miR-122–based therapies select for three distinct resistance mechanisms based on alterations in RNA structure

Hepatitis C virus (HCV) is a positive-sense RNA virus that interacts with a liver-specific microRNA called miR-122. miR-122 binds to two sites in the 5′ untranslated region of the viral genome and promotes HCV RNA accumulation. This interaction is important for viral RNA accumulation in cell culture, and miR-122 inhibitors have been shown to be effective at reducing viral titers in chronic HCV-infected patients. Herein, we analyzed resistance-associated variants that were isolated in cell culture or from patients who underwent miR-122 inhibitor–based therapy and discovered three distinct resistance mechanisms all based on changes to the structure of the viral RNA. Specifically, resistance-associated variants promoted riboswitch activity, genome stability, or positive-strand viral RNA synthesis, all in the absence of miR-122. Taken together, these findings provide insight into the mechanism(s) of miR-122–mediated viral RNA accumulation and provide mechanisms of antiviral resistance mediated by changes in RNA structure.

Hepatitis C virus (HCV) is a positive-sense RNA virus of the Flaviviridae family. The ∼9.6 kb HCV genomic RNA consists of a single open reading frame, which gives rise to the viral polyprotein that is processed into the 10 mature viral proteins flanked by highly structured 5′ and 3′ untranslated regions (UTRs) (1, 2). As a positive-sense RNA virus, the viral genome itself must serve as a template for the different stages of the viral life cycle, including viral translation, replication, and packaging (2). To this end, the viral 5′ and 3′ UTRs contain several cis-acting RNA elements that play important roles in directing the various stages of the viral life cycle (2–4). Specifically, the 5′ UTR contains the viral internal ribosomal entry site (IRES) made up of stem–loops (SL) II-IV, which is required for viral translation, while sequences and SL structures in both the 5′ (SLI-II) and 3′ UTRs (variable region, polyU/UC-tract, and 96-nt X-tail) are required for viral RNA replication (2, 5–8). Additionally, the 5′ terminus of the HCV genome interacts with the highly abundant, liver-specific human microRNA (miRNA), miR-122 (9–12).miR-122 is a highly expressed miRNA in the liver with ∼135,000 copies per hepatocyte (9, 13). While miRNAs typically interact with the 3′ UTRs of their target messenger RNAs (mRNAs) to dampen gene expression, miR-122 interacts to two sites in the 5′ UTR (site 1 and site 2) of the viral genome, and these interactions promote viral RNA accumulation (9, 10, 12). Several recent studies have led to a new model for miR-122:HCV RNA interactions that suggest that miR-122 plays at least three roles in the HCV life cycle (Fig. 1) (14–17). Firstly, the HCV 5′ UTR is thought to initially adopt an energetically favorable conformation (termed SLII^alt), which results in the recruitment of an Ago:miR-122 complex to site 2 of the HCV genome. This results in an RNA chaperone-like switch in conformation, akin to a bacterial riboswitch (18), resulting in the formation of SLII and allowing the viral IRES (SLII-IV) to form (15–17). Secondly, this change in conformation allows the recruitment of Ago:miR-122 to site 1, which protects the 5′ terminus from pyrophosphatase activity and subsequent exoribonuclease-mediated decay (12, 14, 19–21). Finally, the Ago protein bound to site 2 makes direct contact with the viral IRES, promoting HCV IRES-mediated translation (16).Open in a separate windowFig. 1.Model of miR-122 interactions with the HCV genome. The HCV genomic RNA is thought to enter the cell in an energetically stable conformation termed SLII^alt. The recruitment of the first Ago:miR-122 molecule to the accessible (unpaired) site 2 serves as an RNA chaperone, akin to a bacterial riboswitch, which refolds the RNA into the functional SLII conformation and allows the viral IRES (SLII-IV) to form (1). Subsequent recruitment of a second Ago:miR-122 molecule to site 1 promotes genome stability by protecting the 5′ terminus from cellular pyrophosphatases and exoribonuclease-mediated decay (2). In order to accommodate the Ago:miR-122 complex at site 1, the Ago:miR-122 complex at site 2 releases its auxiliary interactions but is likely stabilized by interactions between the Ago protein and SLII of the HCV IRES (3). Collectively, these interactions promote HCV IRES-mediated translation. miR-122 seed and auxiliary binding sequences are indicated (red).Due to the importance of miR-122 in the HCV life cycle, two miR-122 inhibitors (antisense oligonucleotides), the first miRNA-based drugs to enter clinical trials, have been developed and used to treat chronic HCV infection in the clinic (22, 23). Both Miravirsen (Santaris Pharma, a/s) and RG-101 (Regulus Therapeutics) miR-122 inhibitors have completed Phase II or Ib clinical trials, respectively, to investigate their clinical efficacy in chronic HCV infection (22, 23). Excitingly, both treatments led to dose-dependent and sustained reductions in viral loads; and, in the latter study, two patients achieved a sustained virological response (at least up to 76 wk posttreatment) after receiving a single dose of RG-101 (23). Neither treatment was associated with significant adverse events or long-term safety issues, suggesting that antisense targeting of miR-122 may be an effective treatment that could be used in future combination therapies. Interestingly, while no resistance was apparent during treatment, when viral RNA rebounded after the cessation of the inhibitor, several resistance-associated variants (RAVs) were identified in the 5′ UTR of the HCV genome (23). Specifically, C3U (genotype 1) was identified as a RAV in both the Miravirsen and RG-101 trials, while the C2GC3U (genotype 3/4) RAV was identified in the RG-101 trial alone, with both RAVs identified in multiple patients (22, 23). Additional RAVs were identified in cell culture, including studies performed with genotype 1 (A4C) and genotype 2 (U4C, G28A, and C37U), which were identified alone (A4C and G28A) or in combination with other RAVs (i.e., G28A+C37U, U4C+G28A+C37U) (24–26). As the cell culture–based studies identifying RAVs were performed with genotype 2 and the majority of the RAV nucleotide identities are present in genotype 2 (save for A4C, although U4C was deemed an equivalent RAV observed in this genotype), herein, we explored the mechanism of action of these collective RAVs (C2GC3U, C3U, U4C, G28A, and C37U) using genotype 2a (J6/JFH-1) reporter RNAs (Fig. 2A) (22–26). Previous work suggests that the G28A mutation is “riboswitched” and promotes the formation of the functional SLII structure even in the absence of miR-122 (16, 17). Similar to G28A, we hypothesized that the other RAVs also alter the structure of the viral RNA in a manner that negates the requirement for one or more miR-122 activities. Thus, we sought to provide insight into the mechanism(s) of action of the RAVs using RNA structure analysis and assays for viral RNA accumulation and decay. Our analyses suggest that each of the RAVs alter the structure of the viral RNA, and we identify three distinct resistance mechanisms based on unique changes in viral RNA structure.Open in a separate windowFig. 2.RAV accumulation in cell culture. (A) The positions of the RAVs on the 5′ UTR of the HCV RNA. Nucleotides 1 to 45 of the HCV genome (black) and the binding topology of the two miR-122 molecules (teal) are shown. Full-length RLuc HCV genomic reporter RNAs (WT and RAVs) were coelectroporated with a capped Firefly fuciferase (FLuc) mRNA into (B) Huh-7.5 or (C) miR-122 KO cells. Full-length RLuc HCV genomic reporter RNAs containing mutations at (D) site 1 (S1:p3A) or (E) site 2 (S2:p3A) were coelectroporated with a capped FLuc mRNA and compensatory miR-122p3U molecules into miR-122 KO cells. Luciferase activity was measured at the indicated time points post-electroporation. The limit of detection is indicated, and all data are representative of at least three independent replicates. Error bars represent the SD of the mean.