Induction of bona fide DNA genomic amplifications in cancer cells exposed to ionizing radiation or other DNA double strand break-inducing agents.

a, Flow cytometry profile showing the increase in genomic content greater than G2/M (highlighted in red) in FaDu cells exposed to increasing doses of IR. b, Histogram showing the extent of genomic amplifications in Cal27 and FaDu HNSCC cell lines as well as the h-TERT immortalized keratinocytes (OKF) in response to increasing doses of IR. Genomic amplifications were examined 48 h post-IR. c, (left) Representative images of DAPI-stained Fadu nuclei following exposure to 9 Gy. (right) Quantitation of the nuclear size of FaDu cells 72 h following the exposure to 9 Gy. d, Genomic amplifications in HCT116 and p53-null (p53-/-) cells at various doses of IR measured 72 h post-IR. e, BrdU-7AAD FACS profiles of FaDu cells exposed to 9 Gy and pulse-labelled with BrdU for 1 h, 48 h post-IR. f, Genomic amplifications in U2OS cells exposed to 9 Gy, and treated with or without nocodazole at 24 h post-IR. Genomic amplifications were examined 48 h post-IR. g, Experimental workflow of the various treatments of U2OS cells for the detection of IR-induced genomic amplifications by DNA isotope labelling and CsCl ultracentrifugation and fractionation. h, Plot showing the DNA content of each CsCl fraction in cells treated according to the workflow in g. i. Histogram showing the relative amount of DNA detected in the fractions corresponding to non-replicating (L:L), singly replicated (H:L), and doubly replicated (H:H; doubly replicated) DNA as determined by measuring the area under the curve (shown in h) from three independent experiments (n=3). j, Histogram showing the percentage of U2OS cells with rereplication 48 h following the exposure 9 Gy with or without aphidicolin (Aph) treatment according to the schematic shown on top. k, The percentage of U2OS cells with genomic amplifications following the expression of AsiSI restriction endonuclease by 4-OHT treatment for 72 h. l, Histogram showing the percentage of the indicated melanoma cell lines with genomic amplifications measured at 48 and 72 h following exposure to 9 Gy (n=3). Correlation and statistical significance of the induction of genomic amplifications both at 48 and 72 h post-IR and IR-induced cytotoxicity in all the melanoma lines shown, defined as surviving fraction at 10% of the parental population (SF10), is shown on the top-right corner. Right: Representative clonogenic survival curves of a subset of the melanoma lines used. Data in all figures represent the average of three independent experiments ± S.D. p-values were calculated using Student’s t-test: *p < 0.05, **p < 0.01, ***p < 0.001.

DIGA is not mediated by ORC1-2/CDT1-dependent origin licensing, and is suppressed by SET8/SUV4-20H1.

a, Schematic of the various molecules involved in origin licensing in mammalian cells in which SET8 and CDT1 are targeted for degradation by CRL4CDT2 during S-phase to suppress origin re-licensing. b, Immunoblots of CDT1 and SET8 in Cal27 cells exposed to IR (4 Gy) and analysed at the indicated time points post-IR (4 Gy). c, (top) immunoblotting of CDT1 and SET8 in Cal27 following treatment with the indicated siRNAs. (bottom) Histogram showing the percentage of control (siGL2) Cal27 cells or Cal27 cells with genomic amplifications following the depletion of CDT1 or SET8 by siRNA (for 48 h) and subsequent exposure to 4 or 9 Gy (analyzed 48 hr post irradiation). d, Immunoblot showing the impact of CRISPR/Cas9-mediated deletion of SUV4-20H1 or SUV4-20H2 in U2OS cells on histone H4K20 methylation. e, Histogram showing the percentage of U2OS cells undergoing genomic amplifications following the deletion of SUV4-20H1 or SUV4-20H2 (two independent clones each are shown). f, The impact of ORC1 or ORC2 deletion in HCT116 on IR-induced genomic amplifications. The histogram shows the percentage of control HCT116 or HCT116 deleted of ORC1 (1ORC1) or ORC2 (1ORC2) undergoing genomic amplifications 72 h following the exposure to the indicated doses of IR. Data in the histograms represent the average of three independent experiments ± S.D. ns: non-significant, *p < 0.05, **p < 0.01, ***p < 0.001.

The impact of various proteins involved in the repair of DSBs on IR-induced genomic amplifications.

a, Histogram (right) showing the impact of increasing IR doses on the induction of genomic amplifications in control U2OS cells or two independent clones with deletion of DNA-PKcs by CRISPR/Cas9. (left) Western blots illustrating the efficiency of deletion of DNA-PKcs in U2OS cells. b, Histogram showing the percentage of U2OS cells with genomic amplifications 72 h following the exposure to 9 Gy of control, DMSO-treated U2OS cells or U2OS cells treated with the DNA-PKcs inhibitor NU7441 for 24 h prior to IR. c, Immunoblot showing deletion of effectors of the c-NHEJ pathway (LIG4, XRCC4, XLF). d, Histogram showing the impact of deletion of the indicated effectors of c-NHEJ on the induction of genomic amplifications by IR. e, The impact of inhibiting DNA LIG4 on the induction of genomic amplifications by IR in U2OS cells. f, g, Deletion of KTM4B (coding for SUV4-20H1) but not deletion of KTM4C (coding for SUV4-20H2) inhibits the recruitment of 53BP1 to DSBs. f, Representative immunofluorescent images of 53BP1 showing formation of 53BP1 foci following exposure of control (pX330) U2OS cells or U2OS cells deleted of KTM4B or KTM4C to 5 Gy analysed 1 h post-exposure. g, Quantitation of the number of 53BP1 foci per cell in control (pX330) cells or U2OS deleted of KTM4B or KTM4C (two independent clones each). The results represent the average of a minimum of 100 nuclei in each of three independent experiments for each condition ± S.D. ns: non-significant, ***p < 0.001. h, i, Histograms showing the percentage of cells undergoing genomic amplifications following the exposure of control or ATM-deleted U2OS cells (inset: immunoblotting of ATM levels) to 9 Gy (h), or following pharmacological inhibition of ATM in parental U2OS cells by KU55933 24 h prior to IR (i). Genomic amplifications was monitored 72 h post-IR. j, Histogram showing the percentage of control (pX330) U2OS cells or U2OS cells deleted of RNF8 or RNF168 and exposed to 9 Gy. Genomic amplifications was monitored 72 h post-IR. k, l, Histograms showing the percentage of control (pX330) U2OS cells or U2OS cells deleted of 53BP1 (in two independent clones; immunoblots shown in inset (k)) or RIF1 (immunoblots shown in inset (l)) with genomic amplifications following the exposure to 9 Gy. Genomic amplifications was monitored 72 h post-IR. Data in all the histograms represent the average of three independent experiments ± S.D. *p < 0.05, **p < 0.01, ***p < 0.001.

Unprotected, hyper-resected DSBs promote DIGA in mammalian cells through BIR-like mechanism.

a, b, Histogram (b) showing the impact of IR (9 Gy) on the induction of genomic amplifications in control U2OS cells or in U2OS cells depleted of CtIP or EXO1 by siRNA (immunoblot shown in (a)). c, Quantitation of the number of 53BP1 foci per cell in control (pX330) cells or U2OS deleted of SHILD1, SHILD2, or SHILD3. Representative images are shown in the Extended Data Fig. 9a. The results represent the average of a minimum of 100 nuclei in each of three independent experiments for each condition ± S.D. ns: non-significant. d, Histogram showing the percentage of cells with genomic amplifications following the exposure of control (pX330) U2OS cells or individual independent clones of U2OS cells lacking SHILD1, SHILD2, or SHILD3 to 9 Gy. Genomic amplifications were monitored 72 h post IR. Representative FACS profiles shown in Extended Data Fig. 9b. e, Induction of genomic amplifications by IR requires RAD51. Representative FACS (PI) profiles of control U2OS cells or U2OS cells depleted of RAD51 by two different siRNAs and exposed to 9 Gy. Genomic amplifications were monitored 72 h post IR. f, Dose-dependent suppression of strand invasion following incubation of U2OS cells with the RAD51 recombinase inhibitor B02. Representative FACS profiles shown in Extended Data Fig. 9c. g-i, Induction of genomic amplifications by IR requires the break-induced replication factors RAD52, POLD3 and POLD4. Representative FACS (PI) profiles of control U2OS cells or U2OS cells depleted of RAD52 (g), POLD3 (h) or POLD4 (i) by two different siRNAs each and exposed to 9 Gy. Genomic amplifications were monitored 72 h post IR. j, k, Histogram (k) showing the percentage of U2OS cells undergoing genomic amplifications when analysed at 72 h following the exposure of U2OS synchronized at various stages of the cell cycle to 9 Gy. Cells were irradiated at Early (E), Mid (M) or late (L) S-S-phase of the cell cycle following the synchronization by DTB and release (depicted in the schematic in (j; top) or in early (E) or Mid (M) G1 phase of the cell cycle following the synchronization in prometaphase by nocodazole and release (depicted in the schematic in (j; bottom)). Genomic amplifications in all cases was monitored 72 h following the exposure to IR. The synchronization PI-FACS profiles and representative rereplication PI-FACS profiles are shown in Extended data Fig. 10. h, Data in all the histograms represent the average of three independent experiments ± S.D. *p < 0.05, **p < 0.01, ***p < 0.001.

Induction of genomic amplifications in cancer cells by ionizing radiation.

a, Representative flow cytometry (FACS) profiles (PI staining) showing the increase in genomic content greater than G2/M (highlighted in red) in FaDu cells exposed to 0.1 μM MLN4924 or 9 Gy, and examined 72 h post-treatment. b, Representative flow cytometry profiles of the indicated cancer cell lines showing the extent of rereplication as determined by PI-FACS 48 h following exposure to 9 Gy. c, Histogram showing the percentage of the indicated cancer cell lines with genomic amplifications 48 h following exposure to increasing doses of IR. Data represent the average of three independent experiments ± S.D. d, The impact of p53 on the percentage of HCT116 colon cancer cells exhibiting rereplication 48 h after treatment with 0.3 μM MLN4924. Data represent the average of three independent experiments ± S.D. *p < 0.05, ***p < 0.001.

Time-dependent accumulation of heterogenous (sizes) genomic amplifications following the exposure to IR.

a, time-dependent increases in genomic amplifications following the exposure of U2OS cells to 9 Gy. Top Flow cytometry profile alignment to show the impact of IR on the increase of >G2/M DNA content as a function of time. b, Representative flow cytometry profile showing the impact of IR (9G) on cell cycle distribution as a function of time, as indicated. Areas of genomic amplifications (with Percentage) are shown in red. c, Flow cytometry profile of U2OS depicting the gradual increase in DNA copy number (as measured by increased fluorescence (showing divisions to be used as cut offs for quantitation of copy # size in d and e) following the exposure of U2OS to 9 Gy. d, e, Histograms showing the increase in genomic amplifications (stratified by size in c) in U2OS cells exposed to increasing doses of IR (as indicated) and measured 48 (d) and 72 (e) hours post-IR exposure. Data represent the average of three independent experiments ± S.D. *p < 0.05, ***p < 0.001.

DNA double strand breaks (DSBs) induce bona fide genomic amplifications in cancer cells.

a, FACS profiles showing de novo DNA synthesis (BrdU incorporation) in U2OS cells with greater than G2/M DNA content (as assessed by 7AAD staining) 48 or 72 h following exposure to 9 Gy. Cells were pulsed with BrdU for 1 h prior to harvest. b, Representative FACS profiles (PI staining) of U2OS cells 48 h following exposure to 9 Gy with or without nocodazole (Noc) treatment at the 24-48 h time point following IR exposure. Cells with genomic amplifications highlighted in red. Quantitation is shown in Fig. 1f. c, Elisa-based quantitation of BrdU incorporation in the peak fractions of DNA (50 ng each of L:L, H:L, and H:H DNA) isolated from the CsCl ultracentrifugation gradient shown in Fig. 1h. d, Workflow (top) and representative FACS profiles (PI staining; bottom) of U2OS cells left untreated or exposed to IR for 48 h and treated with or without aphidicolin (Aph) added immediately following IR and washed 24 h after (0-24 h) or added 24 h post-IR (24-48 h). Cells with genomic amplifications are highlighted in red, and quantitation of the results is shown in Fig. 1j. e, The appearance of H:H DNA in cells exposed to IR is aphidicolin-sensitive. The line histogram is an extension of the plot shown in Fig. 1h, and shows additional treatment with aphidicolin (Aph) with or without IR treatment as depicted in the experimental workflow shown on top. f, Histogram showing the percentage of U2OS cells with genomic amplifications (as determined by PI-FACS) 72 h following treatment with etoposide (1 μg ml-1), doxorubicin (0.1 μM), or exposure to ultraviolet radiation (UV; 100 J m-2). Data represent the average of three independent experiments ± S.D. g, Work flow (top) and quantitation of the percentage of AsiSI-ER-U2OS cells with genomic amplifications (as determined by PI-FACS) following the induction of DSBs by the addition of 4-OHT for 48 h. Cells were treated with or without aphidicolin (Aph) added together with 4-OHT and washed 24 h after (0-24 h), or with Aph added 24 or 48 h following treatment with 4-OHT and harvested 24 h after Aph (48 and 72 h following treatment with 4-OHT, respectively). Data represent the average of three independent experiments ± S.D. **p < 0.01. h, Representative FACS profiles showing de novo DNA synthesis (BrdU incorporation) in AsiSI-ER-U2OS cells with greater than G2/M DNA content (as assessed by 7AAD staining) 48, 72, or 96 h following treatment with 300 nM 4-hydroxy-tamoxifen (4-OHT). Cells were pulsed with BrdU for 1 h prior to harvest.

Induction of genomic amplifications in melanoma cells by IR correlates with IR-induced toxicity but not with melanoma cell proliferation.

a, Histogram showing the relative growth rate of the indicated melanoma cells determined by cell counting. Data represent the average of three independent experiments ± S.D. b, c, Histograms showing the correlation between the percentage of melanoma cells (13 cell lines) with genomic amplifications induced 48 h (b) or 72 h (c) following exposure to 9 Gy in three independent experiments (Fig. 1l), and the doubling time of the various melanoma cells determined by cell counting in three independent growth curves. d, e, Histograms showing the correlation between the percentage of melanoma cells (13 cell lines) with the extent to which genomic amplifications are increased as determined by PI-FACS at 48 h (d) or 72 h (e) following exposure to 9 Gy in three independent experiments (shown in Fig. 1l), and the surviving fraction (at 10% of the population) determined from clonogenic survival curves in three independent experiments.

Rereplication induction by MLN4924 is CDT1-dependent, and IR-induced genomic amplifications are suppressed by SUV4-20H1.

a, CDT1 is critical for rereplication induction by MLN4924. The Histogram shows the percentage of cells with rereplication (assessed by PI-FACS) following the treatment of control U2OS cells (si-Gl2) and U2OS cells depleted of CDT1 for 72 h prior to treatment with MLN4924 for 48 h. Data represent the average of three independent experiments ± S.D. ***p < 0.001. b, SUV4-20H1 suppresses the induction of genomic amplifications in VMM39 melanoma cells. The histogram shows the percentage of control (pX330) VMM39 cells or VMM39 cells with KMT5B deletion with genomic amplifications determined by PI-FACS 48 h post-exposure to 9 Gy. Data represent the average of three independent experiments ± S.D. **p < 0.01.

The impact of deletion or pharmacological inhibition of various proteins involved in DSB repair on the induction of genomic amplifications by IR.

a-c. DNA-PKcs suppresses IR-induced genomic amplifications in cancer cells. a, Deletion of PRKDC (coding for DNA-PKcs) in 293T cells stimulates the induction of genomic amplifications by IR. The histogram shows the percentage of cells with genomic amplifications (as determined by PI-FACS) in control (pX330) 293T cells or 293T deleted of DNA-PKcs (two independent clones; shown in the immunoblot; inset) following exposure to the indicated doses of IR. Genomic amplifications were assessed 48 h post-IR. b, c, Pharmacological inhibition of DNA-PKcs stimulates the induction of genomic amplifications by IR in 293T cells (b) or DM93 melanoma cells (c). The histograms show the percentage of cells undergoing genomic amplifications (as determined by PI-FACS) 48 h (293T cells) or 72 h (DM93 cells) following exposure to 5 Gy (293T cells) or 9 Gy (DM93 cells). Where indicated, the cells were pre-treated with the DNA-PKcs specific inhibitor NU7441 for 24 h prior to IR exposure. d, Histogram showing the percentage of control U2OS cells (pX330) or U2OS cells deleted of LIG4 or XRCC4 (immunoblots shown in Fig. 3c) undergoing genomic amplifications (as determined by PI-FACS) 72 h following treatment with control DMSO or 0.1 μM MLN4924. e, Histogram showing the percentage of control U2OS cells (pX330) or U2OS cells deleted of LIG1 undergoing genomic amplifications 72 h following exposure to 9 Gy. f, Histogram showing the percentage of control (pX330) U2OS cells or U2OS cells deleted of POLQ (three independent clones; shown in the immunoblot; inset) undergoing genomic amplifications as determined by PI-FACS 72 h following the exposure to 9 Gy. Data in all histograms represent the average of three independent experiments ± S.D. ns: non-significant, *p < 0.05, **p < 0.01, ***p < 0.001.

DIGA, but not rereplication induced by deregulated origin licensing, is stimulated by the deletion of factors involved in the recruitment of 53BP1 to DSB sites.

a, ATM suppresses the induction of genomic amplifications by DSBs. The histogram shows the percentage of AsiSI-ER-U2OS cells with genomic amplifications as determined by PI-FACS 72 h following treatment with 4-OHT. Where indicated, the cells were pre-treated with the ATM specific inhibitor KU3355 5 or 24 h prior to treatment with 4-OHT. The results represent the average of three independent experiments ± S.D. ***p < 0.001. b, ATM inhibition does not stimulate rereplication induction by MLN4924. The histogram shows the percentage of control (pX330) U2OS cells or U2OS cells with ATM deletion (1ATM) with rereplication as determined by PI-FACS 48 h following treatment with 0.1 μM MLN4924. The results represent the average of three independent experiments ± S.D. ns: non-significant. c, d, Deletion of RNF8, RNF168, or RIF1 suppresses 53BP1 recruitment to DSBs. c, Representative immunofluorescence images of 53BP1 showing formation of 53BP1 foci following exposure of control (pX330) U2OS cells or U2OS cells deleted of RNF8, RNF168, or RIF1 to 5 Gy and analysed 1 h post-exposure. d, Quantitation of the number of 53BP1 foci per cell in control (pX330) cells or U2OS deleted of RNF8, RNF168, or RIF1. The results represent the average of a minimum of 100 nuclei in each of three independent experiments for each condition ± S.D. **p < 0.01, ***p < 0.001. g, MDC1 and RNF168 suppress IR-induced genomic amplifications. Histograms showing the percentage of control (pX330) 293T cells or 293T cells deleted of MDC1 or RNF168 with genomic amplifications as determined by PI-FACS 48 h following the exposure to 5 Gy. The results represent the average of three independent experiments ± S.D. **p < 0.01, ***p < 0.001.

53BP1 suppresses IR-induced genomic amplifications but does not impact rereplication induction by deregulated origin licensing.

a, Deletion 53BP1 in 293T cells enhances the induction of genomic amplifications by IR. The histogram shows the percentage of control (pX330) 293T cells or 293T deleted of 53BP1 (two independent clones; shown in the immunoblot; inset) undergoing genomic amplifications as determined by PI-FACS following exposure to the indicated doses of IR. Genomic amplifications were assessed 48 h post-IR exposure. b, Deletion of 53BP1 in VMM39 cells enhances the induction of genomic amplifications by IR. The histogram shows the percentage of control (pX330) VMM39 melanoma cells or VMM39 cells deleted of 53BP1 (two independent clones) undergoing genomic amplifications as determined by PI-FACS 72 h following exposure to 9 Gy. c, Deletion of 53BP1 does not stimulate rereplication induction by MLN4924. The histogram shows the percentage of control (pX330) U2OS cells or U2OS cells with 53BP1 deletion (153BP1) with rereplication as determined by PI-FACS 48 h following treatment with 0.1 μM MLN4924. Data in all histograms represent the average of three independent experiments ± S.D. ns: non-significant, *p < 0.05, **p < 0.01, ***p < 0.001.

The shieldin complex and RAD51 protect cancer cells against IR-induced genomic amplifications.

a, b, Deletion of SHLD1, SHLD2, or SHLD3 in U2OS cells does not impact the recruitment of 53BP1 to DSBs. a, Representative immunofluorescence images of 53BP1 showing formation of 53BP1 foci following exposure of control (pX330) U2OS cells or U2OS cells deleted of SHLD1, SHLD2, or SHLD3 to 5 Gy and analysed 1 h post-exposure. Quantitation of the number of 53BP1 foci per cell in control (pX330) cells or in U2OS deleted of SHLD1, SHLD2, or SHLD3 is shown in Fig. 4c. b, Representative flow cytometry profiles show the increase in genomic content greater than G2/M (highlighted in red) in control (pX330) U2OS cells or in cells deleted of SHLD1, SHLD2, or SHLD3 (multiple independent clones shown) exposed to 9 Gy and examined 72 h post-IR. Quantitation of the data is shown in Fig. 4d. c, The induction of genomic amplifications by IR requires the recombinase activity of RAD51. Representative FACS (PI) profiles of U2OS cells treated with DMSO or increasing concentrations of the RAD51 specific inhibitor B02, and exposed to 9 Gy. Genomic amplifications (highlighted in red, as determined by PI staining) were assessed 72 h post-IR exposure. Quantitation of the results is shown in Fig. 4f.

The impact of cell cycle on rereplication induction by IR.

a, b, Experimental workflow (a) for the synchronization of U2OS cells at the G1/S by DTB and subsequent release. Cells were either harvested for PI-FACS analysis (b; top panel) at the indicated time points or exposed to 9 Gy at these time points and harvested 72 h post-IR (b; bottom panel) to monitor genomic amplifications by PI-FACS. c, d, Experimental workflow (c) for the synchronization of U2OS cells in prometaphase by nocodazole treatment (0.5 μg ml-1) and subsequent release. Cells were either harvested for PI-FACS analysis (c; top panel) at the indicated time points or exposed to 9 Gy at these time points and harvested 72 h post-IR (d; bottom panel) to monitor genomic amplifications by PI-FACS. Cells with genomic amplifications and their percentages are highlighted in red. Representative FACS profiles of each sample (in b and d) are shown. Quantitation from three independent synchronization and treatment experiments is shown in Fig. 4k.