ATM suppresses c-Myc overexpression in the mammary epithelium in response to estrogen

Abstract

ATM gene mutation carriers are predisposed to estrogen-receptor-positive breast cancer (BC). ATM prevents BC oncogenesis by activating p53 in every cell; however, much remains unknown about tissue-specific oncogenesis after ATM loss. Here, we report that ATM controls the early transcriptional response to estrogens. This response depends on topoisomerase II (TOP2), which generates TOP2-DNA double-strand break (DSB) complexes and rejoins the breaks. When TOP2-mediated ligation fails, ATM facilitates DSB repair. After estrogen exposure, TOP2-dependent DSBs arise at the c-MYC enhancer in human BC cells, and their defective repair changes the activation profile of enhancers and induces the overexpression of many genes, including the c-MYC oncogene. CRISPR/Cas9 cleavage at the enhancer also causes c-MYC overexpression, indicating that this DSB causes c-MYC overexpression. Estrogen treatment induced c-Myc protein overexpression in mammary epithelial cells of ATM-deficient mice. In conclusion, ATM suppresses the c-Myc-driven proliferative effects of estrogens, possibly explaining such tissue-specific oncogenesis.

Keywords: CP: Cancer; CP: Cell biology; DNA double-strand break repair; ataxia-telangiectasia (ATM); breast cancer; c-MYC; enhancer; estradiol; topoisomerase II.

Figure 1.. E2 is genotoxic to mammary epithelial cells in Atm-deficient mice

(A) Cross-section of mammary ducts of female B6 mice carrying the indicated genotypes. Representative images of 53BP1-focus-positive (53BP1⁺) mammary epithelial cells at 6 h after an i.p. injection of E2 or solvent. Cytokeratin-8 (CK8) is a marker of epithelial cells. Scale bar, 50 μm. (B–D) Percentage of 53BP1⁺ epithelial cells at 6 h (B) or indicated hours (C and D) after an i.p. injection of E2 into B6 (B), B6;129 (C), and ATMi-treated B6 mice (D) carrying the indicated genotypes. ATMi was injected together with E2. Examples of histological images for (C) and (D) are shown in Figures S1B and S1C, respectively. Data (B–D) represent mean ± standard deviation from triplicates. ***p < 0.005, unpaired two-tailed t test.

Figure 2.. ATM promotes the repair of E2-induced TOP2-dependent DSBs in human BC cells

(A) The experimental protocol to analyze E2-induced DSBs in (B) and (C). Serum-starved cells were exposed to E2 and ATMi for 2 h. ‘Before’ indicates time before the exposure, and ‘2 h’ and ‘22 h’ indicate hours for which cells were incubated in E2-free media with or without ATMi after the exposure. (B) Representative image of 53BP1 foci in wild-type and ATM^−/− MCF-7 cells at G₁ phase. Analysis was performed before (left) and after 2 h of E2 exposure (middle) and after 2 h of additional incubation with E2-free media (right). Scale bar, 25 μm. (C) Percentage of G -phase 53BP1⁺ 1 MCF-7 cells (≥10 foci/cell) carrying the indicated genotypes. Data replotted in boxplots are shown in Figure S2C. (D) The inhibitory effect of fulvestrant on E2-induced 53BP1-focus formation in ATM^−/− cells at ‘2 h’ in (A). Data replotted in boxplots is shown in Figure S2G. Data (C and D) are mean ± standard deviation from triplicates. *p < 0.05 and ***p < 0.005, unpaired two-tailed t test.

Figure 3.. ATM promotes the removal of 5′ TOP2 adducts from DSB ends in the G₁ phase by phosphorylating CtIP at T847/T859

(A and C) The DSB repair kinetics of G₁-phase TK6 (A) and MCF-7 (C) cells after a 0.5-h pulse exposure to etoposide. Cells were cultured in etoposide-free medium after the pulse exposure. Percentage of 53BP1⁺ cells was measured (≥5 foci/cell and ≥10 foci/cell for TK6 and MCF-7 cells, respectively). Representative images of TK6 cells are shown in Figure S3C. (B) Proficient repair of ‘clean’ DSBs generated by the AsiSI restriction enzyme in ATM^−/− cells but not in NHEJ-deficient LIG4^−/− cells. Cells expressing AsiSI fused with ER were treated with 4-OHT for 4 h for DSB induction. The 53BP1 foci were counted in G₁-phase TK6 cells at 0 h and 4 h after the removal of 4-OHT. The average number of foci in untreated cells was subtracted and subtracted values for 0 h were set as 100%. (D) Representative dot blot analysis of TOP2ccs in TK6 cells of the indicated genotypes treated with etoposide for 2 h. Genomic DNA (50 μg) was separated by cesium chloride gradient ultracentrifugation, and individual fractions were blotted onto polyvinylidene fluoride filters followed by dot blotting using an α-TOP2β antibody (see Figure S3D). The first and the second fractions represent free TOP2. The third fraction contains stalled TOP2ccs having intact TOP2. (E–I) Quantification of TOP2ccs. The whole stalled TOP2ccs (E, F, and H) were quantified for the indicated genotypes. Representative blots are shown in (D), Figures S3E and S3M, respectively. The x axis shows the number of etoposide-induced stalled TOP2ccs relative to that in wild-type. Stalled TOP2 having degraded TOP2 was quantified in (G) and (I). Representative western blots were shown in Figures S3E and S3M, respectively. The x axis shows the percentage of degraded TOP2ccs relative to the whole stalled TOP2ccs in (F) and (H), respectively. Data (A–C and E–I) represent mean ± standard deviation from triplicates. **p < 0.05 and ***p < 0.005. NS, not significant. Student’s t test.

Figure 4.. Defective repair of stalled TOP2ccs dysregulates estrogen-dependent activation of potential enhancers

(A) Nascent transcriptome analysis at the indicated time-points after the addition of E2. Heatmaps show eRNA expression levels (z scores) in TDP2^−/− cells expressing TDP2 (TDP2^−/− + TDP2, left) or empty vector (TDP2^−/− + Mock, right). Each row represents individual FANTOM5 enhancers that changed the expression of eRNAs with time, defined by TC-seq (p < 0.05). The enhancers were divided into five groups by TC-seq and clustered hierarchically within the groups. See STAR Methods. (B) Area-proportional Venn diagrams showing overlap of upregulated (top) and downregulated (bottom) all enhancers (all enhancers, left), enhancers located within 1 kb of known ERα binding sites (ER⁺ enhancers, middle) and TSSs (TSSs, right). The proportion of ER⁺ enhancers is shown in Figure S4F. (C) Pie charts showing fractions of E2 responsive enhancers localized within 1 kb of the ER and FOXA1-binding sites. (D) Venn diagrams showing the overlap of differentially expressed enhancers (left) and promoters (right) at the 2-h time point between wild-type + DMSO and wild-type + ATMi. (E and F) Heat maps showing log fold change (log₂) of enhancer (E) and promoter (F) expression levels at 2 h relative to those at 0 h after exposure of MCF-7 cells to the indicated reagents. (G) Pie charts showing fractions of E2 responsive enhancers associated with the ERα and FOXA1-binding sites for wild-type cells treated with DMSO and ATMi.

Figure 5.. ATM loss increases the c-MYC transcriptional response to E2 in ER⁺ human BC and murine mammary epithelial cells

(A and B) Kinetics of E2-induced c-MYC transcription (normalized to TFRC) in G₁-arrested MCF-7 (A) and T47D (B) cells carrying the indicated genotypes. (C) γH2AX ChiP quantifying E2-induced DSBs in the c-MYC E–67 enhancer in G₁-arrested MCF-7 cells. The y axis indicates ChiP fold change over input. (D) 3C analyses of G₁-arrested MCF-7 cells to measure the extent of interactions between the +135 kb enhancer and the promoter of the c-MYC gene under the indicated conditions. The y axis indicates fold changes to DMSO-treated data. (E) Quantification of c-MYC mRNA (normalized to TFRC) after cleavage at the c-MYC +135 kb enhancer or two loci outside the enhancer in G₁-arrested MCF-7 cells. (F–H) Percentage of c-Myc⁺ mammary epithelial cells at 6 h (F) and the indicated hours (G–H) after E2 injection into B6 (F), B6;129 (G), and ATMi-treated B6 (H) mice. ATMi was injected together with E2. Representative images are shown in Figures S5C, S5D and S5F, respectively. Data (A–H) represent mean ± standard deviation from triplicates. ***p < 0.005, **p < 0.05, Student t test.

Figure 6.. ATM loss causes abnormal proliferation following daily injection of E2

(A) The experimental design for (B–G) and Figures S6A–S6F. The indicated chemicals were i.p. into mice every day for 3 days. Mammary glands were isolated at day 4 to quantify cells proliferating for the last 3 days. (B and C) Representative images (B) and quantification (C) of EdU positive (EdU⁺) mammary epithelial cells (CK8⁺) after injection with the indicated chemicals. (D–G) Percentage of PCNA⁺ mammary epithelial cells (D, E, and G). Representative images are shown in Figures S6E, S6F, and (F), respectively. Data (C–E and G) represent mean ± standard deviation from triplicates. **p < 0.05 and ***p < 0.005, unpaired two-tailed Student t test. Scale bar, 50 μm.