- 现金
- 17064 元
- 精华
- 12
- 帖子
- 9399
- 注册时间
- 2007-6-26
- 最后登录
- 2017-11-25
|
We transduced HepG2.2.15 cells with Cas9-2A-Puro lentiviruses encoding Cas9 and individual sgRNAs (sg6, sg17, sg21) chosen based on our initial results, and treated cells with puromycin to select for transduced cells (Fig. 2a). As controls, cells were also transduced with constructs containing sgRNAs and a nuclease deficient Cas9 (D10A/H840A; dCas9) to control for nuclease-independent effects of Cas9 on viral fitness, or WT Cas9 with mutated sgRNAs (gXM) to control for guide sequence-independent effects. Cas9/sgRNAs induced robust suppression of HBV DNA release (77-95% decrease across different sgRNAs), HBeAg secretion, and viral mRNA production (greater than 50%) (Fig S4). We next analyzed the effect of Cas9-mediated cleavage on the abundance of non-integrated viral forms, composed mainly of cccDNA (See Methods). Quantitative PCR showed a robust reduction in total HBV DNA and in cccDNA. Pooling the data from sg6, sg17, and sg21, cccDNA reduction progressed from 71鈥�+鈥�/-7% reduction at day 21 to 92鈥�+鈥�/-4% at day 36 post transduction (Fig. 2b,c, data for individual sgRNAs in Fig S5). These results were confirmed by directly analyzing low molecular weight DNA from transduced cells by Southern blot (Fig. 2d). cccDNA and its deproteinated relaxed circular form (dpRC DNA) precursor were greatly depleted in Cas9/sgRNA transduced cells. In contrast, when total HBV DNA was analyzed, no substantial reduction in the levels of integrated HBV DNA was detected (Fig S6).
We then performed the Surveyor assay on HBV, to directly determine whether the viral DNA was cleaved and repaired via error-prone NHEJ similar to genomic targets of CRISPR/Cas9. Interestingly, analysis of total HBV DNA forms for indel formation, an indirect measure of Cas9-mediated cleavage, revealed a substantial mutagenesis rate (Fig. 2e top). When we performed the same analysis after depleting integrated genomic HBV, we observed a lower rate of indel formation (0% vs 32%, 62% vs 88% and 21% vs 66% for guides sg21, sg17 and sg6, respectively) (Fig. 2e bottom). Notably, however, this analysis method cannot detect Cas9-mediated cleavage of cccDNA followed by exonuclease-mediated degradation from the newly-formed free DNA ends (instead of re-ligation by NHEJ), and may be limited by the very small amount of episomal HBV remaining at late time points (Fig. 2d, Fig S4). Consistent with high levels of indel formation in the core ORF targeted by sg17, immunostaining for HBV core protein (HBc) revealed a robust reduction in HBc levels in sg17-expressing cells as compared to controls (Fig. 2f). Because long-term expression of Cas9 and guide RNAs can lead to off-target cleavage at sites with homology to the target sequence, we then performed next-generation sequencing at several computationally predicted off-target sites for sg6, sg17, and sg21. Within the sensitivity of our assay (<鈥�0.3% based on read depth), we detected no indel formation at the 8 off-target sites that we surveyed after constitutive expression of Cas9 and sgRNAs for over four weeks (Fig. S7a). This observed specificity may be due to the large sequence differences between viral and human genomic DNA (Fig. S7b-d). These encouraging results still did not exclude the possibility that some of the antiviral effects of Cas9 in the HepG2.2.15 system occur through mutations in integrated HBV DNA, thereby reducing the fitness and/or persistence of virions produced from mutated loci rather than acting directly on episomal DNA. Since integration of HBV DNA into the host human genome is not part of the canonical HBV life-cycle, we next evaluated the effects of Cas9 targeting in the context of de novo HBV infection, where episomal cccDNA serves as the only template for viral gene expression and replication.
Cas9 cleaves cccDNA and inhibits de novo HBV infection
To evaluate our anti-HBV CRISPR/Cas9 strategy in a setting of de novo infection, we used HepG2 cells overexpressing the HBV receptor NTCP (Hep-NTCP)14, which are permissive to infection with HBV. Because sg17 showed the highest levels of cccDNA mutagenesis in our HepG2.2.15 experiments, these cells were transduced with Cas9/sg17, Cas9/sg17M, or dCas9/sg17 lentiviruses, co-cultured with HBV producing HepG2.2.15 cells, and selected with puromycin to get rid of non-transduced Hep-NTCP and co-cultured HepG2.2.15 cells (Fig S8 left). Alternatively, Hep-NTCP cells were selected with puromycin following transduction and subsequently infected with HBV-positive patient serum (Fig S8 right). When the transduced Hep-NTCP were infected with cell culture-produced virus, Cas9/sg17 greatly abrogated productive HBV infection, as reflected by reduction in HBsAg and HBV DNA secretion, as well as 3.5kb RNA and cccDNA levels, compared to controls (Fig. 2g); this was confirmed by infection with patient-derived virus (Fig S9). While nuclease-deficient Cas9 also reduced viral 3.5kb RNA abundance in this system, this finding fits with other reports that dCas9 binding can inhibit transcription in mammalian cells15. Surveyor assay performed using DNA from cells infected de novo with HepG2.2.15-derived virus confirmed direct Cas9-mediated mutagenesis of HBV episomal DNA (Fig. 2h). Although some mutagenesis was also detected when the mutated sg17M was used, this most likely was due to low-level cleavage with DNA bulge-containing guide RNAs16. This finding provides direct evidence that Cas9 is capable of targeting episomal forms of the virus, and exerting anti-HBV effects by directly targeting cccDNA.
Discussion
Although largely unexplored in mammalian systems, bacteria and archaea utilize sequence specific DNA nucleases to interfere with viral replication17. Inspired by CRISPR鈥檚 evolutionary origins, we aimed to exploit the antiviral activity of Cas9 to target HBV DNA in mammalian cells. We show that targeting multiple conserved regions of HBV with Cas9 results in robust suppression of viral replication and direct mutagenesis and depletion of cccDNA. While integrated forms of HBV DNA were not depleted by Cas9 cleavage, these forms should not contribute to viral rebound in vivo18, and Cas9-driven mutagenesis of these sequences nonetheless would damage the viability of viral proteins generated from integrants. The unique advantages of the CRISPR/Cas9 system (such as multiplexed targeting) are of interest in developing antiviral applications, and indeed, very recently other groups have published examples of Cas9 cleavage of HBV in multiple model systems19,20,21,22. Our work provides an extension beyond these complementary studies, by demonstrating the anti-HBV effects of sgRNAs specifically targeting highly conserved regions of HBV in vitro and in vivo, by directly confirming mutagenesis in cccDNA in a de novo infection model of HBV, and extending this antiviral activity to patient-derived virus. Additionally, our finding that appropriately chosen virus-targeting sgRNAs can avoid inducing off-target cleavage, even upon sustained Cas9/sgRNA expression, strengthens the case for selecting viral targets as good candidates for CRISPR/Cas9 therapeutic use23. Interestingly, while Cas9/sg17 was efficient in suppressing infection and in directly cleaving nuclear cccDNA, Cas9/sg21 efficiently cleaved only integrated but not episomal DNA, which resulted in a lack of activity for Cas9/sg21 in de novo infection experiments (data not shown). The reason for this is unclear and warrants further study. Cas9 is a large multi-domain protein, and thus one hypothesis is that particular regions of the HBV genome are differentially accessible to Cas9 because of the tightly packed physical architecture of cccDNA. This underscores the importance of using models of authentic cccDNA to investigate therapeutic applications of targeted nucleases for HBV, and suggests that a careful selection of targets and guides will be required to achieve a substantial mutagenesis and depletion of viral DNA. In addition, our proof of concept experiments show that multiplexing sgRNAs can generate stronger antiviral effects (Fig S2), suggesting that this strategy may further maximize CRISPR-mediated restriction of components of the viral life cycle, possibly including cccDNA stability.
This study provides a proof of concept, but clinical translation of CRISPR/Cas9 systems to cure HBV will require some advances over the work described here. First, an exhaustive profiling of possible Cas9 target sites on cccDNA can uncover optimal target sites based on cccDNA accessibility and sgRNA binding properties. Secondly, delivery of Cas9/sgRNA constructs in vivo will require the use of clinically relevant delivery vectors such as AAV, which may require additional modifications such as switching to smaller Cas9 orthologs to save packaging size30. Finally, although we could not find evidence of off-target cutting in our directed sequencing, possibly due to the low homology between viral and human genomic Cas9 targets, an extensive genome-wide profiling of off-target effects is warranted.
The unusual persistence of cccDNA is currently the major obstacle for curing chronic HBV infection. To eliminate the virus and to prevent possible re-activation, it is probably necessary to eliminate all or at least the vast majority of episomal DNA from hepatocytes through a combination of exogenous treatment (presented here) and immune-mediated endogenous clearance. CRISPR/Cas9-mediated therapy may synergize with currently-used RT inhibitors, which should block the formation of new molecules of cccDNA via re-entry of newly synthesized replicative forms to the nucleus. The developments proposed above represent an active area of investigation for groups looking for ways to use CRISPR in a therapeutic fashion more broadly, which may accelerate progress toward an anti-HBV CRISPR therapeutic.
In summary, these results constitute the first example of CRISPR/Cas9 systems directly targeting an authentic pathogenic virus with episomal DNA, and demonstrate the potential for cccDNA-directed antiviral therapy using Cas9, which may represent a significant step towards the cure of chronic HBV infection. The results demonstrated here may also be used to inform the development of CRISPR/Cas9-based therapeutics for other DNA viruses, such as herpesviruses and papillomaviruses that use an episomal DNA as a template for their gene expression and replication.
|
|