An RNA-based system to study hepatitis B virus replication and evaluate antivirals

Abstract

Hepatitis B virus (HBV) chronically infects an estimated 300 million people, and standard treatments are rarely curative. Infection increases the risk of liver cirrhosis and hepatocellular carcinoma, and consequently, nearly 1 million people die each year from chronic hepatitis B. Tools and approaches that bring insights into HBV biology and facilitate the discovery and evaluation of antiviral drugs are in demand. Here, we describe a method to initiate the replication of HBV, a DNA virus, using synthetic RNA. This approach eliminates contaminating background signals from input virus or plasmid DNA that plagues existing systems and can be used to study multiple stages of HBV replication. We further demonstrate that this method can be uniquely applied to identify sequence variants that confer resistance to antiviral drugs.

Fig. 1.. The RNA launch system.

From left to right, the figure shows that in vitro–transcribed pgRNA is (I) translated (HBcAg is detected by immunofluorescence in ~80% of Huh-7.5-NTCP cells), (II) reverse-transcribed [qPCR detects HBV DNA and the signal is HBV polymerase dependent, as a catalytic site mutation (Pol⁻, YMHD) decreases the DNA signal by 10,000-fold], and (III) transported to the nucleus to form cccDNA (Southern blot confirms that cccDNA is produced; see below for lane description); (IV) cccDNA is transcribed to produce viral mRNA and protein [shown by HBsAg chemiluminescent immunoassay (CLIA)]; and (V) in vitro–transcribed pgRNA transfected into mpPHH and engrafted into humanized mice initiates productive virus infection, whereas ribonuclease A (RNase A)–treated pgRNA fails to initiate infection. Red asterisks indicate mice considered HBV-positive based on having at least two consecutive samples with rising HBV DNA levels above the lower limit of quantification (LLOQ) indicated on the graph. LLOQ for CLIA assay = 0.05 IU/ml, corresponding to less than 1% of the untreated control. Values plotted for HBcAg staining, qPCR, and HBsAg CLIA are n = 3 ± SEM. WT, wild type; IU, international units; N.D., not detected; copies per well, one well of a six-well plate; kb, kilobase; Southern blot: lane 1, 1-kb ladder; lane 2, WT HBV DNA; lane 3, WT HBV DNA heated to 85°C; lane 4, WT HBV DNA heated to 85°C and then linearized with Eco RI; lane 5, Pol⁻ HBV DNA; RC, relaxed circular DNA; DSL, double-stranded linear DNA; CCC, covalently closed circular DNA; SS, single-stranded DNA. For mouse experiments, n = 12 pgRNA; n = 10 pgRNA + RNase.

Fig. 2.. RNA launch method can be used to evaluate pan-genotype HBV replication and study multiple classes of anti-HBV drugs.

(A) Left: Unrooted phylogenetic tree of HBV genotypes A to H. The tree was generated using SeaView (46). The scale bar indicates nucleotide substitutions per site. Middle: qPCR of intracellular HBV DNA for HBV genotypes A to H 2 days after transfection with and without 10 μM ETV. Right: HBsAg from HBV genotypes A to H was measured by CLIA 6 days after transfection with and without 10 μM ETV. For both qPCR and HBsAg CLIA, 3 ± SEM. (B) Left to right: The figure shows that in vitro–transcribed pgRNA can be used to test the efficiencies of (I) antisense oligonucleotides (ASOs), (II) capsid assembly modulators (CAMs), (III) nucleotide RT inhibitors (NRTIs), and (IV) interferons (IFNs). Secreted HBsAg was quantified by CLIA and normalized to controls at 4 days (ASOs) or 6 days (CAMs and NTRIs) after transfection with HBV genotype A pgRNA as described in Materials and Methods. Data plotted are n ≥ 3 for NRTIs and n ≥ 6 for other inhibitors. Error bars are ±SEM. LLOQ = 0.05 IU/ml, which corresponds to less than 1% of the untreated control. BPV, bepirovirsen; IRR, irrelevant ASO control; LAM, lamivudine; TDF, tenofovir disoproxil fumarate.

Fig. 3.. RNA launch method can be used to test drug efficiency and identify drug resistance mutations.

(A) HBV DNA copy number by qPCR 2 days after transfecting cells with WT or drug-resistant pgRNA variants in the presence of antivirals. Left, 2.8 μM Bay 41-4109; middle, 400 μM LAM; right, 128:1 molar ratio ASO:pgRNA. Values plotted are mean with each replicate indicated as a separate dot. (B) Fold enrichment of drug-resistant HBV DNA 2 days after transfection. Drug-resistant pgRNAs were mixed with WT at 1:99 ratio with the same drug concentration described in (A). Values plotted are mean with each replicate indicated as a separate dot. (C) Sequencing HBV DNA from cells transfected with WT HBV pgRNA in the presence of 400 μM LAM identifies the two most common amino acid substitutions, i.e., M204V and M204I, found in LAM-resistant patients. Threshold = Bonferroni correction, alpha = 0.05. (D) Positively enriched amino acid substitutions were obtained from sequencing a deep mutational scanning (DMS) library covering the RT domain of HBV Pol in the presence of 400 μM LAM. (E) AlphaFold 2 model of HBV RT domain overlaid on a crystal structure of the HIV RT in complex with dsDNA (Protein Data Bank ID: 1C9R). The HIV protein is hidden. The three common amino acids found mutated in LAM-treated patients are colored red. Other amino acid positions positively enriched (arbitrary cutoff = 14-fold) are colored pink.

Fig. 4.. Nucleotide enrichment pattern in HBV DNA following BPV treatment.

(A) Design of the 3N (left) and 5N (right) deep mutation scanning library used to enrich BPV-resistant HBV sequences. Colors indicate nucleotide modifications in BPV: red, unmodified gap DNA region; blue, 2′MOE-modified RNA wings. Direct repeat 2 (DR2) is indicated with a brown line. (B) Positional enrichment of nucleotide mutants in the 3N (left) and 5N (right) libraries when comparing BPV-treated samples to IRR-treated samples. Each dot represents the combined P value for all three non-WT nucleotides at a given position. (C) Logo plot describes the nucleotide composition at each site. (D) π values for HBV individual sequences with the number of mutations per sequence binned as indicated on the x-axis. The total number of sequences in each bin with π ≥ 3 is indicated on top. (E) Dose-response curve of inhibition for WT and C9A mutant pgRNA in the presence of BPV or IRR ASOs. HBsAg was measured by CLIA 4 days after transfection and normalized to untreated control. R square and 95% confidence intervals are presented in table S2.