Aspirin protects against genotoxicity by promoting genome repair

Dear Editor, Radiation sickness is a major health concern. The quest for radiation countermeasures started in the wake of the devastation witnessed following the nuclear detonations during the Second World War and has continued through the subsequent radiological accidents around the world. A radioprotector is also required for prophylactic use by staff working at radiation sources, pilots, and astronauts at high risk of space radiation, or patients undertaking lengthy radiological procedures. Despite decades of research, a safe, efficient, and cost-effective radioprotector is yet to be unveiled. Acetylsalicylic acid (aspirin) is the oldest drug in the history of medicine. It has been used for over 4000 years for the treatment of pain, inflammation, fever, and more recently for cardiovascular prophylaxis and cancer prevention. Bone marrow failure is the primary cause of mortality following irradiation. Hence, protecting the bone marrow is a primary goal in the development of radiation countermeasures. Inflammation is a key outcome and driver of irradiation-induced tissue injury. Given its anti-inflammatory effects, we inquired whether aspirin could protect against radiation. When inoculated into mice, aspirin protected against irradiation-induced bone marrow ablation (Fig. 1a; Supplementary information, Fig. S1a–d) and suppressed the induction of inflammatory genes including Ifnb1, Mx1 and Tnfa in vivo and in bone marrow-derived monocytes (BMDMos) (Supplementary information, Figs. S1e–g and S2a–c). Pattern recognition receptors (PRRs) including Toll-like receptors (TLRs), RIG-I-like receptors (RLRs), and cytosolic DNA sensors (CDS) are central to the initiation of inflammation and cell death. PRRs signal via key adaptors including MYD88 and TRIF (for TLRs), MAVS (for RLRs), and STING (for CDS) (Supplementary information, Fig. S3a). To assess the impact of aspirin on PRR pathways, we stimulated BMDMos with specific agonists for PRRs including TLRs (TLR2: Pam3CSK4, TLR3: Poly(I:C)), RIG-I (Poly(I:C) transfection), cGAS-STING (poly(dA:dT) or cGAMP transfection) and AIM2 inflammasome (poly(dA:dT)). We found that aspirin inhibited inflammatory gene induction via all these PRRs (Supplementary information, Fig. S3b–f) but not the AIM2 inflammasome (Supplementary information, Fig. S4a–c). To assess if radioprotection by aspirin was due to suppression of PRR-driven inflammation, we compared wild-type (WT) mice with those defective in PRR signaling. Similar to WT (Fig. 1a; Supplementary information, Fig. S1a–d), aspirin protected the bone marrows of triple knockout (TKO) mice that are defective in both TLR and RLR pathways (Myd88TrifMavs) (Fig. 1b; Supplementary information, Fig. S5a–d), or those defective in cytosolic DNA sensing (Sting) against irradiation (Fig. 1c; Supplementary information, Fig. S6a–d), and suppressed inflammatory gene expression (Supplementary information, Figs. S5e–g and S6e–g). This implied that observed bone marrow suppression was independent of PRR-driven inflammation and that radioprotection by aspirin was uncoupled from its anti-inflammatory effects. Double-stranded DNA breaks (DSBs) are the most deleterious outcomes of irradiation. Micronuclei are key aftereffects of DSBs. HEK293 cells are defective in PRR signaling and lack prostaglandinendoperoxide synthases (COX1 and COX2)— also key mediators of inflammation and pain, and the best-known targets of aspirin. Aspirin suppressed irradiation-induced micronuclei generation in HEK293 cells (Supplementary information, Fig. S7a, b), indicating that such effect was independent of its anti-inflammatory activity. When irradiated on ice (to prevent spontaneous repair), then transferred from ice to 37 °C to allow DNA repair to occur, aspirin pre-treated cells repaired DSBs faster (Fig. 1d, e). Aspirin also accelerated the repair of DSBs induced by the anti-cancer drug doxorubicin (Supplementary information, Fig. S8a, b). DSB repair occurs via homologous recombination (HR) and NonHomologous End Joining (NHEJ). Results from GFP-based reporter systems revealed that aspirin promotes the HR but not the NHEJ (Supplementary information, Fig. S9a–e). BRCA1 and 53BP1 are key checkpoint proteins for HR and NHEJ repair, respectively. Aspirin enhanced recruitment of BRCA1 but not the NHEJ repair protein 53BP1 to DNA damage sites (Fig. 1f, g; Supplementary information, Fig. S10). Accordingly, deletion of BRCA1 significantly blunted acceleration of DSB repair by aspirin (Supplementary information, Fig. S11). In contrast, ablation of 53BP1 (Supplementary information, Fig. S12) or inhibition of the NHEJ kinase DNAPKc did not (Supplementary information, Fig. S13). Chromatin decompaction is essential for the recruitment of DNA repair machinery to damage sites. The N-terminal tail of histone H4 is central for inter-nucleosome interaction (Supplementary information, Fig. S15a). Acetylation of histone H4 at lysine K16 (Ac-H4K16) is vital for decreasing the nucleosome–nucleosome stacking and chromatin folding, to permit the recruitment of repair proteins. AcH4K16 also supports the preferential recruitment of BRCA1 over 53BP1 to damage sites, thereby tipping the balance towards HR. Aspirin-treated cells exhibited elevated Ac-H4K16 and recruitment of BRCA1 but not 53BP1 to DNA damage sites (Fig. 1h, I; Supplementary information, Fig. S10a–d). Consistent with direct donation of acetyl groups to targets, aspirin increased Ac-H4K16 in cells treated with the histone acetyltransferase inhibitor (Supplementary information, Fig. S14a, b) or when incubated directly with chromatin isolates (Supplementary information, Fig. S14c). In contrast, acetylated H3K27 was already high at steady state and remained largely unaltered following aspirin (Fig. 1j). Conceivably, given its location at the N-terminal tail of histone H4 where it functions as the first contact point anchoring the H4 tail on the adjacent nucleosome (Supplementary information, Fig S15a), H4K16 is likely more accessible for direct acetylation by aspirin. To examine whether aspirin modulates chromatin compaction, we employed the AO3 cells containing genomic insertions of


Microscopic visualization of chromatin compaction
The AO3 reporter cells 9 cultured in a 1:1 mixture of DME/Ham's F12 medium supplemented with antibiotics and 20% FCS to 70% density were transfected by lipofectamine with the 1 µg/ml mCherry-LacR-stop plasmid 8 . After 4 hours, they were treated with DMSO or indicated concentrations of aspirin (1 or 2 mM). 18 hours later samples were fixed with 4% paraformaldehyde and analyzed my fluorescence microscopy as described previously 10 .

HR and NHEJ reporter assays
To assess the effect of aspirin on homologous recombination (HR) and NHEJ repair, briefly, the pHPRT-DRGFP (HR-reporter plasmid) 6 and the pimEJ5GFP (NHEJ reporter plasmid) 7 were stably transfected into HEK293T cells. 0.5 × 10 6 HEK293T stable reporter cells seeded in 6-well plates were transfected with 2 μg HA-I-SceI expression plasmid (pCBASce) then treated with aspirin or DMSO. 48 hours later, cells were analyzed by flow cytometry for GFP expression. Standard Mean of Error (±SEM) was calculated from three independent experiments.

Immunofluorescence
Cells were seeded and cultured on glass coverslips in 12 well plate and fixed in 4% paraformaldehyde (PFA) in PBS for 20 min at room temperature. Cells were permeabilized in 0.5% Triton X-100 for 10 min, blocked in 5% normal goat serum (NGS) then incubated with primary antibodies diluted in 1% NGS overnight at 4 °C, followed by incubation with indicated secondary antibodies diluted in 1% NGS at RT for 1 h then finally stained with DAPI for 15 min at room temperature. Coverslips were mounted using Dako Fluorescence Mounting Medium (Agilent) and imaged using Nikon confocal microscope (Eclipse C1 Plus). All scoring was performed under blinded conditions. γH2A.X, BRCA1, and 53BP1 foci were counted from 40 microscopic fields containing approx. 300 cells from 3 independent experiments.

Chromatin fractionation and immunoblotting
To isolate the chromatin, we used the Subcellular Protein Fractionation Kit (Thermo Fisher) according to the manufacturer's instructions and as previously described 11 . Proteins were quantified by BCA reagent (Thermo Fisher Scientific, Rockford, IL). Samples were resolved in SDS-PAGE, transferred to nitrocellulose membrane (Amersham Protran 0.45 μm NC) and immunoblotted with specific primary antibodies followed by HRP-conjugated secondary antibodies. Protein bands were detected by Supersignal West Pico or Femto Chemiluminescence kit (Thermo Fisher Scientific).

Inflammasome activation analysis
Analysis of inflammasome activation was done as previously described [12][13][14] . Briefly, BMDMos seeded in the density of 1.5 x 10 6 cells/well were treated with aspirin overnight and then primed with 500 ng/ml LPS for 4 h. Cells were then transfected with 1 μg/ml poly(dA:dT) for 1 h using Lipofectamine 2000 (Invitrogen). Supernatants were collected. Proteins were precipitated using chloroform:methanol extraction and re-suspended in 2 x Laemmli buffer. Cells were lysed in 2 x Leammli buffer. Samples were separated on 13.5% SDS-PAGE gel and analysed for activation of Caspase-1 and IL-1β by immunoblotting, as described in the section above.

In vitro protein acetylation assay by aspirin
Chromatins fractions isolated as described above were incubated with indicated concentration of aspirin in reaction buffer (40 mM Tris-HCl, 5 mM MgCl2, 100 mM NaCl) for 1 hour at 37 °C. The mixture was boiled in loading buffer and analyzed by immunoblotting.

Analysis of DNA repair by Comet assay
Cells were subjected to the indicated doses of γ-irradiation or doxorubicin and chromosome fragmentation was determined by comet assay as previously described 10,11,15 . Briefly, during irradiation cells were kept on ice to stop the DNA repair process. Thereafter, cells were transferred to 37°C to allow DNA repair to occur for indicated duration. Cells were then harvested by brief centrifugation and resuspension in cold PBS. Cells were mixed with 1% low-melting agarose (40°C) at a ratio of 1:3 vol/vol) before pipetting onto CometSlides. Slides were then immersed in prechilled lysis buffer (1.2 M NaCl, 100 mM EDTA, 0.1% sodium lauryl sarcosinate, 0.26M NaOH PH>13) for overnight (18-20 h) lysis at 4°C in the dark. Slides were carefully removed and submerged in room temperature rinse buffer (0.03 M NaOH and 2 mM EDTA, pH > 12) for 20 min in the dark. This washing step was done 2 times. Slides were transferred to a horizontal electrophoresis chamber containing rinse buffer and separated for 25 min at voltage (0.6 V/cm). Finally, slides were washed with distilled water and stained with 10 μg/ml propidium iodide and analyzed by fluorescence microscopy. 20 fields with about 200 cells in each sample were evaluated and quantified by the Fiji software to determine the tail length (tail moment).

Determination of micronuclei
HEK293 cells pre-treated with aspirin (1 mM) then exposed to γ-irradiation (or not) were cultured for 24 hours, then fixed (4% PFA), permeabilized (0.5% Triton X-100), DAPI stained then analyzed by microscopy as described previously 10,11,15 . Micronuclei were defined as discrete DNA aggregates separate from the primary nucleus in cells where interphase primary nuclear morphology was normal. Cells with an apoptotic or necrotic appearance were excluded.

RT-qPCR
Total RNA was extracted using the Trizol (Thermo Fisher) according to the manufacturer's protocol. cDNA was prepared using Maxima H Minus First Strand cDNA Synthesis Kit and random oligomer primers (Thermo Fisher Scientific). Real-time qPCR was performed by using QuantStudio 5. The results were normalized to 18s (reference gene) and expressed as fold change relative to untreated or mock-treated controls using the comparative CT method (ΔΔCT). The following TaqMan Gene Expression Assays (FAM) (Applied Biosystems, Thermo Fisher Scientific) in combination with the TaqMan Gene Expression Master Mix (#4369016; Applied Biosystems, Thermo Fisher Scientific) were applied: Ifnβ (Mm00439552_s1), Mx1 (Mm00487796_m1), Tnfα (Mm00443258_m1) and Rn18s (Mm03928990_g1).

Statistical Analysis
Statistical analysis was performed by GraphPad Prism 5.0 software. All of the data shown in the histograms were the results of at least three independent experiments and are presented as the mean ± SEM or mean ± SD. The sample size (n) for each statistical analysis and statistical methods used to assess significant differences are indicated in figure legends. Differences between values were considered statistically significant when * P < 0.05, * * P < 0.01, * * * P < 0.001, and * * * * P < 0.0001.