Histone ADP-ribosylation promotes resistance to PARP inhibitors by facilitating PARP1 release from DNA lesions

Abstract

Poly(ADP-ribose) polymerase 1 (PARP1) has emerged as a central target for cancer therapies due to the ability of PARP inhibitors to specifically kill tumors deficient for DNA repair by homologous recombination. Upon DNA damage, PARP1 quickly binds to DNA breaks and triggers ADP-ribosylation signaling. ADP-ribosylation is important for the recruitment of various factors to sites of damage, as well as for the timely dissociation of PARP1 from DNA breaks. Indeed, PARP1 becomes trapped at DNA breaks in the presence of PARP inhibitors, a mechanism underlying the cytotoxitiy of these inhibitors. Therefore, any cellular process influencing trapping is thought to impact PARP inhibitor efficiency, potentially leading to acquired resistance in patients treated with these drugs. There are numerous ADP-ribosylation targets after DNA damage, including PARP1 itself as well as histones. While recent findings reported that the automodification of PARP1 promotes its release from the DNA lesions, the potential impact of other ADP-ribosylated proteins on this process remains unknown. Here, we demonstrate that histone ADP-ribosylation is also crucial for the timely dissipation of PARP1 from the lesions, thus contributing to cellular resistance to PARP inhibitors. Considering the crosstalk between ADP-ribosylation and other histone marks, our findings open interesting perspectives for the development of more efficient PARP inhibitor-driven cancer therapies.

Keywords: ADP-ribosylation; DNA repair; PARP1; chromatin; fluorescence microscopy.

Fig. 1.

Increased PARP1 trapping upon loss of HPF1 is not solely due to impaired automodification. (A and B) Representative image time courses (A) and recruitment kinetics (B) of GFP-tagged PARP1 WT and PARP1 3SA after DNA damage by 405 nm laser irradiation in PARP1 KO and PARP1/HPF1 double KO cells. (Scale bar, 5 µm.) (C) PARP1 release speed was assessed for each condition by estimating the characteristic dissipation time, defined as the time needed to dissipate half of the maximum recruitment intensity on the recruitment curves shown in (B).

Fig. 2.

Heteromodification contributes to the mobilisation of PARP1 from the DNA lesions. (A and B) Representative image time courses (A) and recruitment kinetics (B) of mCherry-tagged PARP1 E988K after DNA damage by 405 nm laser irradiation, in PARP1 KO coexpressing, or not, GFP-tagged PARP1 WT. (C) Schematic representation of the fluorescence two-hybrid assay used to monitor the level of PARP1 ADP-ribosylation in living cells. The mCherry-tagged fusion between the LacI and the macrodomain of macroH2A1.1 is tethered to the LacO array, thus appearing as a bright spot in the nucleus. After DNA damage by laser irradiation away from the LacO array, GFP-tagged PARP1 will enrich at the LacO array depending on its ADP-ribosylation status. (D and E) Representative images of UO2S cells bearing LacO repeats and coexpressing a mCherry-tagged fusion between the LacI and the macrodomain of macroH2A1.1 and GFP-tagged PARP1 WT (D) or PARP1 E988K (E). Images were taken prior to damage and 50 s post 405 nm laser irradiation. Insets pseudocolored according to the look-up table displayed below show magnification of the LacO array. (F) Quantification of the PARP1 accumulation at the LacO array from the images shown in (D and E). (G) Schematic representation of the FRAP assay used to monitor the level of PARP1 ADP-ribosylation in living cells. Cells are coexpressing a mCherry-tagged fusion between H2B and the macrodomain of macroH2A1.1 (H2B-macro) and GFP-tagged PARP1. After DNA damage by laser irradiation, the diffusion of ADP-ribosylated PARP1 within the nucleus is slowed down by its binding to the H2B-macro fusion. This slowing-down is assessed by measuring FRAP a region of interest away from the damaged region. (H) Representative time-course images of the FRAP of a circular area (white circle) within the nucleus of U2OS PARP1 KO cells expressing GFP-tagged PARP1 WT alone or coexpressing GFP-tagged PARP1 E988K and mCherry-tagged PARP1 WT before (pre) and 60 s after (post) 405 nm laser irradiation. Images are pseudocolored according to the look-up table shown on panel (D). (I) Normalized fluorescence recovery curves for GFP-tagged PARP1 WT and E988K obtained from the images shown in (H). (J) Characteristic recovery times estimated from the fit of the curves shown in (I). Scale bar, 5 µm for (A, D, E, and H).

Fig. 3.

Spontaneous increase of heteromodification is sufficient to impair PARP1 association with DNA. (A) Normalized FCS autocorrelation curves obtained for GFP-tagged PARP1 WT expressed in ARH3 KO cells treated or not with 20 µM of the PARGi PDD00017273 for 48 h. (B) Residence times within the focal volume estimated from the fit of the autocorrelation curves shown in (A). (C) Residence times within the focal volume estimated from the fit of the autocorrelation curves obtained for GFP-tagged PARP1 S499/507/519A E988K (PARP1 3SA-EK) expressed in ARH3 KO cells treated or not with PARGi.

Fig. 4.

Histone ADP-ribosylation promotes PARP1 release from DNA breaks. (A) Immunoblot of ARH3 KO cells, untransfected, or expressing either GFP-tagged ARH3 WT, ARH3 D77/78N (DD), LANA-ARH3 WT (L-ARH3), or LANA-ARH3 D77/78N (L-ARH3 DD) treated or not with H₂O₂. The blot was stained with a pan-ADPr and an ARH3 antibody. H3 staining was used as loading control. (B) Representative images of UO2S cells bearing LacO repeats and coexpressing a mCherry-tagged fusion between the LacI and the macrodomain of macroH2A1.1, PARP1 WT fused to Halotag stained with JF646, and LANA-ARH3 WT or D77/78N tagged with GFP. Images were taken prior to damage and 50 s post 405 nm laser irradiation. Insets pseudocolored according to the look-up table displayed below show magnification of the LacO array. (C) Quantification of PARP1 accumulation at the LacO array from the images shown in (B). (D–F) Representative image time courses (D), recruitment kinetics (E), and dissipation times (F) of GFP-tagged PARP1 WT after DNA damage by 405 nm laser irradiation, in PARP1/ARH3 KO coexpressing mCherry-tagged ARH3 WT or LANA-ARH3 WT. (G and H) Recruitment kinetics (G) and dissipation times (H) of GFP-tagged PARP1 WT after DNA damage by 405 nm laser irradiation, in PARP1/ARH3 KO coexpressing mCherry-tagged LANA-ARH3 WT or LANA-ARH3 D77/78N. (I and J) Recruitment kinetics (I) and dissipation times (J) of GFP-tagged PARP1 3SA after DNA damage by 405 nm laser irradiation, in PARP1/ARH3 KO coexpressing mCherry-tagged ARH3 WT or LANA-ARH3 WT. Scale bar, 5 µm for (B and D).

Fig. 5.

Histone ADP-ribosylation underlies resistance to the PARP inhibitor olaparib. (A and B) Representative images of the clonogenic assay (A) and cell survival curves (B) for WT cells, ARH3 KO cells, two clones of ARH3 KO cells expressing mCherry-tagged ARH3 WT (ARH3 OE), and two clones of ARH3 KO cells expressing mCherry-tagged LANA-ARH3 WT, upon treatment with the PARPi olaparib.