www.aging-us.com AGING 2020, Vol. 12, No. 24 Research Paper Age-associated deficient recruitment of 53BP1 in G1 cells directs DNA double-strand break repair to BRCA1/CtIP-mediated DNA-end resection Teresa Anglada1, Anna Genescà1,*, Marta Martín1,* 1 Department of Cell Biology, Physiology and Immunology, Universitat Autònoma de Barcelona, Bellaterra 08193, Spain *Equal contribution Correspondence to: Marta Martín, Anna Genescà; email: Marta.Martin@uab.cat, Anna.Genesca@uab.cat Keywords: aging, double-strand break repair, 53BP1, BRCA1, non-homologous end-joining Received: August 17, 2020 Accepted: December 3, 2020 Published: December 27, 2020 Copyright: © 2020 Anglada et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT DNA repair mechanisms play a crucial role in maintaining genome integrity. However, the increased frequency of DNA double-strand breaks (DSBs) and genome rearrangements in aged individuals suggests an age- associated DNA repair deficiency. Previous work from our group revealed a delayed firing of the DNA damage response in human mammary epithelial cells (HMECs) from aged donors. We now report a decreased activity of the main DSB repair pathways, the canonical non-homologous end-joining (c-NHEJ) and the homologous recombination (HR) in these HMECs from older individuals. We describe here a deficient recruitment of 53BP1 to DSB sites in G1 cells, probably influenced by an altered epigenetic regulation. 53BP1 absence at some DSBs is responsible for the age-associated DNA repair defect, as it permits the ectopic formation of BRCA1 foci while still in the G1 phase. CtIP and RPA foci are also formed in G1 cells from aged donors, but RAD51 is not recruited, thus indicating that extensive DNA-end resection occurs in these breaks although HR is not triggered. These results suggest an age-associated switch of DSB repair from canonical to highly mutagenic alternative mechanisms that promote the formation of genome rearrangements, a source of genome instability that might contribute to the aging process. INTRODUCTION several evidences link aging with a decline in DSB repair efficiency that could account for the Aging is a complex process that implies the loss of accumulation of unrepaired DNA damage and increased physiological integrity and affects the DNA of cells. genome instability with age. However, the nature of the Several studies have reported increased frequencies of DSB repair defect underlying the age-related both DNA double-strand breaks (DSBs) and genome accumulation of DNA damage remains an open rearrangements in different tissues of aged organisms question. and senescent cells [1, 2]. Also, mutations in some DSB repair genes have been described as causing a premature There are two main DSB repair pathways operating in aging phenotype [3], thus pointing to a relationship mammalian cells: the canonical non-homologous end- between defective DNA repair and age. In this regard, joining (c-NHEJ), which is able to join scarcely the direct measurement of DSB repair events using resected broken DNA ends that share little or no plasmid constructs has demonstrated an age-associated homology, and the homologous recombination (HR) decrease in the efficiency and fidelity of the repair pathway, which needs extensive DNA resection and function in mice, rats and human cells [4–6]. Thus, uses the sister chromatid as a template for repair. The www.aging-us.com 24872 AGING choice of either DSB repair pathway is highly regulated, efficiency of these pathways with aging, and the cell cycle being a major determinant [7–9]. While the c- factors that eventually influence the interplay between NHEJ is active throughout the cycle and is the DNA repair pathways in this context, remain to be predominant pathway during G1, HR activates from explored. mid-S to G2, when a sister chromatid is available as a template for repair. In a previous work from our group, we measured the frequency of γH2AX foci in G1 human mammary At a molecular level, the balance between 53BP1 and epithelial cells (HMECs) from young and aged donors BRCA1 plays an important role in the DSB repair after exposure to γ-rays [26]. Results of that work pathway choice by modulating end resection [10–12]. showed that an age-associated delay in the firing of the After DNA damage induction, multiple sites of 53BP1 DNA damage response (DDR) was responsible for the are phosphorylated by the ataxia-telangiectasia mutated repair deficiency observed in HMECs from older (ATM) kinase and 53BP1 is rapidly relocated to the women [26]. In order to find out the exact nature of this break site [13]. Phosphorylated 53BP1 directly repair defect, in the present work we evaluated the recognizes the ubiquitination of H2A at lysine 15 activity of the c-NHEJ and HR repair pathways and the (H2AK15ub) [14], but the selective recognition of H4 dynamics of the main proteins involved in these di-methylation at lysine 20 (H4K20me2) is also mechanisms after DSB induction. We observed a required for the focal accumulation of 53BP1 at DSBs deficiency in the recruitment of 53BP1 to DSBs in cells [15]. This epigenetic mark is catalyzed by from aged donors. Absence of 53BP1 from some DSBs methyltransferases SETD8 and SUV4-20 and >80% of in G1 cells allows ectopic access of BRCA1, CtIP and nucleosomes in G1 cells present H4K20me2 [16]. Also, RPA but not RAD51 to these breaks. We propose that 53BP1 binding affinity with H4K20me2 is enhanced by increasing age hampers proper recruitment of 53BP1 to the deacetylation of H4 at lysine 16 [17, 18]. Once G1 DSBs. This defect is probably influenced by 53BP1 accumulates at the DSBs it serves as a platform changes in the epigenetic landscape affecting the aging for the recruitment of other factors that direct repair to genome, such as H4K20 methylation, directly involved the c-NHEJ, while restricting BRCA1 accumulation at in the recruitment of 53BP1. DSBs devoid of 53BP1 are the DSB during the G1 phase of the cell cycle [10]. accessible to the DNA-end resection machinery, thus switching DSB repair from c-NHEJ to error-prone Progression into S phase enables BRCA1 recruitment to mechanisms such as Alt-EJ, which could contribute to DSBs by recognition of the unmethylated lysine 20 of the reported age-associated formation of genomic histone H4 (H4K20me0) on newly synthesized DNA rearrangements. [19]. BRCA1 interacts with CtIP in a cyclin-dependent kinase (CDK)-dependent manner and promotes DNA- RESULTS end resection, thus inhibiting c-NHEJ [20, 21]. The initial steps of resection are directed by the The efficiency of the c-NHEJ and HR DSB repair CtIP/MRE11 complex [22], which, along with other pathways decreases with age nucleases, generate long stretches of single stranded DNA (ssDNA) that are coated by RPA [23]. RAD51 is In a previous study, we described an age-associated a core protein of the HR pathway that progressively increase in the frequency of DNA DSBs in HMECs displaces the RPA from ssDNA to form a nucleoprotein following ionizing radiation (IR) exposure due to an filament that is competent to invade a sister chromatid age-related delayed firing of the DDR [26]. To strand and use it as a template for repair [24]. Finally, investigate the nature of this repair defect, here we used DNA ends are ligated and the original sequence is HMECs derived from mammary tissues of female restored without error. donors classified as young donors (YDs ≤ 27 years old) and aged donors (ADs ≥ 60 years old). Low population In addition, under specific cellular conditions, doubling (PD) cells’ (< 20 PD) were used in all of the alternative highly mutagenic pathways, such as the experiments to rule out any effect of replicative cellular alternative end-joining (Alt-EJ), can repair those DSBs senescence on DNA repair. that have suffered from DNA-end resection. This mechanism promotes ligation of resected DNA ends, First, we evaluated the efficiency of the c-NHEJ and thus increasing the probability of introducing HR pathways by transfecting the cells with reporter alterations in the DNA sequence [25]. In the last few plasmids. Transfection efficiency values for each donor years, there have been great advances in describing the were calculated using a constitutively GFP-expressing mechanisms that regulate the interplay between the plasmid. As expected for primary cells, these values different DSB repair pathways (c-NHEJ, HR and Alt- were low, but there were no statistical differences EJ) and their components. However, the decline in the between the two age groups (Supplementary Table 1; www.aging-us.com 24873 AGING YDs = 9.18%; ADs = 11.51%, t-test; p-value > .05). BRCA1, were evaluated in cells from the two age Then we evaluated the efficiency of the c-NHEJ and groups. To this end, DSBs were induced in localized HR pathways by transfecting the cells with the reporter areas of cell nuclei using the micro-irradiation plasmids pimEJ5GFP or pDRGFP together with an I- technique described by Suzuki et al. [27]. DSBs were SceI enzyme-expressing plasmid (Supplementary identified at 15, 30 and 60 min post-irradiation (pIR) as Figure 1A–1C). ADs showed a statistically significant discrete γH2AX foci. The newly induced DSBs in decrease in the c-NHEJ activity, as shown by the localized nuclear areas could be clearly differentiated normalized frequency of GFP-positive cells (4.50% in from the basal ones (Figure 1C), whose frequency was YDs and 2.32% in ADs; t-test; p-value < .001) (Figure low but significantly increased in aged donors [26]. 1A and Supplementary Table 1). Consistent with this, c- NHEJ activity was negatively correlated with the We first checked the colocalization of 53BP1 and residual number of γH2AX foci scored 24 h after γH2AX at radiation-induced DSBs. To do so, micro- exposure to 1 Gy of γ-rays in each donor in our irradiated nuclear areas were localized, and previous study [26] (R2 = 0.39; p-value = 0.0308) individualized γH2AX foci were scored. Next, the (Supplementary Figure 1D). Therefore, the c-NHEJ coincidence in space of 53BP1 focus with each γH2AX repair pathway is less active with age and this defect foci site was visually determined (Figure 1C) and most probably contributes to the accumulation of referred to as colocalization. The colocalization of residual DSBs in ADs. 53BP1 with γH2AX foci increased with time after irradiation both in YDs and ADs, reaching a plateau at In contrast to the c-NHEJ pathway, the HR pathway 60 min pIR (Figure 1D). However, percentages of only operates in S and G2 phases of the cell cycle. colocalization were significantly lower for the ADs in According to this, all donors showed less HR than c- comparison to the younger ones at all times analyzed NHEJ activity, as shown by the normalized (YDs: 62.59% at 15 min, 76.39% at 30 min and 81.00% frequencies of GFP-positive cells (Figure 1B; at 60 min; ADs: 48.19% at 15 min, 62.42% at 30 min Supplementary Table 1). Our results showed that the and 66.79% at 60 min; two-way ANOVA and HR repair pathway activity was also reduced in ADs Bonferroni multiple correction test; p-value < .0001) compared to YDs (3.14% for YDs and 1.33% for ADs), (Figure 1E; Supplementary Table 2). and the difference was statistically significant (t-test; p-value < .001). A reduced activity of the HR repair An age-related impaired recruitment of 53BP1 to DSBs pathway in ADs could be due to an increased could account for the delayed firing of the DDR and the frequency of ADs’ cells halted in the G1 phase. increased frequency of residual breaks previously However, the cell cycle analysis showed no significant reported in AD cells as well as for the reduced differences in the distribution of the cell cycle phases efficiency of the c-NHEJ pathway observed in the between YD and AD cells (two-way ANOVA and present study. Frequencies of 53BP1 recruitment to Bonferroni multiple correction test; p-value > .05) DSBs for each donor positively correlated with values (Supplementary Figure 1F–1G). Finally, we observed of c-NHEJ efficiency evaluated with the plasmid that the donors with low HR activity also showed low reporter system at all time points analyzed (R2 = 0.54 at c-NHEJ activity (R2 = 0.54, p-value = 0.0065) 15 min; R2 = 0.61 at 30 min and R2 = 0.69 at 60 min (Supplementary Figure 1E) and they mostly pIR) (Figure 1F). Also, 53BP1 recruitment to DSBs at corresponded to aged donors. These results indicate 60 min pIR negatively correlated with the frequencies that the age-related DSB repair defect is not restricted of γH2AX foci in HMECs from the same donors to a single pathway, but instead affects both main DSB reported previously [26] both in non-irradiated samples repair pathways, thus contributing to the accumulation (R2 = 0.76) and 24 h after irradiation (R2 = 0.79) of DNA damage in aged organisms. (Supplementary Figure 2), indicating that proper recruitment of 53BP1 favors γH2AX foci clearance. Age-associated decrease in the recruitment of 53BP1 Finally, data from the 53BP1/γH2AX colocalization at to radiation-induced DSBs the different time points hierarchically clustered donors into two differentiated groups that coincided with The DDR operates in a highly coordinated manner and chronological age (Figure 1G), revealing that the protein recruitment dynamics are essential for the frequency of DSBs signaled by 53BP1 is a good marker proper triggering of the DNA repair mechanisms. Thus, of age. Altogether, these results indicate that 53BP1 we next hypothesized that the repair defect observed in recruitment to DSB sites soon after DNA damage ADs could be due to an impaired recruitment of DNA induction is impaired in the HMECs from aged donors, damage signaling and repair proteins to DSB sites. The and it translates into a lower activity of the c-NHEJ and dynamics of recruitment of the main proteins into an increased frequency of both basal and induced responsible for the repair pathway choice, 53BP1 and γH2AX foci. www.aging-us.com 24874 AGING Figure 1. Decreased efficiency of DNA–DSB repair and deficient recruitment of 53BP1 in aged donor cells. (A, B) Normalized frequency of GFP-positive cells after co-transfection with pimEJ5GFP (A) or pDRGFP (B) and the I-SceI expressing plasmids. Mean and SD are indicated (* p < .001; t-test). (C) Immunofluorescent labeling of cell nuclei (DAPI, blue), γH2AX (Cy3, red) and 53BP1 (A488, green). γH2AX and 53BP1 foci were scored within the irradiated pore area (yellow dotted lines). Scale bar = 10 µm. (D, E) Percentage of 53BP1/γH2AX foci colocalization for five young and five aged donors (D) and summary values for each age group (E). Boxes include data from the upper to the lower quartile and whiskers compile minimum to maximum values (* p < .0001; n is stated in Supplemental Table 2; two-way ANOVA + Bonferroni). (F) Correlation between c-NHEJ efficiency and 53BP1/γH2AX foci colocalization. Best-fit line, 95% confidence bands (dotted lines) 2 and Pearson’s correlation coefficient (R ) are indicated (p < .05). (G) Hierarchical clustering of the ten donors according to the percentage of 53BP1/γH2AX foci colocalization. www.aging-us.com 24875 AGING BRCA1 is efficiently recruited to DSBs in G2 phase repair to the HR, like BRCA1, RPA and RAD51. Again, HMECs from older donors the levels of these proteins showed inter-individual variations, but no age-associated tendency was detected Next, we assessed BRCA1 recruitment to radiation- (Figure 3A). Thus, we conclude that the decline in c- induced DSBs by quantifying colocalization of BRCA1 NHEJ and HR repair and the recruitment defect with γH2AX foci in G2 cells identified by positive observed in ADs is not due to depletion of DNA repair CENPF staining, a protein that progressively proteins. accumulates from the S phase to mitosis (Figure 2A, panel i). BRCA1 recruitment kinetics followed a pattern We next argued that epigenetic alterations associated resembling that of 53BP1: it increased with time after with aging could be influencing 53BP1 recruitment micro-irradiation in both YDs and ADs, reaching a efficiency in AD cells. The focal accumulation of plateau at 60 min pIR. Nevertheless, there were no 53PB1 on DSBs relies on the specific binding of 53BP1 detectable differences in the percentage of to the H4K20me2 [29–31]. Conversely, lysine 16 BRCA1/γH2AX colocalization between the two age acetylation of the same histone H4 (H4K16Ac) groups (YDs: 42.55% at 15 min, 62.67% at 30 min and significantly reduces 53BP1 interaction with 67.42% at 60 min; ADs: 43.04% at 15 min, 61.87% at H4K20me2 [32, 33]. Hence, we measured H4K16 30 min and 64.84% at 60 min; two-way ANOVA and acetylation levels in G1 cells (only 1 pericentrin mark) Bonferroni multiple correction test; p-value > .05) from young and aged donors. Regardless of donor’s age, (Figure 2B; Supplementary Table 3). Given that H4K16ac fluorescence intensity showed high cell-to- BRCA1 recruitment to DSBs in G2 cells from ADs was cell variability amongst cells of the same donor (Figure as effective as in those from YDs, we ruled out an age- 3C), and no statistically significant difference between associated deficiency in BRCA1-mediated DSB age groups was found (Figure 3C, panel ii). We next recognition. However, difficulties related to the hypothesized that an inefficient methylation of H4K20 recruitment of other effector proteins downstream of in cells from ADs could translate into a deficient BRCA1 could account for the reduced activity of the accumulation of 53BP1 at DSBs. The first step to HR repair pathway in cells from older donors. Indeed, methylate H4K20 is the monomethylation by lysine ADs showed reduced percentages of RAD51 methyltransferase SETD8 [34]. Decreased SETD8 colocalization with γH2AX foci in G2 cells at 4 h after mRNA levels have been reported in senescent cells and exposure to 5 Gy of γ-rays (53.37% in YD vs. 40.91% in in vitro aged fibroblasts [35, 36]. Thus, we measured in AD; one-way ANOVA and Tukey multiple the expression of SETD8 by RT-qPCR in HMECs from correction test; p-value < .05) (Figure 2C, 2D). young and aged donors. Interestingly, results revealed a Altogether, the results suggest that an impaired or significant decrease in SETD8 mRNA levels in HMECs delayed RAD51 recruitment to DSBs could affect the from aged donors in comparison to the younger ones proper progression of homologous recombination repair (mean relative expression: YD = 0.96; AD = 0.62; t- with age and account for the decreased efficiency of the test; p-value < .01) (Figure 3D). We speculate that the HR pathway detected with the plasmid reporter assay in observed decrease of SETD8 could translate into an cells from older donors. age-associated increase of H4K20me0 sites to which 53BP1 recruitment would be hindered, while occupancy Decreased mRNA levels of H4K20 methyltransferase by other DNA repair proteins, such as BRCA1, would SETD8 in HMECs from older donors be favored. We next aimed to explore the causes underlying In the absence of 53BP1, BRCA1 is ectopically deficient recruitment of repair proteins in HMECs from recruited to DSBs in G1 older individuals. Our first hypothesis was that protein expression was differentially regulated by age. Since the In G1 cells, 53BP1 recruitment acts as a barrier for HR expression of DNA repair enzymes has been evaluated by preventing DSBs’ end resection, however, BRCA1 is in cells from aged individuals and senescent cells with also expressed at this stage and can form IR-induced inconsistent results [6, 28], we measured 53BP1 gene foci that colocalize with γH2AX foci in 53BP1 depleted expression by RT-qPCR and protein levels by Western cells [10]. Thus, we next aimed to determine whether blot in HMECs from old and young donors. Although 53BP1 deficient recruitment to DSBs in ADs would some variation was detected amongst donors, no allow BRCA1 to ectopically form foci at DSB sites significant differences in 53BP1 mRNA and protein during G1. We first scored, in G1 cells (negative for levels were observed between the two age groups CENPF staining), the γH2AX foci induced 30 min after (Figure 3A, panel i, and 3B). Likewise, Western blot localized DSB induction (Figure 2A, panel ii), and results showed no age-related differences for 53BP1’s checked for BRCA1 foci colocalization inside the pore effector protein Ku70, or for other proteins directing zone. Indeed, BRCA1/γH2AX foci colocalization was www.aging-us.com 24876 AGING Figure 2. Recruitment of BRCA1 to DNA damage sites in G2 is not impaired in cells from aged donors. (A) Immunofluorescent labeling of cell nuclei (DAPI, blue), γH2AX (A594, red), BRCA1 (A488, green) and CENPF (A532, grey). γH2AX and BRCA1 foci were scored within the irradiated pore zone (yellow dotted lines) in CENPF-positive (i) or CENPF-negative (ii) cells. Scale bar = 10 µm. (B) Percentage of BRCA1/γH2AX foci colocalization in CENPF-positive HMECs from four YDs and four ADs. Boxes include data from the upper to the lower quartile and whiskers compile minimum to maximum values (n is stated in Supplementary Table 3). (C) Percentage of RAD51/γH2AX foci colocalization in CENPF-positive cells at 4 h after irradiation (5 Gy, γ-rays). Error bars indicate SEM (* p < .05; n ≥ 500 γH2AX foci/donor; one- way ANOVA + Tukey). (D) Immunofluorescent labeling of cell nuclei (DAPI), γH2AX (A488), RAD51 (A594) and CENPF nuclear staining (A532) at 4 h after exposure to 5 Gy of γ-rays. Arrowheads indicate G2 (CENPF-positive) cells. Scale bar = 20 µm. www.aging-us.com 24877 AGING Figure 3. Expression levels of 53BP1, SETD8 and H4K16ac. (A) Western blot analysis of c-NHEJ and HR factors. Basal levels of (i) high and (ii) low molecular weight DNA repair proteins. Stain-free technology and/or Integrin β1 (ITGB1) have been used for sample normalization and U2OS cells were used as a positive control. (B, D) RT-qPCR analysis of 53BP1 (B) and SETD8 (D). GAPDH and β-actin were used as reference genes (* p < .05, ns p > .05; t-test). (C) H4K16ac analysis. (i) Immunofluorescent labeling of cell nuclei (DAPI, blue), H4K16ac (A488, green) and pericentrin (A594, red). Representative G1 cells with high or low H4K16ac fluorescence intensity are shown. Scale bar = 10 µm. (ii) Fluorescence intensity of H4K16 acetylation in G1 cells (1 perincentrin signal). Each dot corresponds to one cell and the mean and quartiles are indicated (a≠b≠c p < .05; n = 40 cells/donor; Kruskal–Wallis + Dunn). www.aging-us.com 24878 AGING significantly higher in G1 cells from ADs (ADs: Dunn’s multiple correction test; p-value > .05) (Figure 41.66%; YDs: 17.41%; Mann–Whitney test; p-value 4H). Thus, DNA breaks occupied with BRCA1 in G1 < .0001) (Supplementary Table 4; Figure 4A, 4B). cells from older donors are not processed by the HR Moreover, the fraction of BRCA1-free γH2AX foci was pathway. However, the impact of those DSBs suffering lower in ADs (~10%) than in YDs (~40%) (Figure 4A), DNA-end resection in G1 should be reckoned with, as indicating that in G1 cells, BRCA1 can form foci at DNA resected intermediates are able to suppress c- γH2AX-signaled DSBs and that the frequency of this NHEJ-mediated DSB repair to favor alternative and event increases with the donor’s age. Since 53BP1 highly mutagenic DNA repair mechanisms. recruitment is compromised in ADs, it is tempting to speculate that ectopic BRCA1 recruitment is promoted DISCUSSION at the 53BP1-orphan DSBs. In fact, the sum of the two colocalization frequencies (53BP1/γH2AX plus The mechanisms by which older individuals accumulate BRCA1/γH2AX) reaches ~100% in each donor (Figure genome rearrangements and DSBs have remained 4B, 4C), suggesting that either 53BP1 or BRCA1 is unknown for years. In a previous work, we reported a recruited to almost all DSBs within 30 min of their delayed firing of the DDR in G1 HMECs from aged induction. We propose that ectopic BRCA1 foci donors that could contribute to the accumulation of formation at DSBs in G1 cells is an age-related DSBs with age [26]. In the present work, we explored response, most probably related to an attempt to occupy the causes of the postponed ignition of the repair DSBs that have failed to recruit 53BP1. response in aged individuals. We observed a deficient recruitment of 53BP1 to radiation-induced DSBs from CtIP-mediated DNA-end resection of DSBs occurs in old donors, a defect that was previously described by G1 cells from aged donors our group using in vitro aged HMECs [37], suggesting that the deficiency in 53BP1 recruitment could be an To further explore the functional relevance of the age-related characteristic. Using the reporter plasmids BRCA1 ectopic recruitment observed in G1 cells from technique, a decreased activity of the c-NHEJ in aged ADs, we analyzed whether CtIP was also being donors has also been detected. It has been described that recruited. To this end, HMECs from YDs and ADs were 53BP1 acts with fast kinetics [13] and its rapid exposed to 5 Gy of γ-rays and the number of G1 cells positioning at the break site along with its effector positive for CtIP (≥ 5 foci) was scored. At 1 h pIR, ADs protein RIF1 impedes DNA-end resection and promotes showed a significantly increased frequency of CtIP- c-NHEJ repair during G1 [10]. Therefore, the impaired positive cells (< 6% in YDs vs. > 15% in ADs; Fisher’s recruitment of 53BP1 to DSBs described here could exact test; p-value < .05) (Figure 4D, 4E). These results alter the triggering of DSB repair and could translate suggest that, although not all of the γH2AX foci into defective c-NHEJ repair. colocalizing with BRCA1 had efficiently recruited CtIP, a significant fraction of DSBs in AD cells eventually The deficient recruitment of 53BP1 to DSBs is not due did and might undergo end resection. to its depletion, as no age-associated differences were detected between the two age groups neither at mRNA Because RPA binds and stabilizes ssDNA intermediates nor at the protein level. We also discarded factors that arise after DNA processing, we next analyzed its related to an aberrant DSB signaling. A defect in DSB recruitment in G1 cells. At 1 h pIR, the number of RPA detection would translate into an aberrant formation of foci scored in G1 cells was negligible regardless of γH2AX foci but, instead, results from our previous donor’s age (results not shown). We reasoned that an work demonstrated a higher frequency of γH2AX foci extensive resection that would allow RPA recruitment in cells from ADs [26]. However, aging is associated might require more time. At 4 h after irradiation most with changes at the epigenetic level [38], which could YD cells in G1 were still devoid of RPA foci, but the influence 53BP1 recruitment to DSBs. The frequency of RPA-positive cells (≥ 5 foci) in AD cells accumulation of 53BP1 at DSB sites requires both had increased significantly (0% in YDs vs. ≥ 10% in recognition of H2AK15ub via its UDR motif and ADs; Fisher’s exact test; p-value < .05) (Figure 4F, 4G). H4K20me2 via its Tudor domain [14, 33]. H2AK15ub These results suggest that the initial DNA-end resection marking is catalyzed by the RNF pathway that is also elicited by CtIP in G1 cells from older donors translates, required for BRCA1 recruitment to DSBs [39]. Since with time, into an extensive DNA-end resection that BRCA1 showed no deficiency in foci formation at results in single-stranded stretches of DNA that are DSBs in HMECs, we discard that a failure in this effectively coated with RPA. Finally, the colocalization pathway might be the cause of 53BP1 impaired of RAD51 foci with γH2AX foci was extremely low in recruitment in cells from aged women. Similarly, both young and aged donors at 4 h after 5 Gy exposure acetylation of H4K16 has been described to prevent the (2.12% in YDs vs. 2.47% in ADs; Kruskal–Wallis and interaction of 53BP1 with H4K20me2 and consequent www.aging-us.com 24879 AGING Figure 4. BRCA1, CtIP and RPA but no RAD51 are ectopically recruited to DNA DSBs in G1 cells from aged donors. (A) Percentage of BRCA1/γH2AX foci colocalization in CENPF-negative cells at 30 min after irradiation for each age group. Each dot corresponds www.aging-us.com 24880 AGING to the fraction of BRCA1 and γH2AX foci colocalizing within one pore and the mean and quartiles are indicated (* p < .0001; n is stated in Supplementary Table 4; Mann–Whitney test). (B) Summary values for individual donors. Error bars indicate SEM (a≠b p < .05; n is stated in Supplementary Table 4; Kruskal–Wallis + Dunn). (C) Percentage of 53BP1/γH2AX foci colocalization for individual donors at 30 min after irradiation. Error bars indicate SEM (n is stated in Supplementary Table 2). (D) Immunofluorescent labeling of cell nuclei (DAPI), CENPF (A532), γH2AX (A594) and CtIP (A488). Arrowheads indicate G1 (CENPF-negative) cells. Scale bar = 20 µm. (E) Frequency of CtIP-positive HMECs (≥ 5 foci) at 1 h after irradiation (5 Gy, γ-rays). Analysis was restricted to CENPF-negative cells (* p < .05; n ≥ 50 cells/donor; Fisher’s exact test). (F) Immunofluorescent labeling of cell nuclei (DAPI), CENPF (A532), γH2AX (A594) and RPA (A488). Arrowheads indicate G1 (CENPF-negative) cells. Scale bar = 20 µm. (G) Frequency of RPA-positive HMECs (≥ 5 foci) at 4 h after irradiation (5 Gy, γ-rays). Analysis was restricted to CENPF-negative cells (* p < .05; n ≥ 50 cells/donor; Fisher’s exact test). (H) Percentage of RAD51/γH2AX foci colocalization in CENPF-negative cells at 4 h after irradiation (5 Gy, γ-rays). Error bars indicate SEM (ns p > .05; n ≥ 1000 γH2AX foci/donor; Kruskal–Wallis + Dunn). binding to DNA [17], but aged HMECs did not display of CtIP foci. From S to G2 phases, CtIP activation is hyperacetylation of this mark. In line with this, both directly regulated by CDKs, but in G1 cells CtIP can be aging and senescence processes have been linked to phosphorylated by polo-like kinase 3 (PLK3) [42, 43]. H4K16 hypoacetylation [40, 41]. Instead, we found an Additionally, not only CtIP, but also RPA foci age-associated decreased expression of SETD8, the formation in AD cells indicates that long stretches of enzyme responsible for the monomethylation of ssDNA are generated in G1 in the absence of 53BP1, H4K20. This result was previously reported in suggesting that extensive DNA-end resection has taken senescent and in in vitro aged cells [35, 36], place in some DBSs in G1 cells. If CtIP initiates end demonstrating that SETD8 decreased expression is resection in these breaks, c-NHEJ can no longer operate. shared between different aging models. Decreased Launching the HR in G1 cells is not possible due to the levels of methyltransferase SETD8 in HMECs from absence of a homologous repair template, together with aged donors might result in the persistence of BRCA1-PALB2-BRCA2 complex assembly inhibition H4K20me0 sites in G1, compromising DSB recognition by CDKs [44]. In agreement with this, RAD51 was not by 53BP1. In this regard, it has been recently loaded in G1 cells from ADs, discarding the HR as an demonstrated that the H4K20me0 mark is required for active repair pathway to process DNA-end resected BRCA1 recruitment to post-replicative chromatin [19]. DSBs in G1. We speculate that the age-associated Thus, while 53BP1 recruitment to DSB at H4K20me0 ectopic activity of BRCA1/CtIP could account for the sites would be hindered, BRCA1 recruitment would be increased accumulation of genome rearrangements enhanced. typically observed in aged tissues [2, 45]. CtIP has oncogenic properties due to its implication in the Alt-EJ Given the antagonistic role of 53BP1 and BRCA1 in the pathway, and has been associated with chromosomal DSB repair pathway choice, it is not surprising that an instability and the generation of aberrant chromosomal impaired recruitment of 53BP1 would have rearrangements [46]. consequences in BRCA1 dynamics. While the recruitment of BRCA1 was not affected by age in G2 While neither the c-NHEJ nor the HR operated in cells, we observed an increased formation of ectopic 53BP1-free DSBs in G1 cells from ADs, DNA breaks BRCA1 foci in G1 cells from ADs. Ectopic BRCA1 and were actually being repaired, as only a small fraction of CtIP foci formation in G1 after inhibition or depletion DSBs remain unrepaired a long time after irradiation of 53BP1 or its effector protein, RIF1, has been [26]. It is tempting to speculate that these DSBs are previously reported [10, 42]. In the present work 53BP1 being repaired by an alternative repair mechanism such is not depleted from ADs cells, but it is absent in as as the Alt-EJ, which shares with HR the initial steps of much as 40% of the radiation-induced DSBs, giving DNA-end resection mediated by CtIP and MRE11, prior room to BRCA1 to occupy these breaks, most probably to the ligation of the DNA-ends [47]. The Alt-EJ can in an attempt to resolve the otherwise orphan DSBs. All operate both in the G1 and G2 phases of the cell cycle of these results evidence the mutually exclusive [48] and in cells with functional c-NHEJ and HR relationship between BRCA1/CtIP and 53BP1/RIF1 pathways [47]. Contrary to c-NHEJ, Alt-EJ requires a tandems, which ultimately determine the DSB repair greater degree of end resection prior to the DNA-end machinery that will be loaded onto the DSB. Thus, we ligation step, often resulting in extensive losses, propose that the age-associated repair defect in HMECs additions or alterations of the DNA sequence at the relies on the antagonism between 53BP1 and BRCA1 DSB junction. Also, the Alt-EJ functions with slower proteins that probably shift the DNA–DSB repair kinetics, thus increasing the probability of illegitimate pathway choice to more mutagenic repair mechanisms. repair [48]. All these scenarios are compatible with the reported increased frequencies of genome The increased frequency of BRCA1 foci in G1 cells rearrangements and somatic mutations in tissues from from ADs was concomitant with an increased frequency aged individuals [2, 45]. www.aging-us.com 24881 AGING In summary, we propose that 53BP1 deficient by RPA while awaiting ligation. Because the c-NHEJ recruitment to DSBs is a hallmark of age, and we pathway is inhibited by extensive DNA-end resection of present a model for a DSB repair pathway choice in breaks and HR cannot be launched in G1, we speculate HMECs from aged women (Figure 5). Upon induction two possible outcomes for these DSBs: they might of DSBs in YD cells, 53BP1 is recruited in the G1 remain extensively processed and unrepaired until the phase and promotes c-NHEJ. In contrast, AD cells fail cells progress to S phase and HR can be fully launched to recruit 53BP1 to DSBs efficiently, allowing their in the presence of a DNA template or, otherwise, they ectopic occupancy by BRCA1, which is followed by become substrates for alternative and highly mutagenic CtIP-mediated DNA-end resection. If resection is backup mechanisms of DNA repair, such as Alt-EJ extensive enough, single-stranded fragments are coated repair. Figure 5. Model for the age-related shift in the DSB repair pathway choice in G1 cells. In response to DSB induction, 53BP1 is recruited to the break site in G1 cells from YDs and promotes repair by canonical non-homologous end-joining (c-NHEJ). Instead, in G1 AD cells, the deficient recruitment of 53BP1 permits the ectopic recruitment of BRCA1 to some DSBs, followed by CtIP-mediated DNA-end resection and RPA coating of the ssDNA. However, RAD51 loading is inhibited in these G1 cells and homologous recombination (HR) cannot be launched. Thus, DSBs from ADs that have suffered DNA-end resection in G1 become substrates for alternative DSB repair mechanisms, such as the alternative end-joining (Alt-EJ). www.aging-us.com 24882 AGING The present work evidences the connection between acetone for 1 min and blocked with 1x PBS – 1% FBS genome integrity and aging. Future efforts addressing – 5% BSA. For H4K16ac labeling, after the exact mechanism to counteract BRCA1/CtIP permeabilization with triton, cells were incubated for 1 activity during the G1 phase in cells from older h at 60° C with Dako REALTM Target Retrieval individuals could be of extreme interest, as they would Solution (Agilent Technologies, Santa Clara, CA, allow the restoration of DNA repair fidelity during USA). After the blocking step, primary antibodies aging and prevent a rise in genomic instability in older (listed in Supplementary Table 5) were incubated individuals. overnight at 4° C. After three washes with 1x PBS – 0.1% Tween20 or PBS – 1% FBS, the secondary MATERIALS AND METHODS antibodies anti-rabbit A488 and anti-mouse A594 (Supplementary Table 5) were incubated for 1 h at Cell culture room temperature. If a third protein was detected (CENPF), secondary antibodies from the other two Finite lifespan pre-stasis HMECs were obtained from were reincubated in order to occupy the maximum reduction mammoplasty of 11 donors: 48R (16 yo), number of epitopes before anti-CENPF was incubated 240L (19 yo), 168R (19 yo), 184 (21 yo), 59L (23 yo), overnight at 4° C. The secondary antibody for CENPF 123 (27 yo), 153L (60 yo), 112R (61 yo), 122L (66 yo), was then incubated for 20 min at room temperature. 29 (68 yo) and 429ER (72 yo); and one peripheral non- Finally, all samples were washed, briefly rinsed with tumor containing mastectomy tissue of one donor: 353P distilled water, underwent progressive alcohol (72 yo). Donors were classified depending on their age dehydration and counterstained with 4’,6-diamidino-2- into young donors (YDs ≤ 27 years) and aged donors phenylindole (DAPI) at a final concentration of 0.25 (ADs ≥ 60 years). When referring to donors, the age μg/ml in Vectashield Mounting Medium (Vector group is followed by specimen identification and the Laboratories, Inc., Burlingame, CA, USA). Analysis age of the donor. HMECs were cultured as pre-stasis and image acquisition were performed using an strains using M87A medium with cholera toxin and Olympus BX61 epifluorescent microscope (Olympus, oxytocin according to previously reported methods [49], Hamburg, Germany) equipped with a CV-M4+CL with the addition of 100 U/ml penicillin and 100 μg/ml camera (JAI, Grosswallstadt, Germany) and Cytovision streptomycin. U2OS cells were cultured with DMEM software (Applied Imaging, Newcastle, UK). Fiji and Ham’s F10 medium at 1:1 and supplemented with software [50] was used for H4K16ac fluorescence 10% FBS and penicillin/streptomycin. Cells were kept intensity measurement. in the incubator at 37° C and 5% CO2 atmosphere. Western blot Irradiation HMECs were washed with cold PBS and stored at -80° Micro-localized DNA damage was induced as described C upon collection. Cells were lysed in RIPA buffer (50 by Suzuki et al. [27]. Briefly, cells were labelled with mM Tris-HCl pH 8, 1% sodium deoxycholate, 150 mM 10 μm bromodeoxyuridine (BrdU) for 48 h. Just before sodium chloride, 20 mM sodium fluoride, 1% Triton- the irradiation medium was removed, cells were briefly X100, 1 mM EDTA, 0,1% SDS, DTT 1 mM, 20 mM b- rinsed with PBS and covered with a Whatman™ glicerolphosphate, 1 mM Na ortovanadate and protease Cyclopore™ polycarbonate membrane (Whatman, inhibitor cocktail (Roche, Basel, Switzerland) and Maidstone, UK) with pores of 5 μm in diameter. Cells sonicated. The protein concentration was measured were exposed to 15J m-2 of UVC light at a dose rate of using a PierceTM BCA Protein Assay Kit (Thermo 1J m-2 s-1. After exposure, free-BrdU medium was Fisher Scientific, Waltham, MA, USA) and 25 μg of added and cells were incubated at 37° C and 5% CO2 total protein was loaded onto a 7.5% acrylamide TGX for 15–60 min. Whole-cell irradiation was performed by Stain-Free gel or a 10% Bis-Tris gel. After exposing cells to 5 Gy of γ-rays using an IBL-437C R- electrophoresis, proteins were transferred to a PVDF 137 Cs irradiator at a dose rate of 5.10 Gy/min. membrane. After a blocking step with 5% BSA or 5% nonfat milk for 1 h, primary antibodies (Supplementary Immunofluorescence Table 5) were incubated overnight at 4° C. After three washing steps with TTBS (Tris 20 mM, NaCl 150 mM, HMECs were fixed during 15 min with 4% PFA, and 0.1% Tween20), secondary antibodies permeabilized with 1x PBS-0.5% Triton-X-100 (Supplementary Table 5) were incubated for 1 h at room solution for 20 min and blocked with 1x PBS – 0.5% temperature. Chemiluminescent detection was BSA – 0.15% glycine for 15 min. For RAD51 and CtIP performed after incubation with Immobilon Western labelling, cells were also fixed with ice-cold methanol Chemiluminescent HRP Substrate (Millipore, Madrid, for 30 min after PFA, permeabilized with ice-cold Spain) and using a ChemiDocTM Touch Imaging System www.aging-us.com 24883 AGING (Bio-Rad, Hercules, CA, USA). For reprobing, a mild RNA isolation, cDNA synthesis and quantitative stripping solution (1.5% glycine, 0.1% SDS, 1% real-time PCR Tween20) adjusted to pH 2.2 was used for antibodies with the same host species. An incubation in 30% H202 For RNA collection, cells were lysed using TRIzolTM was applied for membrane reprobing with antibodies solution (Thermo Fisher Scientific, Waltham, MA, with different host species. USA) and stored at -80° C until all samples were collected. After thawing, chloroform was added at 1:5 Transfection with DNA repair reporter plasmids (v/v), incubated for 2 minutes and centrifuged at 12000 and flow cytometric analysis g for 15 minutes at 4° C. RNA was purified from the resulting aqueous phase using Maxwell® RSC To measure c-NHEJ and HR repair activity, 40,000 simplyRNA kit (Promega, Madison, WI, USA) and cells from an exponentially growing culture were following manufacturer’s instructions. RNA quantity seeded onto 12-well plates. When properly attached, and purity were measured using NanoDropTM 2000 cells were transfected with 0.5 μg of reporter spectrophotometer (Thermo Fisher Scientific, Waltham, pimEJ5GFP (Addgene #44026) or pDRGFP (Addgene MA, USA). Samples scored purity odds ratios 260/280 #26475) and with 0.5 μg of the I-SceI-expressing and 260/230 between 1.9 and 2.1. RNA integrity was plasmid pCBASceI (Addgene #26477) or with an empty measured with an Agilent 2100 Bioanalyzer (Agilent vector pCAGGS (kind gift from Surrallés’ and Jasin’s Technologies, Santa Clara, CA, USA) and it reached the laboratories) as a negative control. As a positive control, maximum score (RIN = 10). Retrotranscription of 1 μg cells were transfected with 0.5 μg of the GFP- of total RNA was performed using iScript cDNA expressing plasmid NZE-GFP (kind gift from Surrallés’ synthesis kit (Bio-Rad, Hercules, CA, USA) in a final and Jasin’s laboratories). The mix of DNA was volume of 20 μL. Real-time quantitative PCR was prepared in OptiMEM medium and then mixed in a 1:4 performed using SYBR® Green (Bio-Rad, Hercules, proportion with Fugene® HD transfection reagent CA, USA) on a CFX384 thermal cycler with CFX (Promega, Madison, WI, USA) following the Manager software (Bio-Rad, Hercules, CA, USA). For manufacturer’s instructions. Transfection mix was each reaction, 50 ng of cDNA were used and samples added to cells in 1 ml of antibiotic-free medium and were run in triplicates. SETD8 and ACTB primer incubated for 13 h at 37° C and 5% CO2 atmosphere. sequences were obtained from the literature [36, 52]. The percentage of GFP-positive cells was measured by Otherwise, 53BP1 and GAPDH were designed using the flow cytometry with a FACS Calibur cytometer (Becton Primer3 online tool [53]. All primers were purchased Dickinson, Franklin Lakes, NJ, USA) at 40 h after from Condalab (Metabion, Munich, Germany) and are transfection. FlowJo software (v10.0.7; FlowJo LLC, listed in Supplementary Table 6. Data normalization Ashland, OR, USA) was used for gate adjusting and was performed using the geometric mean of GAPDH plot representation. The threshold of GFP-negative cells and ACTB reference genes. Fold change was calculated was determined for each donor with values from cells following the 2-∆∆Ct method [54]. transfected with the pCAGGS empty vector. To compare activity levels of each repair pathway between Statistical analysis donors, the fluorescent value obtained for each reporter plasmid was normalized with the transfection efficiency The data obtained were analyzed using Microsoft Excel value obtained after transfection with the GFP- (Microsoft® Excel® 2011, v14.1, Redmond, expressing plasmid NZE-GFP. At least two independent Washington, USA). The statistical analysis was transfections with the reporter plasmids and their performed using GraphPad Prism 8 (GraphPad Software respective controls were performed for each donor. Inc., San Diego, CA, USA) with methods indicated in the results where applicable. For the hierarchical cluster Cell cycle by flow cytometry analysis, the Ward method was applied [55] using R software (version 3.4.4, Vienna, Austria). A cell cycle analysis was performed as described elsewhere [51]. Briefly, after trypsinization, cells were Abbreviations fixed in ethanol and stored at -20° C. After all samples were collected, cells were stained with PI staining DSBs: DNA double-strand breaks; HMECs: human solution (0.1% Triton-X-100 in PBS, 0.2 mg/ml DNase- mammary epithelial cells; c-NHEJ: canonical non- free RNase A and 0.02 mg/ml of propidium iodide). homologous end-joining; HR: homologous After 30 min of incubation at room temperature, recombination; ATM: ataxia-telangiectasia mutated; fluorescence intensity was measured using a FACS H2AK15ub: ubiquitination of H2A at lysine 15; Calibur cytometer (Becton Dickinson) and analyzed H4K20me2: H4 di-methylation at lysine 20; using FlowJo software. H4K20me0: unmethylated lysine 20 of histone H4; www.aging-us.com 24884 AGING CDK: cyclin-dependent kinase; ssDNA: single stranded 2. Kalfalah F, Seggewiß S, Walter R, Tigges J, Moreno- DNA; Alt-EJ: alternative end-joining; DDR: DNA Villanueva M, Bürkle A, Ohse S, Busch H, Boerries M, damage response; IR: ionizing radiation; YDs: young Hildebrandt B, Royer-Pokora B, Boege F. Structural donors; ADs: aged donors; PD: population doubling; chromosome abnormalities, increased DNA strand pIR: post-irradiation; H4K16ac: H4 acetylation at breaks and DNA strand break repair deficiency in lysine 16; PLK3: polo-like kinase 3; BrdU: dermal fibroblasts from old female human donors. bromodeoxyuridine; DAPI: 4’,6-diamidino-2- Aging (Albany NY). 2015; 7:110–22. phenylindole. https://doi.org/10.18632/aging.100723 PMID:25678531 AUTHOR CONTRIBUTIONS 3. Bohr VA. Human premature aging syndromes and genomic instability. Mech Ageing Dev. 2002; MM and AG conceived the study. TA acquired and 123:987–93. analyzed data. MM, AG and TA critically discussed https://doi.org/10.1016/s0047-6374(02)00039-8 results, wrote the manuscript and approved its final PMID:12044947 version. 4. Vyjayanti VN, Rao KS. DNA double strand break repair ACKNOWLEDGMENTS in brain: reduced NHEJ activity in aging rat neurons. Neurosci Lett. 2006; 393:18–22. Special thanks to Martha Stampfer and Mark LaBarge https://doi.org/10.1016/j.neulet.2005.09.053 (Lawrence Berkeley National Laboratory) for providing PMID:16226837 HMEC samples and advice on cell culture. Our sincere 5. Vaidya A, Mao Z, Tian X, Spencer B, Seluanov A, gratitude to Jordi Surrallés and Jordi Minguillón Gorbunova V. Knock-in reporter mice demonstrate (Universitat Autònoma de Barcelona) for their help that DNA repair by non-homologous end joining with the plasmid reporter assay. We thank Eduard declines with age. PLoS Genet. 2014; 10:e1004511. Escrich (Universitat Autònoma de Barcelona) for the https://doi.org/10.1371/journal.pgen.1004511 material, equipment and advice with the Western blot PMID:25033455 assay. We also acknowledge Pablo Huertas (CABIMER, 6. Li Z, Zhang W, Chen Y, Guo W, Zhang J, Tang H, Xu Z, Sevilla) for providing the CtIP antibody and critical advice. We gratefully thank Ignasi Roig (Universitat Zhang H, Tao Y, Wang F, Jiang Y, Sun FL, Mao Z. Autònoma de Barcelona) for his critical reading of the Impaired DNA double-strand break repair contributes manuscript and Joan Repullés for his help with figure to the age-associated rise of genomic instability in design and creation. humans. Cell Death Differ. 2016; 23:1765–77. https://doi.org/10.1038/cdd.2016.65 PMID:27391797 CONFLICTS OF INTEREST 7. Chapman JR, Taylor MR, Boulton SJ. Playing the end game: DNA double-strand break repair pathway The authors declare that they have no conflicts of choice. Mol Cell. 2012; 47:497–510. interest. https://doi.org/10.1016/j.molcel.2012.07.029 PMID:22920291 FUNDING 8. Hustedt N, Durocher D. The control of DNA repair by the cell cycle. Nat Cell Biol. 2016; 19:1–9. The AG and MM laboratory is funded by the Consejo https://doi.org/10.1038/ncb3452 de Seguridad Nuclear, Generalitat de Catalunya (2017- PMID:28008184 SGR-503) and Fundación Dexeus Salud de la Mujer 2016 grants. TA has been a beneficiary of FI-DGR 9. Her J, Bunting SF. How cells ensure correct repair of 2015-2017 (Generalitat de Catalunya) and Universitat DNA double-strand breaks. J Biol Chem. 2018; Autònoma de Barcelona PhD program fellowships. 293:10502–11. https://doi.org/10.1074/jbc.TM118.000371 REFERENCES PMID:29414795 10. Escribano-Díaz C, Orthwein A, Fradet-Turcotte A, Xing 1. Sedelnikova OA, Horikawa I, Redon C, Nakamura A, M, Young JT, Tkáč J, Cook MA, Rosebrock AP, Munro Zimonjic DB, Popescu NC, Bonner WM. Delayed M, Canny MD, Xu D, Durocher D. A cell cycle- kinetics of DNA double-strand break processing in dependent regulatory circuit composed of 53BP1-RIF1 normal and pathological aging. Aging Cell. 2008; and BRCA1-CtIP controls DNA repair pathway choice. 7:89–100. Mol Cell. 2013; 49:872–83. https://doi.org/10.1111/j.1474-9726.2007.00354.x https://doi.org/10.1016/j.molcel.2013.01.001 PMID:18005250 PMID:23333306 www.aging-us.com 24885 AGING 11. Panier S, Boulton SJ. Double-strand break repair: https://doi.org/10.1038/s41556-019-0282-9 53BP1 comes into focus. Nat Rev Mol Cell Biol. 2014; PMID:30804502 15:7–18. 20. Sartori AA, Lukas C, Coates J, Mistrik M, Fu S, Bartek J, https://doi.org/10.1038/nrm3719 PMID:24326623 Baer R, Lukas J, Jackson SP. Human CtIP promotes DNA 12. Bakr A, Köcher S, Volquardsen J, Reimer R, Borgmann end resection. Nature. 2007; 450:509–14. K, Dikomey E, Rothkamm K, Mansour WY. Functional https://doi.org/10.1038/nature06337 PMID:17965729 crosstalk between DNA damage response proteins 21. Huertas P, Jackson SP. Human CtIP mediates cell cycle 53BP1 and BRCA1 regulates double strand break repair control of DNA end resection and double strand break choice. Radiother Oncol. 2016; 119:276–81. repair. J Biol Chem. 2009; 284:9558–65. https://doi.org/10.1016/j.radonc.2015.11.001 https://doi.org/10.1074/jbc.M808906200 PMID:26615718 PMID:19202191 13. Anderson L, Henderson C, Adachi Y. Phosphorylation 22. Shibata A, Moiani D, Arvai AS, Perry J, Harding SM, and rapid relocalization of 53BP1 to nuclear foci upon Genois MM, Maity R, van Rossum-Fikkert S, Kertokalio DNA damage. Mol Cell Biol. 2001; 21:1719–29. A, Romoli F, Ismail A, Ismalaj E, Petricci E, et al. DNA https://doi.org/10.1128/MCB.21.5.1719-1729.2001 double-strand break repair pathway choice is directed PMID:11238909 by distinct MRE11 nuclease activities. Mol Cell. 2014; 14. Fradet-Turcotte A, Canny MD, Escribano-Díaz C, 53:7–18. Orthwein A, Leung CC, Huang H, Landry MC, Kitevski- https://doi.org/10.1016/j.molcel.2013.11.003 LeBlanc J, Noordermeer SM, Sicheri F, Durocher D. PMID:24316220 53BP1 is a reader of the DNA-damage-induced H2A Lys 23. Symington LS. Mechanism and regulation of DNA end 15 ubiquitin mark. Nature. 2013; 499:50–54. resection in eukaryotes. Crit Rev Biochem Mol Biol. https://doi.org/10.1038/nature12318 PMID:23760478 2016; 51:195–212. 15. Pellegrino S, Michelena J, Teloni F, Imhof R, Altmeyer https://doi.org/10.3109/10409238.2016.1172552 M. Replication-coupled dilution of H4K20me2 guides PMID:27098756 53BP1 to pre-replicative chromatin. Cell Rep. 2017; 24. Heyer WD, Ehmsen KT, Liu J. Regulation of homologous 19:1819–31. recombination in eukaryotes. Annu Rev Genet. 2010; https://doi.org/10.1016/j.celrep.2017.05.016 44:113–39. PMID:28564601 https://doi.org/10.1146/annurev-genet-051710- 16. Saredi G, Huang H, Hammond CM, Alabert C, Bekker- 150955 PMID:20690856 Jensen S, Forne I, Reverón-Gómez N, Foster BM, 25. Ceccaldi R, Rondinelli B, D’Andrea AD. Repair pathway Mlejnkova L, Bartke T, Cejka P, Mailand N, Imhof A, et choices and consequences at the double-strand break. al. H4K20me0 marks post-replicative chromatin and Trends Cell Biol. 2016; 26:52–64. recruits the TONSL–MMS22L DNA repair complex. https://doi.org/10.1016/j.tcb.2015.07.009 Nature. 2016; 534:714–18. PMID:26437586 https://doi.org/10.1038/nature18312 PMID:27338793 26. Anglada T, Repullés J, Espinal A, LaBarge MA, Stampfer 17. Tang J, Cho NW, Cui G, Manion EM, Shanbhag NM, MR, Genescà A, Martín M. Delayed γH2AX foci Botuyan MV, Mer G, Greenberg RA. Acetylation limits disappearance in mammary epithelial cells from aged 53BP1 association with damaged chromatin to women reveals an age-associated DNA repair defect. promote homologous recombination. Nat Struct Mol Aging (Albany NY). 2019; 11:1510–23. Biol. 2013; 20:317–25. https://doi.org/10.18632/aging.101849 https://doi.org/10.1038/nsmb.2499 PMID:23377543 PMID:30875333 18. Hsiao KY, Mizzen CA. Histone H4 deacetylation 27. Suzuki K, Yamauchi M, Oka Y, Suzuki M, Yamashita S. facilitates 53BP1 DNA damage signaling and double- Creating localized DNA double-strand breaks with strand break repair. J Mol Cell Biol. 2013; 5:157–65. microirradiation. Nat Protoc. 2011; 6:134–39. https://doi.org/10.1093/jmcb/mjs066 https://doi.org/10.1038/nprot.2010.183 PMID:23329852 PMID:21293454 19. Nakamura K, Saredi G, Becker JR, Foster BM, Nguyen 28. Seluanov A, Danek J, Hause N, Gorbunova V. Changes NV, Beyer TE, Cesa LC, Faull PA, Lukauskas S, Frimurer in the level and distribution of Ku proteins during T, Chapman JR, Bartke T, Groth A. H4K20me0 cellular senescence. DNA Repair (Amst). 2007; recognition by BRCA1-BARD1 directs homologous 6:1740–48. recombination to sister chromatids. Nat Cell Biol. 2019; https://doi.org/10.1016/j.dnarep.2007.06.010 21:311–18. PMID:17686666 www.aging-us.com 24886 AGING 29. Mailand N, Bekker-Jensen S, Faustrup H, Melander F, DNA damage in mammary epithelial cells aged in vitro. Bartek J, Lukas C, Lukas J. RNF8 ubiquitylates histones PLoS One. 2013; 8:e63052. at DNA double-strand breaks and promotes assembly https://doi.org/10.1371/journal.pone.0063052 of repair proteins. Cell. 2007; 131:887–900. PMID:23667571 https://doi.org/10.1016/j.cell.2007.09.040 38. Booth LN, Brunet A. The aging epigenome. Mol Cell. PMID:18001824 2016; 62:728–44. 30. Doil C, Mailand N, Bekker-Jensen S, Menard P, Larsen https://doi.org/10.1016/j.molcel.2016.05.013 DH, Pepperkok R, Ellenberg J, Panier S, Durocher D, PMID:27259204 Bartek J, Lukas J, Lukas C. RNF168 binds and amplifies 39. Wang B, Matsuoka S, Ballif BA, Zhang D, ubiquitin conjugates on damaged chromosomes to Smogorzewska A, Gygi SP, Elledge SJ. Abraxas and allow accumulation of repair proteins. Cell. 2009; RAP80 form a BRCA1 protein complex required for the 136:435–46. DNA damage response. Science. 2007; 316:1194–98. https://doi.org/10.1016/j.cell.2008.12.041 https://doi.org/10.1126/science.1139476 PMID:19203579 PMID:17525340 31. Misteli T, Soutoglou E. The emerging role of nuclear 40. Krishnan V, Chow MZ, Wang Z, Zhang L, Liu B, Liu X, architecture in DNA repair and genome maintenance. Zhou Z. Histone H4 lysine 16 hypoacetylation is Nat Rev Mol Cell Biol. 2009; 10:243–54. associated with defective DNA repair and premature https://doi.org/10.1038/nrm2651 senescence in Zmpste24-deficient mice. Proc Natl Acad PMID:19277046 Sci USA. 2011; 108:12325–30. 32. Sanders SL, Portoso M, Mata J, Bähler J, Allshire RC, https://doi.org/10.1073/pnas.1102789108 Kouzarides T. Methylation of histone H4 lysine 20 PMID:21746928 controls recruitment of Crb2 to sites of DNA damage. 41. Contrepois K, Thuret JY, Courbeyrette R, Fenaille F, Cell. 2004; 119:603–14. Mann C. Deacetylation of H4-K16Ac and https://doi.org/10.1016/j.cell.2004.11.009 heterochromatin assembly in senescence. Epigenetics PMID:15550243 Chromatin. 2012; 5:15. 33. Botuyan MV, Lee J, Ward IM, Kim JE, Thompson JR, https://doi.org/10.1186/1756-8935-5-15 Chen J, Mer G. Structural basis for the methylation PMID:22932127 state-specific recognition of histone H4-K20 by 53BP1 42. Bakr A, Köcher S, Volquardsen J, Petersen C, Borgmann and Crb2 in DNA repair. Cell. 2006; 127:1361–73. K, Dikomey E, Rothkamm K, Mansour WY. Impaired https://doi.org/10.1016/j.cell.2006.10.043 53BP1/RIF1 DSB mediated end-protection stimulates PMID:17190600 CtIP-dependent end resection and switches the repair 34. Milite C, Feoli A, Viviano M, Rescigno D, Cianciulli A, to PARP1-dependent end joining in G1. Oncotarget. Balzano AL, Mai A, Castellano S, Sbardella G. The 2016; 7:57679–93. emerging role of lysine methyltransferase SETD8 in https://doi.org/10.18632/oncotarget.11023 human diseases. Clin Epigenetics. 2016; 8:102. PMID:27494840 https://doi.org/10.1186/s13148-016-0268-4 43. Barton O, Naumann SC, Diemer-Biehs R, Künzel J, PMID:27688818 Steinlage M, Conrad S, Makharashvili N, Wang J, Feng 35. Tanaka H, Takebayashi SI, Sakamoto A, Igata T, Nakatsu L, Lopez BS, Paull TT, Chen J, Jeggo PA, Löbrich M. Polo- Y, Saitoh N, Hino S, Nakao M. The SETD8/PR-Set7 like kinase 3 regulates CtIP during DNA double-strand methyltransferase functions as a barrier to prevent break repair in G1. J Cell Biol. 2014; 206:877–94. senescence-associated metabolic remodeling. Cell Rep. https://doi.org/10.1083/jcb.201401146 2017; 18:2148–61. PMID:25267294 https://doi.org/10.1016/j.celrep.2017.02.021 44. Orthwein A, Noordermeer SM, Wilson MD, Landry S, PMID:28249161 Enchev RI, Sherker A, Munro M, Pinder J, Salsman J, 36. Shih CT, Chang YF, Chen YT, Ma CP, Chen HW, Yang CC, Dellaire G, Xia B, Peter M, Durocher D. A mechanism Lu JC, Tsai YS, Chen HC, Tan BC. The PPARγ-SETD8 axis for the suppression of homologous recombination in constitutes an epigenetic, p53-independent checkpoint G1 cells. Nature. 2015; 528:422–26. on p21-mediated cellular senescence. Aging Cell. 2017; https://doi.org/10.1038/nature16142 16:797–813. PMID:26649820 https://doi.org/10.1111/acel.12607 PMID:28514051 45. Moskalev AA, Shaposhnikov MV, Plyusnina EN, 37. Hernández L, Terradas M, Martín M, Feijoo P, Soler D, Zhavoronkov A, Budovsky A, Yanai H, Fraifeld VE. The Tusell L, Genescà A. Increased mammogram-induced role of DNA damage and repair in aging through the www.aging-us.com 24887 AGING prism of Koch-like criteria. Ageing Res Rev. 2013; https://doi.org/10.1158/0008-5472.CAN-09-0270 12:661–84. PMID:19773443 https://doi.org/10.1016/j.arr.2012.02.001 50. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, PMID:22353384 Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld S, 46. Zhang Y, Jasin M. An essential role for CtIP in Schmid B, Tinevez JY, White DJ, Hartenstein V, et al. chromosomal translocation formation through an Fiji: an open-source platform for biological-image alternative end-joining pathway. Nat Struct Mol Biol. analysis. Nat Methods. 2012; 9:676–82. 2011; 18:80–84. https://doi.org/10.1038/nmeth.2019 PMID:22743772 https://doi.org/10.1038/nsmb.1940 PMID:21131978 51. Darzynkiewicz Z, Juan G, Bedner E. Determining cell 47. Truong LN, Li Y, Shi LZ, Hwang PY, He J, Wang H, cycle stages by flow cytometry. Curr Protoc Cell Razavian N, Berns MW, Wu X. Microhomology- Biol. 2001; Chapter 8:Unit 8.4. mediated end joining and homologous recombination https://doi.org/10.1002/0471143030.cb0804s01 share the initial end resection step to repair DNA PMID:18228389 double-strand breaks in mammalian cells. Proc Natl 52. Hernandez-Segura A, Rubingh R, Demaria M. Acad Sci USA. 2013; 110:7720–25. Identification of stable senescence-associated https://doi.org/10.1073/pnas.1213431110 reference genes. Aging Cell. 2019; 18:e12911. PMID:23610439 https://doi.org/10.1111/acel.12911 PMID:30710410 48. Iliakis G, Murmann T, Soni A. Alternative end-joining 53. Koressaar T, Remm M. Enhancements and repair pathways are the ultimate backup for abrogated modifications of primer design program Primer3. classical non-homologous end-joining and homologous Bioinformatics. 2007; 23:1289–91. recombination repair: implications for the formation of https://doi.org/10.1093/bioinformatics/btm091 chromosome translocations. Mutat Res Genet Toxicol PMID:17379693 Environ Mutagen. 2015; 793:166–75. https://doi.org/10.1016/j.mrgentox.2015.07.001 54. Livak KJ, Schmittgen TD. Analysis of relative gene PMID:26520387 expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods. 2001; 49. Garbe JC, Bhattacharya S, Merchant B, Bassett E, 25:402–08. Swisshelm K, Feiler HS, Wyrobek AJ, Stampfer MR. https://doi.org/10.1006/meth.2001.1262 Molecular distinctions between stasis and telomere PMID:11846609 attrition senescence barriers shown by long-term culture of normal human mammary epithelial cells. 55. Everitt BS, Landau S, Leese M, Stahl D. Cluster Analysis. Cancer Res. 2009; 69:7557–68. 5th ed. Quality and Quantity. 2011. John Wiley & Sons, Ltd. www.aging-us.com 24888 AGING SUPPLEMENTARY MATERIALS Supplementary Figures Supplementary Figure 1. Reporter plasmids assays and cell cycle analysis with flow cytometry. (A) Structure of the reporter plasmids for the analysis of (i) c-NHEJ and (ii) HR efficiency. (B, C) Representative flow cytometry plots of the frequency of GFP-positive cells www.aging-us.com 24889 AGING after co-transfection with the c-NHEJ reporter plasmid pimEJ5GFP (B) or the HR reporter plasmid pDRGFP (C) and the I-SceI expressing plasmids. (D, E) Correlation between c-NHEJ activity and the residual number of γH2AX foci (D) or between c-NHEJ and HR pathways’ activity 2 (E) in HMECs from YDs and ADs. Best-fit line, 95% confidence bands (dotted lines) and Pearson’s correlation coefficient (R ) are indicated (p < 2 .05). (F) Representative diagrams of cell cycle distribution for YDs and ADs before and at 30 min post-irradiation (15 J/m , UVC, BrdU 2 sensitized cells). (G) Frequency of cells at G1, S and G2/M stages in non-irradiated HMECs and at 15, 30 and 60 min after irradiation (15 J/m , UVC, BrdU sensitized cells). Mean values and SD from three young donors (YD240L(19), YD168R(19) and YD184(21)) and three ADs (AD112R(61), AD122L(66) and AD429ER(72)) are shown. Supplementary Figure 2. Negative correlations between 53BP1/γH2AX foci colocalization and the number of DSBs scored as γH2AX foci in HMECs from YDs and ADs. 53BP1/γH2AX foci colocalization was evaluated at 15, 30 and 60 min pIR. γH2AX foci were 2 scored before irradiation and 1 h and 24 h pIR. Best-fit line, 95% confidence bands (dotted lines) and Pearson’s correlation coefficient (R ) are indicated (p < .05). Most significant values are highlighted in bold letters. www.aging-us.com 24890 AGING Supplementary Tables Supplementary Table 1. Plasmid reporter assays for the c-NHEJ and HR pathways in HMECs. Transfection efficiency c-NHEJ HR Donor % SD Normalized % Normalized % Young donors 9.18 5.75 4.50 3.14 YD48R(16) 10.67 0.40 4.39 3.63 YD240L(19) 4.21 1.01 4.05 3.60 YD168R(19) 11.94 4.40 5.00 2.46 YD184(21) 5.51 1.94 4.87 2.85 YD59L(23) 8.11 0.43 4.37 3.11 YD123(27) 22.35 0.78 4.33 3.17 Aged donors 11.51 7.74 2.32 1.33 AD153L(60) 22.03 3.31 1.07 0.16 AD112R(61) 16.62 9.45 2.09 2.73 AD122L(66) 4.99 0.79 1.24 1.13 AD29(68) 6.16 0.60 2.78 0.82 AD429ER(72) 7.51 5.73 3.99 1.73 AD353P(72) 11.93 1.56 2.77 0.96 Supplementary Table 2. Colocalization of 53BP1 and γH2AX foci at radiation-induced DSBs in HMECs. 15’ pIR 30’ pIR 60’ pIR Donor N % N N % N % N N 53BP1 N N N 53BP1 γH2AX 53BP1/ γH2AX 53BP1 53BP1/ γH2AX 53BP1/ pores foci pores pores foci foci γH2AX foci foci γH2AX foci γH2AX Young donors 440 4250 2660 62.59 451 4193 3203 76.39 453 4227 3424 81.00 YD48R(16) 56 486 318 65.43 54 361 302 83.66 62 527 438 83.11 YD240L(19) 110 1245 776 62.33 108 1086 792 72.93 107 1117 903 80.84 YD168R(19) 109 988 664 67.21 110 1121 928 82.78 104 1061 877 82.66 YD184(21) 107 1023 598 58.46 92 895 669 74.75 94 787 647 82.21 YD59L(23) 58 508 304 59.84 87 730 512 70.14 86 735 559 76.05 Aged donors 467 4692 2261 48.19 464 4510 2815 62.42 411 3701 2472 66.79 AD112R(61) 106 1025 511 49.85 103 1035 665 64.25 105 933 654 70.10 AD122L(66) 105 1126 558 49.56 109 1189 677 56.94 93 894 550 61.52 AD29(68) 110 1242 575 46.30 109 1021 718 70.32 98 878 620 70.62 AD429ER(72) 92 796 374 46.98 50 396 249 62.88 64 491 330 67.21 AD353P(72) 54 503 243 48.31 93 869 506 58.23 51 505 318 62.97 www.aging-us.com 24891 AGING Supplementary Table 3. Colocalization of BRCA1 and γH2AX foci at radiation-induced DSBs in CENPF-positive (G2) HMECs. 15’ pIR 30’ pIR 60’ pIR Donor N N % N N % N N % N N N γH2AX BRCA1 BRCA1/ γH2AX BRCA1 BRCA1/ γH2AX BRCA1 BRCA1/ pores pores pores foci foci γH2AX foci foci γH2AX foci foci γH2AX Young donors 94 1107 471 42.55 86 1050 685 62.67 83 1231 830 67.42 YD48R(16) 15 108 45 41.67 16 124 81 65.32 8 74 51 68.92 YD240L(19) 28 407 178 43.73 31 460 284 61.74 30 549 376 68.49 YD168R(19) 22 256 109 42.58 18 211 135 63.98 25 367 243 66.21 YD184(21) 29 336 139 41.37 21 255 158 61.96 20 241 160 66.39 Aged donors 87 1034 445 43.04 93 1238 766 61.87 106 1439 933 64.84 AD112R(61) 28 340 139 40.88 23 321 202 62.93 32 405 255 62.96 AD29(68) 24 300 131 43.67 26 359 217 60.45 31 425 284 66.82 AD429ER(72) 26 245 108 44.08 30 373 237 63.54 29 389 250 64.27 AD353P(72) 9 149 67 44.97 14 185 110 59.46 14 220 144 65.45 Supplementary Table 4. Colocalization of BRCA1 and γH2AX foci at radiation-induced DSBs in CENPF-negative (G1) HMECs. 30’ pIR Donor N pores N γH2AX foci N BRCA1 foci % BRCA1/ γH2AX Young donors 90 896 156 17.41 YD48R(16) 31 261 46 17.62 YD240L(19) 29 351 63 17.95 YD168R(19) 30 284 47 16.55 Aged donors 86 761 317 41.66 AD112R(61) 27 268 95 35.45 AD122L(66) 26 203 100 49.26 AD429ER(72) 33 290 122 42.07 www.aging-us.com 24892 AGING Supplementary Table 5. List of antibodies used. Working dilution Antibody Host Reference IF WB Primary antibodies Anti-53BP1 Rabbit Abcam, ab21083 1:2000 1:1000 Anti-BRCA1 Mouse Abcam, ab16781, clone MS13 1:500 1:500 Anti-CENPF Rabbit Abcam, ab90, clone 14C10/1D8 1:1000 - Anti-CtIP Mouse Millipore, MABE1060, clone 14-1 1:500 - Anti-Histone H4K16ac Rabbit Active Motif, 39930 1:200 - Anti-Integrin β1 Rabbit Abcam, ab52971 - 1:1000 Anti-Ku70 Rabbit Abcam, ab83501 - 1:3000 Anti-Pericentrin Rabbit Abcam, ab4448 1:2000 - Anti-RAD51 Rabbit Abcam, ab63801 1:15000 1:2000 Anti-RPA32/RPA2 Mouse Abcam, ab2175, clone 9H8 1:500 1:1000 Anti-γH2AX (Ser139) Mouse Millipore, 05-636, clone JBW301 1:1000 - Anti-γH2AX (Ser139) Rabbit Abcam, ab81299, clone EP854(2)Y 1:500 - Secondary antibodies Anti-mouse HRP conjugate Goat Millipore, 12-349 - 1:5000 Anti-mouse Alexa Fluor® 488 Goat Jackson ImmunoResearch Inc., 115-545-205 1:500 - Anti-mouse Cyanine Cy™3 Goat Jackson ImmunoResearch Inc., 115-165-146 1:800 - Anti-rabbit HRP conjugate Goat Millipore, 12-384 - 1:5000 Anti-rabbit Alexa Fluor® 488 Goat Thermo Fisher Scientific, A-11034 1:500 - Anti-rabbit Alexa Fluor® 532 Goat Thermo Fisher Scientific, A-11009 1:1000 - Anti-rabbit Alexa Fluor® 594 Goat Thermo Fisher Scientific, A-11037 1:500 - Supplementary Table 6. Sequences of RT-qPCR primers. Target Forward primer Reverse primer 53BP1 5’-TTCCTCAACATCCTGGCTCT-3’ 5’-ACATTCCCTTGGTGCTGAAG-3’ ACTB 5’-CCAACCGCGAGAAGATGA-3’ 5’-CCAGAGGCGTACAGGGATAG-3’ GAPDH 5’-AGCCACATCGCTCAGACAC-3’ 5’-GCCCAATACGACCAAATCC-3’ SETD8 5’-TCTTGTGATTCCACCAATGC-3’ 5’-GGACAGGGTAGAAATCCGT-3’ www.aging-us.com 24893 AGING