ATR inhibition potentiates ionizing radiation-induced interferon response via cytosolic nucleic acid-sensing pathways

Abstract

Mechanistic understanding of how ionizing radiation induces type I interferon signaling and how to amplify this signaling module should help to maximize the efficacy of radiotherapy. In the current study, we report that inhibitors of the DNA damage response kinase ATR can significantly potentiate ionizing radiation-induced innate immune responses. Using a series of mammalian knockout cell lines, we demonstrate that, surprisingly, both the cGAS/STING-dependent DNA-sensing pathway and the MAVS-dependent RNA-sensing pathway are responsible for type I interferon signaling induced by ionizing radiation in the presence or absence of ATR inhibitors. The relative contributions of these two pathways in type I interferon signaling depend on cell type and/or genetic background. We propose that DNA damage-elicited double-strand DNA breaks releases DNA fragments, which may either activate the cGAS/STING-dependent pathway or-especially in the case of AT-rich DNA sequences-be transcribed and initiate MAVS-dependent RNA sensing and signaling. Together, our results suggest the involvement of two distinct pathways in type I interferon signaling upon DNA damage. Moreover, radiation plus ATR inhibition may be a promising new combination therapy against cancer.

Keywords: ATR; MAVS; cGAS/STING; radiation; type I interferon.

Figure 1. ATR inhibition significantly potentiates ionizing radiation (IR)‐induced inflammatory signals

1. Hierarchical clustering analysis of RNA sequencing data from cells treated with DMSO, 20 Gy IR, 250 nM AZD6738, or AZD6738 + IR (n = 2 biological replicates).

2. Top enriched pathway analysis of differentially expressed genes in each cluster as shown in A.

3. Heat map of 33 representative inflammatory genes regulated by AZD6738 + IR.

4. MCF10A WT cells were pretreated with DMSO or AZD6738 (250 nM) for 2 h, and then irradiated with 20 Gy, cells were harvested at indicated time points, and immunoblotting was performed with indicated antibodies. h, hours; pSTAT1, pSTAT1(Y701); pIRF3, pIRF3(S396); pTBK1, pTBK1(S172).

5. MCF10A WT cells were pretreated with DMSO or AZD6738 (250 nM) for 2 h, and then irradiated with indicated dose. 72 h later, cells were harvested for immunoblotting with indicated antibodies.

6. Same condition as E except the irradiation dose with 20 Gy, mRNA levels of inflammatory genes as indicated were assessed by real‐time quantitative PCR, presented as mean ± standard error of the mean of three biological replicates.

Figure 2. G2/M DNA damage checkpoint abrogation is required for interferon signaling induced by ATR inhibition + ionizing radiation (IR)

1. MCF10A WT cells were pretreated with AZD6738 (250 nM) and RO‐3306 (9 μM) for 2 h, and then irradiated with 20 Gy; 3 days later, cells were harvested for immunoblotting with indicated antibodies.

2. Same condition as described in A, except that RO‐3306 is replaced by PD‐0332991 (1 μM).

3. MCF10A WT cells were pretreated with DMSO or AZD6738 (250 nM) for 2 h, and then irradiated with 20 Gy; 24 h later, cell cycle distribution was measured by flow cytometry. EdU (10 μM) was added into medium 2 h before harvesting for flow cytometry.

4. MCF10A WT cells were pretreated with DMSO or AZD6738 (250 nM) for 2 h, and then irradiated with 20 Gy, and cells were harvested at indicated time points for immunoblotting.

5. MCF10A control cells and CDC25C‐WT/constitutively active mutant CDC25C‐S216A overexpressed cells were pretreated with DMSO or AZD6738 (250 nM) for 2 h, and then irradiated with 20 Gy; 24 h later, cells were harvested for immunoblotting with indicated antibodies.

6. Same condition as indicated in E, except that cells were harvested 72 h later.

Figure 3. The cGAS/STING‐dependent cytosolic dsDNA‐sensing pathway is dispensable for interferon (IFN) signaling induced by ATR inhibition + ionizing radiation (IR)

1. A

   MCF10A wild‐type (WT), cGAS knockout (KO), and STING KO cells were treated with cGAMP at indicated concentrations for 4 h, and cells were harvested for immunoblotting.

2. B

   MCF10A WT, cGAS KO, and STING KO cells were transfected with herring testis DNA (HT‐DNA, 2 μg/ml) for 6 h with Lipofectamine 3000. Immunoblotting was performed with indicated antibodies.

3. C–F

   (C, D, and F) MCF10A WT or two independent clones for cGAS, STING, and IRF3 KO cells were pretreated with DMSO or AZD6738 (250 nM) for 2 h, and then irradiated with 20 Gy; 3 days later, cells were harvested for immunoblotting with indicated antibodies. cGAS, STING, and IRF3 KO were verified by immunoblotting and Sanger DNA sequencing (Appendix Fig S7). (E) Same condition as described in C, except that cells were harvested for real‐time quantitative PCR. IFNB1 mRNA level was measured and statistical analyzed by two‐way ANOVA (*** means P < 0.001). Data are presented as mean ± standard error of the mean of three biological replicates.

Figure 4. The MAVS‐dependent cytosolic RNA‐sensing pathway is indispensable for interferon (IFN) signaling induced by ATR inhibition + ionizing radiation (IR)

1. A–F

   (A, C, D, and F) MCF10A WT or two independent clones for MAVS, RIG‐I, MDA5, and IRF7 KO cells were pretreated with DMSO or AZD6738 (250 nM) for 2 h, and then irradiated with 20 Gy; 3 days later, cells were harvested for immunoblotting with indicated antibodies. MAVS, RIG‐I, MDA5, and IRF7 KO were verified by immunoblotting and Sanger DNA sequencing (Appendix Fig S7). (B and E) Same condition as described in A, except that cells were harvested for real‐time quantitative PCR instead of immunoblotting. IFNB1 mRNA level was measured by real‐time quantitative PCR. Data are presented as mean ± standard error of the mean of three biological replicates.

Source data are available online for this figure.

Figure 5. Cytosolic nucleic acid‐sensing pathways involved in ionizing radiation (IR)‐induced type I interferon (IFN) signaling differ in murine and human cells

1. 4T1 cells were pretreated with AZD6738 at indicated concentrations for 2 h, and then irradiated with 10 Gy. 4 days later, Ifnb1 mRNA level was measured by real‐time quantitative PCR. Data are presented as mean ± standard error of the mean of three biological replicates.

2. Same condition as described in A. Ifnb1 mRNA levels were measured by real‐time quantitative PCR in WT, Sting KO, Mavs KO, and Irf3 KO 4T1 cells. Data are presented as mean ± standard error of the mean of three biological replicates. All KO cells were verified by immunoblotting and Sanger DNA sequencing (Appendix Fig S7).

3. MC38 cells were pretreated with AZD6738 at indicated concentrations for 2 h, and then irradiated with 20 Gy. 3 days later, Ifnb1 mRNA level was measured by real‐time quantitative PCR. Data are presented as mean ± standard error of the mean of three biological replicates.

4. Same condition as described in C. Ifnb1 mRNA levels were measured by real‐time quantitative PCR in WT, Sting KO, Mavs KO, and Irf3 KO MC38 cells. Data are presented as mean ± standard error of the mean of three biological replicates. All KO cells were verified by immunoblotting and Sanger DNA sequencing (Appendix Fig S7).

Figure 6. DNA damage‐elicited AT‐rich DNA mediates type I interferon (IFN) signaling in human cells

1. MCF10A WT, cGAS KO, STING KO, and MAVS KO cells were treated with poly(dA‐dT)/LyoVec at indicated concentrations for 24 h. Immunoblotting was performed with indicated antibodies.

2. 4T1 WT, Sting KO, Mavs KO, and Irf3 KO cells were treated with poly(dA‐dT)/LyoVec at indicated concentrations for 24 h. Immunoblotting was performed with indicated antibodies.

3. END‐seq was used to map double‐strand breaks (DSBs) associated with TA dinucleotide repeats in MCF10A cells with the indicated treatments. ATR inhibitor (ATRi) AZD6738 (250 nM) was used with ionizing radiation (IR, 10 Gy) for 6 h. NT, not treated.

4. MCF10A WT, MDA5 KO, RIG‐I KO, and MAVS KO cells were transfected with AT‐rich oligos (2 μg/ml) with LyoVec for 24 h. Immunoblotting was performed with indicated antibodies.

5. MCF10A WT and STING KO cells were transfected with AT‐rich and non‐AT‐rich oligos (2 μg/ml) with lipofectamine 3000 for 4 h. Immunoblotting was performed with indicated antibodies.

6. 4T1 WT, Sting KO, and Irf3 KO cells were transfected with AT‐rich and non‐AT‐rich oligos (2 μg/ml) with LyoVec for 24 h. Immunoblotting was performed with indicated antibodies.

7. A proposed potential mechanism for the divergent cytosolic nucleic acid‐sensing pathway activated in response to radiation alone or with ATRi among different cell lines.