 0.45 μm PVDF filter. Each clarified lysate was passed twice through a Ni2+-charged 1 ml HiTrap Chelating HP column (Cytiva) equilibrated in Buffer TN (50 mM Tris pH 7.4, 150 mM NaCl, 10% glycerol, 5 mM BME, protease inhibitors) with 10 mM imidazole. Three washes with TN buffer containing 10, 20, and 40 mM imidazole were performed followed by elution with TN buffer containing 400 mM imidazole. Eluted protein was pooled and digested with TEV protease for 16 h at 4°C to remove the 6×His tag. The protein was dialyzed into NETN-500 Buffer (50 mM Tris pH 7.4, 500 mM NaCl, 0.5% NP40, 1 mM EDTA) using an Amicon Ultra 100 kD cutoff centrifugal filter (EMD Millipore). The retained sample was applied to 250 μl of packed M2 anti-FLAG beads (Sigma) which had been equilibrated in NETN-500 buffer. The beads were washed twice with 5 ml of NETN-500 buffer and the protein was eluted twice in 500 μl Storage Buffer (100 mM Tris pH 8.0, 400 mM NaCl, 10% glycerol, 5 mM BME) containing 3× FLAG peptide. Eluted protein was concentrated and dialyzed against Storage Buffer without 3X FLAG peptide and frozen in liquid nitrogen followed by storage at −80°C. Recombinant MRE11 was provided by Dr. Jean-Yves Masson, Laval University Cancer Research Center, Québec City, Canada. Treatment with NSC617145 (5340, Tocris), olaparib (S1060, Selleckchem), HU (H8627, Millipore Sigma), Mirin (M9948, Millipore Sigma), C5 (HY-128729, MedChemExpress), and NU7441 (KU-57788, S2638, Selleckchem) was performed in complete growth medium at 37°C with 5% CO 2 .
Oligonucleotide substrate preparations. The regressed fork, i.e., Chicken-foot (CF) substrate, and replication fork (RF) substrate were prepared using oligonucleotides listed in Supplementary Table 1 as follows: For fork restoration assays, 10 pmol of oligo B was labeled at its 5' end using T4 polynucleotide kinase and 30 µCi γ 32 P-ATP in a 20 ul reaction at 37°C for 40 min. After heat inactivation at 65°C for 20 min, free ATP was removed by applying the sample to a G-25 spin column (Cytiva). 10 pmol of either oligo A (for CF) or oligo D (for RF) was annealed to the labeled oligo B in 50 µl total volume with 10 mM Tris-HCl, pH 7.5 and 50 mM NaCl by heating the oligonucleotides at 95°C for 5 min followed by slow cooling to RT. 50 pmol of Oligo C and either oligo D (for CF) or A (for RF) were annealed similarly in 10 mM Tris, pH 7.5, and 50 mM NaCl. Just prior to the fork restoration assay, the CF substrate was prepared by combining 1 pmol of radiolabeled substrate AB and 2 pmol of unlabeled substrate CD in annealing buffer (10 mM Tris-HCl pH 7.5, 10 mM MgCl 2 , 50 mM NaCl) and incubating at 37°C for 30 min followed by RT for an additional 30 min. For the replication fork, radiolabeled 1 pmol of BD substrate was annealed to 2 pmol of substrate AC in a similar fashion.
In vitro nuclease assay. CF substrate was prepared as described above, however oligonucleotide C was radiolabeled and annealed to oligonucleotide D and oligonucleotides A and B were annealed, followed by annealing of the partial duplex CD to partial duplex AB. The indicated concentrations of WRN-E84A were incubated Fig. 9 WRNi potentiates olaparib cytotoxicity in BRCA2-mutated cancer cells. a Relative survival (%) of colonies in PEO1 (upper panel) and PEO4 (lower panel) cells treated with either graded concentrations of olaparib only or in combination with 0.25 µM NSC617145. Cells were plated on a six-well plate at 1,000 cells/well and exposed to 0.25 µM WRNi and indicated doses of olaparib 4 h after cell seeding. Cells were allowed to form colonies in presence of the drugs for 12 days. Data represent mean ± SD of three (n = 3) independent experiments. b Coefficient of drug interaction (CDI) for indicated concentrations of olaparib and 0.25 µM NSC617145 as described in a. Dashed lines in the graph represent thresholds for synergism (blue) and significant synergism (red). Data represent mean ± SD of three (n = 3) independent experiments. c Excess over Bliss (EOB) scores obtained from olaparib-NSC617145 Bliss synergy analysis are shown with a color heat map. Survival (%) data from three independent experiments were used to calculate CDI and EOB in b and c, respectively. d and e Schematic model of defective replication restart and fork hyper-degradation induced by WRN loss (d) or targeted WRN helicase inhibition (e) in BRCA2-deficient cancer cells. d Model depicting the involvement of WRN in stabilizing replication forks under the condition of BRCA2 deficiency. BRCA2 deficiency renders reversed forks deprotected and vulnerable to MRE11 nuclease-mediated degradation. Following replication fork stalling, WRN helicase-mediated restoration of reversed forks to active replication forks limits MRE11-mediated uncontrolled nascent strand degradation and ensures fork restart, albeit with reduced efficiency. WRN loss results in increased availability of reversed fork substrates for MRE11mediated degradation leading to enhanced nascent DNA degradation and severely compromised replication restart under BRCA2-deficient conditions. The yellow light of the traffic signal icon depicts replication stress. e Pharmacological inhibition of WRN helicase causes increased fork stalling induced by chromatin sequestration of WRN, followed by SNF2 translocase-mediated fork reversal and increased chromatin enrichment of MRE11 nuclease. In the absence of BRCA2, forks stalled by static WRN-DNA complexes are subjected to MRE11-mediated extensive degradation resulting in fork collapse, genomic instability, and cell death. Source data are provided as a Source Data file.
with the CF substrate (0.5 nM) in 10 µl reactions containing modified branch migration reaction salts (35 mM Tris-HCl, pH 7.5, 5 mM MnCl 2 , 1 mM MgCl 2 , 0.1 mg/ml BSA, 2 mM DTT, 15 mM phosphocreatine, 30 U/ml creatine phosphokinase and 5% glycerol) and 2 mM ATP for 10 min at 37°C. After the initial reaction, MRE11 was added to the reaction at a final concentration of 5 nM and incubated for an additional 50 min at 37°C. One microliter of proteinase K (0.8 U) was added to the reaction mixture and incubated at 37°C for 15 min followed by the addition of 10 µl of formamide loading buffer (80% formamide, 0.1% xylene cyanol, 0.1% bromophenol blue, 0.5× TBE). Samples were heated at 95°C for 5 min and then loaded on 12% denaturing polyacrylamide gels with 8 M urea and 0.5× TBE. Gels were electrophoresed for 40 min at 50 W and exposed to a phosphorscreen overnight and imaged using a Typhoon FLA 9500 (Cytiva). Percent exonuclease activity was determined by taking the volume of bands below the intact oligo C for each lane of the gel and dividing by the total volume of the lane using ImageQuant TL (Cytiva).
To measure the fork stability upon genetic knockdown, cells (1 × 10 5 ) were transfected with specific siRNAs 72 h prior to fiber labeling. U2OS and DLD1 WRN KO cells transfected with control or BRCA2 siRNA (80 nM) were retransfected with 2 μg of empty vector (EV), WT WRN, or catalytic dead WRN mutant plasmids 24 h post-siRNA transfection. Transfected cells were subjected to fork stability assay 48 h after plasmid transfection. Following CldU and IdU labeling, cells were washed twice with warm 1× PBS and subsequently incubated with 4 mM HU for 5 h before harvesting. To assess the effect of WRN or DNA2 inhibition on the stability of HU-stalled forks in PEO1 and PEO4 cells, WRNi NSC617145 (4 µM) or DNA2 inhibitor C5 (25 µM) was added to the culture medium concomitantly with 4 mM HU after the IdU labeling. DLD1 cells were treated with WRNi (1 µM) for 1 h and subsequently labeled with CldU and IdU nucleotides before the addition of HU. For assessing the effect of NSC617145 alone on fork stability, PEO1 and PEO4 cells were incubated with 4 µM of NSC617145 for 5 h following sequential labeling with CldU and IdU. To assess the effect of MRE11 or DNA2 inhibition on NSC617145-induced fork degradation, PEO1 and PEO4 cells were optionally pretreated with 50 µM Mirin or 25 µM C5 for 1 h prior to nucleotide labeling and the inhibitors were present throughout the experiment. DNA fibers with contiguous CldU-IdU signals were scored and the length of CldU and IdU tracts of individual fibers was measured manually using ZEN 3.0 (blue edition) software. IdU to CldU tract length ratios (IdU/CldU) of individual DNA fibers were calculated in Excel spreadsheet and scatter dot plots were generated using GraphPad Prism version 5.03.
To assess fork restart efficiency in BRCA2-depleted cells upon WRN complementation, U2OS WRN KO cells (1 × 10 5 ) were transfected with either control or BRCA2 siRNA (80 nM). Twenty-four hours post-siRNA transfection, cells were re-transfected with 2 μg of empty vector (EV), WT WRN, or catalytic dead WRN mutant plasmids. Transfected cells were subjected to fork restart assay 48 h after plasmid transfection. Cells were treated with 2 mM HU during CldU labeling to induce replication stalling. After 60 min incubation with HU, cells were washed three times in warm 1× PBS and incubated further in HU-free medium containing IdU for 45 min. For the fork restart experiments in PEO1 and PEO4 cells, control or WRN siRNA (80 nM) transfected cells were subjected to fork restart assay 72 h post-siRNA transfection. After 60 min incubation with 2 mM HU in the presence of CldU, cells were washed three times in warm 1× PBS and incubated in an HU-free medium containing IdU nucleotides for 60 min. In experiments with the WRNi, PEO4 and PEO1 cells were cotreated with 1 µM NSC617145 and 2 mM HU concomitantly with CldU and IdU labeling. IdU tract lengths of the restarted forks (forks characterized by contiguous CldU-IdU label) were measured to assess fork restart efficiency under different experimental conditions. To negate the possible effect of fork progression defect caused by different experimental conditions on restart efficiency quantitation, the length of an IdU tract was normalized to that of the corresponding CldU tract for a given fork. To analyze stalled fork recovery upon RAD52 knockdown in BRCA2-mutated cells exposed to WRNi, PEO1 cells transfected with either control or RAD52 siRNA (80 nM) were subjected to fork restart assay 72 h post-transfection. After CldU labeling for 20 min, cells were incubated in 2 mM HU for 4 h. After PBS wash, cells were subsequently incubated in HU-free medium containing IdU nucleotides for 20 min. WRNi (1 μM) was added to the medium during CldU labeling and present throughout the experiment. The percentage of stalled forks was determined as a percentage of only red-labeled fibers among the total red labeled (only red + redgreen fibers) DNA tracts.
For fork progression analysis, exponentially growing cells (1 × 10 5 ) were labeled with CldU for 15 min followed by IdU labeling for 60 min. To determine the immediate effect of WRNi on fork progression, cells were treated with 2 µM NSC617145 concomitantly with the IdU labeling. IdU to CldU tract length ratios of individual ongoing forks (fork with continuous CldU-IdU signal) were determined to assess relative fork progression upon WRN helicase inhibition. To determine fork asymmetry upon WRN helicase inhibition, sister tract (IdU) lengths of individual bi-directional fork structures (IdU-CldU-IdU) were measured, and short to long SF ratios were calculated. The relative decrease in the SF ratio serves as a readout of fork asymmetry.
For all fiber experiments, at least 125-300 untangled fibers were scored for each experimental condition. Statistical significance was determined by the Mann-Whitney test.
Chromatin isolation. Chromatin fractionation with CSK buffer containing 0.5% Triton X-100 was done as described previously 73 with some modifications. Approximately 2 × 10 7 cells grown in 150 mm dish were washed twice with 1× icecold PBS and collected by trypsinization. Cells were lysed in 1 ml modified CSK buffer (10 mM PIPES pH 6.8, 175 mM NaCl, 300 mM sucrose, 3 mM MgCl 2 , 1 mM EGTA, 1 mM DTT, 0.5% Triton X-100) supplemented with 1X protease inhibitor cocktail (78429, ThermoFisher Scientific) for 10 min at 0°C. Lysates were centrifuged at 5000×g for 3 min and the supernatant was collected as Triton-soluble fraction (cytosol + nucleosol). The pellet containing insoluble chromatin fractions were washed three times with CSK buffer each for 30 min on a rotator at 4°C. The final pellet was resuspended in 100 µl of CSK buffer (without Triton X-100) supplemented with 1 µl/ml benzonase (E1014, Millipore Sigma) and incubated for 15 min at 37°C. 1× LDS sample buffer (NP0007, ThermoFisher Scientific) was added to the sample and sonicated for 20 s using a probe sonicator (Branson digital sonifier) with 10% amplitude. Samples were clarified by centrifugation at 18,000×g for 10 min and supernatants were collected as chromatin-bound fractions.
Radiolabeling of NSC617145 and 14 C-NSC617145-WRN binding assay. Radiolabeling of NSC617145 was performed under contract by Moravek, Inc. In brief, a condensation reaction was carried out, whereby 2,3-dichloromaleic anhydride was reacted with 14 C-labeled 2,2,-dimethyl-1,3-propanediamine in acetic acid. The solid product was isolated by filtration and subsequently stored as a 0.1 mCi/mL solution in acetonitrile. Radiochemical purity was 98%.
For the in vitro 14 C-NSC617145-WRN binding assay, an equivalent concentration (300 nM) of purified recombinant WRN, FANCJ, or RECQL1 protein was incubated with 10 µM 14 C-NSC617145 in 20 µl binding buffer (100 mM Tris pH 8.0, 400 mM NaCl) for 15 min at 37°C. Using a 96-well Bio-dot microfiltration apparatus (#170-6545, Bio-Rad), the binding incubation samples were applied onto 0.2 µm nitrocellulose membrane pre-soaked in 1× TBS (Tris Buffered Saline) and allowed to filter through the membrane by gravity flow. The membrane was washed three times with 1× TBS by applying the vacuum before exposing it to a phosphorscreen. 14 C signal was detected with Typhoon FLA 9500 (GE Healthcare) and quantified using ImageQuant software (GE Healthcare). To determine 14 C-NSC617145-WRN binding in vivo, WRN −/− U2OS cells (~8 × 10 6 ) were transfected with 30 µg of either FLAG-empty vector or FLAG-WRN expression plasmids using Lipofectamine 2000 (11668019, ThermoFisher Scientific). Fortyeight hours post-transfection, cells were treated with 10 µM 14 C-NSC617145 for 1 h and subsequently lysed in 1 ml lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100) complemented with benzonase (1 µl/ml) and protease inhibitor cocktail. Cell lysates were centrifuged for 10 min at 12,000×g at 4°C and the supernatant was incubated with 25 µl of anti-FLAG M2 affinity beads (A2220, Millipore Sigma) overnight at 4°C with gentle rotation. Beads were washed three times with lysis buffer and 14 C-NSC617145: FLAG-WRN complexes were eluted (twice) by incubating beads with 100 µl of 3× FLAG peptide (250 ng/µl final concentration, F4799, Millipore Sigma) for 45 min on a rotator at 4°C. Eluted samples were subjected to dot blot assay using 96-well Bio-dot microfiltration apparatus as described above.
Neutral comet assay. DSBs upon WRN helicase inhibition were analyzed using neutral comet assay (single-cell gel electrophoresis) according to the Comet SCGE assay kit (Enzo, ADI-900-166) protocol. PEO4 and PEO1 cells (1 × 10 5 cells) were transfected with either control or MUS81 siRNA (80 nM) in a six-well plate. Seventy-two hours post-transfection, cells were treated with WRNi (5 µM) for 6 h. Cells were collected by trypsinization, counted, and resuspended in ice-cold 1× PBS (Ca 2+ and Mg 2+ free) at 1 × 10 5 cells/ml. Cells were mixed with molten lowmelting agarose (at 37°C) at a ratio of 1:10 (v/v) and 75 µl of agarose/cells suspension was immediately spread on the sample area of CometSlides (Enzo). Slides were kept at 4°C in the dark until the agarose solidified. Next, slides were immersed in prechilled 1× lysis buffer (Enzo) at 4°C for 1 h. Slides were washed once in 1× Tris-Borate-EDTA (TBE) buffer and subjected to horizontal gel electrophoresis in ice-cold 1× TBE buffer with 21 V for 20 min at 4°C. Slides were briefly rinsed in deionized water, immersed in 70% ethanol for 10 min, and kept at 37°C for drying. Samples in dried agarose were stained with CYGREEN nucleic acid dye (Enzo, diluted at a ratio of 1:1000) for 30 min at RT in the dark. Slides were briefly rinsed in deionized water and mounted using Permount mounting medium (Fisher Scientific). Comet slides were visualized using a GFP filter of Zeiss Axio Observer.Z1 fluorescence microscope with 10x objective lens. Comet tail moment was analyzed using OpenComet analysis tool 74 of ImageJ. Statistical analysis was performed in GraphPad Prism 5, version 5.03.
Metaphase spread preparation. PEO1 and PEO4 cells (2.5 × 10 6 ) grown in 100 mm dishes were treated with 2 µM NSC617145 for 48 h. To assess the effects of MRE11 or DNA-PK inhibition on WRNi-induced chromosomal instability in BRCA2-mutated cells, PEO1 cells (5 × 10 6 ) were treated with WRNi (5 µM) for 16 h in the presence or absence of Mirin (25 µM) or NU7441 (10 µM), respectively. Cells (~8 × 10 6 ) were arrested in metaphase by incubating with KaryoMAX Colcemid solution (0.1 µg/ml) (15212012, ThermoFisher Scientific) for 3 h, washed twice with 1× ice-cold PBS, trypsinized, and collected in 15 ml tubes by centrifugation at 100×g for 5 min. Cell pellets were gently resuspended in 1 ml ice-cold 1× PBS followed by dropwise addition of 9 ml preheated (37°C) KCl (75 mM) solution while slowly vortexing the tubes. After 20 min incubation in KCl solution at 37°C, cells were pelleted by centrifugation at 100×g for 5 min. Cells were fixed (two times) on ice by adding 2 ml freshly prepared fixative solution (methanol:acetic acid, 3:1) dropwise with mild vortexing. After adding another 4 ml of fixative solution, cells were pelleted by centrifugation at 100×g for 5 min. Cell pellets were resuspended in 1 ml fixative solution and kept at −20°C overnight. The chromosome spreads were prepared by dropping 200 µl of the cell suspension on 40% acetic acid spread on a glass slide (5070, Newcomer Supply). The slides were airdried and mounted with Prolong Gold antifade mountant (P36930, ThermoFisher Scientific). The slides were visualized under Zeiss Axio Observer.Z1/7 microscope with a 6x3 (oil) objective lens and images of metaphase chromosomes were acquired with the color camera attached. Chromosomal aberrations were analyzed using ZEN 3.0 (blue edition) software and graphs were generated in GraphPad Prism, version 5.03.
Colony formation assay. Cells (1 × 10 5 ) were transfected with siRNAs (80 nM) and plated (500-1,500 cells) on 6-well, 12-well or 24-well tissue culture dishes 72 h post-transfection. Drugs were added to the medium 4 h after cell seeding and cells were grown to form colonies in presence of drugs for 10-16 days. The number of cells plated, and colony formation times used for individual cell lines are detailed in the respective figure legends. Colonies were washed twice with 1× PBS, fixed with 100% methanol, and stained with crystal violet staining solution (25% methanol + 0.5% crystal violet). Percent survival upon drug treatment was determined by normalizing the total number of colonies formed with plating efficiency under individual conditions. Graphs were generated using GraphPad Prism, version 5.03. WST1 cytotoxicity assay. Cell viability in response to WRNi treatment was assessed using WST1 cell viability assay reagent (1164480700, Millipore Sigma) according to the manufacturer's protocol. Cells (1,500 cells in 100 μl complete cell culture medium) were plated on a 96-well plate and grown overnight before the addition of WRNi. Drug incubation times for individual experiments are given in the respective figure legends. For WST experiments following siRNA transfection, cells were first transfected with specific siRNAs (80 nM) in a six-well plate and subsequently plated on a 96-well plate (1500 cells) 24 h post-transfection. After 48 h of siRNA transfection, cells were exposed to WRNi for another 48 h before the addition of WST1 reagent (diluted 1:10 in cell culture medium). After 1 h incubation at 37°C, the absorbance of the samples was taken with an ELISA microplate reader at 450 nm. Relative cell viability upon WRNi treatment was determined with respect to untreated cells and line graphs were generated in Excel or GraphPad Prism 5, version 5.03.
In vivo mouse xenograft experiments. Mice were maintained as outlined in the Guide for the Care and Use of Laboratory Animals, under a protocol approved by the Animal Care and Use Committee (ACUC) of NCI-Frederick.
BRCA2-deficient and BRCA2-proficient DLD1 human colorectal adenocarcinoma cells were injected subcutaneously into the flanks of 8-week-old female athymic nude (NCr-nu/nu) mice (NCI-Frederick animal facility) at a concentration of 5 × 10 6 in 200 μl of PBS. The following number of mice were used for four groups: DLD1 BRCA2 +/+ treated with vehicle, n = 7; DLD1 BRCA2 −/− treated with vehicle, n = 7; DLD1 BRCA2+/+ treated with NSC617145, n = 7; and DLD1 BRCA2 −/− treated with NSC617145, n = 6. The resource equation method was used to determine sample sizes. No selection criteria were used for selecting the mice for each group. After the tumor volume reached 100-150 mm 3 , vehicle or NSC617145 at a dose of 20 mg/kg was administered by IP injection every other day for 2 weeks (total 7 treatments). WRNi NSC617145 was dissolved (1.25 mg/ml) in 3% DMA (N, N-dimethylacetamide), 1.8% Tween 80 in D5W (5% Dextrose in water) for IP injection. Tumor size was measured every other day for 3-4 weeks. Tumors were harvested before they reach the maximal size (2000 mm 3 ) permitted by NCI-Frederick Animal Care and Use Committee (ACUC). After harvesting, tumors were flash-frozen in liquid nitrogen for further characterization. Tumor sizes were measured using a digital Vernier caliper and tumor volume (in mm 3 ) was calculated as a product of (length × width 2 )/2. Tumor growth inhibition (TGI) was calculated using the formula, % TGI = (1−(mean volume of treated tumors)/ (mean volume of control tumors))×100. Mice were maintained according to the procedures outlined in the Guide for the Care and Use of Laboratory Animals and a protocol approved by the NCI-Frederick Animal Care and Use Committee (ACUC). Mice were housed as per NCI-Frederick ACUC guidelines as described below: Housing: All mice were group-housed unless scientific justification was approved by the NCI-Frederick ACUC (e.g., the last animal left in a study group). All cages had at least one form of environmental enrichment. All singly housed mice were provided at least two forms of enrichment, such as a plastic hut or additional type of nesting material; Light cycle: Rooms were maintained at a constant light cycle of 12 h on and 12 h off; Temperature: Rooms were maintained at temperatures in the range of 68-79°F; Humidity: Relative-humidity levels were maintained at 30-70%. Animal caretakers checked room humidity by using dedicated digital thermometer/hygrometers during the AM and PM animal and cage observations; Air changes: air exchanges in the animal rooms were in the range of 10-15 air changes per hour (ACH).
Drug interaction analysis. The CDI was determined using the following equation: CDI = AB/(A × B) where AB is the ratio of the viability of cells treated with both NSC617145 and Olaparib to the no treatment cells; A (NSC617145) or B (Olaparib) is the ratio of the cell viability of the cells treated with either NSC617145 or Olaparib to the no treatment cells. A CDI values can be separated into several categories: CDI > 1 (antagonistic), CDI = 1 (additive), CDI < 1 (synergistic), or CDI < 0.7 (significantly synergistic) 75 . CDI values were calculated at different concentrations of Olaparib with a single concentration of NSC617145. Bliss index was calculated using the following equation: BI = (E A + E B −E A *E B )/E AB . The Bliss index was calculated at each Olaparib concentration with constant NSC617145 concentration. The excess over Bliss (EOB) was calculated by subtracting the expected effect of the drug combination (E A + E B −E A *E B ) from the observed effect (E AB ) and multiplying that difference by 100. Here, E A and E B are the calculated effects of NSC617145 and Olaparib individually where the E = 1-% cell viability, and E AB is the actual effect of the two inhibitors in combination. More positive EOB values indicate stronger synergism 76 .
Statistics and reproducibility. Statistical significance was determined by Student's two-tailed t-test, one-way or two-way ANOVA, and Mann-Whitney test using Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
All data of the present study are available within the article and its supplementary information files or available from the corresponding author upon request. Source data are provided with this paper.
Received: 23 November 2020; Accepted: 20 October 2021; Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2021