Hepatitis B virus infection disrupts homologous recombination in hepatocellular carcinoma by stabilizing resection inhibitor ADRM1

Abstract

Many cancers harbor homologous recombination defects (HRDs). A HRD is a therapeutic target that is being successfully utilized in treatment of breast/ovarian cancer via synthetic lethality. However, canonical HRD caused by BRCAness mutations do not prevail in liver cancer. Here we report a subtype of HRD caused by the perturbation of a proteasome variant (CDW19S) in hepatitis B virus-bearing (HBV-bearing) cells. This amalgamate protein complex contained the 19S proteasome decorated with CRL4WDR70 ubiquitin ligase, and assembled at broken chromatin in a PSMD4Rpn10- and ATM-MDC1-RNF8-dependent manner. CDW19S promoted DNA end processing via segregated modules that promote nuclease activities of MRE11 and EXO1. Contrarily, a proteasomal component, ADRM1Rpn13, inhibited resection and was removed by CRL4WDR70-catalyzed ubiquitination upon commitment of extensive resection. HBx interfered with ADRM1Rpn13 degradation, leading to the imposition of ADRM1Rpn13-dependent resection barrier and consequent viral HRD subtype distinguishable from that caused by BRCA1 defect. Finally, we demonstrated that viral HRD in HBV-associated hepatocellular carcinoma can be exploited to restrict tumor progression. Our work clarifies the underlying mechanism of a virus-induced HRD subtype.

Keywords: DNA repair; Hepatology; Liver cancer; Ubiquitin-proteosome system; Virology.

Figure 1. Interference with CRL4^WDR70 by HBx induces a viral HRD.

(A) Repair frequency of indicated pathways in WDR70-knockout or HBx-expressing cells relative to control cells (293T). **P <0.05 by 2-tailed t test. NS, no statistic significance. (B) Left: Example confocal images showing 53BP1 (red) and RPA32 (green) IRIF in HBx-expressing L02 cells 8 hours after IR. Soluble nuclear proteins were preextracted with 0.1% Triton X-100. Scale bar: 10 μm. Right: Pixel intensity (vertical) across the maximal central line of individual IRIFs. Precipitation of red line (53BP1) and rising of green line (RPA32) along the vertical axis indicate the central cavity. (C and D) ChIP assay depicting p-RPA32 chromatin loading at indicated distance from the DSB upon expression of gRNA (g1) targeting the PPP1R12C/p84 locus. WDR70-knockout or control 293T cells (C) and L02 cells expressing HBx (D) were cotransfected with si53BP1 or control siRNA (siScr). (E) Relative HR/SSA efficiency for L02 cells pretreated with HBx, siWDR70, or siBRCA1 and concomitant silencing of 53BP1. (F and G) Representative images (F) and quantification (G) of aberrant chromosomes in the indicated cells cotreated with olaparib (1 μM) and/or sh53BP1. (H) Giemsa staining for colony formation (left) and survival curves (right) of control L02 (HBV^–) and T43 (HBV⁺) cells subjected to olaparib treatment. Survival at endpoints was analyzed for statistical significance. n = 3 biological repeats; error bars indicate SD; t test. (I) Survival curves for L02 and T43 cells treated with siWDR70 (left) or sh53BP1 (right) with simultaneous exposure to indicated concentrations of olaparib. n = 3 experimental repeats; error bars indicate SD; P values by t test are shown for indicated groups.

Figure 2. CDW19S engages DSB-proximal chromatin.

(A) TAP-affinity-purified spWdr70-interacting proteins separated by gradient SDS-PAGE. Proteins identified by MALDI-TOF mass spectrometry are shown on the right. See peptide coverage in Supplemental Table 1. Subdomains of interface (Int), PCI, MPN, and ATPase (ATP) are indicated for RP subunits using human and yeast nomenclatures. IgH, heavy chain of rabbit IgG; TEV, tobacco etch virus endopeptidase. (B) Immunoblotting for p-RPA32 and H2B monoubiquitination (uH2B) in L02 cells with indicated siRNA and CPT treatment. (C) Left: Illustration for XbaI-based resection assay at the selected AsiSI-dependent DSB site (chromosome 1: 89,458,595–89,458,603). Right: Example of monitoring of ssDNA production by semiquantitative PCR. (D) Quantification for XbaI-based resection assay showing DSB processing in DIvA cells depleted for the indicated CDW19S subunits. Inset: Excessive p-RPA32 immunostaining implies hypersection in siADRM1^Rpn13 RPE1 cells. Data normalized to control (siScr) with 4-OHT induction. Original magnification (inset): ×400. (E) ChIP assay 2.5 kb distal to an AsiSI-induced DSB showing break association of FLAG-tagged CDW19S subunits upon PSMD4^Rpn10 ablation relative to control transfection. DIvA cells all treated with 4-OHT. (F) Left: ChIP assay for FLAG-tagged WDR70/DDB1 at indicated distances from AsiSI-induced DSB ends. Right: Representative PCR products. (G) ChIP assay of FLAG-PSMD4^Rpn10 2.5 kb from an AsiSI-induced DSB upon PSMD4^WT or PSMD4^dUIM expression. Anti-FLAG immunoblotting is shown in the right panel. (H) Equivalent ChIP assay for indicated CDW19S subunits in the presence of PSMD4^WT or PSMD4^dUIM expression. (I) Top: Representative images of p-RPA32 and BRCA1 IRIF in the presence of PSMD4^Rpn10 or PSMD4^dUIM (4 hours after IR). Nuclei counterstained with DAPI. Scale bars: 10 μm. Bottom: Quantification of fluorescent intensity or foci numbers. In H and I, PSMD4 plasmids are FLAG-less and siRNA resistant, and cells were cotransfected with siPSMD4^Rpn10. (J and K) PSMD4^Rpn10 enrichment upon 4-OHT induction at 0.5 kb from an AsiSI-induced DSB after treatment with the indicated siRNA or inhibitors.

Figure 3. Separate CDW19S modules regulate MRE11 and EXO1 activation.

(A and B) ChIP assay showing DSB loading of p-RPA32 at 2.5 kb (A) or 0.5 kb (B) distal to an AsiSI-induced DSB upon silencing of indicated CDW19S subunits. Concomitant 53BP1 knockdown was performed in A. (C) Enumeration of MRE11 foci upon silencing of individual CDW19S subunits in RPE1 cells. Immunofluorescence was carried out 30 minutes after IR. n = 3 biological repeats, 50 cells counted for each repeat. Error bars indicate SD. P values by t test are shown. (D) ChIP assay showing loading of EXO1 at 2.5 kb distal to an AsiSI-induced DSB.

Figure 4. The docking platform of CRL4^WDR70 on 19S RP.

(A) ChIP assay for FLAG-tagged WDR70 at 2.5 kb distal to an AsiSI-induced DSB following siRNA treatment of 19S subunits. Data normalized to siScramble with 4-OHT induction. (B) Coimmunoprecipitation of FLAG-PSMD4^Rpn10 and WDR70 from chromatin fractions of CPT-treated HEK293T cells with or without ablation of indicated RP components. (C) Schematic showing the high-salt procedure for screening 19S components mediating direct engagement with WDR70. (D) Pull-down assay using purified WDR70 and PSMD5^Hsm3. (E and F) In vitro pull-down assay for purified WDR70 (0.5 μg) and 19S proteasome (2 μg), the latter containing FLAG-UCHL5. Recombinant PSMD5^Hsm3 (1 μg, E) or specific antibodies (0.5 μg, F) were added into the reaction. (G and H) ChIP assay for loading of FLAG-tagged PSMD5^Hsm3 (G) or DDB1/WDR70 (H) 2.5 kb distal to an AsiSI-induced DSB following siRNA treatments. (I) Immunoprecipitation for endogenous WDR70 and FLAG-tagged 19S subunits with or without PSMD5^Hsm3 silencing.

Figure 5. CRL4^WDR70 targets ADRM1^Rpn13 for UPS degradation.

(A) Left: ChIP assay showing DSB loading of p-RPA32 at 2.5 kb distal to DSB upon silencing of ADRM1^Rpn13 (2 different siRNAs) or WDR70. Right: Equivalent assay expressing si001-resistant wild-type (WT) or mIFD mutant ADRM1^Rpn13. (B) Protein abundance of ADRM1^Rpn13 in chromatin fraction of 293T cells as measured by immunoblotting (see also Supplemental Figure 5E). WDR70 was ablated, followed by CPT insult (2 μM) for 1 hour and release into drug-free medium. (C) Immunoblotting for ADRM1^Rpn13 in cells cotreated with CPT and MG132. (D–F) Ubiquitin pull-down assay to identify polyubiquitinated species of FLAG-ADRM1^Rpn13 upon treatment with siWDR70 (D), expression of ubiquitin variants (E), or ADRM1^Rpn13 K>R mutants (F). Cells were challenged with 2 μM CPT for 2 hours. (G) Immunoblotting for CPT-induced p-RPA32 upon expression of WT, K99R, or K99-only mutant of FLAG-ADRM1^Rpn13 that is siADRM1^Rpn13 resistant. (H) Reconstitution of FLAG-ADRM1^Rpn13 ubiquitination catalyzed by purified proteins. (I) Equivalent reconstitution using WT or K99R versions. GF-ADRM1, GST/FLAG-tagged ADRM1. (J and K) ADRM1^Rpn13 ubiquitination reconstituted with addition of purified 19S and/or His-PSMD5^HSM3 (J), or in the presence of anti-WDR70 or anti-PSMD5^HSM3 (K). (L) ChIP assay for DSB-associated ADRM1^Rpn13 (2.5 kb from an AsiSI-induced DSB; top panel) and immunoblotting for chromatin-bound ADRM1^Rpn13 following CPT treatment with or without PSMD5^HSM3 silencing (bottom panel). (M) Schematic for chromatin regulation of DSB repair by coordinative action of CDW19S modules.

Figure 6. Torso CDW19S and ADRM1^Rpn13 accumulation marks HBV-induced HRD subtype.

(A) ChIP assay (left) for FLAG-tagged CDW19S subunits 2.5 kb from DSB in the presence or absence of HA-tagged HBx expression (right). Quantification was normalized to DDB1 value without HBx expression. (B) Chromatin and soluble nuclear fractionation of indicated proteins upon CPT insult with or without HBx-HA expression. Densitometry for DDB1 from 3 repeats is shown on the right. Results were obtained from identical biological samples immunoblotted from different concentrations of PAGE gels. (C) Enrichment of EXO1 or p-RPA32 loading 2.5 kb distal from DSB with HBx expression or siADRM1^Rpn13. (D) HR/SSA repair assay in the presence of HBx or siWDR70 in L02 cells, with or without concomitant siADRM1^Rpn13 treatment. (E) Schematic showing the HBx-induced “torso” CDW19S and consequent failure of ADRM1^Rpn13 removal. (F) Representative images (left) and counting (right) of 53BP1 IRIF (8 hours after IR) in WDR70-ablated cells. Simultaneous ADRM1^Rpn13 silencing was performed as indicated. Scale bar: 10 μm. (G) ChIP showing the inability of siADRM1^Rpn13 to restore the DSB loading of p-RPA32 and EXO1 in BRCA1-depleted cells. (H) Parallel comparison of HR/SSA improvement by control siRNA (green) or siADRM1^Rpn13 (orange) in BRCA1- and WDR70-depleted cells. P values for multiple-group comparison in B–D, F, and G were calculated by 2-way ANOVA test.

Figure 7. HBV-induced HRD subtype sensitizes HBVHCC to PARP inhibition.

(A) Responses of T43 xenografts to monotreatment with olaparib or vehicle. Error bars indicate SD; t test. (B and C) Responses of T43 cells (B, 3 biological repeats) and xenografts (C, 6 littermates included) to conjunctive administration of olaparib and cisplatin. Tumor volumes are presented as means ± SD. DMSO: equivalent amount of solvent solution. Error bars indicate SD; P values were calculated by t test (B) and by 2-way ANOVA test (C). (D) Schematic of PARPi administration to HCC engraftment in NOD-SCID mice (top) and tumor responses (bottom). Tumor volumes of 4 HBVHCCs (patients 16, 17, 19, and 23) and 1 HBV-free HCC (patient 76, progressive disease) are shown at indicated days after inoculation. Olaparib: 33.3 mg/kg/d; cisplatin: 0.5 mg/kg/2 days (O/C). Horizontal axis, days after tumor transplantation; arrows, starting date of medication. Numbers of animals were as follows: patient 16 (vehicle, n = 4; treatment, n = 4), patient 17 (vehicle, n = 2; treatment, n = 2), patient 19 (vehicle, n = 3; treatment, n = 2), patient 23 (vehicle, n = 3; treatment, n = 2), patient 76 (vehicle, n = 3; treatment, n = 3). (E) Tumor response for HBVHCC patient-derived xenograft (PDX) sublines treated with vehicle or O/C at week 2–3. Graphs show mean ± SEM, analyzed with 2-sided unpaired Student’s t test. (F) Kaplan-Meier plot indicating progression-free survival of HBVHCC sublines. The y axis is the percentage of animals whose tumor volumes were smaller than 300 mm³. P value was calculated by log-rank test.