Unrestrained poly-ADP-ribosylation provides insights into chromatin regulation and human disease

Abstract

ARH3/ADPRHL2 and PARG are the primary enzymes reversing ADP-ribosylation in vertebrates, yet their functions in vivo remain unclear. ARH3 is the only hydrolase able to remove serine-linked mono(ADP-ribose) (MAR) but is much less efficient than PARG against poly(ADP-ribose) (PAR) chains in vitro. Here, by using ARH3-deficient cells, we demonstrate that endogenous MARylation persists on chromatin throughout the cell cycle, including mitosis, and is surprisingly well tolerated. Conversely, persistent PARylation is highly toxic and has distinct physiological effects, in particular on active transcription histone marks such as H3K9ac and H3K27ac. Furthermore, we reveal a synthetic lethal interaction between ARH3 and PARG and identify loss of ARH3 as a mechanism of PARP inhibitor resistance, both of which can be exploited in cancer therapy. Finally, we extend our findings to neurodegeneration, suggesting that patients with inherited ARH3 deficiency suffer from stress-induced pathogenic increase in PARylation that can be mitigated by PARP inhibition.

Keywords: ADP-ribosylation; ARH3/ADPRHL2; BRCA; DNA damage; PARG; PARP inhibitor; cancer; chromatin; neurodegeneration; telomere.

Graphical abstract

Figure 1

Loss of ARH3 leads to the enrichment of chromatin-associated MARylation throughout the cell cycle (A) Cells were subjected to subcellular fractionation. ADPr signals were analyzed using western blotting. (B and C) Pan-ADPr (B) and MARylation (C) signals were detected by immunofluorescence in detergent pre-extracted cells at the indicated cell cycle stages. Scale bars, 10 μm. See also Figure S1.

Figure 2

Suppression of PARG activity leads to the accumulation and persistence of PARylation in ARH3-deficient cells (A) Cells were pre-treated with DMSO, 10 μM olaparib, or 10 μM PARGi for 1 h followed by 2 mM H₂O₂ treatment for the indicated time in the presence of the drugs. ADPr signals were analyzed using western blotting. (B) Cells were treated with DMSO for 8 days or with 25 μM PARGi for the indicated number of days. ADPr and γH2AX levels were analyzed using western blotting. (C–E) Levels of pan-ADPr (C), PARylation (D), or MARylation (E) were analyzed using immunofluorescence in detergent pre-extracted cells treated with DMSO or 25 μM PARGi for 6 days. Scale bars, 10 μm. (F and G) Quantification of (C)–(E) for 6 day DMSO (F) or 25 μM PARGi treatment (G). Data represent fold change in mean intensity per nucleus relative to DMSO-treated control cells and are shown as mean ± SEM; at least 300 cells were analyzed per condition. (H) Levels of PAR in U2OS cells treated with DMSO or 25 μM PARGi for 4 days were quantified by UPLC-MS/MS analysis. Ribosyl-adenosine (R-Ado) is representative for the overall content of PAR. Data are shown as mean ± SEM, n = 4; ^∗∗∗p < 0.001 (one-way ANOVA followed by Tukey post-test). (I) Radioactive ADP-ribosylation assay of unmodified H3 peptide or H3 peptide with S10-linked MAR (H3S10MAR). See also Figure S2.

Figure 3

ARH3 deficiency is synthetically lethal with PARG suppression and renders cancer cells resistant to PARP inhibition (A and B) Representative images (top) and quantification (bottom) of colony formation assay with control and ARH3-KO cells (A and B) and ARH3-KO cells complemented with ARH3 WT or catalytically inactive D77/78N mutant (B) treated with DMSO or as indicated. (C) Cells were treated with 25 μM PARGi for 4 days. ADPr signals were analyzed using western blotting. (D) Quantification of cell cycle analysis by flow cytometry of EdU- and DAPI-stained cells after 6 day exposure to DMSO or indicated treatment and 1 h EdU pulse. (E and F) Cell proliferation and DNA synthesis after exposure to DMSO or indicated treatment for 6 days and 1 h (E) or 24 h (F) EdU pulse. (G and H) Quantification of colony formation assay with U2OS cells transfected with BRCA1 or BRCA2 siRNA (G) or with SUM149PT cells (H) treated with DMSO or olaparib. Data are shown as mean ± SD, n = 3 (A, B, G, and H), or as mean ± SEM, n = 4 (E), n = 2 (F); ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001 (two-tailed Student’s t test). See also Figure S3.

Figure 4

Simultaneous loss of ARH3 and PARG activity causes dysregulation of chromatin modification and transcription profiles (A) NAD⁺ quantification assay in cells treated with DMSO or 25 μM PARGi for 4 days or 10 nM FK866 for 1 day. Data are shown as mean ± SD, n = 2; ^∗∗p < 0.01 (two-tailed Student’s t test). (B and C) Quantification (B) and representative images (C) of ALT-associated PML bodies (APBs) after DMSO or 5 μM PARGi treatment for 72 h. Data in (C) are shown as mean ± SEM, n = 5; at least 1,300 cells were analyzed per condition; ^∗∗p < 0.01 (two-tailed Student’s t test). Scale bars, 5 μm. (D and G) Cells were treated with DMSO or 25 μM PARGi for 6 (D) or 4 days (G). ADPr and protein levels were analyzed using western blotting. (E) MA plot showing differentially expressed genes (upregulated in red, downregulated in blue) in ARH3-KO U2OS cells treated with 25 μM PARGi for 4 days against DMSO control. n = 3, adjusted p < 0.05, absolute fold change > 1.5. (F) Summary of significantly upregulated and downregulated pathways from gene set enrichment analysis (GSEA) in ARH3-KO U2OS cells treated with 25 μM PARGi for 4 days against DMSO control. n = 3, q < 0.05 (in red), q < 0.1 (in orange). See also Figure S4.

Figure 5

Loss of ARH3 activity in patient-derived primary fibroblasts and glioblastoma cells leads to the accumulation of ADPr, increased PARGi sensitivity, and PARPi resistance (A) Pan-ADPr signals were detected using immunofluorescence in detergent pre-extracted control and ARH3 C26F mutant patient cells at different cell cycle stages. Scale bars, 10 μm. (B) Cells were treated with DMSO, 25 μM PARGi or 25 μM PARGi, and 1 μM olaparib for 10 days. ADPr, H3 modification, and PARP1 levels were analyzed using western blotting. (C) Levels of pan-ADPr were analyzed using immunofluorescence in detergent pre-extracted cells treated with DMSO or 25 μM PARGi for 4 days. Scale bars, 10 μm. (D) Quantification of crystal violet assay with cells treated with DMSO or as indicated. (E and F) Quantification of colony formation assay with U251 cells treated with DMSO or as indicated. Data are shown as mean ± SD, n = 3; ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001 (two-tailed Student’s t test). See also Figure S5.

Figure 6

Models of two-step ADPr reaction and stress-induced pathogenesis in ARH3-deficient neurodegeneration patients (A) ADPr reaction consists of distinctly regulated initiation and elongation steps. Initiation of ADPr at serine residues requires PARP1 activation (e.g., by endogenous DNA damage) and a cofactor protein HPF1. Elongation of initial ADPr attachments can be performed by PARP1 alone. ARH3 and PARG hydrolases predominantly reverse initiation and elongation steps, respectively. (B) ARH3 deficiency results in the accumulation of MAR and short PAR initiation sites, which cannot be efficiently elongated, because of the presence of PARG activity. (C) Because of the presence of ARH3 that removes the initiation sites necessary for the subsequent elongation step, suppression of PARG results only in a slight enrichment of elongation products, composed mainly of PARP1 autoPARylation but not chromatin-associated PARylation. (D) Combined ARH3 deficiency and PARG suppression results in both initiation and elongation steps proceeding without reversal. The accumulated initiation sites are uncontrollably extended to long PAR chains, which are toxic to the cell and eventually lead to PARP-dependent cell death. (E) ARH3-deficient neural cells accumulate MAR and short PAR initiation sites, and are “primed” for rapid elongation upon encountering additional stressors, in particular viral infection, that could specifically downregulate PARG, promoting pathogenic accumulation of PARylation and thus neurodegeneration (top). PARPi treatment can alleviate stress-induced neurodegeneration in ARH3-deficient neural cells by preventing the formation of both initiation and elongation products (bottom).