Targeting neddylation sensitizes colorectal cancer to topoisomerase I inhibitors by inactivating the DCAF13-CRL4 ubiquitin ligase complex

Abstract

Colorectal cancers (CRCs) are prevalent worldwide, yet current treatments remain inadequate. Using chemical genetic screens, we identify that co-inhibition of topoisomerase I (TOP1) and NEDD8 is synergistically cytotoxic in human CRC cells. Combination of the TOP1 inhibitor irinotecan or its bioactive metabolite SN38 with the NEDD8-activating enzyme inhibitor pevonedistat exhibits synergy in CRC patient-derived organoids and xenografts. Mechanistically, we show that pevonedistat blocks the ubiquitin/proteasome-dependent repair of TOP1 DNA-protein crosslinks (TOP1-DPCs) induced by TOP1 inhibitors and that the CUL4-RBX1 complex (CRL4) is a prominent ubiquitin ligase acting on TOP1-DPCs for proteasomal degradation upon auto-NEDD8 modification during replication. We identify DCAF13, a DDB1 and Cullin Associated Factor, as the receptor of TOP1-DPCs for CRL4. Our study not only uncovers a replication-coupled ubiquitin-proteasome pathway for the repair of TOP1-DPCs but also provides molecular and translational rationale for combining TOP1 inhibitors and pevonedistat for CRC and other types of cancers.

Fig. 1. High-throughput screening identifies pevonedistat (PEV) with TOP1 inhibitors as synergistic combination in CRC cells.

a Dose-response curve of PEV derived from MIPE 5.0 library screen in HCT116 CRC cells. b Scheme of 6 × 6 matrix screening to examine PEV in combination with the MIPE5.0 library in HCT116 cells. In response matric, red indicates strong response whereas black indicates poor response. In excess over the Highest Single Agent (ExcessHSA or ΔHSA) metric, orange indicates synergy whereas blue indicates resistance. c Drug-target enrichment analysis plots highlighting the synergy of PEV with TOP1 inhibitors. PEV-TOP1 inhibitor pairs ranked using the ExcessHSA metric. d Response (top panels) and ΔHSA (bottom panels) heatmaps for the combination of PEV with camptothecins irinotecan and SN38 (the bioactive metabolite of irinotecan) and indenoisoquinolines LMP776 (indimitecan) and LM744 across defined concentration ranges in HCT116 cells. e Top panels: viability curves for 72 h treatments with SN38 at defined concentrations in the indicated CRC cell lines (mean ± SD, N = 3 biologically independent experiments) using ATPlite Luminescence Assay. Cells were treated with PEV at defined concentrations 4 h before SN38. Bottom panels: SN38-PEV pairs ranked by ExcessHSA metric using Combenefit, an interactive platform for the analysis of drug combinations.

Fig. 2. Irinotecan in combination with PEV exhibits synergy in preclinical CRC models.

a Top panels: Three 3D organoids generated from liver-dominant metastatic CRC patients (mCRC PDOs #1, #2 and #3) were treated with SN38 plus PEV as indicated for 72 h for CellTiter-Glo Luminescent Cell Viability Assay (mean ± SD, N = 3 biologically independent experiments). Bottom panels: SN38-PEV pairs ranked by ExcessHSA metric using Combenefit. b Brightfield images of mCRC PDO #1 treated with mock (no treatment), 2.5 µM PEV, 1.25 µM SN38 or 2.5 µM PEV + 1.25 µM SN38. The scale bar represents 50 µm. c HCT116 tumor-bearing athymic nude mice were treated either with vehicle, PEV (2.5 mg/kg), irinotecan (20 mg/kg) or a combination of PEV plus irinotecan. Treatments were initiated 10 days after tumor inoculation when mean tumor size reached 95.7 mm³. Tumor growth is shown as fold-change versus first day of treatment (n = 10 per treatment group, except n = 8 for PEV 2.5 mg/kg). d Kaplan–Meier survival curve of mice following the three treatment cycles (n = 10 in vehicle and PEV group, n = 7 all other groups). Survival after treatment with low dose PEV (2.5 mg/kg) in combination with irinotecan. p value determined by Mantel–Cox test. e Kaplan–Meier survival curve of mice following the three treatment cycles (n = 10 in vehicle and PEV group, n = 7 all other groups). Survival after treatment with high dose (15 mg/kg) in combination with irinotecan. p value determined by Mantel–Cox test. f Left panel: representative bioluminescence images 21 days after treatment start; Right panel: Quantification of bioluminescence imaging (n = 10 per treatment group, except n = 8 for PEV 2.5 mg/kg). Data are shown as mean ± SEM. p values determined by two-tailed Student’s t test.

Fig. 3. Neddylation fosters the repair of TOP1-DPCs in a manner epistatic to the ubiquitin-proteasome pathway.

a Scheme of the in vivo complex of enzyme (ICE) bioassay. b Left panel: ICE bioassay shows that SN38-induced TOP1 were removed after 4 h exposure to SN38 in HCT116 cells, and that pre-treatment with 10 µM PEV for 1 h blocked the removal of TOP1-DPCs. DNA loading of each sample was confirmed using anti-DNA antibody. Right panel: densitometric analyses of TOP1-DPCs from triplicate experiments including blots in the left panel. Density of TOP1-DPC/density of DNA of each group was normalized to that of cells treated with 500 nM SN38 for 1 h. Data are presented as mean ± SD, N = 3 biologically independent experiments. The p values in this figure were calculated using two-tailed Student’s t test. c Top panels: 20 s filming of TOP1-HaloTag single molecules in HCT116 cells. The cells were divided into indicated treatments. Middle panels: plots of tracks of TOP1-HaloTag single molecules as shown in the top panels. Bottom panels: count of jumps of TOP1-HaloTag single molecules derived from the top-panel films. The bin size is 0.1 μm. d Left panel: ICE bioassay in HCT116 cells treated with the indicated drug combinations for 2 h. Right panel: densitometric analyses of TOP1-DPCs. Density of TOP1-DPC/density of DNA of each group was normalized to that of cells treated with SN38 alone. Data are presented as mean ± SD, N = 3 biologically independent experiments. e HCT116 cells were treated with SN38 (10 μM) and collected at indicated time points for Western blotting. Cells were then lysed with the neutral lysis procedure. The scale bar represents 300 μm. f Left panel: HCT116 were treated with SN38 (10 μM) for 2 h in the presence and absence of indicated inhibitors. Cells were then subjected to neutral comet assay. Right panel: quantitation of tail moments for comet assay samples using OpenComet. Data are presented as mean ± SD, n = 180 total cells. Biological independent experiments were repeated three times. g HCT116 cells were treated with SN38 (10 μM) for 2 h. PEV (10 μM) and BTZ (1 μM) were added 1 h prior to the SN38 treatment. Cells were then subjected to Western blotting using indicated antibodies.

Fig. 4. CRL4 is activated by NEDD8 in response to TOP1-DPCs and target TOP1-DPCs for ubiquitylation.

a HCT116 cells were transfected with the indicated Myc-tag cullin overexpression plasmids plus RBX1-FLAG overexpression plasmid for 48 h before CPT treatment (20 μM, 30 min). The cells were subjected to the DUST assay for immunodetection of ubiquitylated TOP1-DPC and total TOP1-DPC using anti-ubiquitin (Ub) and anti-TOP1 antibodies. Total DNA was detected using anti-DNA antibody as loading control. The order of DNA slot blots has been altered, as indicated by the line, and that uncropped labeled blots can be found in the Source Data file. b In vitro ubiquitylation assay with recombinant TOP1-DPC (generated using a suicidal DNA substrate) and CUL4A-RBX1 complex. TOP1 was tested for ubiquitin conjugation in the presence of Ub E1, Ub E2 UbcH5a and the indicated concentrations of CUL4A-RBX1 complex. Reaction products were separated by SDS-PAGE and monitored by IB using anti-ubiquitin antibody. c HCT116 cells were transfected with CUL4A siRNA or Myc-CUL4A and RBX1-FLAG overexpression plasmids (CRL4) for 48 h before PEV pre-treatment (10 μM, 1 h) then co-treatment with CPT (20 μM, 30 min). The cells were subjected to DUST assay for immunodetection of ubiquitylated TOP1-DPC and total TOP1-DPC using anti-ubiquitin and anti-TOP1 antibodies. The order of DNA slot blots has been altered, as indicated by the line, and that uncropped labeled blots can be found in the Source Data file. d HCT116 cells WT, CUL4B KO cells and CUL4B KO cells replenished with Myc-CUL4B overexpression plasmid were pre-treated with BTZ (1 μM, 4 h) or PEV (10 μM, 4 h) then co-treated with CPT (20 μM, 30 min). The cells were then subjected to DUST assay for immunodetection of ubiquitylated TOP1-DPC and total TOP1-DPC using anti-ubiquitin and anti-TOP1 antibodies. The order of DNA slot blots has been altered, as indicated by the line, and that uncropped labeled blots can be found in the Source Data file. e Myc-CUL4A WT or K705R (KR) overexpressing HCT116 cells were transfected with HA-NEDD8 overexpression plasmid, followed by treatments with indicated inhibitors. Immunoprecipitation (IP) using anti-Myc tag antibody was performed after the treatments. IP samples and cell lysates (input) were subjected to immunoblotting (IB) with indicated antibodies. f Myc-CUL4B WT or K859R (KR) overexpressing HCT116 cells were transfected with HA-NEDD8 overexpression plasmid, followed by treatments with indicated inhibitors. IP using anti-Myc tag antibody was performed after the treatments. IP samples and input were subjected to IB with indicated antibodies. g RBX1-FLAG overexpressing HCT116 cells were transfected with indicated CUL4 overexpression plasmids, followed by CPT treatment (20 μM, 30 min) for DUST assay for immunodetection of ubiquitylated TOP1-DPC and total TOP1-DPC using anti-ubiquitin and anti-TOP1 antibodies. The order of DNA slot blots has been altered, as indicated by the line, and that uncropped labeled blots can be found in the Source Data file.

Fig. 5. CRL4 ubiquitylates TOP1-DPCs for proteasomal degradation in a replication-dependent manner.

a HCT116 cells were transfected with empty vector (EV), Myc-CUL4A + RBX1-FLAG overexpression plasmids (CRL4A) or Myc-CUL4B + RBX1-FLAG overexpression plasmids (CRL4B) for 48 h. The cells were subjected to 1 h pre-treatment with replication inhibitor aphidicolin (APH, 10 μM) or CDK7/transcription inhibitor THZ1 (10 μM), followed by co-treatment with CPT (20 μM, 30 min). The cells were subjected to DUST assay for immunodetection of ubiquitylated TOP1-DPC and total TOP1-DPC using anti-ubiquitin and anti-TOP1 antibodies. The order of DNA slot blots has been altered, as indicated by the line, and that uncropped labeled blots can be found in the Source Data file. b HCT116 cells were transfected with the Ub K48 or K63 single lysine overexpression plasmid and Myc-CUL4A or B + RBX1-FLAG overexpression plasmids for 48 h before CPT treatment (20 μM, 30 min). The cells were then subjected to DUST assay for immunodetection of ubiquitylated TOP1-DPC and total TOP1-DPC using anti-ubiquitin and anti-TOP1 antibodies. The order of DNA slot blots has been altered, as indicated by the line, and that uncropped labeled blots can be found in the Source Data file. c HCT116 transfected with EV or Myc-CUL4A + RBX1-FLAG overexpression plasmids were pre-treated with BTZ (1 μM) or APH (10 μM) for 1 h before co-treatment with CPT (500 nM) for 2 h. The cells were then subjected to ICE assay for immunodetection of TOP1-DPC using anti-TOP1 antibody. The order of DNA slot blots has been altered, as indicated by the line, and that uncropped labeled blots can be found in the Source Data file. d Densitometric analyses of TOP1-DPCs from triplicate experiments including blots in panel C. Density of TOP1-DPC/density of DNA of each group was normalized to that of cells transfected with EV only. Data are presented as mean ± SD, N = 3 biologically independent experiments. The p value was calculated using two-tailed Student’s t test. e HCT116 transfected with EV or Myc-CUL4A + RBX1-FLAG overexpression plasmids were pre-treated with BTZ (1 μM) or APH (10 μM) for 1 h before co-treatment with CPT (500 nM) for 2 h. The cells were then subjected to ICE assay for immunodetection of TOP1-DPC using anti-TOP1 antibody. f Densitometric analyses of TOP1-DPCs from triplicate experiments including blots in (e). Density of TOP1-DPC/density of DNA of each group was normalized to that of cells transfected with EV only. Data are presented as mean ± SD, N = 3 biologically independent experiments. The p value was calculated using two-tailed Student’s t test.

Fig. 6. DCAF13 connects TOP1-DPCs with CRL4 for ubiquitylation.

a A model for the assembly of the RBX1-CUL4-DDB1-DCAF ubiquitin complex. N8 NEDD8, U ubiquitin. b Scheme of 6×His-TOP1 pull-down in HCT116 cells for LC-MS/MS. Three DCAF member proteins were enriched in the 6×His-TOP1 overexpressing sample. PSM peptide spectrum match. c HCT116 cells were transfected with indicated plasmids and control siRNA (siControl) or DCAF13 siRNA, followed by CPT treatment (20 μM, 30 min) for DUST assay immunodetection of ubiquitylated TOP1-DPC and total TOP1-DPC using anti-ubiquitin and anti-TOP1 antibodies. d Immunofluorescence in IdU-labeled DCAF13-FLAG overexpressing HCT116 cells. Cells were pre-extraction and DCAF13 foci and IdU foci were monitored by instant structured illumination microscope (iSIM) using anti-FLAG and anti-BrdU antibodies. The scale bar represents 10 μm. Biological independent experiments were repeated two times. e 6×His-tagged TOP1-expressing HCT116 cells were transfected with DCAF13-FLAG overexpression plasmid or empty vector, followed by treatments with DMSO or CPT (20 µM) ± APH (10 μM, pre-treatment for 2 h) for 30 min. His-pull-down samples and cell lysates (input) were subjected to immunoblotting (IB) with indicated antibodies. f Top panel: HCT116 cells were transfected with the indicated plasmids before DMSO or CPT treatment (20 μM, 30 min), followed by proximity ligation assay (PLA) using anti-His antibody and anti-FLAG antibody. The scale bar represents 15 μm. Bottom panel: quantitation of foci per cells of each treatment group as shown in the upper panel. Data are presented as mean ± SD, n = 222 total cells. The p value was calculated using two-tailed Student’s t test. Experiments were repeated three times. g DCAF13-FLAG expressing HCT116 cells were pulse-labeled with EdU for 10 min in the absence or presence of CPT (1 μM), then chased with thymidine for 40 min, followed by iPOND analysis. iPOND pull-down samples and cellular lysates (IP) were subjected to IB with indicated antibodies. In no-click samples, desthiobiotin-TEG azide was replaced by DMSO.

Fig. 7. DCAF13 interacts with the core domain of TOP1 using its putative WD40 domains.

a Domain schematics of human TOP1. b HCT116 cells were co-transfected with DCAF13-FLAG overexpression plasmid and indicated 6×His TOP1 constructs, followed by CPT (20 μM, 30 min) in the presence of BTZ (1 μM) for His-tag pull-down. The pull-down samples and cellular lysates (input) were subjected to IB using indicated antibodies. c Model structure of human DCAF13 predicted by AlphaFold2 (yellow, green, red) superimposed with SOF1 (yeast ortholog of human DCAF13 PDB: 6ZQB_46, red, magenta, blue). d HCT116 cells were transfected with the indicated DCAF13-FLAG constructs, followed by CPT (20 μM, 30 min) in presence of BTZ (1 μM) for FLAG-IP. The IP samples and cellular lysates (input) were subjected to IB using indicated antibodies. e HCT116 cells were transfected with the indicated DCAF13-FLAG constructs followed by CPT (20 μM, 30 min) for DUST assay for immunodetection of ubiquitylated TOP1-DPC and total TOP1-DPCs using anti-ubiquitin and anti-TOP1 antibodies. The order of DNA slot blots has been altered, as indicated by the line, and that uncropped labeled blots can be found in the Source Data file. N = 1. f TOP1-DPC arrest the replication forks upon their collision and signal the DCAF13-DDB1-CUL4-RBX1 complex for its recruitment to the DPCs. DCAF13 the substrate receptor that binds TOP1-DPCs and links the DPCs to the CRL4 complexes. CUL4 is activated by mono-neddylation by RBX1, which facilitates the transfer of ubiquitin from E2 and RBX1 to TOP1-DPC. The ubiquitylation leads to proteasomal degradation of TOP1-DPCs at arrested replication forks, leading to exposure of the otherwise concealed seDSB. N8 NEDD8, U ubiquitin.