CYLD induces high oxidative stress and DNA damage through class I HDACs to promote radiosensitivity in nasopharyngeal carcinoma

Abstract

Abnormal expression of Cylindromatosis (CYLD), a tumor suppressor molecule, plays an important role in tumor development and treatment. In this work, we found that CYLD binds to class I histone deacetylases (HDAC1 and HDAC2) through its N-terminal domain and inhibits HDAC1 activity. RNA sequencing showed that CYLD-HDAC axis regulates cellular antioxidant response via Nrf2 and its target genes. Then we revealed a mechanism that class I HDACs mediate redox abnormalities in CYLD low-expressing tumors. HDACs are central players in the DNA damage signaling. We further confirmed that CYLD regulates radiation-induced DNA damage and repair response through inhibiting class I HDACs. Furthermore, CYLD mediates nasopharyngeal carcinoma cell radiosensitivity through class I HDACs. Thus, we identified the function of the CYLD-HDAC axis in radiotherapy and blocking HDACs by Chidamide can increase the sensitivity of cancer cells and tumors to radiation therapy both in vitro and in vivo. In addition, ChIP and luciferase reporter assays revealed that CYLD could be transcriptionally regulated by zinc finger protein 202 (ZNF202). Our findings offer novel insight into the function of CYLD in tumor and uncover important roles for CYLD-HDAC axis in radiosensitivity, which provide new molecular target and therapeutic strategy for tumor radiotherapy.

Fig. 1. CYLD interacts and inhibits the activity of class I HDACs.

a A volcano of differentially expressed genes identified by mRNA-seq in control and CYLD overexpressing NPC cells. b GSEA analysis of CYLD relative pathways. c HDACs activity measurement in control and CYLD knockdown cells. d HDACs activity measurement in control, CYLD, and CYLD-C601A overexpressing cells. e Proximity ligation assay indicating the interaction of CYLD and HDAC1/2 in HK1 cells (red: PLA positive signal; blue: DAPI, scale bar = 10 μm). f HONE1 cells were disrupted. The cell lysates were subjected to immunoprecipitation (IP) with anti-HDAC1/2. g HONE1-EBV cells transfected with the indicated plasmids, acetyl-Lys (Ac-K) was immunoprecipitated and probed with HDAC1 antibodies. The whole cell extract was used as input. h Schematic representation of the Flag-tagged full length CYLD (FL) and its various deletion mutants, including N-terminal 1-303 deletion (△N), middle domain deletion (△M), and C-terminal deletion(△C) constructs. i 293 T cells transfected with the indicated constructs were disrupted. The cell lysates were subjected to immunoprecipitation with anti-Flag.

Fig. 2. CYLD induces high oxidative stress by regulating HDACs.

a Venn plot showing the overlap of differential genes from different treatment groups. b differential genes from Venn plots were enriched by GO. c Representative IHC photographs showing the expression of CYLD and 8-OHdG in consecutive sections of NPC microarrays. d 8-OHdG level is calculated based on CYLD expression in NPC microarrays. e ROS levels of control and CYLD overexpressing HONE1-EBV cells were detected by FCM by using CellROX Deep Red. f ROS levels of HONE1-EBV cells with or without Chidamide treatment (0.5 μM) were detected by FCM by using CellROX Deep Red. g ROS levels of CYLD knockdown HONE1 cells with or without Chidamide treatment (0.5 μM) were detected by FCM by using CellROX Deep Red.

Fig. 3. CYLD-HDAC axis regulates cell antioxidant activity.

a Total glutathione (GSH) and oxidized glutathione (GSSG) of control and CYLD overexpressing HONE1-EBV cells were measured by using a GSH/GSSG assay kit. b Intracellular NADP + /NADPH levels of control and CYLD overexpressing HONE1-EBV cells were assayed by using an NADP + /NADPH assay kit. c Total glutathione (GSH) and oxidized glutathione (GSSG) of HONE1-EBV cells with or without Chidamide treatment (0.5 μM) for 24 h were measured by GSH/GSSG assay kit. d Intracellular NADP + /NADPH levels of HONE1-EBV cells with or without Chidamide treatment (0.5 μM) for 24 h were assayed by using an NADP + /NADPH assay kit. e A heat map showing the expression of the Redox-related genes across CYLD overexpression and control groups. f mRNA and h Protein level of indicate genes were detected after CYLD overexpressing in HONE1-EBV cells, β-actin was used as a control. g mRNA and i protein expression of indicated genes were detected with or without Chidamide treatment (0.5 μM) for 24 h in NPC cells. β-actin was used as a control.

Fig. 4. CYLD induces DNA damage through HDAC1/2.

a–c Analysis of γH2AX foci at different times after 4 Gy radiation. Immunofluorescence staining for γ-H2AX followed by confocal microscopy was performed. Cells displaying 10 or more foci were counted as positive. Representative images of γH2AX foci and the percentage of HK1-EBV and HONE1-EBV cells displaying γH2AX foci are shown. d, e Immunofluorescence staining of γH2AX followed by confocal microscopy was performed. HONE1 cells by CYLD knockdown with or without Chidamide (0.5 μM) treatment for 24 h. f Detection of γH2AX protein level in NPC cells by CYLD overexpression with or without iTSA-1 (10 μM) for 24 h treatment.

Fig. 5. Downregulation of CYLD contributes to cancer radioresistance.

a Immunoblot and b real-time PCR analysis of CYLD expression levels in radiation-resistant cells (CNE2-IR and HK1-IR) compared with radiation-responsive cells (CNE2 and HK1). β-Actin was used as a control. c–f Colony formation assay showing survival fractions of CYLD overexpression cells treated or not treated with 4 Gy irradiation, surviving fractions were calculated by comparing the colony number of each treatment group with untreated groups (0 Gy). (CYLD wt: full-length CYLD plasmid; CYLD-C601A: c601 mutant CYLD plasmid lacking enzyme function). The relative SF (survival fraction) is plotted below the results, Results are plotted as the mean surviving fraction ± SEM of 3 independent experiments.

Fig. 6. CYLD mediates radiation sensitivity through class I HDACs.

a, b Colony formation assay showing survival fractions of CYLD overexpressing cells (CYLD wt: full-length CYLD plasmid; CYLD △N: N-terminal deletion CYLD plasmid) treated with irradiation or untreated; surviving fractions were calculated by comparing the colony number of each treatment group with untreated groups (0 Gy). c Colony formation assay showing survival fractions of CYLD overexpressing cells treated with iTSA (10 μM) or untreated at 24 h before irradiation; surviving fractions were calculated by comparing the colony number of each treatment group with untreated groups (0 Gy). d Colony formation assay showing survival fractions of CYLD knockdown cells, CHI (Chidamide: 0.5 μM) were treated 24 h before irradiation; surviving fractions were calculated by comparing the colony number of each treatment group with untreated groups (0 Gy). Results are plotted as the mean surviving fraction ± SEM of 3 independent experiments.

Fig. 7. HDACi increases radiosensitivity in vivo.

a The overall diagram of the study design. The NPC xenograft model was established using HONE1-EBV cells. b Representative images of xenografts from different treatment groups. Control: saline vehicle; Chidamide: Chidamide 5 mg/kg; IR irradiation with 4 Gy; IR + Chidamide: Chidamide-irradiation combination. c The tumor volume of xenograft from the indicated treatment group (n = 5). d Tumor weight of HONE1-EBV-derived xenografts from the indicated treatment group (n = 5). e–g Tumor sections were stained with hematoxylin and eosin (H&E) and subjected to immunohistochemistry detection for 8-OHdG and γH2AX (scale bar, 500 μm). Results are plotted as the mean surviving fraction ± SEM.

Fig. 8. ZNF202 mediates transcriptional expression of CYLD.

a Total RNA was isolated and subjected to real-time PCR analysis of ZNF202 in EBV-positive (HK1-EBV and HONE1-EBV) cells compared with EBV-negative (HK1 and HONE1) cells. b Total RNA was isolated and subjected to real-time PCR analysis of ZNF202 in radiation-resistant cells (CNE2-IR and HK1-IR) compared with radiation-responsive cells (CNE2 and HK1) cells. c Immunoblot analysis of ZNF202 in EBV-positive (HK1-EBV and HONE1-EBV) cells compared with EBV negative (HK1 and HONE1) cells and ZNF202 protein expression levels in radiation-resistant cells (CNE2-IR and HK1-IR) compared with radiation-responsive cells (CNE2 and HK1) cells. β-Actin was used as a control. HK1-EBV cells transfected with ZNF202 siRNAs: d Cell lysates were then extracted and subjected to Western blotting. β-Actin was used as a control. e total RNA from cells was isolated and subjected to real-time PCR. f The level of ZNF202 binding to the CYLD promoter in HONE1 and HONE1-EBV cells was analyzed by using ChIP followed by RT-PCR of 3 specific regions (n = 3). g Schematic illustration of the CYLD promoter and 3 potential binding sites of ZNF202. h 293 T cells were or were not co-transfected with ZNF202 and the luciferase reporter driven by the wild-type CYLD promoter or mutant CYLD promoter, together with a PLR-TK construct. Results are plotted as the mean surviving fraction ± SEM of 3 independent experiments. i Representative IHC staining of ZNF202 and CYLD expression from pathological sections of nasopharyngeal squamous cell carcinoma patients. j The CYLD protein expression level was calculated according to ZNF202 expression of nasopharyngeal squamous cell carcinoma patients. High- and low-expressing groups were classified according to median score.

Fig. 9. CYLD induces high oxidative stress and DNA damage through class I HDACs to promote radiosensitivity in nasopharyngeal carcinoma.

We discovered a mechanism which CYLD binds to and inhibits class I HDACs enzyme functions by inducing HDAC1 acetylation. While class I HDACs mediate redox abnormalities and DNA damage repair, which leading to radiotherapy resistance in CYLD low-expressing tumors. Blocking HDACs by class I HDACs inhibitor Chidamide could effectively decrease radioresistance in vitro and in vivo. HDACi Chidamide could be a promising therapeutic strategy in CYLD low-expressing tumors to increase tumor radiotherapy sensitivity. (Ac acetylation, HDAC1(blue background: inactive; red background: active)).