Combinatorial RNA Interference Therapy Prevents Selection of Pre-existing HBV Variants in Human Liver Chimeric Mice

Abstract

Selection of escape mutants with mutations within the target sequence could abolish the antiviral RNA interference activity. Here, we investigated the impact of a pre-existing shRNA-resistant HBV variant on the efficacy of shRNA therapy. We previously identified a highly potent shRNA, S1, which, when delivered by an adeno-associated viral vector, effectively inhibits HBV replication in HBV transgenic mice. We applied the "PICKY" software to systemically screen the HBV genome, then used hydrodynamic transfection and HBV transgenic mice to identify additional six highly potent shRNAs. Human liver chimeric mice were infected with a mixture of wild-type and T472C HBV, a S1-resistant HBV variant, and then treated with a single or combined shRNAs. The presence of T472C mutant compromised the therapeutic efficacy of S1 and resulted in replacement of serum wild-type HBV by T472C HBV. In contrast, combinatorial therapy using S1 and P28, one of six potent shRNAs, markedly reduced titers for both wild-type and T472C HBV. Interestingly, treatment with P28 alone led to the emergence of escape mutants with mutations in the P28 target region. Our results demonstrate that combinatorial RNAi therapy can minimize the escape of resistant viral mutants in chronic HBV patients.

Figure 1. T472C HBV is partially resistant to S1 shRNA.

(a) The nucleotide sequence and corresponding amino acid sequence within the S1 target site of wild type (WT) and T472C HBV genome. The triplet codons for surface and polymerse open reading frame (ORF) are illustrated by lines located above and below the sequence, respectively. The mutated nucleotide in T472C HBV is shown in red, which does not change the amino acid sequence in either surface or polymerase genes. (b) Reduction in HBsAg levels in cell culture medium by shRNAs. Huh-7 cells were co-transfected with shRNA- and wild-type HBV (WT, pHBV1.3) or T472C mutant HBV-encoding plasmids. Mock transfection served as negative controls. Three days later, HBsAg levels in the supernatant were measured by ELISA and are presented as a percentage of those with mock transfection (each value represents the mean ± SD of three independent transfection experiments). **P < 0.01. (c) Schematic diagram of hydrodynamic co-injection used to analyze the S1-resistance of T472C HBV. C57BL/6 mice (n = 7 per group) were hydrodynamically co-injected with equal amounts of WT and T472C HBV plasmids and a plasmid encoding either S1 or GL2 shRNA, then, 3 days later, serum HBV DNA was extracted for PCR amplification and the ratio of the two HBV strains estimated by sequencing the PCR product corresponding to the S1 target region. * indicates the position of nt 472. (Picture credit: Y.M.S.)

Figure 2. Screening of PICKY-predicted shRNAs by co-transfection of HBV- and shRNA-encoding plasmids.

(a) Schematic representation of the HBV genome (bold line), the four transcripts (thin lines), and the open reading frames (ORF, boxes). The approximate locations of the target sequences predicted by the PICKY software are indicated. (b) Huh-7 cells were transfected with wild-type HBV plasmid alone or together with a plasmid encoding the indicated shRNA (each value represents the mean ± SD of three independent transfection experiments) or (c) C57BL/6 mice (n = 5–7) were hydrodynamically injected with wild-type HBV plasmid alone or together with a plasmid encoding the indicated shRNA, then, 3 days later, HBsAg levels in the culture supernatant (b) or serum (c) were measured. The data are presented as a percentage of the level produced by cells transfected with HBV plasmid alone (b) or mice injected with HBV plasmid alone (mock) (c) (mean ± SD).

Figure 3. Screening of PICKY-predicted shRNAs in HBV transgenic mice.

ICR/HBV mice (n = 3–5) were injected i.v. with 1 × 10¹² vg per mouse of AAV8 vector encoding the indicated shRNA or saline. (a) Serum samples were collected before treatment and at 2 or 6 weeks after treatment and the HBV DNA titer determined (mean ± SD). (b,c) The mice were sacrificed at week 6 and liver tissues collected for (b) Southern or (c) Northern blot analysis. (b) Total liver DNA was analyzed for HBV replicative intermediates (upper panel) and AAV vector genomes (lower panel). Bands corresponding to the integrated transgene (Tg) and the relaxed circular (RC) or single-stranded (SS) linear HBV DNA replicative form are indicated. The integrated transgene was used to normalize the amount of DNA loaded on the gel. (c) Total liver RNA was analyzed for the 3.5 kb and 2.4/2.1 kb HBV transcripts (upper panel); ethidium bromide-stained 18 S and 28 S ribosomal RNA (rRNA) served as the loading controls (lower panel).

Figure 4. Evaluation of the therapeutic efficacy of AAV-mediated shRNAs in HBV-infected hu-FRG mice.

(a–d) HBV-infected hu-FRG mice (n = 3–4) were injected i.v. with 1 × 10¹² vg per mouse of AAV8 vector encoding (a) GL2, (b) P28, (c) S1, or (d) S1 + P28 shRNA, then serum samples were collected at different times and the HBV DNA titer determined by qPCR. The identification numbers of the animals in each group are shown in the lower left corner of each panel. (e,f) Levels of HBsAg (e) and HBeAg (f) in the serum samples were determined before treatment and at 2 or 5 weeks after treatment (mean ± SD).

Figure 5. Effect of shRNA treatment on intrahepatic HBcAg in HBV-infected hu-FRG mice.

HBV-infected hu-FRG mice treated with different AAV/shRNAs as described in Fig. 4 were sacrificed at week 5 post AAV injection and liver samples obtained for immunofluorescent staining. Human hepatocytes were visualized by red staining of human fumarylacetoacetate hydrolase (FAH) (middle panels), HBcAg staining was shown in green (left panels), and DAPI (blue) was used as nuclear counterstaining. The merged pictures show HBcAg expression in human hepatocytes (yellow; right panels). Original magnification, × 400; bar, 100 μm.

Figure 6. HBV variant analysis in AAV8/shRNA-treated hu-FRG mice.

(a) Left panel: Setting up of the PCR-RFLP assay for rapid discrimination of wild-type and T472C HBV. The wild-type and T472C HBV plasmids were mixed at the indicated ratios and digested with BccI to generate different ratios of the S1-wt (195 bp) and S1-T472C (129 bp) fragments for use as a standard. Right panel: Serum samples collected at 5 weeks after AAV8/shRNA treatment from the huFRG mice described in Fig. 4 were analyzed by PCR-RFLP assay and the DNA fragments separated on a 4% agarose gel. The faster migrating band indicates the T472C HBV sequence (S1-T472C) and the slower migrating band indicates the wild-type HBV sequence S1-wt. (b) Sequence analysis of the P28 target regions in the serum HBV genome from mouse 4T-3 at 5 weeks after (top panel) and before (bottom panel) AAV8/P28 treatment. (c) Analysis of intrahepatic cccDNA. Liver samples collected at 5 weeks post AAV8/shRNA treatment from the huFRG mice described in Fig. 4 were used in a real-time qPCR assay, using selective cccDNA primers and FRET hybridization probe, to quantify the intrahepatic cccDNA levels. The amount of human mitochondria DNA was quantified using specific primers and used to normalize cccDNA copies (mean ± SD) per human hepatocyte (expressed as human genome equivalents) determined in chimeric livers (top panel). Sequence analysis of the shRNA target regions in the intrahepatic cccDNA extracted from AAV8/S1 or AAV8/P28 treated mice. The number of clones with wild-type and mutations in the S1 and P28 regions are shown (bottom panel).