Human serum from patients with chronic HBV/HDV infection followed at the Strasbourg University Hospitals, Strasbourg, France was obtained with informed consent. PHHs were obtained from liver tissue from patients undergoing liver resection for liver metastasis at the Strasbourg University Hospitals with informed consent. Protocols were approved by the local Ethics Committee of the Strasbourg University Hospitals (CPP) and the Ministry of Higher Education and Research of France (DC-2016-2616). Human samples from HBV-infected patients followed at the Chang Gung Memorial Hospital (Taipei, Taiwan) were obtained with informed consent. Protocols were approved by the local Ethics Committee (Institutional Review Board 102-3825C).
Cell lines and viruses
NTCP-overexpressing Huh-106 and HepG2-NTCP cell lines5,57 as well as human embryonic kidney 293T (HEK 293T)58 cell line have been described. PHHs were isolated and cultured as described58. Recombinant HDV production5,57 as well as purification of infectious HBV particles from the inducible human hepatoblastoma HepAD38 has been described5,59,60.
Reagents and plasmids
DMSO, polybrene, and PEG 8000 (polyethylene glycol) were obtained from Sigma-Aldrich (Merck). DNA and RNA transfection at the indicated concentrations was performed using the CalPhos Mammalian Transfection Kit (Clonetech) and Lipofectamine RNAiMAX (Thermo Scientific) according to the manufacturers’ instructions, respectively. The ORF-encoding lentivirus constructs for validations were obtained from the RNAi Platform, Broad Institute of MIT and Harvard (Cambridge, MA, USA). Cell viability/proliferation was assessed using PrestoBlue Cell Viability Reagent (Invitrogen) according to the manufacturer’s instructions. Cell toxicity was assessed using LDH-Glo cytotoxicity assay (Promega) in the supernatant according to the manufacturer’s instructions. Palbociclib and LEE011 (Ribociclib) were obtained from Synkinase and Sellekchem, respectively.
The binding of HBV virions at the cell surface was assessed as described5. In brief, cells were incubated with pretreated HBV in the presence of 4% PEG for 24 h at 16 °C. Unbound virions were removed by three washes with phosphate-buffered saline (PBS), and cells and bound virions were lysed. HBV total DNA was quantified by qPCR using a standard curve generated from known HBV genome copies.
HBV and HDV infections
For HBV infection, NTCP-overexpressing cell lines and PHHs were infected by recombinant HBV in the presence of 4% of PEG-8000 (GEq 500 or 1000 per cell)5,60. After infection, Huh7-106 and HepG2-NTCP cells were washed and culture in PMM medium with 2% or 3.5% of DMSO, respectively for 10 days. HBV infection was assessed 10 dpi by IF using a mouse monoclonal antibody (Ab) targeting HBsAg (Bio-Techne, clone 1044/329, 1:100) and Alexa Fluor 647-labeled secondary Ab targeting mouse immunoglobulin G (IgG; Jackson Research, 1:200). Cell nuclei were stained with 4,6-diamidino-2-phenylindole. Fluorescent imaging was performed using an Axio Observer Z1 microscope (Carl Zeiss, Germany). Alternatively, cells were lysed and total RNA was extracted using the ReliaPrep RNA Miniprep Systems (Promega) and quantitative reverse transcriptase PCR (qRT-PCR) quantification of HBV pgRNA was assessed as described5,60,61. HBsAg and HBeAg secretion were quantified by chemiluminescence immunoassay (Autobio) following the manufacturer’s instructions. Southern blot detection of HBV cccDNA was performed using digoxigenin (DIG)-labeled (Roche) specific probes as described62. Total DNA from HBV-infected cells was extracted using the Hirt method as described63. Specific DIG-labeled probes for the detection of HBV and mitochondrial probes for the detection of HBV and mitochondrial DNAs were synthetized using the PCR DIG Probe Synthesis Kit (Roche) and the primers as described60. HBV total RNAs were detected by northern blot. Total RNA was purified using ReliaPrep RNA Miniprep Systems (Promega). Five µg of total RNA was subjected to electrophoresis through 2.2 M formaldehyde and 1% agarose gel and transferred to a nylon membrane positively charged (Roche). The membrane-bound RNA was hybridized to a 32P-labeled RNA probe specific for detection of HBV RNA of 1200–1944 bp of viral genome (3.5–2.1 kbp). Quantification of HBV DNA and RNA bands from blots was performed using Image Lab Version 5.2.1 (Bio-Rad). For HDV infection, NTCP-overexpressing cell lines were infected with recombinant HDV (GEq 100 per cell) as described5,60. HDV infection was assessed 7 days after infection by IF using an Ab targeting the hepatitis delta antigen (1:200) purified from serum of an HBV/HDV co-infected patient64 and AF647-labeled secondary Ab targeting human IgG (Jackson Research, 1:200) as described5,65.
Genome-scale lentiviral expression library and gain-of function screen
hORFeome V8.1 library (Broad Institute of MIT and Harvard, Cambridge, MA, USA) containing a pool of 16,172 clonal ORFs (mapping 13,833 human genes) was cloned into a pLX_TRC317 vector. The establishment of the genome-scale ORFeome library has been described18. Thirty million Huh-106 cells were transduced with the lentiviral ORFeome library in duplicate in the presence of polybrene (4 µg/mL). To avoid a cumulative effect of multiple ORFs, the lentivirus volume was optimized to obtain 30% of transduced cells. Cells were then selected with puromycin (0.9 µg/mL) for 3 days. After amplification, transduced cells were infected with recombinant HBV at an MOI of 1000 GEq/cell or mock-infected. At 10 dpi, cells were stained for HBsAg expression and sorted by flow cytometry.
Gene expression analysis in HBV-infected Huh-106 after ORFeome transduction
HBV-infected cells were fixed in 100% methanol for at least 20 min at −20 °C. Cells were then blocked and permeabilized using PBS–0.5% bovine serum albumin (BSA) and 0.05% saponin for 1 h at room temperature (RT). Cells were stained using an AF647-conjugated mouse monoclonal anti-HBsAg Ab (Bio-Techne, clone 1044/329) and resuspended in 0.5% BSA. HBsAg-positive cells were sorted by FACS (BD FACSAria Flow Cytometer). Twenty million cells were taken from HBV-infected sample as pre-sort control and total genomic DNA (gDNA) was extracted from cell pellets using Qiagen kits according to the manufacturer’s protocol (Qiagen). In addition, gDNA was extracted from 20 million HBV-positive sorted cells from two biological replicates, named HBV sorted. Extracted DNA was used as a template for PCR to amplify the barcode sequences that accompany every ORF in the library. The unique barcode associated with each ORF construct was determined by Sanger sequencing in an arrayed collection of all the ORF constructs prior to pooling. PCR and sequencing were performed as previously described66,67. The details of the PCR primers and conditions can be found here: https://portals.broadinstitute.org/gpp/public/resources/protocols. Samples were sequenced on a HiSeq2000 (Illumina). The resulting reads were matched to their barcodes and their associated ORFs using PoolQ (see https://portals.broadinstitute.org/gpp/public/resources/protocols for more information on PoolQ). For analysis, the read counts were normalized to reads-per-million and then log2 transformed. Log2FC of each ORF was determined relative to the initial time point for each biological replicate. Ninety hits with Log2FC values above the threshold set at 1.5 were selected as candidates.
For further flow cytometric analysis of HBV-infected cells, cells were fixed in 100% methanol for at least 20 min at −20 °C. Cells were then blocked and permeabilized using PBS–1% fetal bovine serum and 0.05% saponin for 30 min at RT. HBsAg was stained using a mouse monoclonal anti-HBsAg Ab (Bio-Techne, clone 1044/329, 1:1000) for 30 min at 4 °C and then with an AF647-labeled secondary Ab targeting mouse IgG (Jackson Research, 1:1000) for 30 min at 4 °C. For flow cytometric analysis of DNA content, cells were fixed in ice-cold 75% ethanol in water for 30 min at 4 °C. Cells were washed and resuspended and incubated in PBS 50 µg/mL propidium iodide (Invitrogen) and 50 µg/mL Ribonuclease A (Sigma-Aldrich, Merck) for 30 min at RT. Cells were subsequently washed and resuspended in PBS–5 µM EDTA prior to sorting through a CytoFLEX flow cytometer system (Beckman Coulter). The gating strategy is presented in Supplementary Fig. 7.
Candidate selection from the primary screen
The impact of gene overexpression on HBV infection was defined by a specific enrichment in cDNA sequences in HBV-positive sorted cells compared to the pre-sort population. For hit selection, a functional threshold of Log2FC = 1.5 compared to pre-sorted cells was applied, leading to a total of 90 candidates (Supplementary Table 1, Fig. 2c, d). As multiple ORF sequences for one given gene are sometimes present in the library, individual sequences were analyzed. Candidate genes with multiple associated ORFs were selected only if clones presented significant differences in their sequences (truncations in Cter or Nter of the proteins) or if at least two identical ORFs exhibited a Log2FC > 1. Candidate gene expression in the liver was then assessed through the Human Protein Atlas (available from www.proteinatlas.org)68. Candidates with liver expression <0.1 transcript per million were removed from the analysis, leading to a final selection of 47 candidates (Supplementary Table 1). Forty-seven ORF-containing lentiviruses were then obtained for individual validations, 35 of which met internal quality control based on lentiviral titration. In addition, lentiviruses encoding GFP, KRT80, and CPA1 cDNA sequences were obtained as negative controls from the primary screen.
Hit validation in Huh-106 cells and PHHs
Individual ORFs were expressed from pLX-Blast-V5 (lentiviral) expression plasmids. Lentivirus particles were produced in HEK 293T cells by cotransfection of plasmids expressing the human immunodeficiency virus gap-pol, the vesicular stomatitis virus glycoprotein, and the pLX-Blast-V5-ORF plasmids in the ratio of 10:3:10, using the CalPhos Mammalian Transfection Kit as described58. Three days after transfection, supernatants were collected, pooled, and clarified using 0.45-µm pore filters. Huh-106 were individually transduced with the 38 ORF-expressing lentivirus constructions and selected with 6 µg/mL of blasticidin 48 h prior to HBV infection. HBV infection was assessed after 10 days by quantification of HBeAg and HBsAg expression in the supernatant of infected cells as described above. For further validations, PHH and Huh-106 were transduced with individual ORF-containing lentivirus prior to HBV infection. Infection was assessed after 10 days by Southern blot detection of HBV DNA, northern blot and qRT-PCR detection of HBV RNAs, immunodetection of HBsAg, and quantification of HBeAg as described above.
CDKN2C HepG2-NTCP KO generation
To generate clonal HepG2-NTCP CDKN2C KOs, the following primers corresponding to guide RNAs targeting CDKN2C exons were cloned into the Zhang laboratory-generated Cas9 expressing pX458 plasmid (Addgene plasmid #48138): guide 1; Fw: 5’-CACCGACACCGCCTGTGATTTGGCC-3’, Re: 5’-AAACGGCCAAATCACAGGCGGTGTC-3’. guide 2; Fw: 5’-CACCGCACAGGCGGTGTCCCCCTTA-3’, Re: 5’-AAACTAAGGGGGACACCGCCTGTGC-3’. pX458 plasmids encoding guide RNAs against CDKN2C were transfected into HepG2-NTCP cells using Lipofectamine 3000 (Life Technologies) according to the manufacture’s guidelines. Transfected cells were single cell sorted based on +GFP expression into 96-well plates using the SONY SH800S cell sorter. Individual clones were expanded, and four clonal cell lines were eventually selected for further characterization.
RNAi loss-of-function studies
ON-TARGETplus siRNA pools (Dharmacon) targeting the transcripts of CDKN2C and SLC10A1 (NTCP) were reverse-transfected into HepG2-NTCP cells with Lipofectamine RNAiMAX (Invitrogen) as described. RNA was purified from cells harvested 2 days after transfection, and gene expression was analyzed by qRT-PCR. For silencing of CDKN2C expression in PHHs, PHHs were transduced with lentiviral vectors containing CDKN2C-targeting shRNA (target sequence: GATGTTAACATCGAGGATAAT) or a scrambled shRNA control (target sequence: CCTAAGGTTAAGTCGCCCTCG) obtained from VectorBuilder. RNA was purified from PHHs harvested 3 days after transduction, and gene expression was analyzed by qRT-PCR.
Comparative analysis of gene expression in Huh-106 and HepG2-NTCP cells
Huh-106 and HepG2-NTCP cells were lysed and total RNA from three biological replicates per cell line was then extracted as described above. Microarray analysis of gene expression in both cell lines was performed at the IGBMC GenomEast platform (Illkirch, France). Biotinylated single-strand cDNA targets were prepared from 200 ng of total RNA using the Ambion WT Expression Kit (Cat # 4411974) and the Affymetrix GeneChip® WT Terminal Labeling Kit (Cat # 900671) according to Affymetrix recommendations. Following fragmentation and end-labeling, 3 μg of cDNAs were hybridized for 16 h at 45 °C on GeneChip® Human Gene 2.0 ST arrays (Affymetrix) interrogating >400,000 RefSeq transcripts and ~11,000 long non-coding RNAs. The chips were washed and stained in the GeneChip® Fluidics Station 450 (Affymetrix) and scanned with the GeneChip® Scanner 3000 7G (Affymetrix) at a resolution of 0.7 µm. Raw data (CEL Intensity files) were extracted from the scanned images using the Affymetrix GeneChip® Command Console (AGCC) version 4.1.2. CEL files were further processed with the Affymetrix Expression Console software version 1.4.1 to calculate probe set signal intensities using Robust Multi-array Average algorithms with default settings. Modulated molecular pathways were determined by using gene set enrichment analysis69. Individual differential gene expression of the selected candidates was evaluated through the Z score transformation. The dataset is publicly available in the NCBI Gene Expression Omnibus database (accession number GSE132638).
Analysis of gene expression using qRT-PCR
RNA was extracted as described above, and gene expression was assessed by qRT-PCR as described60. Gene expression was normalized to GADPH expression. Primers and TaqMan® probes for quantification of GAPDH, CDKN2C, and SLC10A1 mRNA expression were obtained from ThermoFisher (TaqMan® Gene Expression Assays). Gene expression was quantified using iTaq Universal Probes Supermix (Bio-Rad). Primers for quantification of HNF4A (Fw: 5’-ACATTCGGCAAGAAGATT-3’; Re: ACTTGGCCCACTCAACGAG-3’), HLF (Fw: 5’CACCACGAAGACGATTTAG-3’; Re: 5’-CAAAAACTCCTCCAGGTCCA-3’), PPARA (Fw: 5’-GAGGGTCTCCACTGACGTG-3’; Re: 5’-ACACTGTGTATGGCTGAGAAG-3’), and GAPDH expression (Fw: 5’-GTCTCCTCTGACTTCAACAGCG-3’; Re: 5’-ACCACCCTGTTGCTGTAGCCAA-3’) were obtained from Sigma-Aldrich (Merck). Gene expression was quantified using iTaq Universal SYBR Green Supermix (Bio-Rad).
The expression of CDKN2C and β-tubulin was assessed by western blot as described5 using a monoclonal rabbit anti-CDKN2C Ab (anti-p18 INK4c, ab192239, Abcam, 1:1000), a rabbit polyclonal anti-β-tubulin Ab (GTX101279, Gentex, 1:3000), and a rabbit polyclonal anti-glyceraldehyde 3-phosphate dehydrogenase (anti-GAPDH; ab9485, Abcam, 1:2500), respectively. Peroxidase-AffiniPure Goat Anti-Rabbit IgG (H+L) (Jackson Research 111-035-144, 1:10,000) was used as a secondary Ab. Protein expression was assessed using the ChemiDoc™ Imaging System (BioRad).
Analysis of nascent HBV RNA synthesis
Run-on assays were performed using the Click-iT™ Nascent RNA Capture Kit from ThermoFisher Scientific according to the manufacturer’s instructions. HBV total and nascent RNA expression was assessed from HBV-infected Huh-106 cells overexpressing either GFP or CDKN2C by qRT-PCR 4 days after virus inoculation with 2 h of ethynyl uridine (EU) labeling. Actinomycin D (ActD, Sigma-Aldrich, Merck) was used as a negative control. Cells were pretreated with ActD at 10 mg/mL for 20 min prior to EU labeling in the presence of ActD. Specific primers and TaqMan® probes for total HBV RNAs (Pa03453406_s1) were purchased from Life Technologies. HBV RNA levels were normalized to GUSB expression using primers and TaqMan® probes from Life Technologies (Hs99999908_m1).
Analysis of CDKN2C expression in patients
For the analysis of CDKN2C mRNA expression in patients, CDKN2C mRNA expression was assessed in control healthy patients (n = 6), HBV-infected patients with no detectable HBV DNA (n = 32), and HBV-infected patients with detectable HBV DNA (n = 90) from GSE8314870. Similarly, CDKN2C mRNA expression was assessed in HBV patients at different stages of virus infection, including immune tolerant phase (n = 22), immune clearance phase (n = 50), and inactive carrier phase (n = 11) from GSE65359. Alternatively, total RNA was extracted from liver tissue of nine HBV-infected patients by using the High Pure RNA Paraffin Kit (Roche) according to the manufacturer’s instruction, and gene expression analysis was performed by RNA-seq as previously reported71. To analyze the correlation between CDKN2C expression and the progression of liver disease in HBV-infected patients, CDKN2C mRNA expression was assessed in HBV-related liver fibrosis patients of different stages from GSE8404472 (n = 37 score 0, n = 33 score 1, n = 34 score 2, n = 15 score 3). Finally, CDKN2C expression in HBV-induced HCC patients was assessed from GSE6548573 (n = 50 tumor tissue, n = 5 non-tumor tissue) and from GSE1452074 (n = 221 tumor tissue, n = 199 non-tumor tissue). CDKN2C mRNA expression is shown as signal intensity values. For survival analysis, liver expression level of CDKN2C and survival data were derived from The Cancer Genome Atlas (TCGA; https://www.cancer.gov/about-nci/organization/ccg/research/structural-genomics/tcga) TCGA-LIHC database75. To analyze CDKN2C expression in liver tissue of patients with chronic liver disease, fragments per kilobase of transcript per million mapped read values and clinical data were retrieved from TCGA. This dataset includes samples from HCV-infected patients (34 tumor samples including 5 paired tumor/non-tumor samples), HBV-infected patients (76 tumor samples including 7 paired tumor/non-tumor samples), patients with ALD (72 tumor samples including 8 paired tumor/non-tumor samples), and patients with NAFLD (11 tumor samples including 2 paired tumor/non-tumor samples).
Statistics and reproducibility
Individual experiments were reproduced three times in an independent manner with similar results except otherwise stated. The precise number (n) of biologically independent samples used to derive statistics is indicated in the figure legends. For in vitro experiments, statistical analyses were performed using a two-tailed Mann–Whitney U test; p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***) were considered statistically significant. Significant p values are indicated by asterisks in the individual figures and figure legends. The exact p values are provided in the Source Data file. For n < 10, the corresponding data points are presented with the bar charts. For microarray analyses, two-tailed unpaired Student’s t test was performed by comparing the values from three biological replicates per cell line. p < 0.01 was considered statistically significant. For clinical data, Mann–Whitney U test was used when comparing two groups (Fig. 9d, e). For multiple group comparison (Fig. 9b, c), Kruskal–Wallis H test adjusted for multiple comparisons was used. Correlation between CDKN2C expression and HBV viral load in patients was assessed using Spearman’s rank correlation coefficient (Spearman’s rho). Survival functions depending on CDKN2C expression were obtained using the Kaplan–Meier estimator. p value was calculated using log-rank test for comparisons of Kaplan–Meier survival. p < 0.01 was considered statistically significant. Representative graphs and pictures presented in Figs. 1a, c, e; 4c; 5a, d, f; and 7f are representative of three independent experiments with similar results. Representative graph presented in Fig. 3e is representative of two independent experiments with similar results. Graphs were designed using GraphPad PRISM 6 for Windows and Microsoft Excel for Microsoft Office 365 ProPlus (version 1911).
Further information on research design is available in the Nature Research Reporting Summary linked to this article.